From b.j.smith at ieee.org Tue Aug 1 01:26:37 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:17:58 2006 Subject: AMD / Intel CPU's Was: [Pc_Support] Re: AMD's L1 cache is 4x larger than Intel's In-Reply-To: <44CEAE62.9050206@HiWAAY.net> References: <20060716170859.50864.qmail@web32908.mail.mud.yahoo.com> <1153492860.10798.2.camel@x2.ultrasound.home> <1153580589.2926.53.camel@bert64.oviedo.smithconcepts.com> <1153581601.2926.68.camel@bert64.oviedo.smithconcepts.com> <1758.71.252.176.10.1153622045.squirrel@qtmail.dgnal.net> <1153621400.2923.24.camel@bert64.oviedo.smithconcepts.com> <44CEAE62.9050206@HiWAAY.net> Message-ID: <1154409997.3047.32.camel@bert64.oviedo.smithconcepts.com> On Mon, 2006-07-31 at 20:29 -0500, William A. Mahaffey III wrote: > Was this price cut supposed to includes S939 opterons as well ? Nope. No Opteron cuts, not even Socket-939. You might see a slight drop, but nothing like the 25-50+% cuts on consumer models. -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From b.j.smith at ieee.org Tue Aug 1 01:36:15 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:17:58 2006 Subject: [Pc_Support] Re: WARNING: ZipZoomFly playing games on Athlon 64 x2 sales ... UPDATE In-Reply-To: <20060729201605.29596.qmail@web32912.mail.mud.yahoo.com> References: <20060729201605.29596.qmail@web32912.mail.mud.yahoo.com> Message-ID: <1154410575.3047.38.camel@bert64.oviedo.smithconcepts.com> On Sat, 2006-07-29 at 13:16 -0700, Bryan J. Smith wrote: > ... I found ZipZoomFly was $260 for the 4600+ and said "in stock" > -- even if not the other models. So I went ahead and bought it for > $260 on Monday ... On Tuesday they said it was "packed" and > ready-to-ship and I could not cancel ... On Friday, I just got the > following e-mail. I remember checking all throughout the week on > various models -- including the x2 4600+ -- and they said "in stock." Okay, here's a "slight" update. First off, ZipZoomFly.COM _did_ finally update their pages for the 4600+ as "backordered" over the weekend. Secondly, ZipZoomFly.COM is still the _cheapest_ place to buy a 4600+ at $260. NewEgg.COM wants $329.99 ($70 more) and Monarch Computer wants $289.99 ($30 more), although both have them in stock. So, at this point, I can only assume one of the following ... 1. ZipZoomFly.COM is honest and is trying to get stock and will honor the orders made -- especially those of us who ordered the very first day. 2. ZipZoomFly.COM realizes it sold too low, and is going to go the full 2 weeks and then cancel the order. If this is the case, I will fully demand they they honor the $260 price with free shipping ($1.99 2-day shipping) whenever they have stock again. I am assuming #1. However, I am more than ready to demand #2 come next Monday. -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From b.j.smith at ieee.org Tue Aug 1 01:39:25 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:17:58 2006 Subject: [Pc_Support] Re: WARNING: ZipZoomFly playing games on Athlon 64 x2 sales ... UPDATE In-Reply-To: <1154410575.3047.38.camel@bert64.oviedo.smithconcepts.com> References: <20060729201605.29596.qmail@web32912.mail.mud.yahoo.com> <1154410575.3047.38.camel@bert64.oviedo.smithconcepts.com> Message-ID: <1154410765.3047.42.camel@bert64.oviedo.smithconcepts.com> On Tue, 2006-08-01 at 01:36 -0400, Bryan J. Smith wrote: > Secondly, ZipZoomFly.COM is still the _cheapest_ place to buy a 4600+ at > $260. NewEgg.COM wants $329.99 ($70 more) and Monarch Computer wants > $289.99 ($30 more), although both have them in stock ... cut ... > 2. ZipZoomFly.COM realizes it sold too low, and is going to go the full > 2 weeks and then cancel the order. If this is the case, I will fully > demand they they honor the $260 price with free shipping ($1.99 2-day > shipping) whenever they have stock again. Just FYI ... I _carefully_ followed the stock of all existing processors, basically everything from the x2 3800+ to 4800+ (only the 5000+ is new). Virtually _everyone_ "had stock" and have been playing games on stock the _second_ they dropped pricing. I assume this is because resellers were negotiating a "refund" or other "buy back/discount" from distributors for unsold stock when the price cut hit. I understand that and have made that my assumption in the case of ZipZoomFly.COM. But with Monarch Computer only being $30 more and having stock _now_, they had better honor the price by next Monday. If I see them drop my order while claiming stock next Monday, there's going to be consumer hell to pay. Again, I'm sticking with being positive and benign by default until I see this. -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From thomas at tecsplace.com Tue Aug 1 02:38:29 2006 From: thomas at tecsplace.com (Thomas Carlson) Date: Tue Oct 31 13:17:58 2006 Subject: [Pc_Support] GB switch? Worth it? Message-ID: <1154414309.2477.7.camel@localhost.localdomain> Folks, I saw this GB 8 port switch and thinking of buying it. http://dealmac.com/deals/Netgear-GS108-8-Port-Copper-Gigabit-Ethernet-Switch-for-60-shipped-after-rebate/126897.html?redir=1&ref=txt_dealmacdaily Here's the specs... http://www.netgear.com/Products/Switches/DesktopSwitches/GS108.aspx?detail=Specifications In the reviews, one person states that the 9k jumbo frames didn't work, and that seems to be a major factor affecting performance. http://www.buy.com/prod/NETGEAR_GS108_Switch_8_port_s_10Base_T_100Base_TX_1000Base_T_1_Gbps_EN/q/loc/101/10332287.html?adid=17662 Thomas From wam at HiWAAY.net Tue Aug 1 08:32:23 2006 From: wam at HiWAAY.net (William A. Mahaffey III) Date: Tue Oct 31 13:17:58 2006 Subject: AMD / Intel CPU's Was: [Pc_Support] Re: AMD's L1 cache is 4x larger than Intel's In-Reply-To: <1154409997.3047.32.camel@bert64.oviedo.smithconcepts.com> References: <20060716170859.50864.qmail@web32908.mail.mud.yahoo.com> <1153492860.10798.2.camel@x2.ultrasound.home> <1153580589.2926.53.camel@bert64.oviedo.smithconcepts.com> <1153581601.2926.68.camel@bert64.oviedo.smithconcepts.com> <1758.71.252.176.10.1153622045.squirrel@qtmail.dgnal.net> <1153621400.2923.24.camel@bert64.oviedo.smithconcepts.com> <44CEAE62.9050206@HiWAAY.net> <1154409997.3047.32.camel@bert64.oviedo.smithconcepts.com> Message-ID: <44CF49D7.9000005@HiWAAY.net> Bryan J. Smith wrote: >On Mon, 2006-07-31 at 20:29 -0500, William A. Mahaffey III wrote: > > >>Was this price cut supposed to includes S939 opterons as well ? >> >> > >Nope. No Opteron cuts, not even Socket-939. You might see a slight >drop, but nothing like the 25-50+% cuts on consumer models. > > CR*P !!!! & here I was waiting patiently :-) .... -- William A. Mahaffey III ---------------------------------------------------------------------- "The M1 Garand is without doubt the finest implement of war ever devised by man." -- Gen. George S. Patton -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.matrixlist.com/pipermail/pc_support/attachments/20060801/d2c7d40e/attachment.html From readg at nfl.jaguars.com Wed Aug 2 10:59:21 2006 From: readg at nfl.jaguars.com (Read, Greg) Date: Tue Oct 31 13:17:58 2006 Subject: [Pc_Support] AGP 4X/8X Message-ID: Well, I changed my mind and ordered the card. I found somewhere on XFX's site that the card required a 4x or 8x slot. Someone also pointed me to this link: http://www.directron.com/15agpguide.html#what I plugged it in last night and it worked fine. The DVI out is noticeably sharper w/ better color. Thanks guys. Greg > -----Original Message----- > From: pc_support-bounces@matrixlist.com [mailto:pc_support- > bounces@matrixlist.com] On Behalf Of Read, Greg > Sent: Thursday, July 20, 2006 3:14 PM > To: This is the PC Support list. > Subject: RE: [Pc_Support] AGP 4X/8X > > Thanks everyone, > > I can't find the voltage to the AGP slot listed anywhere, nor can I find > the voltage to the card listed on XFX's site. > > Jason's experience with XFX is strike two. > > I'll wait until I upgrade the MB. > > > -----Original Message----- > > From: pc_support-bounces@matrixlist.com > > [mailto:pc_support-bounces@matrixlist.com] On Behalf Of Bryan J. Smith > > Sent: Wednesday, July 19, 2006 11:39 PM > > To: This is the PC Support list. > > Subject: Re: [Pc_Support] AGP 4X/8X > > > > On Wed, 2006-07-19 at 15:41 -0400, Read, Greg wrote: > > > I have an old system, Athlon 1700XP and it has an AGP 4X slot; I'm > > > currently using a NVidia GeForce 3 Ti200 in it. > > > I bough a new LCD monitor that has a DVI input and I wanted > > to get a > > > cheap NVidia graphics card that has DVI. > > > I found one on Buy.com, XFX GeForce 6200 128MB DDR AGP Video Card ( > > > DVI TV-Out VGA ) - PVT44ARA it's $28 after rebate. > > > The catch is, it says it's AGP 8X, can I use it in my 4X slot? > > > > The NV44 (GeForce 6200) kicks about all pre-NV40 ass -- > > including the non-TurboCache (which includes AGP) being 4-6x > > faster than a NV34 (GeForce FX5200/5500): > > http://thebs413.blogspot.com/2006/02/geforce-6-and-7-series-va > > riants-nuts.html > > > > > I've seen similar eVGA cards on Newegg that say they are AGP4X/8X, > > > should I go with one of those? > > > > Depends. On my blog, I listed the AGP 1.0 (1x/2x), 2.0 > > (2x/4x) and 3.0 > > (4x/8x) voltages: > > http://thebs413.blogspot.com/2005/11/agp-agp-pro-pci-and-pci-x > > -voltage.html > > > > Many newer AGP 3.0, 0.8V cards are _not_ always 1.5V > > tolerant. As such, they do _not_ often work in AGP 2.0 slots > > that can only do 1.5V (and are possibly 3.3V tolerant). This > > has been a major issue with compatibility. > > > > To make matters worse, if you buy the card and put it in the > > slot, you _could_ blow it and/or the slot. Typically the > > problem is when your older card uses a higher voltage than > > the newer slot of lower voltage, and a newer card with lower > > voltage won't blow the older slow of higher voltage. > > > > But you never know. And the mainboard as well as the cards > > do a _poor_ job of telling you. > > > > I don't know if nVidia did a good job with the NV44 (6200) > > and 1.5V tolerance -- let alone if vendors are honoring them. > > But I can tell you that did _not_ do a good job with the > > NV43 (6600) when it came to 1.5V tolerance. They typically > > _only_ work in AGP 3.0 (4x/8x) slots and _not_ older AGP 2.0 > > (2x/4x) slots. > > > > > > > > -- > > Bryan J. Smith Professional, technical annoyance > > mailto:b.j.smith@ieee.org http://thebs413.blogspot.com > > --------------------------------------------------------- > > The world is in need of solutions. Unfortunately, people > > seem to be more interested in blindly aligning themselves > > with one of only two viewponts -- an "us v. them" debate that > > has nothing to do with finding an actual solution. > > > > > > _______________________________________________ > > Pc_support mailing list > > Pc_support@matrixlist.com > > http://lists.matrixlist.com/mailman/listinfo/pc_support > > > > _______________________________________________ > Pc_support mailing list > Pc_support@matrixlist.com > http://lists.matrixlist.com/mailman/listinfo/pc_support From b.j.smith at ieee.org Wed Aug 2 12:01:27 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:17:58 2006 Subject: [Pc_Support] AGP 4X/8X In-Reply-To: References: Message-ID: <1154534487.2961.78.camel@bert64.oviedo.smithconcepts.com> On Wed, 2006-08-02 at 10:59 -0400, Read, Greg wrote: > Well, I changed my mind and ordered the card. > I found somewhere on XFX's site that the card required a 4x or 8x slot. > Someone also pointed me to this link: > http://www.directron.com/15agpguide.html#what That information isn't accurate. First off, you can physically put a AGP 1.0 (3.3V) card in a newer AGP 3.0 (0.8V and/or 1.5V) slot. That's because most of those slots don't have any keys to protect you voltage-wise. Secondly, there are AGP 3.0 cards that are 0.8V-only, and not 1.5V tolerant. Those can bite you as well. > I plugged it in last night and it worked fine. That's good. You found a AGP 3.0 card that also does 1.5V and/or your mainboard is capable of 0.8V. > The DVI out is noticeably sharper w/ better color. Yep. DVI is definitely ideal. -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From dave at dgnal.net Thu Aug 3 02:27:28 2006 From: dave at dgnal.net (David Simmons) Date: Tue Oct 31 13:17:58 2006 Subject: [Pc_Support] The Joomla! CMS & the FacileForms module In-Reply-To: <1153842523.2943.15.camel@bert64.oviedo.smithconcepts.com> References: <44C4F10C.50604@sunstatemartialarts.com> <1153834225.3120.14.camel@bert64.oviedo.smithconcepts.com> <1153839017.7672.2.camel@x2.ultrasound.home> <1153842523.2943.15.camel@bert64.oviedo.smithconcepts.com> Message-ID: <10209.71.252.176.10.1154586448.squirrel@qtmail.dgnal.net> Guys, Was wondering if there's anyone on the list that uses the CMS Joomla! (which I guess would also mean to say any current version Mambo users)...that are familiar with the Module/Component entitled, "FacileForms"? If so, please contact me offlist Thanks! - dave From b.j.smith at ieee.org Thu Aug 3 03:18:55 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:17:58 2006 Subject: [Pc_Support] Seagate's new 24x7 rated line is the Barracuda ES ... Message-ID: <1154589535.2938.33.camel@bert64.oviedo.smithconcepts.com> Just wanted to drop a line that Seagate's new, 24x7 rated line is the "Barracuda ES" line, and no longer the "NL35" line. "Barracuda ES" 24x7 rated drives, like the "NL35" before it, are the drives from the stock "Barracuda 7200.x" series that test to the best tolerances and minimal vibrations. They come off the same line. I've personally found that Seagate's quality has gone to junk after the Barracuda 7200.7 series -- about the same time they started outsourcing the Maxtor (who they more recently merged with). The 7200.8 has now bitten me several times, and others (so hearsay/2nd hand) have stated the 7200.9 has been a bad idea as well. Can't comment on the 7200.10 series. So now more than ever, I recommend the "NL35" and "Barracuda ES" lines instead of the OEM/retail "Barracuda 7200.8+" models. Of course, you'll pay up to a +100% premium for them though -- typically +50%. Hitachi is similar on its T7K rated 24x7 line over their standard OEM/retail 7K line. The only reseller that seems to minimize markup of their 24x7 rated line over the standard OEM/retail is Western Digital. The Caviar RE[2] is typically only +20% more than the Caviar SE[2]. I have several of the Western Digital 320GB, 107GB/platter generation of ATA/SATA in use -- and I implicitly recommend the $120 WD3200SD Caviar RE because of its price/performance in a 24x7 rated option. I have not tested the native SATA/3Gbps "YS" line yet, and bewar of some of the 320GB, but older 80GB/platter ATA/SATA Caviar SE16 drives on the market. As always, find reviews of the _exact_model_ you are considering. *NEVER* go on brand name. You could trust Seagate before 2003 -- but even they have given into outsourcing. -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From b.j.smith at ieee.org Sat Aug 5 16:09:51 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:17:58 2006 Subject: [Pc_Support] Re: VMware pre-made virtual guests In-Reply-To: <44D4D017.6080002@airmail.net> Message-ID: <20060805200951.97099.qmail@web32912.mail.mud.yahoo.com> Chris Cox wrote: > Somebody at the meeting asked about these: > http://www.vmware.com/vmtn/appliances/directory/ I already have a file server (dual-P3, 1GB of RAM, 3Ware RAID-10). But I have an extra system (Sempron 2800+ with 1GB of RAM and 3Ware RAID-10 storage) -- along with several dual-NIC PCI cards. So I was thinking about loading up CentOS 3 on it and then running both IPCop (Firewall) and either Fedora Core 6 or CentOS 4 with Fedora Directory Server 1.02 (Auth/Dir/Name Server). At some point in the future I'm going to go OpenFiler as well -- all on one box (probably my old dual-Athlon MP2400+). So 3 VMs: - Firewall - Auth/Dir/Name Server - File Server My only worry is that since the underlying host for the VM actually has to enable the interfaces for the underlying guest IPCop box, if there wasn't some worry of level-2 hacking going on at the host before it gets to the IPCop VM. I know I'm probably being over-anal on that, but it's still a concern. And yes, I'd put the other VM guest LAN facilities on their own network interface separate from the ones IPCop uses. But it's still a concern of mine, since the host still opens all those interfaces to itself. -- Bryan J. Smith Professional, Technical Annoyance b.j.smith@ieee.org http://thebs413.blogspot.com -------------------------------------------------- Fission Power: An Inconvenient Solution From hubbardjw at charter.net Sat Aug 5 23:47:18 2006 From: hubbardjw at charter.net (Jerry Hubbard) Date: Tue Oct 31 13:17:58 2006 Subject: Drive hacks (Was:Re: [Pc_Support] Re: VMware pre-made virtual guests) In-Reply-To: <20060805200951.97099.qmail@web32912.mail.mud.yahoo.com> References: <20060805200951.97099.qmail@web32912.mail.mud.yahoo.com> Message-ID: <44D56646.1000200@charter.net> Bryan J. Smith wrote: > Chris Cox wrote: > >> Somebody at the meeting asked about these: >> http://www.vmware.com/vmtn/appliances/directory/ >> > > I already have a file server (dual-P3, 1GB of RAM, 3Ware RAID-10). > > But I have an extra system (Sempron 2800+ with 1GB of RAM and 3Ware > RAID-10 storage) -- along with several dual-NIC PCI cards. > > So I was thinking about loading up CentOS 3 on it and then running > both IPCop (Firewall) and either Fedora Core 6 or CentOS 4 with > Fedora Directory Server 1.02 (Auth/Dir/Name Server). > > At some point in the future I'm going to go OpenFiler as well -- all > on one box (probably my old dual-Athlon MP2400+). So 3 VMs: > - Firewall > - Auth/Dir/Name Server > - File Server > > My only worry is that since the underlying host for the VM actually > has to enable the interfaces for the underlying guest IPCop box, if > there wasn't some worry of level-2 hacking going on at the host > before it gets to the IPCop VM. I know I'm probably being over-anal > on that, but it's still a concern. > > And yes, I'd put the other VM guest LAN facilities on their own > network interface separate from the ones IPCop uses. But it's still > a concern of mine, since the host still opens all those interfaces to > itself. > > > Bryan, Are NIC driver hacks a concern? Since hearing about the Black Hat demo of a wireless driver hack, I have been wondering about this. http://blog.washingtonpost.com/securityfix/2006/08/hijacking_a_macbook_in_60_seco.html -- Jerry Hubbard hubbardjw@charter.net From b.j.smith at ieee.org Sun Aug 6 00:48:22 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:17:58 2006 Subject: Drive hacks (Was:Re: [Pc_Support] Re: VMware pre-made virtual guests) In-Reply-To: <44D56646.1000200@charter.net> Message-ID: <20060806044822.93148.qmail@web32907.mail.mud.yahoo.com> Jerry Hubbard wrote: > Bryan, > Are NIC driver hacks a concern? Since hearing about the Black Hat > demo of a wireless driver hack, I have been wondering about this. Well, I'm just concerned about kernel hacks in general. I mean, I've basically got to leave the network interface "open" on the host OS, so the guest OS can use it. -- Bryan J. Smith Professional, Technical Annoyance b.j.smith@ieee.org http://thebs413.blogspot.com -------------------------------------------------- Fission Power: An Inconvenient Solution From carter at carter.cc Sun Aug 6 12:09:19 2006 From: carter at carter.cc (Carter Manucy) Date: Tue Oct 31 13:17:58 2006 Subject: Drive hacks (Was:Re: [Pc_Support] Re: VMware pre-made virtual guests) In-Reply-To: <20060806044822.93148.qmail@web32907.mail.mud.yahoo.com> References: <20060806044822.93148.qmail@web32907.mail.mud.yahoo.com> Message-ID: <44D6142F.7020602@carter.cc> Bryan J. Smith wrote: > Well, I'm just concerned about kernel hacks in general. > > I mean, I've basically got to leave the network interface "open" on > the host OS, so the guest OS can use it. > What if you don't bind TCP/IP to the host? I do this routinely... either that, or just give the host some kind of bogus IP on the NIC, then let the guest actually assign the 'correct' IP. Obviously if you don't have individual cards for the different interface or if you're not able to VLANs, this could be a bigger issue... but I don't think there's really anything you can do if the host isn't bound to the same NIC. I've yet to see even a hint of someone being able to break out of a VM... and although VMWare is no IBM, the IBM "VM's" on their AS/400's have been around for a long, long time, with nary a worry (so far as I know) about being able to break out of one VM and either get to the host or get to another VM. As a side note, in ESX, you get your own Layer-2 virtual switch 'built in' to the OS that handles all of the traffic. -Carter From carter at carter.cc Sun Aug 6 12:28:41 2006 From: carter at carter.cc (Carter Manucy) Date: Tue Oct 31 13:17:58 2006 Subject: Drive hacks (Was:Re: [Pc_Support] Re: VMware pre-made virtual guests) In-Reply-To: <44D56646.1000200@charter.net> References: <20060805200951.97099.qmail@web32912.mail.mud.yahoo.com> <44D56646.1000200@charter.net> Message-ID: <44D618B9.9090205@carter.cc> Jerry Hubbard wrote: > > Are NIC driver hacks a concern? Since hearing about the Black Hat demo > of a wireless driver hack, I have been wondering about this. > > http://blog.washingtonpost.com/securityfix/2006/08/hijacking_a_macbook_in_60_seco.html > One thing to keep in mind is that wireless (802.11) is, by some accounts, a overly-complicated protocol. This is one of the main reasons you're seeing this wireless exploit at the driver level vs. a 'tried and true' Ethernet (802.3) exploit. I'm not saying it's not possible (as I really don't know the details on this 802.11 hack), but I will say it's MUCH less likely. All the more reason to keep that radio in the "OFF" position while you're not using it! This hack is pretty much OS independent, although I haven't seen or heard of any Linux-based exploits on it, I'm sure they are possible given the protocol. -Carter From b.j.smith at ieee.org Sun Aug 6 14:34:08 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:17:58 2006 Subject: Drive hacks (Was:Re: [Pc_Support] Re: VMware pre-made virtual guests) In-Reply-To: <44D6142F.7020602@carter.cc> Message-ID: <20060806183408.33985.qmail@web32903.mail.mud.yahoo.com> Carter Manucy wrote: > What if you don't bind TCP/IP to the host? I do this routinely... > either that, or just give the host some kind of bogus IP on the > NIC, then let the guest actually assign the 'correct' IP. First off, you don't have to bind TCP/IP on the host so the guest can use it? If that is the case, great! I guess that makes sense -- just never put it together. Secondly, even if TCP/IP isn't bound, the Ethernet link is up. That still bothers me, but I guess since Linux doesn't do bridging by default, it's tolerable. I'm going to be running CentOS, so I'll get updates for a long time. The guest OSes can be Fedora Core or something more fast-moving. > Obviously if you don't have individual cards for the different > interface or if you're not able to VLANs, VLANs? Is that the proper term? I don't think so. ;-> With that said, what mode should the NIC be in? Bridged to host I assume so it has full layer-2 control? > this could be a bigger issue... but I don't think there's really > anything you can do if the host isn't bound to the same NIC. Right. That's what I don't like. I was hoping there was a way to directly use hardware -- kinda like VMWare does allow a hard disk slice -- but I guess that's still not the same. > I've yet to see even a hint of someone being able to break out > of a VM... and although VMWare is no IBM, the IBM "VM's" on > their AS/400's have been around for a long, long time, with > nary a worry (so far as I know) about being able to break out > of one VM and either get to the host or get to another VM. > As a side note, in ESX, you get your own Layer-2 virtual switch > 'built in' to the OS that handles all of the traffic. Right, because the host OS is VMWare's Linux-based OS. Yes, I know, I've deployed it. Just not too familiar with VMWare Server until it became free. -- Bryan J. Smith Professional, Technical Annoyance b.j.smith@ieee.org http://thebs413.blogspot.com -------------------------------------------------- Fission Power: An Inconvenient Solution From carter at carter.cc Sun Aug 6 15:41:30 2006 From: carter at carter.cc (Carter Manucy) Date: Tue Oct 31 13:17:58 2006 Subject: Drive hacks (Was:Re: [Pc_Support] Re: VMware pre-made virtual guests) In-Reply-To: <20060806183408.33985.qmail@web32903.mail.mud.yahoo.com> References: <20060806183408.33985.qmail@web32903.mail.mud.yahoo.com> Message-ID: <44D645EA.4040104@carter.cc> Bryan J. Smith wrote: > First off, you don't have to bind TCP/IP on the host so the guest can > use it? If that is the case, great! I guess that makes sense -- > just never put it together. > > Secondly, even if TCP/IP isn't bound, the Ethernet link is up. That > still bothers me, but I guess since Linux doesn't do bridging by > default, it's tolerable. > I agree. That still 'keeps me up at night' a bit as well. With a Windows host, the only thing you need bound to the card is the "VMWare Bridge Protocol" - all other items (TCP/IP, File & Printer Sharing, etc) can be (and SHOULD be) unbound. With Linux, all you need is to be able to 'turn up' the interface - as you say, get the Ethernet link up. > I'm going to be running CentOS, so I'll get updates for a long time. > The guest OSes can be Fedora Core or something more fast-moving. > I use that for my base OS at home for VMWare Server, as well as for a few other locations for GSX Server. Works very well. > VLANs? Is that the proper term? I don't think so. ;-> > > With that said, what mode should the NIC be in? Bridged to host I > assume so it has full layer-2 control? > Yes, VLANs. If you have a managed switch, you can VLAN a few ports off to further contain the connections. With ESX (as I'm sure you know), it natively supports VLANs so you can just use tagged ports (or trunked, depending on which mfg you're used to listening to) and ESX will take care of the traffic in its own internal 'switch'. You really don't have to worry about placing the NIC in a certain mode with VMWare GSX/Server. With Windows, it adds a binding protocol... with Linux, I don't remember how it interfaces directly... I know it makes its own interfaces (VMNet) and the hosts bind to those. > I was hoping there was a way to directly use hardware -- kinda like > VMWare does allow a hard disk slice -- but I guess that's still not > the same. > No, because VMWare still uses its own drivers to bind to the VMNet interface (used to be VMLance and another AMD driver, now with VMWare Server you don't get a choice). In thinking about it, I could imagine folks could get themselves in a lot of 'trouble' if you gave them direct access to the NIC... for some of us it'd be a great thing, but for others it'd be a nightmare. > Right, because the host OS is VMWare's Linux-based OS. > Yes, I know, I've deployed it. > > Just not too familiar with VMWare Server until it became free. > It is, ya know! Go get yourself a serial # or two (heck, get 100 - they give you the option on their site). The thing that still irks me about VMWare is getting kernel updates and having to re-compile the whole bloody thing after using the new kernel. But other than that minor issue, I love it. And so does my power bill! -Carter From bigjohn at midwest.net Mon Aug 7 16:49:12 2006 From: bigjohn at midwest.net (JohnH) Date: Tue Oct 31 13:17:58 2006 Subject: [Pc_Support] OT? Virus Concern Message-ID: <000b01c6ba62$f11367e0$6401a8c0@3a5ah6vqcd> My VistaPrint Electronic Business CardI have gotten email with text that starts out with " R0lGOD " and then nothing but garbage after that. It also has "sentences" that say absolutely nothing. Is this a virus in disguise? Or am I just being paranoid? John H. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.matrixlist.com/pipermail/pc_support/attachments/20060807/9dd32f26/attachment.html From wam at HiWAAY.net Mon Aug 7 17:20:33 2006 From: wam at HiWAAY.net (William A. Mahaffey III) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] OT? Virus Concern In-Reply-To: <000b01c6ba62$f11367e0$6401a8c0@3a5ah6vqcd> References: <000b01c6ba62$f11367e0$6401a8c0@3a5ah6vqcd> Message-ID: <44D7AEA1.4090107@HiWAAY.net> JohnH wrote: > I have gotten email with text that starts out with " R0lGOD " and then > nothing but garbage after that. > It also has "sentences" that say absolutely nothing. > > Is this a virus in disguise? > Or am I just being paranoid? > > John H. > >------------------------------------------------------------------------ > >_______________________________________________ >Pc_support mailing list >Pc_support@matrixlist.com >http://lists.matrixlist.com/mailman/listinfo/pc_support > > Almost certainly virus or other mal-ware. I get these regularly. They can also be SPAM-mongers fishing for valid addresses. If your ISP doesn't filter them out, you might have a go at doing so yourself :-). -- William A. Mahaffey III ---------------------------------------------------------------------- "The M1 Garand is without doubt the finest implement of war ever devised by man." -- Gen. George S. Patton Jr. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.matrixlist.com/pipermail/pc_support/attachments/20060807/df0d6f5e/attachment.html From b.j.smith at ieee.org Sun Aug 13 14:26:54 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] ASRock 775Dual-VSTA: Core 2 Duo memory/video choice for $55 Message-ID: <1155493614.1706.15.camel@bert64.oviedo.smithconcepts.com> AnandTech had a review of the VIA PT880Pro chipset-based ASRock 775Dual-VSTA. It sports _both_ an AGP and PCIe x4 (physical x16) slot, as well as _both_ 184-pin DDR and 240-pin DDR2 slots. If you only want to upgrade your mainboard for $55 plus a Core 2 Duo CPU for $200, recycling your memory and video, then this is a solid, intermediate solution. http://www.anandtech.com/mb/showdoc.aspx?i=2810 For those running Linux, you might want to hold off as peripheral drivers for the VT8237A/VT8251 are just becoming available in newer kernels and support varies. -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From b.j.smith at ieee.org Sun Aug 13 14:31:53 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] eVGA 7600GS 256MB PCIe (12/5@400, 128b@800) for $89 at Circuit City Message-ID: <1155493913.1706.21.camel@bert64.oviedo.smithconcepts.com> Circuit City has the eVGA 7600GS 256MB PCIe for $149.99 - $60 rebate = $89.99. The 7600GS (12/5 pixel/vertex at 400MHz, 128-bit GDDR2 at 800MHz effective) is not nearly as fast as a GeForce 7600GT and is slower than a GeForce 6800GS/GT, but it's typically faster than a 6600GT. It is also passively cooled and lower power -- with all the benefits of the 7000 series for video. http://www.circuitcity.com/rpsm/oid/152206/rpem/ccd/productDetailSpecification.do -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From jasonb at edseek.com Sun Aug 13 14:37:40 2006 From: jasonb at edseek.com (Jason Boxman) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] eVGA 7600GS 256MB PCIe (12/5@400, 128b@800) for $89 at Circuit City In-Reply-To: <1155493913.1706.21.camel@bert64.oviedo.smithconcepts.com> References: <1155493913.1706.21.camel@bert64.oviedo.smithconcepts.com> Message-ID: <200608131437.40317.jasonb@edseek.com> On Sunday 13 August 2006 14:31, Bryan J. Smith wrote: > Circuit City has the eVGA 7600GS 256MB PCIe for $149.99 - $60 rebate = > $89.99. The 7600GS (12/5 pixel/vertex at 400MHz, 128-bit GDDR2 at > 800MHz effective) is not nearly as fast as a GeForce 7600GT and is > slower than a GeForce 6800GS/GT, but it's typically faster than a > 6600GT. It is also passively cooled and lower power -- with all the > benefits of the 7000 series for video. > > http://www.circuitcity.com/rpsm/oid/152206/rpem/ccd/productDetailSpecificat >ion.do Passive Cooling == awesome! If only I needed to upgrade... sigh. I'll probably buy something like this when I finally do instead of spending money replacing the noisy HSF on my current 6600GT eVGA with a quieter HSF. -- Jason Boxman http://edseek.com/ - Linux and FOSS stuff From b.j.smith at ieee.org Sun Aug 13 15:21:09 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] eVGA 7600GS 256MB PCIe (12/5@400, 128b@800) for $89 at Circuit City In-Reply-To: <200608131437.40317.jasonb@edseek.com> References: <1155493913.1706.21.camel@bert64.oviedo.smithconcepts.com> <200608131437.40317.jasonb@edseek.com> Message-ID: <1155496869.1706.26.camel@bert64.oviedo.smithconcepts.com> On Sun, 2006-08-13 at 14:37 -0400, Jason Boxman wrote: > Passive Cooling == awesome! Yeah, an excellent option for set-top or other usage, especially being a 7000 series and HDTV out. > If only I needed to upgrade... sigh. I'll probably buy something like this > when I finally do instead of spending money replacing the noisy HSF on my > current 6600GT eVGA with a quieter HSF. They'll come down in months as well. But after rebate, it's still $20 cheaper than the cheapest option from NewEgg. -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From b.j.smith at ieee.org Sun Aug 13 21:59:31 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] Low-cost, but solid: TRENDnet TEW-610APB (AP), TEW-441PC (CardBus), TEW-443PI (PCI) Message-ID: <1155520771.3223.1.camel@bert64.oviedo.smithconcepts.com> Wrestling with wireless for the last month, sometimes cheaper is not so bad -- as long as you find the right products: http://thebs413.blogspot.com/2006/08/trendnet-low-cost-but-solid-designs.html -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From damien at mc-kenna.com Sun Aug 13 22:24:43 2006 From: damien at mc-kenna.com (Damien McKenna) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] Low-cost, but solid: TRENDnet TEW-610APB (AP), TEW-441PC (CardBus), TEW-443PI (PCI) In-Reply-To: <1155520771.3223.1.camel@bert64.oviedo.smithconcepts.com> References: <1155520771.3223.1.camel@bert64.oviedo.smithconcepts.com> Message-ID: <44DFDEEB.7050209@mc-kenna.com> Bryan J. Smith wrote: > Wrestling with wireless for the last month, sometimes cheaper is not so > bad -- as long as you find the right products: As soon as we get a chance I'm going to pick one up for general use at home, I'm sick of being stuck with WEP 'cause my card sucketh verily. One question, is WPA or 11g in any way CPU-bound, or does all the hooplah actually happen in hardware? I've just noticed that if our existing Linksys 11b WEP-only card in a P3/500 laptop running Windows2000 looses the connection the CPU gets eaten up quite thoroughly until it finds it again - at that point I tend to pop out the card rather than deal with a machine running at half speed. Thanks for the tip. -- Damien McKenna, husband, father, geek. damien@mc-kenna.com - http://www.mc-kenna.com/ From jasonb at edseek.com Sun Aug 13 23:05:57 2006 From: jasonb at edseek.com (Jason Boxman) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] Low-cost, but solid: TRENDnet TEW-610APB (AP), TEW-441PC (CardBus), TEW-443PI (PCI) In-Reply-To: <1155520771.3223.1.camel@bert64.oviedo.smithconcepts.com> References: <1155520771.3223.1.camel@bert64.oviedo.smithconcepts.com> Message-ID: <200608132305.57128.jasonb@edseek.com> On Sunday 13 August 2006 21:59, Bryan J. Smith wrote: > Wrestling with wireless for the last month, sometimes cheaper is not so > bad -- as long as you find the right products: > http://thebs413.blogspot.com/2006/08/trendnet-low-cost-but-solid-designs.ht >ml "In some stupid, random decision when I went to buy a UPS from CompUSA, I decided to try out a sub-$15 CompUSA 54g cards on clearance." And what did we learn from this exercise? Say it with me now: CompUSA sucks. -- Jason Boxman http://edseek.com/ - Linux and FOSS stuff From dmckenna at thelimucompany.com Tue Aug 15 16:24:36 2006 From: dmckenna at thelimucompany.com (Damien McKenna) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] New PCI-E 24 port (!) SATA cards Message-ID: <1BC454F5-B2E1-441D-A4D8-32876C864E1B@thelimucompany.com> http://www.dailytech.com/article.aspx?newsid=3787 Areca have just announced some new PCI-E SATA cards going up to 24 ports! Yowzers! -- Damien McKenna - Web Developer - dmckenna@thelimucompany.com The Limu Company - http://www.thelimucompany.com/ - 407-804-1014 #include -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.matrixlist.com/pipermail/pc_support/attachments/20060815/c9192ebb/attachment.html From dmckenna at thelimucompany.com Tue Aug 15 17:09:48 2006 From: dmckenna at thelimucompany.com (Damien McKenna) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] New Nvidia Opteron chipset Message-ID: <8D040EE6-A3C4-4317-B6EE-5A1A688F0AB2@thelimucompany.com> http://www.dailytech.com/article.aspx?newsid=3803 New nForce Professional 3000 series to match the new Socket F Opteron. Looks like some good stuff there. -- Damien McKenna - Web Developer - dmckenna@thelimucompany.com The Limu Company - http://www.thelimucompany.com/ - 407-804-1014 #include From b.j.smith at ieee.org Wed Aug 16 04:16:23 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] offtopic / database schema -- COSINE/X.500 and modern LDAP schema ... Message-ID: <1155716183.2987.70.camel@bert64.oviedo.smithconcepts.com> First off, if you want a "short list," then IETF RFC2218 is a dated, but "good start" as it defines "dumb strings" for various, recommended records based on existing X.500/LDAP Pilot schema of the time: http://www.apps.ietf.org/rfc/rfc2218.html --General Attributes -- Field Name Size Syntax Email 360 Mailbox Cert 4000 Certificate Home Page 128 URI Common Name 64 WhitepageString Given Name 48 WhitepageString Surname 48 WhitepageString Organization 64 WhitepageString Locality 20 WhitepageString Country 2 WhitepageString (ISO 3166) Language Spoken 128 WhitepageString (RFC 1766) --Personal Attributes Personal Phone 30 PrintableString Personal Fax 30 PrintableString Personal Mobile Phone 30 PrintableString Personal Pager Number 30 PrintableString Personal Postal Address 255 Address Description 255 WhitepageString --Organizational Attributes Title 64 WhitepageString Office Phone 30 PrintableString Office Fax 30 PrintableString Office Mobile Phone 30 PrintableString Office Pager 30 PrintableString Office Postal Address 255 Address --Ancillary Creation Date 24 GeneralizedTime Creator Name 255 URI Modified Date 24 GeneralizedTime Modifier Name 255 URI I don't know how well it maps to modern LDAP but, and it definitely is _not_ an efficient, binary record solution, but it's a "good start." In the early days of X.500 directory implementations/adoption, COSINE was a pilot set of schema for various, basic tables and logic. This was before LDAP was postured and the Pilot schema is covered in IETF RFC1274: http://www.apps.ietf.org/rfc/rfc1274.html The more "modern" implementation used by most LDAP solutions today comes from IETF RFC2256: http://www.apps.ietf.org/rfc/rfc2256.html And a focus on the schema for an "Internet Person" in IETF RFC2798: http://www.apps.ietf.org/rfc/rfc2798.html These IETF RFCs are fairly "tough to chew" with a lot of coverage of various attributes, logic, hierarchy, etc... right down to the explicit OID references. But RFC they are heavily influenced by the best enterprise LDAP solution by the late '90s (yes, _before_ ADS ;-), Netscape Directory Server. The Sun One documentation (based on Netscape Directory Server) does a good job of introducing the basic records you want on the top half of this page: http://docs.sun.com/source/816-6699-10/schemaov.html If you really want to dive into modern Netscape Directory Server, Red Hat's documentation on Red Hat Directory Server 7.1 (which _is_ Netscape Directory Server) has the full schema here: http://www.redhat.com/docs/manuals/dir-server/schema/7.1/schemaTOC.html How far you want to go is up to you. I think 2218 gives you a good list to start. If you want more records, then follow RFC2256/RFC2798, referencing the RHDS 7.1 docs for better explanations. -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From b.j.smith at ieee.org Wed Aug 16 04:25:56 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] offtopic / database schema -- COSINE/X.500 and modern LDAP schema ... In-Reply-To: <1155716183.2987.70.camel@bert64.oviedo.smithconcepts.com> References: <1155716183.2987.70.camel@bert64.oviedo.smithconcepts.com> Message-ID: <1155716756.2987.75.camel@bert64.oviedo.smithconcepts.com> On Wed, 2006-08-16 at 04:16 -0400, Bryan J. Smith wrote: > The more "modern" implementation used by most LDAP solutions today comes > from IETF RFC2256: > http://www.apps.ietf.org/rfc/rfc2256.html > And a focus on the schema for an "Internet Person" in IETF RFC2798: > http://www.apps.ietf.org/rfc/rfc2798.html BTW, what code are you going to be setting up tables with? Accessing via? Etc...??? There's probably already code to handle the RFC2798 schema for LDAP, so you just need to use it for your database as well. BTW, the YoLinux docs have an excellent table of all the attributes in RFC2798 half-way down in its LDAP setup tutorial (yes, I found this page via Google'ing, I'll readily admit -- but my prior response was _not_ found by just Googling -- other than the RFC2218 IWPS, didn't know about that before) ... http://yolinux.com/TUTORIALS/LinuxTutorialLDAP.html inetOrgPerson object attributes: Requires: objectClass: organizationalPerson objectClass: person (Inherited from object organizationalPerson) objectClass: top (Inherited from object person) sn (Surename/Last Name - Inherited from object person) cn (Common Name - Inherited from object person) May have: o (Organization Name) displayName (RFC2798: Preferred name of a person to be used when displaying entries) audio businessCategory carLicense departmentNumber employeeNumber employeeType (i.e. "Contractor", "Employee", "Intern", "Temp", "External", "Unknown", etc...) givenName homePhone homePostalAddress (After street number and name use line separator "$" in LDIF file: street$ st postalCode) initials (MS/Outlook considers this to be the middle name) jpegPhoto (See the OpenLDAP FAQ: Turn a jpeg into ldif format) labeledURI mail (e-Mail address) manager (Specify dn entry of manager) mobile pager photo roomNumber secretary (Specify dn entry of secretary) uid userCertificate x500uniqueIdentifier preferredLanguage userSMIMECertificate (RFC2633: A PKCS#7 [RFC2315] SignedData) userPKCS12 (PKCS #12 [PKCS12] provides a format for exchange of personal identity information.) Attributes inherited from object organizationalPerson: ou (Organization unit) title x121Address registeredAddress destinationIndicator preferredDeliveryMethod telexNumber teletexTerminalIdentifier telephoneNumber (MS/Outlook considers this to be the "Business Phone") internationaliSDNNumber facsimileTelephoneNumber postOfficeBox postalAddress (MS/Outlook and Netscape both use this for the business address.) physicalDeliveryOfficeName (MS/Outlook considers this to be the field "Office") street (Don't use "street" because Netscape can't use it. Use "postalAddress".) l (Locality/City/Town) st (State/Province) postalCode (Zip code) Attributes inherited from object person: userPassword telephoneNumber (work phone) seeAlso (URL for more info) description -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From jasonb at edseek.com Wed Aug 16 15:56:49 2006 From: jasonb at edseek.com (Jason Boxman) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] offtopic / database schema -- COSINE/X.500 and modern LDAP schema ... In-Reply-To: <1155716756.2987.75.camel@bert64.oviedo.smithconcepts.com> References: <1155716183.2987.70.camel@bert64.oviedo.smithconcepts.com> <1155716756.2987.75.camel@bert64.oviedo.smithconcepts.com> Message-ID: <34892.216.134.200.78.1155758209.squirrel@nebula.internal.foo> Bryan J. Smith wrote: > On Wed, 2006-08-16 at 04:16 -0400, Bryan J. Smith wrote: >> The more "modern" implementation used by most LDAP solutions today comes >> from IETF RFC2256: >> http://www.apps.ietf.org/rfc/rfc2256.html >> And a focus on the schema for an "Internet Person" in IETF RFC2798: >> http://www.apps.ietf.org/rfc/rfc2798.html > > BTW, what code are you going to be setting up tables with? Accessing > via? Etc...??? "For the purposes of this question, I'm not especially interested in normalization related issues. Whether I keep the physical address information in the same table as a user's name or it's referenced using a foreign key shouldn't have much impact on the appropriate length of the fields. (Except it does, but I'd like to ignore data size issues as related to the amount of storage consumed.) "My task ought to be fairly commom. Unique invidivuals with vastly unique addresses, phone numbers, and other contact information will create accounts. These people may exchange currency in the form of U.S. dollars for service. I may need to mail each customer information in return. I need columns robust enough that I can accurately and confidently store information necessary to mail these customers, in the event that is necessary, on a per user basis. "I had hoped there would be some basic guideslines for the above web-store-style address collection task. "Moreover, I'd like to be able to robustly store several contact phone numbers including possible international numbers. "I'd wing it, but this isn't a situation where I feel comfortable taking a best guess as to what field lengths should be -- especially so when this has been dozen 1,000s of times before, I would think. "Certainly, the degree of normalization will vary depending on my specific application, but I think I can handle that part. It's column lengths and related discussion about properly parsing addressing, phone, and first name, last name, and so forth I seek. (Is there some best practice for properly capitalizing names? Do you just store it all UPPER'd or lower'd?)" The backend is the PostgreSQL RDBMS. I'll be accessing it via Ruby and the Rails framework. The output target will be, obviously, a Web browser. Mostly I just didn't want to badly choose column sizes and get bitten by it later. From thebs413 at gmail.com Thu Aug 17 19:34:53 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] HP 50G Calculator for $112.99 shipped ... Message-ID: For those of you that still may be in high school/college and want a solid calculator, the HP 50G just came out. It's $149.99 list, Amazon.COM has it for $129.99, but Buy.COM has it cheapest -- $117.99 - $5 for new customers, shipped free: http://dealnews.com/deals/HP-50-G-Graphing-Calculator-from-113-shipped/129250.html I've been waiting for this thing to come out. I got through college on a $100 HP 48G that is now 15 years-old and still works. I bought a HP 49G+ in 2004 and it was a was a piece of junk, breaking just after the 1 year warranty period expired. From my understanding, HP got out of the calculator market circa 2002-2003 and outsourced some of the HP 49 series final design. Other than RPN and some basic, pre-calc stuff, they didn't include a lot of the 48G's capabilities. I heard the HP 50G is a return to the 48 series style with hardened plastic molding. A number of engineering/surveying solutions are being built around it, like the 48 (but not the 49). It has full Saturn emulation -- the 8-bit, 2x4-bit nibble/BCD processor of the 48 series -- so it includes the full 48 series software ROM and runs all of its goodies. It's a 206MHz ARM underclocked to 75MHz to improve battery life, and can be targetted by GCC. HP added another battery to give it a 6V input, and many other improvements over the 49 series -- or as most say, what the 49 "should have been." I'll find out soon enough as I've ordered one. I'm glad that HP decided to return to calculators. I really don't like anything Casio, TI or anyone else offers, and while Galculator is "tolerable," virtually all other computer-based calculators are not IMHO. And yes, I know about the HP48 emulator -- not the same as one in your hand. ;) Wikipedia articles ... HP 48 series: http://en.wikipedia.org/wiki/HP-48 HP 50 (part of 49 page): http://en.wikipedia.org/wiki/HP-49#HP_50G HP Calc site: http://www.hpcalc.org/ From damien at mc-kenna.com Thu Aug 17 21:53:37 2006 From: damien at mc-kenna.com (Damien McKenna) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] HP 50G Calculator for $112.99 shipped ... In-Reply-To: References: Message-ID: <44E51DA1.5080303@mc-kenna.com> Bryan J. Smith wrote: > For those of you that still may be in high school/college and want a > solid calculator, the HP 50G just came out. So why buy one of these versus the TI jobbies that are dime a dozen used? -- Damien McKenna, husband, father, geek. damien@mc-kenna.com - http://www.mc-kenna.com/ From paulf at quillandmouse.com Fri Aug 18 00:20:10 2006 From: paulf at quillandmouse.com (Paul M Foster) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] HP 50G Calculator for $112.99 shipped ... In-Reply-To: <44E51DA1.5080303@mc-kenna.com> References: <44E51DA1.5080303@mc-kenna.com> Message-ID: <44E53FFA.1040703@quillandmouse.com> Damien McKenna wrote: > Bryan J. Smith wrote: >> For those of you that still may be in high school/college and want a >> solid calculator, the HP 50G just came out. > > So why buy one of these versus the TI jobbies that are dime a dozen used? > Reverse Polish Notation, the *only* way a calculator should be used. -- Paul M. Foster From thebs413 at gmail.com Fri Aug 18 01:25:51 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] Re: HP 50G Calculator for $112.99 shipped ... and why are used TIs cheap and HPs not? Message-ID: Damien McKenna wrote: > So why buy one of these versus the TI jobbies that are dime a dozen used? Let's reverse that question ... "Why are the 10 year-old, HP 48 series still going for $100 USED and the TI jobbies are dime-a-dozen USED? Hmmm?" Other than the HP 39/49G/49G+ (which sucked, hard, long story, glad HP reversed their decision to exit the market), HP's scientific/engineering calculators are _rarely_ resold because their users love them. They last nearly forever (those 4MHz Saturn processors went 3+ years on 3xAAA batteries) and my 48G is now 15 years-old (even well dented from the usage ;-). Paul M Foster wrote: > Reverse Polish Notation, the *only* way a calculator should be used. While I personally agree with you on postfix (aka RPN) entry for myself, the 49/50 series offers both infix (aka common arithmetic/algebraic) and CAS (computer algebraic system) entry as well. On the older 48 series, you use to have to use a leading single quote to enter an infix line -- some people liked that, although most people did not. Now regarding my use of infix, it only took me 1 physics assignment to learn postfix entry. After that, I never wanted to use infix ever again. No "store" or "write down" so-called "temporary values" to re-use later. With postfix, you _never_ have to store anything -- it all goes on the stack. That's why I use it. And it's why engineers learn it, because they get their work done in half (or less) of the time using an infix calculator -- even those that offer CAS or "parenthesis" for order-of-operations. You literally and quickly learn to do order-of-operations in postfix directly, and people are amazed how quick you can punch out answers with such a RPN calculator. From thebs413 at gmail.com Fri Aug 18 02:03:00 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] Realtek 8169 -- sound typical for CIFS/SMB on RTL81x9 series at 100Mbps Message-ID: tonyb wrote: > Is anyone using these successfully between Linux and XP? I've avoided the RTL8169 in general. Although there are options to provide it 8KiB TX + 64KiB RX SRAM, most of the on-board ones I've seen only offer 2KiB TX + 2KiB RX on ASIC/MAC. I believe it is not capable of 9000 byte Jumbo frames in _any_ configuration. At Gigabit Ethernet (GbE), don't expect much beyond 240Mbps (30MBps) actual. > I have two with a cross-wired cat6 cable. Cross-wired how? Understand that 1000Base-T is 4-pair (8-conductor) and does _not_ work with standard 568A/B cross-over wiring. The 4, 5, 7 and 8 conductors must be crossed as well. If they are not, then you get 100Mbps Fast Ethernet (FE), _not_ 1000Mbps GbE. A good page on this is here: http://logout.sh/computers/net/gigabit/ AutoMDI-X GbE NICs are _uncommon_. I.e., they won't use straight-through cables. > When I'm running Linux on both boxes I get 10+MB throughput. Which is typically for 100Mbps FE. > Linux on one and XP on the other and throughput is less than 2MB. What protocol? Microsoft CIFS/SMB is TCP and has a massive amount of overhead. I typically can't break 3-4MBps over 100Mbps Fast Ethernet. On those crappy RTL8169, 2MBps doesn't sound out-of-the-ordinary. > I know...I know... Get rid of XP. Just as soon as I solve the video capture > problem from the previous post. Most people have been going with the Happauge WinTV-PVR-150, 250 or 350 products that start at sub-$100 with _hardware_ MPEG-2 capture supported with the ivtv driver. A few versions of the units are not ivtv compatible -- e.g., some of the select Media Center Editions (MCE) -- so check the _exact_ model number before you buy. From thebs413 at gmail.com Fri Aug 18 02:16:23 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] Re: Realtek 8169 -- sound typical for CIFS/SMB on RTL81x9 series at 100Mbps In-Reply-To: References: Message-ID: Chris wrote: > Aren't these supposed to be GigaBit cards? 2MB does seem a > bit shy of scorching, doesn't it... I'll bet you he's getting 100Mbps (see my comments below). See my previous post to PC_Support on why ... (as well as other reasons) ... http://lists.leap-cf.org/pipermail/pc_support/2006-August/002382.html > MTU issues? (I think these support "jumbo" buffers?) I did some more digging. It looks like that most RTL8169 ICs do have the 8KiB TX + 64KiB RX SRAM standard. But the 8KiB TX only allows 7422 byte frames. That's hardly "standard." But at least it _does_ seem to have 64KiB RX on-ASIC/MAC. Again, I'm trying to verify that all RTL8169 products have that though. I want to say some (earlier?) RTL8169 GbE ASIC/MACs only had a 2KiB RX + 2KiB TX SRAM. But maybe I'm thinking of another vendor? > On the dedicated Linux box, does ifconfig look any different when > the other box is booted into XP? Duplex mode - Any TX/RX errors? Agreed! Hit "netstat -i eth0" and look at the statistics for the interface. NT 5.x (XP/200x) has an equivalent command (with slightly different options). With 64KiB RX SRAM, you shouldn't be quite that low, more like 3-4MBps for CIFS/SMB (again, what's your protocol/usage?). Although the RTL8169 driver under Windows might be crap as well. From tim at mcdonough.net Fri Aug 18 08:18:42 2006 From: tim at mcdonough.net (Tim McDonough) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] Re: HP 50G Calculator for $112.99 shipped ... and why are used TIs cheap and HPs not? In-Reply-To: References: Message-ID: <44E5B022.4000108@mcdonough.net> Bryan J. Smith wrote: > Let's reverse that question ... > "Why are the 10 year-old, HP 48 series still going for $100 USED > and the TI jobbies are dime-a-dozen USED? Hmmm?" > > Other than the HP 39/49G/49G+ (which sucked, hard, long story, glad HP > reversed their decision to exit the market), HP's > scientific/engineering calculators are _rarely_ resold because their > users love them. They last nearly forever (those 4MHz Saturn > processors went 3+ years on 3xAAA batteries) and my 48G is now 15 > years-old (even well dented from the usage ;-). > > Paul M Foster wrote: >> Reverse Polish Notation, the *only* way a calculator should be used. I love my HP 32S II calculator. It's still on it's original batteries even with daily use. I was very concerned a few years ago -- it was on a workbench and I accidentally knocked a spool of solder off a shelf. It put a nasty dent just below the "32S II" and the LCD but thankfully did no internal damage. I got started with HP calculators and RPN back around 1976. It seemed like only about 1/2 the students at the local college had calculators and it was a constant "can I borrow yours" situation in the library or cafeteria. One of my professors showed me how to use his RPN model and I initially bought it because very few people would borrow it since they couldn't operate it. It took me about a day to fall in love with the operation and I've owned one ever since. Tim From paulf at quillandmouse.com Fri Aug 18 13:29:31 2006 From: paulf at quillandmouse.com (Paul M Foster) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] Re: HP 50G Calculator for $112.99 shipped ... and why are used TIs cheap and HPs not? In-Reply-To: <44E5B022.4000108@mcdonough.net> References: <44E5B022.4000108@mcdonough.net> Message-ID: <44E5F8FB.8060502@quillandmouse.com> Tim McDonough wrote: >> Paul M Foster wrote: >>> Reverse Polish Notation, the *only* way a calculator should be used. > > I love my HP 32S II calculator. It's still on it's original batteries > even with daily use. I was very concerned a few years ago -- it was on > a workbench and I accidentally knocked a spool of solder off a shelf. > It put a nasty dent just below the "32S II" and the LCD but thankfully > did no internal damage. > > I got started with HP calculators and RPN back around 1976. It seemed > like only about 1/2 the students at the local college had calculators > and it was a constant "can I borrow yours" situation in the library or > cafeteria. One of my professors showed me how to use his RPN model and I > initially bought it because very few people would borrow it since they > couldn't operate it. It took me about a day to fall in love with the > operation and I've owned one ever since. > Heh. I graduated high school in '75, and asked for an HP calculator for my graduation present (remember, back then calculators were rare and expensive). That calculator (HP25?) is gone, but years ago I bought an HP11c. People since have borrowed it and been completely unable to figure it out. Gives me a warm feeling. ;-} -- Paul M. Foster From dmckenna at thelimucompany.com Fri Aug 18 15:52:24 2006 From: dmckenna at thelimucompany.com (Damien McKenna) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] Re: HP 50G Calculator for $112.99 shipped ... and why are used TIs cheap and HPs not? In-Reply-To: <44E5F8FB.8060502@quillandmouse.com> References: <44E5B022.4000108@mcdonough.net> <44E5F8FB.8060502@quillandmouse.com> Message-ID: On Aug 18, 2006, at 1:29 PM, Paul M Foster wrote: > Heh. I graduated high school in '75, I was born in '75. Urk. > People since have borrowed it and been completely unable to figure > it out. Gives me a warm feeling. ;-} That's what I say about my code... -- Damien McKenna - Web Developer - dmckenna@thelimucompany.com The Limu Company - http://www.thelimucompany.com/ - 407-804-1014 #include -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.matrixlist.com/pipermail/pc_support/attachments/20060818/a8c9ad78/attachment.html From work at sprynet.com Fri Aug 18 22:28:50 2006 From: work at sprynet.com (John Hayden) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] EIKI LC-7000 In-Reply-To: Message-ID: Skipped content of type multipart/alternative-------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/x-pkcs7-signature Size: 4483 bytes Desc: not available Url : http://lists.matrixlist.com/pipermail/pc_support/attachments/20060818/59969fb2/smime.bin From b.j.smith at ieee.org Sat Aug 19 00:06:15 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] EIKI LC-7000 In-Reply-To: References: Message-ID: <1155960375.3052.3.camel@bert64.oviedo.smithconcepts.com> On Fri, 2006-08-18 at 22:28 -0400, John Hayden wrote: > Hello all, > I have an EIKI LC-7000 projector. I can not seem to get > the sync rates correct. Does anyone know what I can do to ?tweak? this > so that I can run a power point from a windows XP-Pro Laptop? Or > better yet have an open source program that will run MS-PowerPoint? I've been running StarImpress since the mid-'90s instead of PowerPower. It's in the OpenOffice.org suite as well as StarOffice, although the converters/fonts in OpenOffice.org might not be quite as complete. > -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From paulf at quillandmouse.com Sat Aug 19 17:43:57 2006 From: paulf at quillandmouse.com (Paul M Foster) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] Heat Sink Framework broken Message-ID: <44E7861D.8050309@quillandmouse.com> I've got a P4 box, my main box, where the frame for the CPU heat sink broke off, and I'm wondering if it's possible to get replacements for that. If you have a P4, you know it's about the size of a 386 CPU, but on top of it sits this massive heatsink/fan block. At least on mine, there are two steel tension clips that clip to a plastic framework and hold the heatsink/fan down onto the CPU. That plastic framework is about an inch to an inch-and-a-half tall, with four plastic "posts" which the spring clips can hook to. Well, the end of one of the posts broke off, and the heatsink/fan came off. Right now, I've got a string of four zip ties holding the fan down on that side. But I'm wondering if anyone's ever heard of this problem, and if a new plastic frame gizmo can be had to replace the old one. Anyone know? -- Paul M. Foster From carter at carter.cc Sat Aug 19 20:29:45 2006 From: carter at carter.cc (Carter Manucy) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] Heat Sink Framework broken In-Reply-To: <44E7861D.8050309@quillandmouse.com> References: <44E7861D.8050309@quillandmouse.com> Message-ID: <44E7ACF9.2050709@carter.cc> Paul M Foster wrote: > I've got a P4 box, my main box, where the frame for the CPU heat sink > broke off, and I'm wondering if it's possible to get replacements for > that. If you have a P4, you know it's about the size of a 386 CPU, but > on top of it sits this massive heatsink/fan block. At least on mine, > there are two steel tension clips that clip to a plastic framework and > hold the heatsink/fan down onto the CPU. That plastic framework is > about an inch to an inch-and-a-half tall, with four plastic "posts" > which the spring clips can hook to. Well, the end of one of the posts > broke off, and the heatsink/fan came off. > > Right now, I've got a string of four zip ties holding the fan down on > that side. But I'm wondering if anyone's ever heard of this problem, > and if a new plastic frame gizmo can be had to replace the old one. > Anyone know? > If you're talking about the ZIF socket on the motherboard, you've got one of four choices so far as I can see it. One is to use your zip ties. Some motherboards have holes in the board right around the socket area - I've had this same thing happen to me in the past, and found a well-placed, good quality set of zip ties kept enough pressure on the socket to keep the CPU cool. Second is to get some heat sink 'glue'. This is usually not a good choice, as it'll bond the heat sink to the CPU, and make it near impossible to get apart without damaging something. Third is to get yourself a new motherboard :) And lastly, if you have the four round holes surrounding the socket, you can probably get another heat sink type that will 'wrap' the motherboard, and thus bypass the white plastic tabs that have broken off. - Carter From b.j.smith at ieee.org Sun Aug 20 13:08:23 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] Web hosting provider with credit card ordering ... Message-ID: <1156093703.2989.16.camel@bert64.oviedo.smithconcepts.com> My father is a land surveyor and is looking to setup a simple commerce site where he can take orders for surveys and other jobs (pretty much all local to Florida). His content would be rather simple**, other than credit card ordering. I used Pair Networks a long time ago but it's been awhile since I did a good glance at various providers. A list of several options you've personally used would be nice. -- Bryan **NOTE: For providing client-specific content (like completed surveys in PDF format), I'm planning on setting up a local SFTP server where his clients can login and it can be controlled. That content would not go on the web hosting provider's server. In fact, I'm likely going to push him towards getting both cable and DSL and using one for his LAN connectivity and the other, completely separate, for the DMZ with this SFTP server. P.S. Being that I'm in New England, I can't support my father much other than setup on the weekends. As such, I'm also looking for some consultants who can assist my father with IT issues when he has them. I think his current IT guy doesn't know much about security. -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From lists at brianrose.net Sun Aug 20 13:26:54 2006 From: lists at brianrose.net (Brian Rose) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] Web hosting provider with credit card ordering ... In-Reply-To: <1156093703.2989.16.camel@bert64.oviedo.smithconcepts.com> References: <1156093703.2989.16.camel@bert64.oviedo.smithconcepts.com> Message-ID: <44E89B5E.1090506@brianrose.net> Bryan J. Smith wrote: > My father is a land surveyor and is looking to setup a simple commerce > site where he can take orders for surveys and other jobs (pretty much > all local to Florida). His content would be rather simple**, other than > credit card ordering. I used Pair Networks a long time ago but it's > been awhile since I did a good glance at various providers. > > A list of several options you've personally used would be nice. > I use Hurricane Electric (http://www.he.net) and they have been very reliable. You can get either a Linux or Windows virtual server. IIRC, their plans start at $9.95/mo and go up from there. From b.j.smith at ieee.org Sun Aug 20 13:38:21 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] Web hosting provider with credit card ordering ... In-Reply-To: <44E89B5E.1090506@brianrose.net> References: <1156093703.2989.16.camel@bert64.oviedo.smithconcepts.com> <44E89B5E.1090506@brianrose.net> Message-ID: <1156095501.2989.41.camel@bert64.oviedo.smithconcepts.com> On Sun, 2006-08-20 at 13:26 -0400, Brian Rose wrote: > I use Hurricane Electric (http://www.he.net) and they have been very > reliable. You can get either a Linux or Windows virtual server. IIRC, > their plans start at $9.95/mo and go up from there. Do they offer a "canned" e-commerce site? Understand my father just needs basic info/contact content and then a "simple store" for ordering. No customization here -- 100% static content and a "simple store." -- Bryan P.S. Again, all other client content that is login-based with be a local server. -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From carter at carter.cc Sun Aug 20 15:44:27 2006 From: carter at carter.cc (Carter Manucy) Date: Tue Oct 31 13:17:59 2006 Subject: [Pc_Support] Web hosting provider with credit card ordering ... In-Reply-To: <1156093703.2989.16.camel@bert64.oviedo.smithconcepts.com> References: <1156093703.2989.16.camel@bert64.oviedo.smithconcepts.com> Message-ID: <44E8BB9B.1030307@carter.cc> Bryan J. Smith wrote: > My father is a land surveyor and is looking to setup a simple commerce > site where he can take orders for surveys and other jobs (pretty much > all local to Florida). His content would be rather simple**, other than > credit card ordering. I used Pair Networks a long time ago but it's > been awhile since I did a good glance at various providers. > > A list of several options you've personally used would be nice. > > Have you looked at GoDaddy's e-commerce solutions? Reasonably priced, secure... great fits for most folks looking for a simple solution. I haven't used their services (so far as e-commerce goes), but I have been using other services of their for years... virtual servers, domains, etc. From damien at mc-kenna.com Mon Aug 21 07:05:21 2006 From: damien at mc-kenna.com (Damien McKenna) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Web hosting provider with credit card ordering ... In-Reply-To: <1156093703.2989.16.camel@bert64.oviedo.smithconcepts.com> References: <1156093703.2989.16.camel@bert64.oviedo.smithconcepts.com> Message-ID: <44E99371.6090906@mc-kenna.com> Bryan J. Smith wrote: > A list of several options you've personally used would be nice. I'm with Dreamhost.com. Their control panel has one-click installers for many apps including ZenCart (http://www.zencart.com/), which is taking over from oscommerce as the OSS shopping cart of choice. The basic site starts at $10/mo, a static IP address (for SSL) is another $4, then your certificate itself and credit card gateway will add extra. It isn't strictly a canned solution, but with a little bit of setup you'd be flying. -- Damien McKenna, husband, father, geek. damien@mc-kenna.com - http://www.mc-kenna.com/ From b.j.smith at ieee.org Mon Aug 21 10:40:44 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Web hosting provider with credit card ordering ... In-Reply-To: <44E99371.6090906@mc-kenna.com> References: <1156093703.2989.16.camel@bert64.oviedo.smithconcepts.com> <44E99371.6090906@mc-kenna.com> Message-ID: <1156171244.2989.88.camel@bert64.oviedo.smithconcepts.com> On Mon, 2006-08-21 at 07:05 -0400, Damien McKenna wrote: > It isn't strictly a canned solution, but with a little bit of > setup you'd be flying. How much setup? Any code? Or is it all "point'n click"? -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From damien at mc-kenna.com Mon Aug 21 11:02:45 2006 From: damien at mc-kenna.com (Damien McKenna) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Web hosting provider with credit card ordering ... In-Reply-To: <1156171244.2989.88.camel@bert64.oviedo.smithconcepts.com> References: <1156093703.2989.16.camel@bert64.oviedo.smithconcepts.com> <44E99371.6090906@mc-kenna.com> <1156171244.2989.88.camel@bert64.oviedo.smithconcepts.com> Message-ID: On Aug 21, 2006, at 10:40 AM, Bryan J. Smith wrote: > On Mon, 2006-08-21 at 07:05 -0400, Damien McKenna wrote: >> It isn't strictly a canned solution, but with a little bit of >> setup you'd be flying. > > How much setup? Any code? Or is it all "point'n click"? With Dreamhost's one click installer its mostly one click to get the software copied in place then some final tweaking to get it fully running. After that its just point 'n click to get the store set the way its wanted and items added. Then there's the SSL cert, which is a little tricky but not too bad really. Does he have a payment gateway account yet, e.g. Authorize.net or Paypal? That's more clicks. All told, it can be done with no code, but it might get a bit overwhelming for someone new to web hosting. -- Damien McKenna, husband, father, geek. damien@mc-kenna.com - http://www.mc-kenna.com/ From jasonb at edseek.com Mon Aug 21 12:12:12 2006 From: jasonb at edseek.com (Jason Boxman) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Web hosting provider with credit card ordering ... In-Reply-To: References: <1156093703.2989.16.camel@bert64.oviedo.smithconcepts.com> <1156171244.2989.88.camel@bert64.oviedo.smithconcepts.com> Message-ID: <200608211212.12826.jasonb@edseek.com> On Monday 21 August 2006 11:02, Damien McKenna wrote: > On Aug 21, 2006, at 10:40 AM, Bryan J. Smith wrote: > > On Mon, 2006-08-21 at 07:05 -0400, Damien McKenna wrote: > >> It isn't strictly a canned solution, but with a little bit of > >> setup you'd be flying. > > > > How much setup? Any code? Or is it all "point'n click"? > > With Dreamhost's one click installer its mostly one click to get the > software copied in place then some final tweaking to get it fully > running. After that its just point 'n click to get the store set the > way its wanted and items added. Then there's the SSL cert, which is > a little tricky but not too bad really. Does he have a payment > gateway account yet, e.g. Authorize.net or Paypal? That's more clicks. I keep hearing back things about DH from RoR people that host there. Apparently their RoR setup is awful. No idea about anything else as I haven't personally used it. -- Jason Boxman http://edseek.com/ - Linux and FOSS stuff From dmckenna at thelimucompany.com Mon Aug 21 13:21:36 2006 From: dmckenna at thelimucompany.com (Damien McKenna) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Web hosting provider with credit card ordering ... In-Reply-To: <200608211212.12826.jasonb@edseek.com> References: <1156093703.2989.16.camel@bert64.oviedo.smithconcepts.com> <1156171244.2989.88.camel@bert64.oviedo.smithconcepts.com> <200608211212.12826.jasonb@edseek.com> Message-ID: <94092DA0-BCB2-4733-8D70-A8CCAB727F5B@thelimucompany.com> On Aug 21, 2006, at 12:12 PM, Jason Boxman wrote: > I keep hearing back things about DH from RoR people that host there. > Apparently their RoR setup is awful. I can't comment on RoR as I've yet to really get into it. I will say, though, that RoR itself does not look like the best designed system for server integration, there are so many variances between web server plugins, which seem to change every few months, that from a web hosting point of view I could see it being a PITA. Point in case, currently the best / only reliable way of getting it working on IIS is by setting up a proxy with a *second* http daemon running in the background - what's up with that? > No idea about anything else as I haven't personally used it. I've used them for a few months, and aside from the major power outage the other week in Los Angeles, they've been very good. My current apps are all PHP and have worked well. They have been very open about everything (I mentioned their blog and status site previously), and are doing their best to work through it each problem as it happens. My only current gripe is that they really need a better spam system - to use their current setup you have to log into *their* webmail interface to access the spam folder, which is just unworkable for most folks. They really need a good one-click-installer for the spam software onto your personal account to so you can control the configuration as wanted. -- Damien McKenna - Web Developer - dmckenna@thelimucompany.com The Limu Company - http://www.thelimucompany.com/ - 407-804-1014 #include -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.matrixlist.com/pipermail/pc_support/attachments/20060821/8f1f003a/attachment.html From b.j.smith at ieee.org Mon Aug 21 20:33:27 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] All my father needs (I wish I wasn't 1, 100 miles away) Message-ID: <1156206807.3017.49.camel@bert64.oviedo.smithconcepts.com> Here's the deal guys, my father doesn't need "pet projects." It's just: 1. Office: himself with maybe 1-2 others 2. Field: survey crews getting their tasks at start of day, and uploading their work at the end of the day (they never come to the office) 3. Clients: accessing their completed files (they never come to the office) 4. Web: so new clients can find him and place orders He has a domain (x). I see it as simple as follows ... A) www.x.com static web site with simple (5-10 item) ordering He doesn't need anything fancy. He's going to sell land surveys (by acreage/type) and options and that's it. Doesn't need "other people who bought Y also bought Z" type logic. Just a tie-in to Visa/MC for billing. The static content has already been created. I really want to find someone who has used a hosting provider before that has a "canned" solution who can set this up in 4-8 hours _max_ (including DNS). B) orlando.x.com remote crew access to his office His crews, unlike his web site, may need to access several MBs at the start of the day (from a total of several GBs) and then upload the same at the end of the day. I don't need document management or other non-sense right now -- just a simple way to access the files. E.g., SFTP using a GUI client like SSH.COM's, web-based file access in a "canned" appliance solution, whatever. I would like to start with a low-cost Linux-based all-in-one appliance. It would be the firewall, file server and remotely accessed file server. It could be a VPN or a hardened (I'd rather not use HTTPS, but SFTP instead) interface to the server. Yes, if I was running the company -- I'd put in IPCop, a dedicated file server, a DMZ server for the remote access, etc... But he's not an IT person and he doesn't have time. I've seen countless, sub-$500 Internet/LAN combinational appliances that do this. His LAN PCs would then be named (name).orlando.x.com and connect to the LAN port on it. C) clients.x.com client access to files After surveys are completed, clients will want a way to get access to their files (likely PDF format). I think we don't need anything but a canned appliance for this either. Again, since the files could easily total 100s of MBs, possibly 1+ GB, within months, this should also be on a local system. I suggested he get a 2nd Internet connection to the office (e.g., SDSL) for this if and when the number of clients hitting his server starts going up. With that said ... I'm at the point that the responses I have received leave me no choice but to do it myself. Again, no "pet projects" here. This stuff should be so simple that I could do it in a weekend (and I'm seriously considering spending the next 2 doing it). Please be considerate of what I'm asking. It's fine if you think I'm a jerk for responding this way, but this is really just a simple solution that would take me no time over 2 weekends -- I'm just 1,100 miles away right now. But I'm starting to look at various hosting providers (his registrar is GoDaddy.COM) as well as little, sub-$500 appliances. The $250 Itian he's looking at doesn't look half-bad, but there are virtually no reviews on it so I don't know about its security. -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From b.j.smith at ieee.org Mon Aug 21 21:32:23 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Small business e-commerce sites ... Message-ID: <1156210343.3017.90.camel@bert64.oviedo.smithconcepts.com> A little Google'ing turned up E-Commerce Guide: http://www.ecommerce-guide.com/ Although it's a year old, they had a 2-part article on "Choosing a Hosted E-Commerce Solution" and they review the "Big 4" (Interland, 1&1, GoDaddy and Hostway). It's more geared towards those who buy, sell and ship on-line, but it's still a good introduction: http://www.ecommerce-guide.com/solutions/building/article.php/3529356 http://www.ecommerce-guide.com/solutions/building/article.php/3529856 They also have a forum for small businesses: http://forums.smallbusinesscomputing.com/forumdisplay.php?s=&forumid=2 The reviewer seemed to like Interland (yes, I know it's MS): http://www.interland.com/website-design/make-a-website.aspx Interland's cost are low for entry-level, even only $19.95/month with PayPal support and _unlimited_ transfers with 5-10GB of storage. But if you want Visa/MC/AmEx payment (Versign's services), it starts at $49.95. But that was still cheaper than most when it came to Merchant Account functionality. -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From b.j.smith at ieee.org Tue Aug 22 22:52:25 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Synology DS-106 w/FTPS (explicit) access In-Reply-To: <44EBB549.5030707@mc-kenna.com> References: <1156174966.2989.121.camel@bert64.oviedo.smithconcepts.com> <1156175268.2989.128.camel@bert64.oviedo.smithconcepts.com> <3C24AB1F-22DB-4092-99F5-49B4396268D2@mc-kenna.com> <1156204319.3017.10.camel@bert64.oviedo.smithconcepts.com> <8CCAE2A4-B430-4EE0-BB76-13335B045B8A@mc-kenna.com> <1156289647.3022.7.camel@bert64.oviedo.smithconcepts.com> <44EBA7EB.6000102@mc-kenna.com> <44EBA938.90002@mc-kenna.com> <1156296745.3022.28.camel@bert64.oviedo.smithconcepts.com> <44EBB549.5030707@mc-kenna.com> Message-ID: <1156301545.3022.39.camel@bert64.oviedo.smithconcepts.com> On Tue, 2006-08-22 at 21:54 -0400, Damien McKenna wrote: > http://www.synology.com/enu/products/index.php Ding ding! We may have a winner! I've pretty much given up on finding an all-in-one SOHO Firewall/NAS/Remote Gateway. But thanx to Damien McKenna, I have found a solid NAS device with FTP-SSL/TLS** access: http://www.synology.com/enu/products/DS106/spec.php **FileZilla makes an excellent, easy-to-use, FTP-SSL/TLS client. They are available for $279 from NewEgg.COM: http://www.newegg.com/Product/Product.asp?Item=N82E16822108003 Note, the $50 cheaper DS-106e/j versions are _not_ the same, they do not offer FTPS support, only FTP. It also offers both USB and eSATA drive support, as well as the internal SATA. And it has a GbE. They only offer 100 shared folders and 128 user accounts, but that's fine to start IMHO. Ironically enough, not much more is offered in their $600, 4 drive versions (1,024 user accounts, but still only 100 shared folders). I think I'll just pair this with a cheap firewall (or maybe a dual WAN firewall) to start. -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From b.j.smith at ieee.org Tue Aug 22 23:04:49 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Re: Synology DS-106 w/FTPS (explicit) access -- has Mac support In-Reply-To: <1156301545.3022.39.camel@bert64.oviedo.smithconcepts.com> References: <1156174966.2989.121.camel@bert64.oviedo.smithconcepts.com> <1156175268.2989.128.camel@bert64.oviedo.smithconcepts.com> <3C24AB1F-22DB-4092-99F5-49B4396268D2@mc-kenna.com> <1156204319.3017.10.camel@bert64.oviedo.smithconcepts.com> <8CCAE2A4-B430-4EE0-BB76-13335B045B8A@mc-kenna.com> <1156289647.3022.7.camel@bert64.oviedo.smithconcepts.com> <44EBA7EB.6000102@mc-kenna.com> <44EBA938.90002@mc-kenna.com> <1156296745.3022.28.camel@bert64.oviedo.smithconcepts.com> <44EBB549.5030707@mc-kenna.com> <1156301545.3022.39.camel@bert64.oviedo.smithconcepts.com> Message-ID: <1156302289.3022.44.camel@bert64.oviedo.smithconcepts.com> For those of you Mac users, you might like the fact that it has AppleTalk and other Mac-centric client support too! Some other cool protocol support is also included. Tom's Hardware did a review of the DS-106e (cheaper, non-FTPS enabled, version) in June: http://www.tomsnetworking.com/2006/06/28/synology_ds106e_full_featured_nas/ On Tue, 2006-08-22 at 22:52 -0400, Bryan J. Smith wrote: > On Tue, 2006-08-22 at 21:54 -0400, Damien McKenna wrote: > > http://www.synology.com/enu/products/index.php > > Ding ding! We may have a winner! > > I've pretty much given up on finding an all-in-one SOHO > Firewall/NAS/Remote Gateway. But thanx to Damien McKenna, I have found > a solid NAS device with FTP-SSL/TLS** access: > http://www.synology.com/enu/products/DS106/spec.php > > **FileZilla makes an excellent, easy-to-use, FTP-SSL/TLS client. > > They are available for $279 from NewEgg.COM: > http://www.newegg.com/Product/Product.asp?Item=N82E16822108003 > > Note, the $50 cheaper DS-106e/j versions are _not_ the same, they do not > offer FTPS support, only FTP. It also offers both USB and eSATA drive > support, as well as the internal SATA. And it has a GbE. > > They only offer 100 shared folders and 128 user accounts, but that's > fine to start IMHO. Ironically enough, not much more is offered in > their $600, 4 drive versions (1,024 user accounts, but still only 100 > shared folders). > > I think I'll just pair this with a cheap firewall (or maybe a dual WAN > firewall) to start. > > -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From damien at mc-kenna.com Tue Aug 22 23:10:08 2006 From: damien at mc-kenna.com (Damien McKenna) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Re: Synology DS-106 w/FTPS (explicit) access In-Reply-To: <1156301545.3022.39.camel@bert64.oviedo.smithconcepts.com> References: <1156174966.2989.121.camel@bert64.oviedo.smithconcepts.com> <1156175268.2989.128.camel@bert64.oviedo.smithconcepts.com> <3C24AB1F-22DB-4092-99F5-49B4396268D2@mc-kenna.com> <1156204319.3017.10.camel@bert64.oviedo.smithconcepts.com> <8CCAE2A4-B430-4EE0-BB76-13335B045B8A@mc-kenna.com> <1156289647.3022.7.camel@bert64.oviedo.smithconcepts.com> <44EBA7EB.6000102@mc-kenna.com> <44EBA938.90002@mc-kenna.com> <1156296745.3022.28.camel@bert64.oviedo.smithconcepts.com> <44EBB549.5030707@mc-kenna.com> <1156301545.3022.39.camel@bert64.oviedo.smithconcepts.com> Message-ID: <44EBC710.8050003@mc-kenna.com> Bryan J. Smith wrote: > I've pretty much given up on finding an all-in-one SOHO > Firewall/NAS/Remote Gateway. But thanx to Damien McKenna, I have found > a solid NAS device with FTP-SSL/TLS** access: > http://www.synology.com/enu/products/DS106/spec.php > It sure looks pretty nifty. I'm also looking at the *Thecus N2050 series which are eSATA drives with built-in RAID 0 and 1, those would be pretty darn wonderful for easy & reliable external expansion on miniature systems (think: Mac Mini).* > **FileZilla makes an excellent, easy-to-use, FTP-SSL/TLS client. > Too bad they don't have plain-jane SSH & SFTP, would have been better IMHO. Ah well, maybe in a ROM upgrade.. > They are available for $279 from NewEgg.COM: > http://www.newegg.com/Product/Product.asp?Item=N82E16822108003 > A bit expensive though. > Note, the $50 cheaper DS-106e/j versions are _not_ the same, they do not > offer FTPS support, only FTP. It also offers both USB and eSATA drive > support, as well as the internal SATA. And it has a GbE. > Good to know, I noticed their site *really* needs a comparison table. > They only offer 100 shared folders and 128 user accounts, but that's > fine to start IMHO. Ironically enough, not much more is offered in > their $600, 4 drive versions (1,024 user accounts, but still only 100 > shared folders). > That would be 100 *separate* *shares*, obviously you could share one folder and add sub-directories. > For those of you Mac users, you might like the fact that it has > AppleTalk and other Mac-centric client support too! Some other cool > protocol support is also included. > Also... because it has MySQL and PHP you could install Gallery2 and use the Gallery2 exporter for iPhoto to put your photos in a network-shared album. I do wish, though, that they used Gallery2 by default rather than their own system, it would give it much more flexibility. -- Damien McKenna, husband, father, geek. damien@mc-kenna.com - http://www.mc-kenna.com/ From jasonb at edseek.com Tue Aug 22 23:13:54 2006 From: jasonb at edseek.com (Jason Boxman) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Re: Synology DS-106 w/FTPS (explicit) access In-Reply-To: <44EBC710.8050003@mc-kenna.com> References: <1156174966.2989.121.camel@bert64.oviedo.smithconcepts.com> <1156301545.3022.39.camel@bert64.oviedo.smithconcepts.com> <44EBC710.8050003@mc-kenna.com> Message-ID: <200608222313.54282.jasonb@edseek.com> On Tuesday 22 August 2006 23:10, Damien McKenna wrote: > Bryan J. Smith wrote: > > I've pretty much given up on finding an all-in-one SOHO > > Firewall/NAS/Remote Gateway. But thanx to Damien McKenna, I have found > > a solid NAS device with FTP-SSL/TLS** access: > > http://www.synology.com/enu/products/DS106/spec.php > > It sure looks pretty nifty. I'm also looking at the *Thecus N2050 > series which are eSATA drives with built-in RAID 0 and 1, those would be > pretty darn wonderful for easy & reliable external expansion on > miniature systems (think: Mac Mini).* Interesting. At work we just USB2 two external enclosures and RAID 1 that on the Mac Minis. Seems to work well enough. > > **FileZilla makes an excellent, easy-to-use, FTP-SSL/TLS client. > > Too bad they don't have plain-jane SSH & SFTP, would have been better > IMHO. Ah well, maybe in a ROM upgrade.. Or WebDAV... -- Jason Boxman http://edseek.com/ - Linux and FOSS stuff From damien at mc-kenna.com Tue Aug 22 23:19:09 2006 From: damien at mc-kenna.com (Damien McKenna) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Re: Synology DS-106 w/FTPS (explicit) access In-Reply-To: <200608222313.54282.jasonb@edseek.com> References: <1156174966.2989.121.camel@bert64.oviedo.smithconcepts.com> <1156301545.3022.39.camel@bert64.oviedo.smithconcepts.com> <44EBC710.8050003@mc-kenna.com> <200608222313.54282.jasonb@edseek.com> Message-ID: <44EBC92D.3040704@mc-kenna.com> Jason Boxman wrote: >> I'm also looking at the Thecus N2050 series which are eSATA drives with built-in RAID 0 and 1, those would be pretty darn wonderful for easy & reliable external expansion on miniature systems (think: Mac Mini). >> > Interesting. Should be more reliable than many other options, and a good deal faster. > At work we just USB2 two external enclosures and RAID 1 that on > the Mac Minis. Seems to work well enough. > Hitting the SATA directly will give a tremendous speed boost, the onboard RAID will be faster than CPU & system I/O-driven RAID, and be more reliable than Firewire. I've seen lots of comments of Firewire-based RAID systems dying because one drive gets unplugged, etc. It just makes me nervous. >> Too bad they don't have plain-jane SSH & SFTP, would have been better >> IMHO. Ah well, maybe in a ROM upgrade.. >> > Or WebDAV... > That too, it should be an easy add-on for the (presumed) onboard Apache. Now if only Apple would do something like that. -- Damien McKenna, husband, father, geek. damien@mc-kenna.com - http://www.mc-kenna.com/ From b.j.smith at ieee.org Wed Aug 23 00:23:58 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Re: Synology DS-106 w/FTPS (explicit) access In-Reply-To: <44EBC710.8050003@mc-kenna.com> References: <1156174966.2989.121.camel@bert64.oviedo.smithconcepts.com> <1156175268.2989.128.camel@bert64.oviedo.smithconcepts.com> <3C24AB1F-22DB-4092-99F5-49B4396268D2@mc-kenna.com> <1156204319.3017.10.camel@bert64.oviedo.smithconcepts.com> <8CCAE2A4-B430-4EE0-BB76-13335B045B8A@mc-kenna.com> <1156289647.3022.7.camel@bert64.oviedo.smithconcepts.com> <44EBA7EB.6000102@mc-kenna.com> <44EBA938.90002@mc-kenna.com> <1156296745.3022.28.camel@bert64.oviedo.smithconcepts.com> <44EBB549.5030707@mc-kenna.com> <1156301545.3022.39.camel@bert64.oviedo.smithconcepts.com> <44EBC710.8050003@mc-kenna.com> Message-ID: <1156307038.3029.29.camel@bert64.oviedo.smithconcepts.com> On Tue, 2006-08-22 at 23:10 -0400, Damien McKenna wrote: > It sure looks pretty nifty. It's GbE interface has full jumbo frame support and whips the Buffalo Gigabit LinkStation in performance according to Tom's. The on-board 266MHz FreeScale processor might have a bit to do with that -- I think the Buffalos are using older XScale (not sure). And that's before the fact that the Buffalo does _not_ offer FTP-SSL/TLS (FTPS) access. > I'm also looking at the *Thecus N2050 ... cut ... I'll answer that in another e-mail, since it's not a NAS. > Too bad they don't have plain-jane SSH & SFTP, would have been better > IMHO. Ah well, maybe in a ROM upgrade.. Don't knock FTPS, it works well, using proven FTP mechanisms and then the explicit SSL/TLS options for authentication and/or transfer. It's nice that the Synology does _full_ IETF standard explicit mode, and not just an implicit (separate port 990), which makes it more compatible with firewalls and whatnot. Furthermore, SSH/SFTP can be a PITA to support. After doing embedded and secure file transfer work over the last 2 years, I really have a _dislike_ for the OpenSSH codebase -- even OpenSSH 4.x. The client _and_ server are _both_ very _buggy_ and/or _limited_ when it comes to options, error codes, etc... The codebase really needs a serious overhaul IMPO -- seems very "hacked." > A bit expensive though. Not really. > That would be 100 *separate* *shares*, obviously you could share one > folder and add sub-directories. But it seems it only has per-user share security, like virtually all NAS devices. At least that's what it seems at first glance. > Also... because it has MySQL and PHP you could install Gallery2 and use > the Gallery2 exporter for iPhoto to put your photos in a network-shared > album. I do wish, though, that they used Gallery2 by default rather > than their own system, it would give it much more flexibility. There is photo support and other uPnP goodies -- although some things are clearly Windows-only. -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From b.j.smith at ieee.org Wed Aug 23 00:41:35 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Re: eSATA RAID-1, putting the "IOP" outside the box -- WAS: Synology DS-106 w/FTPS (explicit) access In-Reply-To: <44EBC710.8050003@mc-kenna.com> References: <1156174966.2989.121.camel@bert64.oviedo.smithconcepts.com> <1156175268.2989.128.camel@bert64.oviedo.smithconcepts.com> <3C24AB1F-22DB-4092-99F5-49B4396268D2@mc-kenna.com> <1156204319.3017.10.camel@bert64.oviedo.smithconcepts.com> <8CCAE2A4-B430-4EE0-BB76-13335B045B8A@mc-kenna.com> <1156289647.3022.7.camel@bert64.oviedo.smithconcepts.com> <44EBA7EB.6000102@mc-kenna.com> <44EBA938.90002@mc-kenna.com> <1156296745.3022.28.camel@bert64.oviedo.smithconcepts.com> <44EBB549.5030707@mc-kenna.com> <1156301545.3022.39.camel@bert64.oviedo.smithconcepts.com> <44EBC710.8050003@mc-kenna.com> Message-ID: <1156308095.3029.46.camel@bert64.oviedo.smithconcepts.com> On Tue, 2006-08-22 at 23:10 -0400, Damien McKenna wrote: > I'm also looking at the *Thecus N2050 series which are eSATA drives > with built-in RAID 0 and 1, those would be pretty darn wonderful for > easy & reliable external expansion on miniature systems (think: Mac > Mini).* On Tue, 2006-08-22 at 23:13 -0400, Jason Boxman wrote: > Interesting. At work we just USB2 two external enclosures and RAID 1 > that on the Mac Minis. Seems to work well enough. Actually, eSATA RAID-1 would be nice. Here's why ... 1. Most of these eSATA devices have FreeScale or XScale 200-400MHz RISC microcontrollers aka I/O Processors (IOP) -- designed in-line I/O Processing with near-0 wait state operations 2. No NAS/general OS overhead on the processor, just pushing I/O data _in-line_ -- which is what the IOPs are designed for, and not all that additional overhead like a cheap NAS which only detract from near-0 wait state I/O operations 3. You don't need much DRAM, firmware, etc... to do this when you do only I/O Processing, further saving on embedded costs 4. SATA is a "direct, dumb block" protocol, unlike USB, TCP/IP or even Ethernet for that matter (sorry, ATA over Ethernet is _not_ viable -- don't get me started ;-). It's basically SAS (Serial Attached SCSI) without the flexibility, but also less overhead too. If you read up on the few reviews, the performance isn't as good as a traditional 3Ware ASIC+SRAM card, but it's still a bit better than most NAS and USB devices too ... http://www.hexus.net/content/item.php?item=4817 http://www.trustedreviews.com/article.aspx?art=3279 They are getting ~50MBps RAID-1, ~80MBps RAID-0, using commodity SATA disks. Far from best, but also much, much better than other options. And I'd personally like to see how it fares against Linux's MD (software) RAID-1 -- I think most would be surprised. Now if the prices came down to $100, I'd buy one. I can only put two (2) SATA drives in my MicroATX cube (assming I'm using the 2x5.25" and 1x3.5" external) and I'd like to have additional, but RAID-1, storage. -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From b.j.smith at ieee.org Wed Aug 23 00:48:52 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Re: eSATA RAID-1, putting the "IOP" outside the box -- WAS: Synology DS-106 w/FTPS (explicit) access In-Reply-To: <44EBC92D.3040704@mc-kenna.com> References: <1156174966.2989.121.camel@bert64.oviedo.smithconcepts.com> <1156301545.3022.39.camel@bert64.oviedo.smithconcepts.com> <44EBC710.8050003@mc-kenna.com> <200608222313.54282.jasonb@edseek.com> <44EBC92D.3040704@mc-kenna.com> Message-ID: <1156308532.3029.54.camel@bert64.oviedo.smithconcepts.com> On Tue, 2006-08-22 at 23:19 -0400, Damien McKenna wrote: > Hitting the SATA directly will give a tremendous speed boost, the > onboard RAID will be faster than CPU & system I/O-driven RAID, Or even if a CPU-I/O interconnect uses PCIe x1 channels for SATA so software RAID-1 can "keep up" with similar performance at least you're not putting that "duplication" on the CPU-I/O interconnect. > and be more reliable than Firewire. > I've seen lots of comments of Firewire-based RAID systems dying > because one drive gets unplugged, etc. It just makes me nervous. Er, I think you're comparing apples'n oranges. First off, 95% of FireWire "RAID" out there is still _host-based_ and _software_ driven. If someone made a target-based hardware FireWire RAID, then it wouldn't be as much of an issue. Secondly, FireWire is compared to USB, _not_ SATA. FireWire is a completely different protocol than SATA. But FireWire has a protocol that is superior to USB. I.e., FireWire has a superior/superset protocol to USB SAS has a superior/superset protocol to SATA SCSI has a superior/superset protocol to ATA (aka IDE) Etc... -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From philb at philb.us Wed Aug 23 01:21:49 2006 From: philb at philb.us (Phil Barnett) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Woot has a nice deal on a video card Message-ID: <200608230121.49593.philb@philb.us> http://www.woot.com -- My other computer is your Windows machine From philb at philb.us Wed Aug 23 02:03:33 2006 From: philb at philb.us (Phil Barnett) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Linux Installfest Postponed due to Holiday Weekend. Message-ID: <200608230203.33909.philb@philb.us> The first Saturday in September, the 2nd, falls on a national holiday weekend. Because of this, the regular installfest date will be moved forward 1 week to September 8th, 2006. Please mark you calendars with the new date! -- My other computer is your Windows machine From b.j.smith at ieee.org Wed Aug 23 03:51:44 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Woot has a nice deal on a video card -- false specs (7900GT)? In-Reply-To: <200608230121.49593.philb@philb.us> References: <200608230121.49593.philb@philb.us> Message-ID: <1156319504.3029.68.camel@bert64.oviedo.smithconcepts.com> On Wed, 2006-08-23 at 01:21 -0400, Phil Barnett wrote: > http://www.woot.com They are quoting the 7900GT specifications -- 450MHz core, 1320MHz effective GDDR3 memory, 24 pixel, 8 vertex. The "desktop" 7900GS just came out (the "mobile" GeForce Go 7900GS has been around awhile). It drops to 20 pixel, 7 vertex, like the older 7800GT. And from what I've read, clocks drop to 375MHz core, 1000MHz effective DDR2? GDDR3? memory. Now that does make it (20/7/375/1000-256b) better than either of the older 6800GT (16/6/350/1000-256b) or 7600GT (12/5/560/1400-128b), which are -- at best -- the same price after rebate/specials. Sounds like the 7900GS GPU might be from 7900GT units that aren't "making the grade" in clock and/or if they have a few pixel or one vertex unit bad. Would make sense, especially since people claim it can run at higher clocks. I'm still better with my older 7800GTX though (24/8/430/1200-256b). I'm tempted to buy one for my wife and just push her to PCIe. FYI ... BS Blog: GeForce 6 and 7 series variants ... nuts! http://thebs413.blogspot.com/2006/02/geforce-6-and-7-series-variants-nuts.html -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From b.j.smith at ieee.org Wed Aug 23 03:57:52 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Woot has a nice deal on a video card -- false specs (7900GT)? In-Reply-To: <1156319504.3029.68.camel@bert64.oviedo.smithconcepts.com> References: <200608230121.49593.philb@philb.us> <1156319504.3029.68.camel@bert64.oviedo.smithconcepts.com> Message-ID: <1156319872.3029.70.camel@bert64.oviedo.smithconcepts.com> On Wed, 2006-08-23 at 03:51 -0400, Bryan J. Smith wrote: > The "desktop" 7900GS just came out (the "mobile" GeForce Go 7900GS has > been around awhile). It drops to 20 pixel, 7 vertex, like the older > 7800GT. And from what I've read, clocks drop to 375MHz core, 1000MHz > effective DDR2? GDDR3? memory. > Now that does make it (20/7/375/1000-256b) better than either of the > older 6800GT (16/6/350/1000-256b) or 7600GT (12/5/560/1400-128b), which > are -- at best -- the same price after rebate/specials. Here's AnandTech's coverage of the Dell leak: http://www.dailytech.com/article.aspx?newsid=2329 -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From dmckenna at thelimucompany.com Wed Aug 23 09:48:42 2006 From: dmckenna at thelimucompany.com (Damien McKenna) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Synology DS-106 w/FTPS (explicit) access In-Reply-To: <1156301545.3022.39.camel@bert64.oviedo.smithconcepts.com> References: <1156174966.2989.121.camel@bert64.oviedo.smithconcepts.com> <1156175268.2989.128.camel@bert64.oviedo.smithconcepts.com> <3C24AB1F-22DB-4092-99F5-49B4396268D2@mc-kenna.com> <1156204319.3017.10.camel@bert64.oviedo.smithconcepts.com> <8CCAE2A4-B430-4EE0-BB76-13335B045B8A@mc-kenna.com> <1156289647.3022.7.camel@bert64.oviedo.smithconcepts.com> <44EBA7EB.6000102@mc-kenna.com> <44EBA938.90002@mc-kenna.com> <1156296745.3022.28.camel@bert64.oviedo.smithconcepts.com> <44EBB549.5030707@mc-kenna.com> <1156301545.3022.39.camel@bert64.oviedo.smithconcepts.com> Message-ID: <07CE710E-0189-42C4-89B5-544658BD0BC3@thelimucompany.com> On Aug 22, 2006, at 10:52 PM, Bryan J. Smith wrote: > http://www.synology.com/enu/products/DS106/spec.php This is an eSATA device that comes with a 33MHz/32bit PCI card. Do you think its performance would improve if it was connected using a 66MHz/64bit card instead? -- Damien McKenna - Web Developer - dmckenna@thelimucompany.com The Limu Company - http://www.thelimucompany.com/ - 407-804-1014 #include -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.matrixlist.com/pipermail/pc_support/attachments/20060823/a96b2df0/attachment.html From dmckenna at thelimucompany.com Wed Aug 23 09:52:04 2006 From: dmckenna at thelimucompany.com (Damien McKenna) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Synology DS-106 w/FTPS (explicit) access In-Reply-To: <07CE710E-0189-42C4-89B5-544658BD0BC3@thelimucompany.com> References: <1156174966.2989.121.camel@bert64.oviedo.smithconcepts.com> <1156175268.2989.128.camel@bert64.oviedo.smithconcepts.com> <3C24AB1F-22DB-4092-99F5-49B4396268D2@mc-kenna.com> <1156204319.3017.10.camel@bert64.oviedo.smithconcepts.com> <8CCAE2A4-B430-4EE0-BB76-13335B045B8A@mc-kenna.com> <1156289647.3022.7.camel@bert64.oviedo.smithconcepts.com> <44EBA7EB.6000102@mc-kenna.com> <44EBA938.90002@mc-kenna.com> <1156296745.3022.28.camel@bert64.oviedo.smithconcepts.com> <44EBB549.5030707@mc-kenna.com> <1156301545.3022.39.camel@bert64.oviedo.smithconcepts.com> <07CE710E-0189-42C4-89B5-544658BD0BC3@thelimucompany.com> Message-ID: <9F4A6D51-607D-4DB0-BD30-196EC00590F6@thelimucompany.com> On Aug 23, 2006, at 9:48 AM, Damien McKenna wrote: > This is an eSATA device that comes with a 33MHz/32bit PCI card. Do > you think its performance would improve if it was connected using a > 66MHz/64bit card instead? I was actually thinking of the Thecus N2050, not the Synology NAS: http://www.thecus.com/products_over.php?cid=1&pid=3 -- Damien McKenna - Web Developer - dmckenna@thelimucompany.com The Limu Company - http://www.thelimucompany.com/ - 407-804-1014 #include -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.matrixlist.com/pipermail/pc_support/attachments/20060823/9a6bd587/attachment.html From dmckenna at thelimucompany.com Wed Aug 23 10:18:55 2006 From: dmckenna at thelimucompany.com (Damien McKenna) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Synology DS-106 w/FTPS (explicit) access In-Reply-To: <9F4A6D51-607D-4DB0-BD30-196EC00590F6@thelimucompany.com> References: <1156174966.2989.121.camel@bert64.oviedo.smithconcepts.com> <1156175268.2989.128.camel@bert64.oviedo.smithconcepts.com> <3C24AB1F-22DB-4092-99F5-49B4396268D2@mc-kenna.com> <1156204319.3017.10.camel@bert64.oviedo.smithconcepts.com> <8CCAE2A4-B430-4EE0-BB76-13335B045B8A@mc-kenna.com> <1156289647.3022.7.camel@bert64.oviedo.smithconcepts.com> <44EBA7EB.6000102@mc-kenna.com> <44EBA938.90002@mc-kenna.com> <1156296745.3022.28.camel@bert64.oviedo.smithconcepts.com> <44EBB549.5030707@mc-kenna.com> <1156301545.3022.39.camel@bert64.oviedo.smithconcepts.com> <07CE710E-0189-42C4-89B5-544658BD0BC3@thelimucompany.com> <9F4A6D51-607D-4DB0-BD30-196EC00590F6@thelimucompany.com> Message-ID: <9E7566D1-128F-4865-9678-6AFE8FE2DEF5@thelimucompany.com> On Aug 23, 2006, at 9:52 AM, Damien McKenna wrote: > http://www.thecus.com/products_over.php?cid=1&pid=3 Further, is there any reason why this would not work when connected to a plain SATA connector rather than an e-SATA connector (using a converter cable)? Thanks. -- Damien McKenna - Web Developer - dmckenna@thelimucompany.com The Limu Company - http://www.thelimucompany.com/ - 407-804-1014 #include -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.matrixlist.com/pipermail/pc_support/attachments/20060823/ad71587d/attachment.html From thebs413 at gmail.com Wed Aug 23 13:48:45 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Re: eSATA RAID-1, putting the "IOP" outside the box -- IOPs, PCI[64|e|X] support and SATA cabling ... Message-ID: Damien McKenna wrote: > I was actually thinking of the Thecus N2050, not the Synology NAS: > http://www.thecus.com/products_over.php?cid=1&pid=3 Yes, I know. I've been changing the subject as appropriate in my past correspondence too. Damien McKenna wrote: > This is an eSATA device that comes with a 33MHz/32bit PCI card. Do > you think its performance would improve if it was connected using a > 66MHz/64bit card instead? But what "commodity" SATA controllers do you know support PCI64 or PCI-X? The overwhelming majority of SATA controllers are either PCI32 or PCIe x1. Which brings us to the I/O processor ... I don't know what I/O Processor they are using. Their full-up NAS product seems to use a 400MHz version of the 80219 X-Scale (Superscalar ARM). Now if you look at Intel's _current_ IOP line-up, pretty much all of them support PCI64 or PCI-X. Their "top-tier" IOP332/333 offer an internal PCIe x8 to PCI-X bridge so you can use PCIe devices: http://www.intel.com/design/iio/index.htm [ NOTE: The clocks listed are _maximum_ for the product, _not_ typical/commodity ] So, again, given the fact that pretty much every "commodity" SATA controller is PCI32 or PCIe x1, they are stuck with PCI32. Unless they want to fork out the dough for a IOP332/333. Yes, it might cost sub-$30 in quantity, but that's a far cry in margins from the sub-$10 cost of other IOPs. Damien McKenna wrote: > Further, is there any reason why this would not work when connected > to a plain SATA connector rather than an e-SATA connector (using a > converter cable)? Nope, as long as you consider the fact that SATA can't go farther than 1m -- at least at 1.5Gbps. SATA 3.0Gbps was supposed to be twisted-pair and not "flat", so it might even be shorter in lenght -- like half. And that's before considering interference/signal loss in connectors. eSATA is supposed to be better shielded for 2x length. Don't know since SATA is still used internally. SAS uses differential serial logic to go 8m. You can also trunk 2, 4 or 8 pair for 6, 12 and 24Gbps aggregate. SAS can fall back to SATA as well. You can think of SAS like LVD SCSI and SATA like SE SCSI -- almost literally. From Bruce.Metcalf at figzu.com Wed Aug 23 14:58:35 2006 From: Bruce.Metcalf at figzu.com (Bruce Metcalf) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Re: [Exec] Linux Installfest Postponed due to Holiday Weekend. In-Reply-To: <200608230203.33909.philb@philb.us> References: <200608230203.33909.philb@philb.us> Message-ID: <7.0.1.0.0.20060823145745.03783e90@figzu.com> At 02:03 AM 8/23/2006, Phil Barnett wrote: >The first Saturday in September, the 2nd, falls on a national >holiday weekend. >Because of this, the regular installfest date will be moved forward 1 week to >September 8th, 2006. I think Phil means Saturday, September 9th. Bruce From thebs413 at gmail.com Thu Aug 24 00:23:52 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Woot has a nice deal on a video card -- 6600GT << 6800GT~7600GT << 7800GT~7900GS < 7800GTX~7900GT < 7900GTX Message-ID: Just to follow-up today's discussion, I posted this to the Woot board ... http://www.woot.com/Forums/ViewPost.aspx?PostID=679757 Generic relationship ... 6600GT << 6800GT~7600GT << 7800GT~7900GS < 7800GTX~7900GT < 7900GTX This 7900GS (20/7/375?/1000?/256b) card should be similar to a 7800GT (20/7/400/1000/256b), although not quite as fast as a 7800GTX (24/8/430/1200/256b) or 7900GT (24/8/450/1320/256b). But it does mean much faster than $150+ 7600GT (12/5/560-580/1400-1500/128b) or 6800GT (16/6/350/1000/256b). I'm going to replace my wife's 6800GT with one of these. From thebs413 at gmail.com Thu Aug 24 13:29:18 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Swap on an NTFS partition? -- pre-allocated blocks Message-ID: Darren Humphrey wrote: > Problem is, the machines are all Windows XP and have NTFS partitions. > Slax does support the newer ntfsmount that offers read-write support on NTFS, > but I am having difficulty getting it to work. Understand that kernel-space (module) NTFS filesystem support has only 2 modes: 1. read-only 2. modify-only -- no change in size/attributes (only "safe" way to write to NTFS) I assume you're trying to do #2, with a pre-allocated set of file blocks. If not, you need to pre-allocate the blocks inside of XP itself. Once you do that, you _should_ be able to use the blocks. If you want to use user-space (Captive) NTFS filesystem support, it'll be dog slow (really slow). > So far I have tried: > umount /mnt/hda2 (originally mounted readonly) > ntfsmount /mnt/hda2 (now mounted read write) > mkswap /mnt/hda2/pagefile.sys (reports success) > Any ideas? I would really like to use a swap file so I don't have to permanently > take 1-2Gb of every sales laptop just to run this app once in a while. I _always_ create a 8-32GB FAT32 partition on _all_ NT5.x (200x/XP) systems. I highly recommend you do this from now on. PartitionMagic is always a Godsend, especially for just taking back the last 8-32GB of a disk. I used to create FAT32 as the C: (system**) drive and the boot** drive/install would go on D:). That was so I could recover the system easier as well as dual-boot with MS-DOS 7.1+ (Win95OSR2+) But now I create the FAT32 as D: (nothing special), and the NTFS is C: and it provides both start-up functions (system** and boot**). BTW, _never_ create a FAT32 C: drive that is over 32GB or you will often have geometry issues -- especially in NT5.1SP2 (XPSP2). **NOTE: This terminology, which is Microsoft's, is ass-backwards. "system" means where the bootstrap and NTLDR are at and must be BIOS disk 80h (C:), although "active" is not required for NT (unlike DOS). "boot" means where the NTKERNEL.EXE and \WINNT directory is at, which can be on a completely different partition or even drive. The opposite makes far more sense, but I try to stick with MS' terminology. From thebs413 at gmail.com Thu Aug 24 13:32:37 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Re: Swap on an NTFS partition? -- pre-allocated blocks (forgot a comment) Message-ID: Darren Humphrey wrote: > So far I have tried: > umount /mnt/hda2 (originally mounted readonly) > ntfsmount /mnt/hda2 (now mounted read write) > mkswap /mnt/hda2/pagefile.sys (reports success) Oops, forgot a comment. "mkswap" is rather silent, even when it fails. Try it with the "-c" option to check bad blocks to verify you can actually write to the file. I don't think your problem is "swapon," but "mkswap." Also, really verify /mnt/hda2 is mounted read/write -- although that's hard to do without creating a new file, something you should _never_ do with NTFS because it modifies filesystem attributes. From damien at mc-kenna.com Fri Aug 25 13:25:00 2006 From: damien at mc-kenna.com (Damien McKenna) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] Brother HL-2040 laser printer for $60AR Message-ID: <8BFA24B4-2228-45CE-9728-55F47770566A@mc-kenna.com> OfficeMax have the brother HL-2040 laser printer on sale this week for $79.98 with no rebates. brother.com has a $20 rebate on this printer so you can get it for $59.98 end price. Tasty deal! Details on the printer: http://www.brother-usa.com/printer/ printer_detail_AREA=PRINTER_1&PRODUCTID=HL2040.aspx OfficeMax: http://www.officemax.com/max/solutions/product/prodBlock.jsp? BV_UseBVCookie=yes&expansionOID=-536907398&prodBlockOID=537442680 Rebate: http://www.brother-usa.com/IMS_DOCS/ 44/44EA51F50DAA04BFE1000000CD8620C7.pdf -- Damien McKenna, husband, father, geek. damien@mc-kenna.com - http://www.mc-kenna.com/ From dave at dgnal.net Fri Aug 25 17:43:57 2006 From: dave at dgnal.net (David Simmons) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] WinXP 64-Bit Software Raid Setup Howto? Message-ID: <42557.192.104.67.222.1156542237.squirrel@qtmail.dgnal.net> Anyone got one? Have two identical SATA drives. Am hoping to stripe part of the drive for performance for video work. Am anticipating setup to be: Disk 1: 100Gb OS Partition 200GB Raid-0 Part I Disk 2: 200GB Raid-0 Part II Extra 100Gb of space I have the partitioning setup...but for the life of me can't figure out how to create a raid array using those two partitions!? Here's where Linux REALLY shines...much easier/intuitive! Bottom line - what am i missing? Help was, of course, no help...and googling shows alot of HowTo's for motherboard based RAIDing setup... Thanks in advance, dave From b.j.smith at ieee.org Sat Aug 26 18:39:35 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] WinXP 64-Bit Software Raid Setup Howto? In-Reply-To: <42557.192.104.67.222.1156542237.squirrel@qtmail.dgnal.net> Message-ID: <20060826223935.28810.qmail@web32914.mail.mud.yahoo.com> First off, NT5.x (200x/XP) requires use a LDM Disk Label (Dynamic Disc) for any volume management. Under Linux, I actually recommend using the LVM Disk Label when using any MD volumes as well. Long story short, because the legacy BIOS/DOS Disk Label (Basic Disc) has absolutely no reservations for volume meta-data, it's easy to "lose a volume." It used to happen to me all-the-time under NT 4.0 and it has under any Linux system I've encountered without LVM/LVM2. Secondly, IIRC, _no_ version of Windows XP, not even Pro, supports mirroring. Only the server editions of 2000/2003 do. I'm not sure about Windows 2000 Professional, it might. Only striping and spanning (concating) is supported on non-server versions. You set it up via Disk Manager. --- David Simmons wrote: > Anyone got one? > > Have two identical SATA drives. Am hoping to stripe part of the > drive for > performance for video work. > > Am anticipating setup to be: > > Disk 1: 100Gb OS Partition 200GB Raid-0 Part I > Disk 2: 200GB Raid-0 Part II Extra 100Gb of space > > I have the partitioning setup...but for the life of me can't figure > out > how to create a raid array using those two partitions!? Here's > where > Linux REALLY shines...much easier/intuitive! > > Bottom line - what am i missing? Help was, of course, no > help...and > googling shows alot of HowTo's for motherboard based RAIDing > setup... > > Thanks in advance, > > dave > > _______________________________________________ > Pc_support mailing list > Pc_support@matrixlist.com > http://lists.matrixlist.com/mailman/listinfo/pc_support > > -- Bryan J. Smith Professional, Technical Annoyance b.j.smith@ieee.org http://thebs413.blogspot.com -------------------------------------------------- Fission Power: An Inconvenient Solution From carter at carter.cc Sat Aug 26 19:33:16 2006 From: carter at carter.cc (carter@carter.cc) Date: Tue Oct 31 13:18:00 2006 Subject: [Pc_Support] WinXP 64-Bit Software Raid Setup Howto? In-Reply-To: <20060826223935.28810.qmail@web32914.mail.mud.yahoo.com> References: <42557.192.104.67.222.1156542237.squirrel@qtmail.dgnal.net> <20060826223935.28810.qmail@web32914.mail.mud.yahoo.com> Message-ID: <1571.24.233.138.31.1156635196.squirrel@carter.cc> > Secondly, IIRC, _no_ version of Windows XP, not even Pro, supports > mirroring. Only the server editions of 2000/2003 do. I'm not sure > about Windows 2000 Professional, it might. Only striping and > spanning (concating) is supported on non-server versions. You set it > up via Disk Manager. You're quite correct... no mirroring/striping with XP or 2K Pro. 2000 Server, 2003 Server are the only recent OS's that support it. So if you don't have it available on-board (FRAID) or via a real hardware controller, you're SOL in Windows Land. -Carter From b.j.smith at ieee.org Sun Aug 27 15:38:07 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] WinXP 64-Bit Software Raid Setup Howto? -- nForce 4/4x0 FRAID-1 works fine ... In-Reply-To: <1571.24.233.138.31.1156635196.squirrel@carter.cc> Message-ID: <20060827193807.74433.qmail@web32905.mail.mud.yahoo.com> carter@carter.cc wrote: > You're quite correct... no mirroring/striping with XP or 2K Pro. > 2000 Server, 2003 Server are the only recent OS's that support it. > So if you don't have it available on-board (FRAID) or via a real > hardware controller, you're SOL in Windows Land. Actually, this weekend I just went FRAID-1 on my wife's computer and I'm considering it for mine as well. I'm using nForce 4 and 410/430 series mainboards with Fedora Core 5 and Windows XP Home as well as Pro. I converted her from an older ViA KM266 (Athlon XP2600+) to a nVidia GeForce6100/nForce410 (Athlon64 3200+). It took a number of intermedia steps to get the Windows XP Home dual-boot working, but it was flawless for Fedora Core 5 -- including copying the original Windows XP Home install from a 3Ware 6200 RAID-1 volume to SATA drives on the nVidia FRAID-1 volume. Fedora Core 5's devicemapper automagically creates the /dev/mapper/nvidia* entries and then symlinks them to /dev/dm-0 (full disk), /dev/dm-1 (first partition), /dev/dm-2, etc... Again, these worked flawlessly for even copying the Windows partitions (including a raw "dd" of the 32MB C:/boot/system drive). GRUB and Linux boot fine from the FRAID-1 volume, and even prevent me from accidentally writing to the "regular" /dev/sda and /dev/sdb device when the BIOS has been set to [F]RAID. I'm quite impressed, at least with nVidia chipsets and the Fedora Core 5 installer/kernel/utilities (including GRUB support). -- Bryan J. Smith Professional, Technical Annoyance b.j.smith@ieee.org http://thebs413.blogspot.com -------------------------------------------------- Fission Power: An Inconvenient Solution From jasonb at edseek.com Sun Aug 27 15:45:16 2006 From: jasonb at edseek.com (Jason Boxman) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] WinXP 64-Bit Software Raid Setup Howto? -- nForce 4/4x0 FRAID-1 works fine ... In-Reply-To: <20060827193807.74433.qmail@web32905.mail.mud.yahoo.com> References: <20060827193807.74433.qmail@web32905.mail.mud.yahoo.com> Message-ID: <200608271545.16523.jasonb@edseek.com> On Sunday 27 August 2006 15:38, Bryan J. Smith wrote: > carter@carter.cc wrote: > > You're quite correct... no mirroring/striping with XP or 2K Pro. > > 2000 Server, 2003 Server are the only recent OS's that support it. > > So if you don't have it available on-board (FRAID) or via a real > > hardware controller, you're SOL in Windows Land. > > Fedora Core 5's devicemapper automagically creates the > /dev/mapper/nvidia* entries and then symlinks them to /dev/dm-0 (full > disk), /dev/dm-1 (first partition), /dev/dm-2, etc... Again, these > worked flawlessly for even copying the Windows partitions (including > a raw "dd" of the 32MB C:/boot/system drive). > > GRUB and Linux boot fine from the FRAID-1 volume, and even prevent me > from accidentally writing to the "regular" /dev/sda and /dev/sdb > device when the BIOS has been set to [F]RAID. Nice. I was finally considering moving to either software RAID 1 or mainboard RAID 1 for future systems when I finally make the jump to SATA. I am pleased to see F.RAID 1 worked for you. That gives me hope. I really can't afford a 3Ware 8000-2 or whatever for hardware SATA RAID 1 for home systems and I don't need hotswap anyway. -- Jason Boxman http://edseek.com/ - Linux and FOSS stuff From carter at carter.cc Sun Aug 27 18:49:51 2006 From: carter at carter.cc (carter@carter.cc) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] WinXP 64-Bit Software Raid Setup Howto? -- nForce 4/4x0 FRAID-1 works fine ... In-Reply-To: <20060827193807.74433.qmail@web32905.mail.mud.yahoo.com> References: <1571.24.233.138.31.1156635196.squirrel@carter.cc> <20060827193807.74433.qmail@web32905.mail.mud.yahoo.com> Message-ID: <34150.24.233.138.31.1156718991.squirrel@www.carter.cc> > Fedora Core 5's devicemapper automagically creates the > /dev/mapper/nvidia* entries and then symlinks them to /dev/dm-0 (full > disk), /dev/dm-1 (first partition), /dev/dm-2, etc... Again, these > worked flawlessly for even copying the Windows partitions (including > a raw "dd" of the 32MB C:/boot/system drive). Wow.... I haven't even bothered with trying to mix on-board stuff with Linux ever since I was burnt by it (my own fault) a long time ago on a BX-133 board. As a matter of fact, that's when I started figuring out what Linux software RAID was all about... and haven't looked back since. It's nice to know there are supported FRAID systems 'out of the box'... maybe I need to give this a look again :) -Carter From b.j.smith at ieee.org Sun Aug 27 23:32:18 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Re: WinXP 64-Bit Software Raid Setup Howto? -- How Linux DeviceMapper handles FRAID ... In-Reply-To: <34150.24.233.138.31.1156718991.squirrel@www.carter.cc> Message-ID: <20060828033218.7268.qmail@web32915.mail.mud.yahoo.com> carter@carter.cc wrote: > Wow.... I haven't even bothered with trying to mix on-board stuff > with Linux ever since I was burnt by it (my own fault) a long time > ago on a BX-133 board. > As a matter of fact, that's when I started figuring out what Linux > software RAID was all about... and haven't looked back since. > It's nice to know there are supported FRAID systems 'out of the > box'... maybe I need to give this a look again :) There have been several approaches to software RAID in Linux, including FRAID card support. 1. MultiDisk (MD) in 2.4/2.6 2. ATARAID (w/HPTRAID, PDCRAID, SILRAID) in 2.4 3. DeviceMapper in 2.6 #3 is what Red Hat is using. #1 is still commonly used for Linux Software RAID, but _not_ with FRAID cards. Atop of basic BIOS/DOS Disk Labels, you'll have the same problems as NT 4.0 did -- "sudden loss" of the volume/boot-time issues. Why? Because there is no reservation of parts of the disk for meta-data -- although at least the newer mdadm tools can "re-scan" and find MD volumes (although that doesn't help you at boot-time ;-). Putting MD volumes in LVM (2.4) and LVM2 (2.6) disk labels solves the latter problem, just like Logical Disk Manager (LDM) Disk Labels (Dynamic Discs) is now Microsoft does in NT 5.0 (2000+). #2 was a stop-gap attempt to build a common FRAID Logic in a single driver (ATARAID), and then have vendor-specific interface drivers (HPTRAID for HighPoint Technologies, PDCRAID for Promise Data Corp and SILRAID for Silicon Image) for their cards. It was largely an utter disaster and major PITA. #3 is the new DeviceMapper component of the Linux 2.6 kernel and also works with the LVM2 subsystem. DeviceMapper is extremely flexible, although portions of it are _not_ in the stock kernel -- at least not yet. It was really an ambitious attempt by Red Hat/Sistina, so not everything is perfected yet -- e.g., Snapshots (which was in the older 2.4/LVM implementation). Striping (RAID-0) and spanning (concatting) went in stock 2.6 and Mirroring (RAID-1) went in later, including more recent Fedora Core releases. The great thing about _true_ volume management and DeviceMapper is that it's not just a "slap atop" approach -- unlike MD. It addresses the _full_ spectrum of hardware to firmware to kernel (including _full_ hotplug support) to the device/user-space detail. That means that DeviceMapper can work with the GRUB boot loader perfectly, including boot/root being on a DeviceMapper volume. That also means DeviceMapper can "understand" many common firmware disk organizations -- including FRAID cards. Apparently, from my experience, the Fedora Core 5 Anaconda installer can recognize (possibly via the nv_sata driver) that the nVidia firmware is set to "RAID" mode (instead of non-RAID/BIOS compatibility). So it knows how to map the drives -- devices, kernel and GRUB boot-time support. And that's exactly what it did -- at least for my Mirrored (RAID-1) volume. Once the devices are transparently mapped -- including their data organization -- the kernel's native (via LVM2/DeviceMapper) mirroring/RAID-1 logic is used -- instead of the vendor driver, but fully emulating it (again, via DeviceMapper's transparent mapping to the kernel RAID-1 logic). So I'm getting the full kernel RAID-1 support, and not the vendor's limited capability. As long as DeviceMapper handles the organization correctly, I shouldn't have any data loss. We'll see. -- Bryan J. Smith Professional, Technical Annoyance b.j.smith@ieee.org http://thebs413.blogspot.com -------------------------------------------------- Fission Power: An Inconvenient Solution From b.j.smith at ieee.org Sun Aug 27 23:42:08 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Re: WinXP 64-Bit Software Raid Setup Howto? -- Linux-Windows XP compatibility In-Reply-To: <20060828033218.7268.qmail@web32915.mail.mud.yahoo.com> Message-ID: <20060828034208.8789.qmail@web32901.mail.mud.yahoo.com> "Bryan J. Smith" wrote: > So I'm getting the full kernel RAID-1 support, and not the vendor's > limited capability. As long as DeviceMapper handles the > organization correctly, I shouldn't have any data loss. We'll see. And when I say "handles the organization correctly," I mean I am not only accessing Linux filesystems but also _FAT32_filesystems_ being read/written to by Windows XP Home/Pro! I literally used Linux's DeviceMapper to copy over existing FAT32 volumes, including dd'ing the C: drive a booting Windows XP copy! The scenario was ... 1. Install new nForce 4/410 mainboard with SATA drives (new RAID-1) and install Fedora Core 5, reserve C: and slice/format D: -- automagically sets up GRUB, etc... 2. Temporarily unplug SATA drives, put in 3Ware card and ATA drives (old RAID-1) into new nForce 4/410 mainboard and load nForce drivers under Windows XP Home/Pro 3. Reconnect SATA drives, 3Ware card and ATA drives are now secondary hard disk (SATA boots first), boot Fedora Core 5 and use dd to copy C: and then find|cpio D:. 4. Remove 3Ware card and ATA drives and boot Windows XP Home/Pro from new SATA volume (device detection, re-activation PITA, etc... but it works) Linux LVM2/DeviceMapper software RAID-1 and the nVidia nForce software RAID-1 driver are working perfectly together in a dual-boot. -- Bryan J. Smith Professional, Technical Annoyance b.j.smith@ieee.org http://thebs413.blogspot.com -------------------------------------------------- Fission Power: An Inconvenient Solution From jasonb at edseek.com Sun Aug 27 23:55:21 2006 From: jasonb at edseek.com (Jason Boxman) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Re: WinXP 64-Bit Software Raid Setup Howto? -- How Linux DeviceMapper handles FRAID ... In-Reply-To: <20060828033218.7268.qmail@web32915.mail.mud.yahoo.com> References: <20060828033218.7268.qmail@web32915.mail.mud.yahoo.com> Message-ID: <200608272355.21933.jasonb@edseek.com> On Sunday 27 August 2006 23:32, Bryan J. Smith wrote: > #1 is still commonly used for Linux Software RAID, but _not_ with > FRAID cards. Atop of basic BIOS/DOS Disk Labels, you'll have the > same problems as NT 4.0 did -- "sudden loss" of the volume/boot-time > issues. Why? Because there is no reservation of parts of the disk > for meta-data -- although at least the newer mdadm tools can > "re-scan" and find MD volumes (although that doesn't help you at > boot-time ;-). Putting MD volumes in LVM (2.4) and LVM2 (2.6) disk > labels solves the latter problem, just like Logical Disk Manager > (LDM) Disk Labels (Dynamic Discs) is now Microsoft does in NT 5.0 > (2000+). You speak of this often, but I don't understand. I haven't personally experienced lost MD volumes, so I don't understand how that will happen to me? Thanks. -- Jason Boxman http://edseek.com/ - Linux and FOSS stuff From b.j.smith at ieee.org Mon Aug 28 03:44:32 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Re: WinXP 64-Bit Software Raid Setup Howto? -- How Linux DeviceMapper handles FRAID ... In-Reply-To: <200608272355.21933.jasonb@edseek.com> Message-ID: <20060828074432.8090.qmail@web32906.mail.mud.yahoo.com> Jason Boxman wrote: > You speak of this often, but I don't understand. I haven't > personally experienced lost MD volumes, so I don't understand how > that will happen to me? Well I have. It is especially notorious if a drive does not come up once or twice and then is reconnected, or a drive is connected on a different channel, etc... In a nutshell, there is no "dynamic way" to keep track of a slice and its volume association outside of the filesystem itself. The volume is then "lost" -- e.g., a "broken mirror" in the case of RAID-1 or "lost stripe" in the case of RAID-0. Hence the "common name" Microsoft chose for the Logical Disk Manager "LDM" disk label -- the "Dynamic Disk." It allows for all sorts of meta-data to be stored outside of the filesystem as part of the disk label itself and associated "dynamically" by a booting kernel that has no other info or interogation software. The legacy BIOS/DOS disk does not, and relies on statically configured information. Again, the mdadm tools have helped quite a bit over the legacy raidtools for MD, especially when it comes to re-scanning finding those "lost/broken volumes." But for boot-time operations, that's not an option. Then again, MD is difficult to use with boot-time operations -- hence why most /boot filesystems are not located on them for safety/usability. In Linux, the Logical Volume Manager (LVM) disk label offers all sorts of meta-data store like LDM disk labels, although unlike Microsoft in NT5+, Linux doesn't make its use mandatory. The kernel itself can associate slices in LVM volumes with MD volumes, hence why I always use MD atop of LVM. Going the other way, LVM atop of MD is a nightmare -- especialy if the MD disk cannot associate with other, required disks in the volume. As "nicer/easier" it seems to make one big MD volume and then put LVM atop of it, it's the least stable/reliable configuration. That's why LVM+MD gets a bad name, from LVM atop of MD. In reality, MD atop of LVM is the recommended solution in my book -- never "lost" a MD volume myself that way. Although DeviceMapper now removes the requirement for the separate MD, and it's completely dynamic, bootable and recoverable right in GRUB/kernel itself -- no userspace tools required. -- Bryan J. Smith Professional, Technical Annoyance b.j.smith@ieee.org http://thebs413.blogspot.com -------------------------------------------------- Fission Power: An Inconvenient Solution From carter at carter.cc Mon Aug 28 08:53:38 2006 From: carter at carter.cc (Carter Manucy) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Re: WinXP 64-Bit Software Raid Setup Howto? -- Linux-Windows XP compatibility In-Reply-To: <20060828034208.8789.qmail@web32901.mail.mud.yahoo.com> References: <20060828034208.8789.qmail@web32901.mail.mud.yahoo.com> Message-ID: <44F2E752.4010703@carter.cc> Bryan J. Smith wrote: >The scenario was ... > >1. Install new nForce 4/410 mainboard with SATA drives (new RAID-1) >and install Fedora Core 5, reserve C: and slice/format D: -- >automagically sets up GRUB, etc... > >2. Temporarily unplug SATA drives, put in 3Ware card and ATA drives >(old RAID-1) into new nForce 4/410 mainboard and load nForce drivers >under Windows XP Home/Pro > >3. Reconnect SATA drives, 3Ware card and ATA drives are now >secondary hard disk (SATA boots first), boot Fedora Core 5 and use dd >to copy C: and then find|cpio D:. > >4. Remove 3Ware card and ATA drives and boot Windows XP Home/Pro >from new SATA volume > Have you ever attempeted to use the BartPE boot disk (modified) to inject drivers into Windows instead of bothering with the whole boot-this, install-that, reboot? -Carter From wam at HiWAAY.net Mon Aug 28 09:56:36 2006 From: wam at HiWAAY.net (William A. Mahaffey III) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] [OT for Linux] Problems upgrading BIOS on GTW WIN2K box .... Message-ID: <44F2F614.10000@HiWAAY.net> .... I have a GTW box, 1.6 GHz P4, w/ WIN2K, SP4 (upgraded this A.M.). I am trying to upgrade the BIOS so that it will recognize an also-new 200 GB HDD that I added recently. When I 1st added the drive, WIN2K SP1 would only recognize ~131 GB of it, not the whole thing. At that time, I just setup a ~131 GB partition & moved on. Now I want the whole 200 GB, so I upgraded WIN2K from SP1 to SP4 (supposedly necessary to get > 131 GB HDD's recognized). I tried to setup the new drive from the newly upgraded WIN2K SP4 by deleting the old partition, but it (the drive setup wizard) still only said ~131 GB HDD, not 200 GB. Then I tried the BIOS upgrade. I followed the directions from GateWay exactly to create an upgrade floppy, but when I try to reboot, the screen says: "non-system disk or blank disk. Install new one & hit any key to restart". The floppy was/is a brand new, empty Sony floppy, just written to this A.M. for this purpose. Does this ring any bells w/ anyone ? TIA for any help .... -- William A. Mahaffey III ---------------------------------------------------------------------- "The M1 Garand is without doubt the finest implement of war ever devised by man." -- Gen. George S. Patton Jr. From jasonb at edseek.com Mon Aug 28 10:26:42 2006 From: jasonb at edseek.com (Jason Boxman) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Re: WinXP 64-Bit Software Raid Setup Howto? -- How Linux DeviceMapper handles FRAID ... In-Reply-To: <20060828074432.8090.qmail@web32906.mail.mud.yahoo.com> References: <200608272355.21933.jasonb@edseek.com> <20060828074432.8090.qmail@web32906.mail.mud.yahoo.com> Message-ID: <40366.216.134.200.78.1156775202.squirrel@nebula.internal.foo> Bryan J. Smith wrote: > Jason Boxman wrote: >> You speak of this often, but I don't understand. I haven't >> personally experienced lost MD volumes, so I don't understand how >> that will happen to me? > > Well I have. > > It is especially notorious if a drive does not come up once or twice > and then is reconnected, or a drive is connected on a different > channel, etc... In a nutshell, there is no "dynamic way" to keep > track of a slice and its volume association outside of the filesystem > itself. The volume is then "lost" -- e.g., a "broken mirror" in the > case of RAID-1 or "lost stripe" in the case of RAID-0. Okay, I understand what you're talking about now. I've been booting with an initrd, so I have mdadm proping drives for me on boot. Otherwise, I see how I'd lose my system when disks get lost or moved about and the kernel can't determine what has happened. > Hence the "common name" Microsoft chose for the Logical Disk Manager > "LDM" disk label -- the "Dynamic Disk." It allows for all sorts of > meta-data to be stored outside of the filesystem as part of the disk > label itself and associated "dynamically" by a booting kernel that > has no other info or interogation software. The legacy BIOS/DOS disk > does not, and relies on statically configured information. > > Again, the mdadm tools have helped quite a bit over the legacy > raidtools for MD, especially when it comes to re-scanning finding > those "lost/broken volumes." But for boot-time operations, that's > not an option. Then again, MD is difficult to use with boot-time > operations -- hence why most /boot filesystems are not located on > them for safety/usability. Do you not use mdadm in an initrd at boot time? With most distro kernels being of the initrd variety, I'd be surprised to find mdadm wasn't being included in a kernel if MD support was offered? Thanks. From tim at mcdonough.net Mon Aug 28 11:25:21 2006 From: tim at mcdonough.net (Tim McDonough) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Linksys WRT54G Config Question Message-ID: <44F30AE1.5030205@mcdonough.net> I know the WRT54G is a popular device but I don't have one and am trying to help a co-worker. I have what's probably a simple configuration question but the answer isn't apparent from their online manual. On my D-Link device I can edit what they call a "Virtual Server" so that anything that comes to a specific port on the WAN side gets routed to a specif IP address and a different port number on the internal network. For example if something connects to Port 2022 on the WAN side I can have it routed to 192.168.0.2, Port 22. How is this configured on the Linksys product? I see how to specify a certain machine for a certain service like web, ssh, etc. but it's not obvious (to me anyway) how to redirect to a different port as well. Tim From dmckenna at thelimucompany.com Mon Aug 28 12:51:32 2006 From: dmckenna at thelimucompany.com (Damien McKenna) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] FRAIDy cat In-Reply-To: <20060827193807.74433.qmail@web32905.mail.mud.yahoo.com> References: <20060827193807.74433.qmail@web32905.mail.mud.yahoo.com> Message-ID: <79835B19-C1C0-40BB-979E-68A5F098A401@thelimucompany.com> On Aug 27, 2006, at 3:38 PM, Bryan J. Smith wrote: > Actually, this weekend I just went FRAID-1 on my wife's computer In the past you have strongly recommended against using FRAID, why the change of heart? Is it just because of the transparent support through devicemapper? -- Damien McKenna - Web Developer - dmckenna@thelimucompany.com The Limu Company - http://www.thelimucompany.com/ - 407-804-1014 #include From thebs413 at gmail.com Mon Aug 28 18:21:34 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Using OS software RAID with FRAID organization, and when not to use it ... Message-ID: Damien McKenna wrote: > In the past you have strongly recommended against using FRAID, > why the change of heart? First off, there's been 0 change of heart. FRAID still _sucks_ for RAID-10, 3 and 5. In those configurations, you're talking 3-4+ disks and it's worth the money for a true hardware RAID card. Secondly, I _still_ advocate on _servers_ that you get a _real_ $200+ server mainboard with PCI64/-X and put in a $100 3Ware Escalade 8506-2LP card for "piece-of-mind." If you need hot-swap, there's _no_ other option for less money. But, third, several months ago, I _admitted_ for _desktops_ that PCIe x1 SATA channels now make software RAID-1 for 2 discs viable, as the impact on the rest of I/O is more minimized. I've played with using MD RAID-1 atop of LVM for some time now in 2 disc configuration. I generally avoid it for desktops because of the dual-boot considerations. > Is it just because of the transparent support through devicemapper? It's more than that. Again, narrowing the focus to _only_ 2 discs in RAID-1 for a _desktop_ system ... 1) First and foremost: you're using Linux kernel software RAID, not the FRAID driver 2) You not only solve the firmware/BIOS boot issue, but the GRUB/device mapping issue 3) You can dual-boot and not have issues, as Linux and Windows use the same organization Now I typically didn't even bother to try any solution out there. I had heard about DeviceMapper supporting the Intel ICH5-7 FRAID organization and the nVidia MCP4/4x0 FRAID organization more recently, but didn't want to bother. But I was putting in a new mainboard for my wife and the discs these days really overload the 32-bit PCI bus, all while those SATA channels are attached to their own, dedicated PCIe x1 channel. I still would _not_ use this on a server. I would _not_ trust it to handle single sector errors/remapping and notifications well either. You _want_ the "piece of mind" out of a 3Ware Escalade 8506-2LP for that. But if you just want something that gives you a mirrored run-time of your data so if a drive goes down you've still got your data, then it does that. Which means it's nice for desktops, where you can be down. Jason Boxman: > Okay, I understand what you're talking about now. > I've been booting with an initrd, so I have mdadm proping drives for me on > boot. Otherwise, I see how I'd lose my system when disks get lost or moved > about and the kernel can't determine what has happened. mdadm does fine when you haven't moved disks around or your disks haven't taken issue with organization. Once they do, it's far more manual of a process -- especially the manual scan you have to do with mdadm to re-write a new config file (and you're not always in a state that offers such ;). That's why I like to use MD atop of LVM. LVM stores information that can be used without a full scale (and not always "automagic") mdadm scan. I've yet to have a "lost" volume in MD since using LVM, and it's well worth it. Again, 99.9% of the problems you heard of LVM+MD is due to people putting LVM atop of MD -- which is still MD on a BIOS/DOS Disk Label (traditional partition table). *NEVER* do that. But with LVM2+DeviceMapper, the need for MD is pretty much dying. And DeviceMapper lets you leverage the firmware/BIOS boot of FRAID -- all while being OS software RAID-1 once Linux boots. It's far from perfect and I'd _never_ use it on a server -- but for a desktop that dual-boots, it's fine when I don't have PCI-X slots but a SATA controller SATA channels on dedicated PCIe x1 channels. From thebs413 at gmail.com Mon Aug 28 18:31:45 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Re: Using OS software RAID with FRAID organization, and when not to use it (in case I wasn't clear) ... Message-ID: On 8/28/06, Bryan J. Smith wrote: > First off, there's been 0 change of heart. FRAID still _sucks_ for > RAID-10, 3 and 5... > Again, narrowing the focus to _only_ 2 discs in RAID-1 for a _desktop_ system ... > I had heard about DeviceMapper supporting the Intel ICH5-7 FRAID > organization and the nVidia MCP4/4x0 FRAID organization more recently, This is for a 2-disc SATA RAID-1 volume, using the mainboard FRAID for firmware/boot, then relying on the DeviceMapper in Linux to match the FRAID driver in Windows. Test before you attempt. If you even attempt to use RAID-0, RAID-10, RAID-5, etc... you will probably run into stripe size issues/assumptions between DeviceMapper in Linux and the FRAID driver in Windows. I would not trust it for a moment because the FRAID vendor _might_ change things at any time in any firmware/BIOS update or Windows driver update. But for RAID-1, it's typically _not_ striped -- just "dumb" mirroring. At the most, the FRAID logic might reserve cylinder 0 for its own housekeeping, and then cylinder 1 actually becomes the Master Boot Record (MBR). Or, more likely, it's the last cylinder used for housekeeping, which doesn't do anything but reduce the total capacity by 1 cylinder. (I personally need to look at the low-level of the disk to find out what nVidia is doing, Intel if I have more time). > I still would _not_ use this on a server. I would _not_ trust it to > handle single sector errors/remapping and notifications well either. This is _key_ here. You do _not_ get sector remapping or fully transparent error handling. You now rely on the mapping to get such. In other words, be prepared for your system to _crash_ as a result of a device disconnect. That is, and has always been, the #1 issue with software RAID -- _period_ -- even if you are using native Linux MD or Windows LDM volume management. As such, this is regulated to uses where you want a _desktop_ system to maintain a mirror of its data. There is _no_ guarantee of consistency or "uptime" -- just that bytes on one disk are being mirrored to another. It's really "software RAID" with boot-time support via the FRAID firmware/BIOS -- and it's really limited to an application of 2 SATA disks in RAID-1 -- *NO* striping (that means no RAID-1e or RAID-10 with 2 discs either). If you even attempt RAID-5 with FRAID, not only don't I expect it to _not_ work, but you've basically turned your high-speed DMA transfer into sub-20MBps Programmed I/O (PIO) operations -- killing everything in your system. The benchmarks don't lie, do *NOT* use FRAID-5 on the Intel ICH7+ or nVidia MCP51 (4x0) _period_ -- not even under Windows. > You _want_ the "piece of mind" out of a 3Ware Escalade 8506-2LP for > that. On a server, the extra $100 for a mainboard with PCI-X and another $100 for a 3Ware Escalade 8506-2LP will save you time and headaches (as well as I/O waste -- although for 2-disc RAID-1, it's minor when your SATA is on its own PCIe x1). From bigjohn at midwest.net Mon Aug 28 18:33:16 2006 From: bigjohn at midwest.net (JohnH) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] fdisk recovery Message-ID: <00ac01c6caf1$f519ddc0$6401a8c0@3a5ah6vqcd> My VistaPrint Electronic Business CardPardon my short memory but............. I just ran a program called Kill FDisk and I lost everything on the wrong drive. I have used this program before but it's been so long I can't remember all the switches (if any). I though I could use a "killfdsk d:\" switch but it still KILLED both drives. Can I get my OS (98) back the way it was????? JohnH bigjohn@midwest.net -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.matrixlist.com/pipermail/pc_support/attachments/20060828/cf4c3cc4/attachment.html From jasonb at edseek.com Mon Aug 28 19:04:11 2006 From: jasonb at edseek.com (Jason Boxman) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Re: Using OS software RAID with FRAID organization, and when not to use it (in case I wasn't clear) ... In-Reply-To: References: Message-ID: <200608281904.11924.jasonb@edseek.com> On Monday 28 August 2006 18:31, Bryan J. Smith wrote: > > I still would _not_ use this on a server. I would _not_ trust it to > > handle single sector errors/remapping and notifications well either. > > This is _key_ here. You do _not_ get sector remapping or fully > transparent error handling. You now rely on the mapping to get such. > In other words, be prepared for your system to _crash_ as a result of > a device disconnect. That is, and has always been, the #1 issue with > software RAID -- _period_ -- even if you are using native Linux MD or > Windows LDM volume management. Sector remapping? -- Jason Boxman http://edseek.com/ - Linux and FOSS stuff From thebs413 at gmail.com Mon Aug 28 19:54:21 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Re: Using OS software RAID with FRAID organization, and when not to use it (in case I wasn't clear) ... Message-ID: Jason Boxman wrote: > Sector remapping? It's an ubiquitous SCSI concept that has always been applied to true hardware RAID cards. SCSI disks don't use all their tracks/sectors. They reserve several tracks and sectors for every cylinder as "extras." When sectors have to be re-read and that re-reading reaches a threshold (typically just a few times), the SCSI host adapter will remap that sector to an extra and mark the original sector bad. Early SCSI required you to remap in a manual procedure, whereas latter SCSI-2 does it automagically during run-time (that's a mega-oversimplification, and I'm a bit ignorant and/or forgot all of my SCSI-2 protocol functions/basics). NOTE: This is _below_ the filesystem. So from the standpoint of the OS filesystem itself, no sectors are bad. Virtually all true hardware RAID cards do the same with their blocking (multiple sectors grouped together, which is how the card spans disks), and more enterprise software RAID does as well. Because you're going to have sectors go bad on one disk but not exactly in the equivalent of the other. The true hardware RAID hides this from you, including reserving portions of the disk at volume creation. Hence why many people new to 3Ware cards complain about it not using the whole disk and shaving off a few hundred MBs. I haven't seen a FRAID card yet that does sector remapping. In a nutshell, most RAID-1 implementations -- software and hardware -- don't do blocking at all, so they don't offer remapping anyway. I know 3Ware cards do remapping, even on RAID-1, but I'm fairly certain they avoid it if at all possible and then do it at the end of the volume where they also store meta-data. I don't know if the DeviceMapper will mount a 3Ware RAID-1 volume with remapped sectors, but since the DeviceMapper does mount 3Ware RAID-10 volumes which does use blocking, it might support it. I want to say Linux's MD actually does remapping, because it's really just an extension of spanning. Although I might be thinking of LVM2 w/DeviceMapper, and one of the advantages of DeviceMapper over MD. I need to do some more research, I'm largely ignorant of what is and isn't supported these days in various software RAID implementations. But I just know every FRAID I've looked at doesn't, and relies on the OS' filesystem to mark its sectors/blocks bad. Bad sectors are far less common than in year's past. But your chances of having one are exponential in any striped/mirrored array by the number of disks -- and then that's a problem. Especially for FRAID, which is using hte "raw" disks and the OS is forced to mark all of the offset sectors bad (even when some are not). From thebs413 at gmail.com Mon Aug 28 20:08:29 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] AMD Turion x2 (Xpress 1100) v. Intel Core (945GM) ... surprising! Message-ID: Although it's not entirely fair comparing the newer Turion x2 and newer Xpress 1100 to the slightly aged Intel Core Duo with slightly aged 945GM chipset, the Turion x2 on a Xpress 1100 or GeForce 61x0 (possibly even a 7200) is typically cheaper or no more expensive than an Intel Core Duo on a measily Intel 945/950G solution today. So it's fair from a price standpoint. I'd personally like to see the GeForce 61x0 (or 7200) mobile too, but I'll take these Xpress 1100 benchmarks for now. AMD surprises, despite the smaller L2 cache (remember, it has 4x the L1 cache): http://www.tomshardware.com/2006/08/22/amd_dual_core_laptops_have_arrived/ Intel wins on power and select multi-tasking -- especially when combining in DVD playback. I don't know if that's an immaturity of the drivers of the Xpress 1100 or Intel's power focus on the chipset -- probably the latter as any chance of the former. Intel's tight integration is always going to be better in that regard, although with AMD buying out ATI, it does signal AMD's focus to deliver an integrated equivalent -- especially with more and more notebooks being sold. BTW, AnandTech _did_ find that even the new Core 2 Duo notebooks on a newer Intel chipset doesn't do much over the Core Duo: http://www.anandtech.com/mobile/showdoc.aspx?i=2663 Really all the Core 2 Duo for notebooks does is give you the Core Duo at higher clocks, but at a little bit of a boost in price. Although Intel does seem to be bringing the prices down, and the CPUs for notebooks should follow the desktop trends too. From damien at mc-kenna.com Mon Aug 28 20:22:28 2006 From: damien at mc-kenna.com (Damien McKenna) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] AMD Turion x2 (Xpress 1100) v. Intel Core (945GM) ... surprising! In-Reply-To: References: Message-ID: <44F388C4.8070304@mc-kenna.com> Bryan J. Smith wrote: > BTW, AnandTech _did_ find that even the new Core 2 Duo notebooks on a > newer Intel chipset doesn't do much over the Core Duo: It seems the key reason is that the Core2Duo really benefits from the increased system bandwidth that the desktop chipset, which the laptop chipset won't support until early next year. -- Damien McKenna, husband, father, geek. damien@mc-kenna.com - http://www.mc-kenna.com/ From damien at mc-kenna.com Mon Aug 28 20:23:20 2006 From: damien at mc-kenna.com (Damien McKenna) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Mac RAID In-Reply-To: References: Message-ID: <44F388F8.1090904@mc-kenna.com> Seeing as we're on the topic, based purely on a features list, what do you think of this: http://www.softraid.com/specs.html -- Damien McKenna, husband, father, geek. damien@mc-kenna.com - http://www.mc-kenna.com/ From thebs413 at gmail.com Mon Aug 28 21:23:26 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Mac RAID Message-ID: Damien McKenna wrote: > Seeing as we're on the topic, based purely on a features list, what do > you think of this: http://www.softraid.com/specs.html Ignoring this software product for the moment, understand Mac's -- like HP PA-RISC, SGI MIPS, Sun SPARCs, etc... -- are much, much easier to support with various software-based volume management because the vendor controls the _firmware_ as well as the OS. I've deployed countless HP, SGI and Sun software-based RAID solutions on their native, non-PC platforms before with great success (although there are a few exceptions). But getting back to this product, unless Apple sanctions it, I'd stick with what Apple offers natively. I don't know how much "software value add" I'd consider -- unless they have the partnerships like FalconStor, Veritas, etc... do. With that said, focusing on storage interconnect, IEEE1394 FireWire -- even on the Mac -- is _not_ a true replacement for enterprise SCSI. As a general, enterprise storage interconnect, it just suffers from disconnects and countless other issues -- and it was _never_ intended to be a multi-targettable storage interconnect, despite the Oracle hack. Now IEEE1394 a great, great consumer storage interconnect, and far, far better than USB, don't get me wrong. But SAS (Serial Attached SCSI) is the true enterprise SCSI replacement, not IEEE 1394 FireWire. So I'd only use it with SATA or SAS externally. *Avoid* FireWire and USB except as consumer external storage. From thebs413 at gmail.com Mon Aug 28 21:28:01 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] AMD Turion x2 (Xpress 1100) v. Intel Core (945GM) ... surprising! Message-ID: Damien McKenna wrote: > It seems the key reason is that the Core2Duo really benefits from the > increased system bandwidth that the desktop chipset, which the laptop > chipset won't support until early next year. Yeah, even the 950 doesn't help at all, which makes the Turion x2 a solid buy right now -- unless you're using your notebook for largely DVD playback (like on an airplane) which Intel's integrated power/video advantage clearly shows. I guess there are some tricks to the 965/975 in reducing memory latency and other interconnect benefits. Then again, I'd like to see this done again, on Linux. The Xpress 1100 is supposed to be better supported in Linux just after its release than the Xpress 200 was -- especially with the new southbridge peripheral logic. If ATI keeps it up and AMD integrates more and more, I might have to switch my preference to ATI from nVidia. We'll see, especially if AMD is able to solve some of the IP issues with its _true_ system interconnect with GPU with peer CPU architecture and open up the kernel driver -- which is a distinct possibility, unlike Intel's "software hacks" to emulate GPU as a CPU peer. From damien at mc-kenna.com Mon Aug 28 21:45:08 2006 From: damien at mc-kenna.com (Damien McKenna) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] AMD Turion x2 (Xpress 1100) v. Intel Core (945GM) ... surprising! In-Reply-To: References: Message-ID: <44F39C24.40004@mc-kenna.com> Bryan J. Smith wrote: > If ATI keeps it up and AMD integrates more and more, I might have to > switch my preference to ATI from nVidia. There was a press release / official statement from ATI/AMD recently regarding their drivers to the effect that they weren't going to boost their GPL support, and it instead suggested tightening the reigns on them. I'm having a hard time finding the URL though, I'll try to find it. -- Damien McKenna, husband, father, geek. damien@mc-kenna.com - http://www.mc-kenna.com/ From thebs413 at gmail.com Mon Aug 28 23:19:16 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] AMD Turion x2 (Xpress 1100) v. Intel Core (945GM) ... surprising! Message-ID: Damien McKenna wrote: > There was a press release / official statement from ATI/AMD recently > regarding their drivers to the effect that they weren't going to boost > their GPL support, and it instead suggested tightening the reigns on > them. I'm having a hard time finding the URL though, I'll try to find it. But _which_ drivers? 1) kernel-memory management? 2) LibGL? 3) X11? #2 and #3 is already _proprietary_ from not only ATI and nVidia, but Matrox and even some parts of Intel's own drivers. Yes, people don't realize not everything Intel has released is GPL or MIT -- and some of it is sacked with IP that is _incompatible_ with the GPOL. #1 is the major kernel issue, and part of it is Intel's IP/software hacks for non-CPU peripherals (i.e., AGP/PCIe connected). For AMD's interconnect with native HTX (non-peripheral AGP/PCIe), you don't need that Intel IP. Intel does _not_ include it in their own drivers (hence why the performance sucks on Linux versus Windows). So it wouldn't surprise me one bit if AMD/ATI had a GPL driver for the kernel, but then everything else was 100% proprietary. I sure with Linux bigots would stop and look at the details, instead of just screaming "I need it 100% GPL" constantly. Many companies release select things GPL and LGPL, other things under an OSD-compliant license and yet other things non-free software. In fact, I constantly defend companies that release most of their things GPL, LGPL and other OSD-compliant -- even if they have some proprietary for various/obvious reasons. E.g., nVidia. The only companies that bother me are the ones that release nothing GPL at all, and barely any OSD-compliant -- e.g., IBM. From thebs413 at gmail.com Wed Aug 30 10:40:50 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] tf@linuxceptional.com, justinkz@gmail.com, dmckenna@thelimucompany.com Message-ID: Tommy Foster wrote: > I just got a 250 gig Hitachi hard drive ... > This drive will be for backups, mostly (maybe exclusively). Just remember that with 3.5" commodity disk drives, you should keep them spindled regularly and do your best to avoid shock. 2.5" notebook and 3.5" enterprise disk drives can take more downtime and uptime, as well as shock (especially 2.5"). > I was going to stick it in one of those usb external drive enclosures. You can pretty much find enclosures these days with both FireWire and USB. FireWire has many advantages, not only in reliability but the fact that you don't have to "hunt" to guarantee you have an EHCI port for full speed USB (only about 1 out of every 3-5 ports is EHCI). But better yet now, SATA is also becoming more available in external enclosures. That gives you more options and far, far _superior_ performance. Simple bracket internal-to-external adapters are under $10. Is the drive SATA or ATA? > If I do, is there anything particularly wrong with having one xfs partition on it? Other than the fact that Windows can't read it, no, I don't see anything. But I haven't tested XFS with regular disconnects which could happen with an external device -- especially USB. > I also have an extra box. I could put a minimal os on the drive (would I even need > swap space?) and one huge directory named Twinkle Toes, or something. Writing to > the enclosure would be slow; rsyncing to the extra box would be much faster. It also gives you the advantage of being able to access it from any OS. That's why the 1-drive NAS appliances are getting popular versus external drives. I haven't done research but when I get time I'm going to start investigating all the cheap 1-drive NAS appliances out there, which ones could be flashed with a custom Linux version in EEPROM, or possibly just use the drive to store the OS (less ideal), etc... There's gotta be several projects out there to take advantage of these sub-$100 ARM/XScale, PowerPC, etc... driven enclosures. I just haven't researched them. I'm starting to like their portable nature after messing with a Synology DS-106 recently (a $279 unit, price w/o hard drive -- not one of the cheaper ones) and desiring more flexibility (especially SFTP access -- the Synology at least offers FTP-SSL/TLS). > I'm talking myself into using the extra box. any hints, warnings, ideas? > How about ten gigs for os + swap, and the rest for Twinkle Toes? LVM always makes resizing easy. > Or should I use the enclosure? what do you think? If you have an extra box, go for it. One of my 3Ware Escalade 6410 (older 4-channel ATA RAID-10) card finally went belly up after over 5 years (it regularly doesn't boot) in my dedicated home LAN backup server. From thebs413 at gmail.com Wed Aug 30 10:47:27 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] a new hard drive. so... Message-ID: Tommy Foster wrote: > I just got a 250 gig Hitachi hard drive ... > This drive will be for backups, mostly (maybe exclusively). Just remember that with 3.5" commodity disk drives, you should keep them spindled regularly and do your best to avoid shock. 2.5" notebook and 3.5" enterprise disk drives can take more downtime and uptime, as well as shock (especially 2.5"). > I was going to stick it in one of those usb external drive enclosures. You can pretty much find enclosures these days with both FireWire and USB. FireWire has many advantages, not only in reliability but the fact that you don't have to "hunt" to guarantee you have an EHCI port for full speed USB (only about 1 out of every 3-5 ports is EHCI). But better yet now, SATA is also becoming more available in external enclosures. That gives you more options and far, far _superior_ performance. Simple bracket internal-to-external adapters are under $10. Is the drive SATA or ATA? > If I do, is there anything particularly wrong with having one xfs partition on it? Other than the fact that Windows can't read it, no, I don't see anything. But I haven't tested XFS with regular disconnects which could happen with an external device -- especially USB. > I also have an extra box. I could put a minimal os on the drive (would I even need > swap space?) and one huge directory named Twinkle Toes, or something. Writing to > the enclosure would be slow; rsyncing to the extra box would be much faster. It also gives you the advantage of being able to access it from any OS. That's why the 1-drive NAS appliances are getting popular versus external drives. I haven't done research but when I get time I'm going to start investigating all the cheap 1-drive NAS appliances out there, which ones could be flashed with a custom Linux version in EEPROM, or possibly just use the drive to store the OS (less ideal), etc... There's gotta be several projects out there to take advantage of these sub-$100 ARM/XScale, PowerPC, etc... driven enclosures. I just haven't researched them. I'm starting to like their portable nature after messing with a Synology DS-106 recently (a $279 unit, price w/o hard drive -- not one of the cheaper ones) and desiring more flexibility (especially SFTP access -- the Synology at least offers FTP-SSL/TLS). > I'm talking myself into using the extra box. any hints, warnings, ideas? > How about ten gigs for os + swap, and the rest for Twinkle Toes? LVM always makes resizing easy. > Or should I use the enclosure? what do you think? If you have an extra box, go for it. One of my 3Ware Escalade 6410 (older 4-channel ATA RAID-10) card finally went belly up after over 5 years (it regularly doesn't boot) in my dedicated home LAN backup server. From Bruce.Metcalf at figzu.com Wed Aug 30 18:04:10 2006 From: Bruce.Metcalf at figzu.com (Bruce Metcalf) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Spindled? In-Reply-To: References: Message-ID: <7.0.1.0.0.20060830180214.037a8160@figzu.com> At 10:40 AM 8/30/2006, Bryan J. Smith wrote: >Tommy Foster wrote: >>I just got a 250 gig Hitachi hard drive ... >>This drive will be for backups, mostly (maybe exclusively). > >Just remember that with 3.5" commodity disk drives, you should keep >them spindled regularly and do your best to avoid shock. Would someone please explain the term "spindled" in this context? Last time I "spindled" something it was a Hollerith card for a keypunch. I'm not at all sure I'd want to do that to a disk drive, even if I *was* strong enough to bend it into a circle. Bruce Metcalf From thebs413 at gmail.com Wed Aug 30 20:48:23 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Spindled? Message-ID: Bruce.Metcalf wrote: > Would someone please explain the term "spindled" in this context? Commodity disk drives should be spun up on a regular basis -- like no less than once every few days in the absolute worst case. From thebs413 at gmail.com Wed Aug 30 20:59:09 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] a new hard drive. so... -- I avoid retail (consumer) hard drives, only buy 24x7 enterprise ... Message-ID: In general, I've been avoiding _all_ retail hard drives (sans one**) because they are _consumer_ grade hard drives. No, I don't buy consumer OEM hard drives either, I buy _enterprise_ OEM hard drives. What are those? Basically they are the cream of the consumer models, tested to better tolerances, thermal, vibration, etc... for 24x7 operation (instead of the 8x5 of their consumer models). Here are some example models ... - Hitachi T7K (based on consumer-grade 7K) - Seagate NL35/Barracuda ES (based on consumer-grade Barracuda 7200.x) - Western Digital Caviar RE/RE2 (based on consumer-grade SE/SE2) Hitachi charges a premium on the T7K. Seagate charges even more of a premium on the NL35 which is now being replaced by the Barracuda ES that isn't so pricey. But by far the best buy is the Western Digital Caviar RE/RE2 series -- typically only 15-30% more than the consumer SE/SE2. You can get a 320GB Western Digital Caviar RE/RE2 for just over $100 -- no rebates, no non-sense. Enterprise-rated 24x7. -- Bryan **NOTE: I still _do_ buy the consumer-grade, retail 200GB Seagate 7200.7 ST3200822A when I see it, and have been for over 2 years. I now own 14 of them (8 on a 3Ware Escalade 7800 and 5 on a NetCell SR5000). Nearly all of them had rebates -- typically from Seagate, possibly via Circuit City or CompUSA. They have _all_ been honored (although I'm just filling one out now and 2 more are coming). As long as I don't buy them the same week/under the same rebate offer, they have honor them -- even if it's on-sale subsequent weeks. From Bruce.Metcalf at figzu.com Wed Aug 30 22:42:48 2006 From: Bruce.Metcalf at figzu.com (Bruce Metcalf) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Re: Spindled? In-Reply-To: References: Message-ID: <7.0.1.0.0.20060830224024.037f4910@figzu.com> Bryan wrote: >>Would someone please explain the term "spindled" in this context? > >Commodity disk drives should be spun up on a regular basis -- like >no less than once every few days in the absolute worst case. Aha, that makes sense! I shall have to look into obtaining a small utility that does that. My present collection of drives were all purchased before Leap lead me to understand that there were two classes of drive. Thanks! Bruce From b.j.smith at ieee.org Thu Aug 31 01:11:02 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Re: Spindled? In-Reply-To: <7.0.1.0.0.20060830224024.037f4910@figzu.com> References: <7.0.1.0.0.20060830224024.037f4910@figzu.com> Message-ID: <1157001062.2995.11.camel@bert64.oviedo.smithconcepts.com> On Wed, 2006-08-30 at 22:42 -0400, Bruce Metcalf wrote: > Aha, that makes sense! > I shall have to look into obtaining a small utility that does that. > My present collection of drives were all purchased before Leap lead > me to understand that there were two classes of drive. Actually, there are many classes of drives. - Commodity and Near-Line Enterprise First off, what you get in the stores are Commodity Consumer 8x5 disks. They are the "middle barrel" which are those that tested well enough to offer a Consumer retail warranty (typically 3-5 years) for 8x5 operation. Some vendors even keep statistics in firmware to see if you have used them regularly for more than 14x5, ambient, etc... Secondly, many of the newer 24x7 "Near-Line Rated" or "RAID" disks still come off the _exact_same_ assembly lines as the "Commodity" 8x5 disks. They are just tested to better tolerances. But they are certainly the "top of the barrel." You can _only_ find them in OEM flavors, but they _always_ have a 5-year warranty and using them 24x7 does _not_ void it. Third, desktops from tier-1 OEMs (like Dell) get the "bottom of the barrel." Expect a _significant_portion_ of them to go "belly up" rather quickly. - True Enterprise *NOW* for *TRUE* "Enterprise Rated" disks, you typically need to spend $1+/GB. These are the 73, 146GB and newer 292GB (sometimes labeled 300GB) capacities and have spindle speeds of 10+Krpm. They have easily 5-10x less vibration due to their superior construction and materials -- at the added cost. They are typically available in only SCSI, FC or SAS flavors, although a few vendors offer them as non-SAS SATA. E.g., the SATA version of the Hitachi UltraStar 10k is the Western Digital Raptor 10k. -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From b.j.smith at ieee.org Thu Aug 31 01:14:28 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Re: Spindled? -- Commodity v. Near-Line v. Enterprise (continued) In-Reply-To: <1157001062.2995.11.camel@bert64.oviedo.smithconcepts.com> References: <7.0.1.0.0.20060830224024.037f4910@figzu.com> <1157001062.2995.11.camel@bert64.oviedo.smithconcepts.com> Message-ID: <1157001268.2995.15.camel@bert64.oviedo.smithconcepts.com> On Thu, 2006-08-31 at 01:11 -0400, Bryan J. Smith wrote: > Actually, there are many classes of drives. I gave a brief intro to these manufacturing realities in a sidebar here: http://www.samag.com/documents/sam0509a/0509a_s1.htm With associated table: http://www.samag.com/documents/sam0509a/0509a_t2.htm It was part of a greater article on virtual tape libraries: http://www.samag.com/documents/sam0509a/ Including this table on comparing different types of media for _off-line_ storage: http://www.samag.com/documents/sam0509a/0509a_t1.htm -- Bryan J. Smith Professional, technical annoyance mailto:b.j.smith@ieee.org http://thebs413.blogspot.com --------------------------------------------------------- The world is in need of solutions. Unfortunately, people seem to be more interested in blindly aligning themselves with one of only two viewponts -- an "us v. them" debate that has nothing to do with finding an actual solution. From thebs413 at gmail.com Thu Aug 31 14:11:58 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] LCD 19" 8ms DVI+VGA for $205 - $100 rebate with free shipping and 3-piece luggage set ... Message-ID: Okay, this is a wacko deal, but still pretty awesome overall for $205 plus tax (free shipping) and then $100 off in a mail-in rebate ... http://dealnews.com/deals/Hanns-G-HU-196-D-19-LCD-Monitor-w-DVI-luggage-set-for-105-shipped-after-rebate/130829.html Have _no_idea_ about the quality of the LCD, but it's 8ms response and has _both_ DVI and VGA inputs with the typical 19" 1280x1024 resolution. From justinkz at gmail.com Thu Aug 31 17:14:37 2006 From: justinkz at gmail.com (Justin M. Keyes) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Is a SATA/1.5 RAID controller compatible with SATA/3.0 drives? Message-ID: <53b562310608311414ja61b8f8o7f82c62fed6c7b6@mail.gmail.com> I've been meaning to fix my backup situation at home, and now seems like the perfect time to take advantage of the information flying around the list lately. So, I like this Western Digital RE SATA/3.0 HDD (my computer has legacy ATX power supply): http://www.newegg.com/Product/Product.asp?item=N82E16822136055 and this 3ware controller (my computer only has PCI): http://www.newegg.com/Product/Product.asp?Item=N82E16816116030 but I am new to this SATA stuff. Would that 3ware card work with that hard drive? Thanks, Justin M. Keyes From dave at dgnal.net Thu Aug 31 17:52:03 2006 From: dave at dgnal.net (David Simmons) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Re: WinXP 64-Bit Software Raid Setup Howto? -- How Linux DeviceMapper handles FRAID ... In-Reply-To: <20060828074432.8090.qmail@web32906.mail.mud.yahoo.com> References: <200608272355.21933.jasonb@edseek.com> <20060828074432.8090.qmail@web32906.mail.mud.yahoo.com> Message-ID: <18249.192.104.67.222.1157061123.squirrel@qtmail.dgnal.net> Ok...to (complete) the answer my own question..grin (as the responses have helped me to realize what needed to happen). Seemed also like everyone was saying what OS could do it EXCEPT WinXP 64-Bit...so maybe this will help someone doing the same. Whereas Windows' Help was pretty useless (typed in RAID for a search string - returned nothing!?)...I was able to enter the string 'Striped' and it found instructions on how to do it. In a nutshell - my problem was that I already has the space partitioned...it needed unpartitioned/empty space on DYNAMIC disks (not the normal BASIC type). Hope that helps to complete the thread - thanks to the many thorough responses. dave From justinkz at gmail.com Thu Aug 31 18:37:11 2006 From: justinkz at gmail.com (Justin M. Keyes) Date: Tue Oct 31 13:18:01 2006 Subject: [Pc_Support] Re: Is a SATA/1.5 RAID controller compatible with SATA/3.0 drives? In-Reply-To: <53b562310608311414ja61b8f8o7f82c62fed6c7b6@mail.gmail.com> References: <53b562310608311414ja61b8f8o7f82c62fed6c7b6@mail.gmail.com> Message-ID: <53b562310608311537ucc8f46fpe4fb50ef67d5dbbb@mail.gmail.com> On 8/31/06, Justin M. Keyes wrote: > I've been meaning to fix my backup situation at home, and now seems > like the perfect time to take advantage of the information flying > around the list lately. So, I like this Western Digital RE SATA/3.0 > HDD (my computer has legacy ATX power supply): > http://www.newegg.com/Product/Product.asp?item=N82E16822136055 > > and this 3ware controller (my computer only has PCI): > http://www.newegg.com/Product/Product.asp?Item=N82E16816116030 > > but I am new to this SATA stuff. Would that 3ware card work with that > hard drive? Ah, nevermind, it looks like SATA/3.0 is backward compatible with SATA/1.5: http://en.wikipedia.org/wiki/SATA_II Thanks. -- Justin M. Keyes From thebs413 at gmail.com Thu Aug 31 19:04:33 2006 From: thebs413 at gmail.com (Bryan J. Smith) Date: Tue Oct 31 13:18:02 2006 Subject: [Pc_Support] Re: WinXP 64-Bit Software Raid Setup Howto? -- did I not answer these immediately? Message-ID: David Simmons wrote: > Ok...to (complete) the answer my own question..grin (as the responses have > helped me to realize what needed to happen). Huh? http://lists.leap-cf.org/pipermail/pc_support/2006-August/002424.html > Seemed also like everyone was saying what OS could do it EXCEPT WinXP > 64-Bit...so maybe this will help someone doing the same. Yes, I have _never_ messed with Windows XP x64. Although it is well-known that many FRAID cards _lack_ Windows XP x64 software drivers. > Whereas Windows' Help was pretty useless (typed in RAID for a search > string - returned nothing!?)...I was able to enter the string 'Striped' > and it found instructions on how to do it. Again, from: http://lists.leap-cf.org/pipermail/pc_support/2006-August/002424.html To end the e-mail ... "Only striping and spanning (concating) is supported on non-server versions. You set it up via Disk Manager." > In a nutshell - my problem was that I already has the space > partitioned...it needed unpartitioned/empty space on DYNAMIC disks (not > the normal BASIC type). Again, from: http://lists.leap-cf.org/pipermail/pc_support/2006-August/002424.html At the _very_head_ of that e-mail ... "First off, NT5.x (200x/XP) requires use a LDM Disk Label (Dynamic Disc) for any volume management." I don't know how much more direct I can be than that. > Hope that helps to complete the thread - thanks to the many thorough > responses. I thought I answered your questions immediately after you asked them. Maybe I just don't know how to communicate well via e-mail? From dave at dgnal.net Thu Aug 31 23:42:39 2006 From: dave at dgnal.net (David Simmons) Date: Tue Oct 31 13:18:02 2006 Subject: [Pc_Support] Re: WinXP 64-Bit Software Raid Setup Howto? -- did I not answer these immediately? In-Reply-To: References: Message-ID: <29478.71.252.176.10.1157082159.squirrel@qtmail.dgnal.net> >> Ok...to (complete) the answer my own question..grin (as the responses >> have helped me to realize what needed to happen). > > Huh? > http://lists.leap-cf.org/pipermail/pc_support/2006-August/002424.html Bryan - I realized (and was grateful) for the previous postings...unfortunately - it was still voodoo for me...Guess what I needed was a here's what NOT to do (as I took two steps forward but needed to take one back to make any progress) > To end the e-mail ... > "Only striping and spanning (concating) is supported on non-server > versions. You set it up via Disk Manager." And that's what really frustrated me!? As I was sure you were right - I just couldn't find it! And the problem is that the raid/stiping options just don't show up at all!? (I would have thought they would have been greyed out or something - then I would have felt better). And having NOTHING show up in the help when you enter the word 'RAID' seems kinda ascinine?! It didn't even have links to mirroring and striping - so you have to know first - but if you knew, why would you need help? Guess I was thinking it would be as friendly as Linux - as 'help raid' supplies quite a few GREAT nuggests to start! > At the _very_head_ of that e-mail ... > "First off, NT5.x (200x/XP) requires use a LDM Disk Label (Dynamic > Disc) for any volume management." > > I don't know how much more direct I can be than that. Except for the fact I thought you were talking about LVM....saying that I needed to create a LVM in linux and then create in WinXP64...seemed kinda backwards - but then that was my whole problem. > Maybe I just don't know how to communicate well via e-mail? I don't think it's that - as I've enjoyed your technical/hardware emails for years! Guess I needed some hints as to what NOT to do - like if it said, "Make sure your Disk Type is NOT basic - but converted to Dynamic" that would've turned some lights on....but like I said - once I read from your emails it was there - that gave me confidence to keep horsing with it (when in town - which was only about 3 days the past 3 weeks). Thanks again! - dave