From gmk at runlevelzero.net Fri Nov 4 11:20:03 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Fri, 4 Nov 2005 11:20:03 -0800 Subject: [cAos] caos2 repository updates Message-ID: <20051104192003.GA17754@titan.runlevelzero.net> Lots of updates including e17 importation, httpd/apr security update, and warewulf maintaince update: current: + SRPMS/apr-0.9.7-3.src.rpm + SRPMS/apr-util-0.9.7-5.src.rpm + SRPMS/e_modules-0.0.1-1.20051017.src.rpm + SRPMS/e_utils-0.0.1-1.20051017.src.rpm + SRPMS/ecore-0.9.9.018-1.20051017.src.rpm + SRPMS/edb-1.0.5.004-1.20051017.src.rpm + SRPMS/edje-0.5.0.018-1.20051017.src.rpm + SRPMS/eet-0.9.10.018-1.20051017.src.rpm + SRPMS/elation-0.0.1.001-1.20051017.src.rpm + SRPMS/elicit-0.9.0-1.20051017.src.rpm + SRPMS/embryo-0.9.1.018-1.20051017.src.rpm + SRPMS/emotion-0.0.1.004-1.20051017.src.rpm + SRPMS/engage-0.0.9-1.20051017.src.rpm + SRPMS/engrave-0.1.0-1.20051017.src.rpm + SRPMS/enlightenment-0.16.999.018-1.20051017.src.rpm + SRPMS/enscribe-0.1-1.20051017.src.rpm + SRPMS/enterminus-0.1-1.src.rpm + SRPMS/entice-0.9.3.004-1.20051017.src.rpm + SRPMS/entrance-0.9.0.004-2.20051017.src.rpm + SRPMS/epeg-0.9.0.004-1.20051017.src.rpm + SRPMS/epsilon-0.3.0.004-1.20051017.src.rpm + SRPMS/esmart-0.9.0.004-1.20051017.src.rpm + SRPMS/etox-0.9.0.004-1.20051017.src.rpm + SRPMS/evas-0.9.9.019-1.20051017.src.rpm + SRPMS/evidence-0.9.8-1.20051017.src.rpm + SRPMS/ewl-0.0.4.004-1.20051017.src.rpm + SRPMS/examine-0.0.1-1.20051017.src.rpm + SRPMS/express-0.0.0-1.20051017.src.rpm + SRPMS/gnuplot-4.0.0-7.caos.src.rpm + SRPMS/httpd-2.0.55-11.src.rpm + SRPMS/iconbar-0.9.1-1.20051017.src.rpm + SRPMS/imlib2-1.2.1.009-1.20051017.src.rpm + SRPMS/imlib2_loaders-1.2.1.010-1.20051017.src.rpm + SRPMS/perl-Tk-804.027-3.caos.src.rpm + SRPMS/sysfsutils-1.3.0-1.src.rpm + SRPMS/warewulf-2.4.5-1.src.rpm - SRPMS/apr-0.9.6-1.src.rpm - SRPMS/apr-0.9.6-2.src.rpm - SRPMS/apr-util-0.9.6-1.src.rpm - SRPMS/apr-util-0.9.6-4.src.rpm - SRPMS/edb-1.0.5-2.20041124.src.rpm - SRPMS/enlightenment-0.16.7.2-2.caos.src.rpm - SRPMS/epeg-0.9.0-1.20050224.src.rpm - SRPMS/epeg-0.9.0.004-1.20050830.src.rpm - SRPMS/evoak-0.0.1-1.20041124.src.rpm - SRPMS/gnuplot-4.0.0-6.caos.src.rpm - SRPMS/httpd-2.0.53-10.src.rpm - SRPMS/perl-Tk-804.027-2.caos.src.rpm - SRPMS/warewulf-2.4.4-1.src.rpm stable: + SRPMS/apr-0.9.7-3.src.rpm + SRPMS/apr-util-0.9.7-5.src.rpm + SRPMS/e_modules-0.0.1-1.20051017.src.rpm + SRPMS/e_utils-0.0.1-1.20051017.src.rpm + SRPMS/edb-1.0.5.004-1.20051017.src.rpm + SRPMS/elation-0.0.1.001-1.20051017.src.rpm + SRPMS/elicit-0.9.0-1.20051017.src.rpm + SRPMS/engrave-0.1.0-1.20051017.src.rpm + SRPMS/enlightenment-0.16.999.018-1.20051017.src.rpm + SRPMS/enscribe-0.1-1.20051017.src.rpm + SRPMS/enterminus-0.1-1.src.rpm + SRPMS/entrance-0.9.0.004-2.20051017.src.rpm + SRPMS/epeg-0.9.0.004-1.20051017.src.rpm + SRPMS/evidence-0.9.8-1.20051017.src.rpm + SRPMS/examine-0.0.1-1.20051017.src.rpm + SRPMS/express-0.0.0-1.20051017.src.rpm + SRPMS/gnuplot-4.0.0-7.caos.src.rpm + SRPMS/httpd-2.0.55-11.src.rpm + SRPMS/iconbar-0.9.1-1.20051017.src.rpm + SRPMS/perl-Tk-804.027-3.caos.src.rpm + SRPMS/sysfsutils-1.3.0-1.src.rpm + SRPMS/warewulf-2.4.5-1.src.rpm - SRPMS/apr-0.9.6-1.src.rpm - SRPMS/apr-util-0.9.6-1.src.rpm - SRPMS/edb-1.0.5-2.20041124.src.rpm - SRPMS/enlightenment-0.16.7.2-2.caos.src.rpm - SRPMS/epeg-0.9.0-1.20050224.src.rpm - SRPMS/evoak-0.0.1-1.20041124.src.rpm - SRPMS/gnuplot-4.0.0-6.caos.src.rpm - SRPMS/httpd-2.0.53-10.src.rpm - SRPMS/iconbar-0.9.1-2.20050224.src.rpm - SRPMS/perl-Tk-804.027-2.caos.src.rpm - SRPMS/warewulf-2.4.4-1.src.rpm -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From gmk at runlevelzero.net Fri Nov 11 12:44:31 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Fri, 11 Nov 2005 12:44:31 -0800 Subject: [cAos] caos2 repository update Message-ID: <20051111204431.GB26837@titan.runlevelzero.net> current: - SRPMS/openssh-3.8.1p1-8.caos.src.rpm stable: + SRPMS/openssh-4.2p1-2.caos.src.rpm - SRPMS/openssh-3.8.1p1-8.caos.src.rpm Coming soon to a mirror near you! -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From gmk at runlevelzero.net Fri Nov 11 17:24:04 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Fri, 11 Nov 2005 17:24:04 -0800 Subject: [cAos] caos2 repository update Message-ID: <20051112012404.GC28028@titan.runlevelzero.net> current: + SRPMS/apache-1.3.34-1.caos.src.rpm - SRPMS/apache-1.3.33-2.caos.src.rpm stable: + SRPMS/apache-1.3.34-1.caos.src.rpm - SRPMS/apache-1.3.33-2.caos.src.rpm Please don't make a mess. -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From gmk at runlevelzero.net Fri Nov 11 21:08:29 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Fri, 11 Nov 2005 21:08:29 -0800 Subject: [cAos] caos2 repository updates Message-ID: <20051112050829.GB28630@titan.runlevelzero.net> current: - SRPMS/curl-7.12.1-2.caos.src.rpm stable: + SRPMS/curl-7.15.0-1.caos.src.rpm - SRPMS/curl-7.12.1-2.caos.src.rpm Enjoy. -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From gmk at runlevelzero.net Sat Nov 12 07:29:40 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Sat, 12 Nov 2005 07:29:40 -0800 Subject: [cAos] caos2 repository update Message-ID: <20051112152940.GA30185@titan.runlevelzero.net> current: + SRPMS/bind-9.3.1-1.caos.src.rpm + SRPMS/wget-1.9.1-22.src.rpm - SRPMS/bind-9.2.3-4.caos.src.rpm - SRPMS/wget-1.9.1-7.caos.src.rpm stable: + SRPMS/bind-9.3.1-1.caos.src.rpm + SRPMS/wget-1.9.1-22.src.rpm - SRPMS/bind-9.2.3-4.caos.src.rpm - SRPMS/wget-1.9.1-7.caos.src.rpm -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From gmk at runlevelzero.net Sat Nov 12 21:38:48 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Sat, 12 Nov 2005 21:38:48 -0800 Subject: [cAos] caos2 repository updates Message-ID: <20051113053848.GB31518@titan.runlevelzero.net> current: + SRPMS/sudo-1.6.8p12-32.caos.src.rpm - SRPMS/sudo-1.6.8-31.caos.src.rpm stable: + SRPMS/sudo-1.6.8p12-32.caos.src.rpm - SRPMS/sudo-1.6.8-31.caos.src.rpm -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From gmk at runlevelzero.net Mon Nov 14 21:32:36 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Mon, 14 Nov 2005 21:32:36 -0800 Subject: [cAos] caos2 repository update Message-ID: <20051115053236.GB23762@titan.runlevelzero.net> current: + SRPMS/ports-0.2-2.caos.src.rpm - SRPMS/ports-0.2-1.caos.src.rpm stable: + SRPMS/ports-0.2-2.caos.src.rpm - SRPMS/ports-0.2-1.caos.src.rpm Coming sooon, to a repository near you! -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From benjamin at pythagoras.no-ip.org Fri Nov 18 04:19:57 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Fri, 18 Nov 2005 12:19:57 +0000 Subject: [cAos] Quick question... Message-ID: <32d4d46a605edfe9ee175af2c98c28da@www.pythagoras.no-ip.org> Will cAos work with multiple processors? I'm about to take delivery of a quad Xeon box and would like to give cAos a go! Many thanks, -- Benjamin benjamin at pythagoras.no-ip.org From stefano.bridi at gmail.com Fri Nov 18 05:58:17 2005 From: stefano.bridi at gmail.com (Stefano Bridi) Date: Fri, 18 Nov 2005 14:58:17 +0100 Subject: [cAos] Quick question... In-Reply-To: <32d4d46a605edfe9ee175af2c98c28da@www.pythagoras.no-ip.org> References: <32d4d46a605edfe9ee175af2c98c28da@www.pythagoras.no-ip.org> Message-ID: 2005/11/18, Benjamin Donnachie : > Will cAos work with multiple processors? I'm about to take delivery of a > quad Xeon box and would like to give cAos a go! > AFAIK cAos work well with multiple cpu, but i have tried only machine with 2 Xeon (with HT enabled) that appears as 4 cpu. The only limitation i have found in the cAos kernel is the support to only 4GB of memory. I have recompiled it with the 64GB support and all works fine. bye stef From gmk at runlevelzero.net Fri Nov 18 06:43:20 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Fri, 18 Nov 2005 06:43:20 -0800 Subject: [cAos] Quick question... In-Reply-To: References: <32d4d46a605edfe9ee175af2c98c28da@www.pythagoras.no-ip.org> Message-ID: <20051118144319.GA3947@titan.runlevelzero.net> On Fri, Nov 18, 2005 at 02:58:17PM +0100, Stefano Bridi wrote: > 2005/11/18, Benjamin Donnachie : > > Will cAos work with multiple processors? I'm about to take delivery of a > > quad Xeon box and would like to give cAos a go! > > cAos will run fine on that. :) I have it running very well on multisocket, multicore and NUMA architectures. For that system, be sure you install with the 'smp' option or after install 'yum install linux-smp'. This system is x86_64, right? > AFAIK cAos work well with multiple cpu, but i have tried only machine > with 2 Xeon (with HT enabled) that appears as 4 cpu. > The only limitation i have found in the cAos kernel is the support to > only 4GB of memory. > I have recompiled it with the 64GB support and all works fine. IIRC, the 32 bit kernels are limited, but the x86_64 kernels are not. -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From stefano.bridi at gmail.com Fri Nov 18 06:58:52 2005 From: stefano.bridi at gmail.com (Stefano Bridi) Date: Fri, 18 Nov 2005 15:58:52 +0100 Subject: [cAos] Quick question... In-Reply-To: <20051118144319.GA3947@titan.runlevelzero.net> References: <32d4d46a605edfe9ee175af2c98c28da@www.pythagoras.no-ip.org> <20051118144319.GA3947@titan.runlevelzero.net> Message-ID: 2005/11/18, Greg M. Kurtzer : > > AFAIK cAos work well with multiple cpu, but i have tried only machine > > with 2 Xeon (with HT enabled) that appears as 4 cpu. > > The only limitation i have found in the cAos kernel is the support to > > only 4GB of memory. > > I have recompiled it with the 64GB support and all works fine. > > IIRC, the 32 bit kernels are limited, but the x86_64 kernels are not. Many thanks for the info! I suppose that but an official voice is better :-) bye stef From astevens at gravitypark.com Fri Nov 18 12:08:09 2005 From: astevens at gravitypark.com (Arthur Stevens) Date: Fri, 18 Nov 2005 12:08:09 -0800 Subject: [cAos] Quick question... References: <32d4d46a605edfe9ee175af2c98c28da@www.pythagoras.no-ip.org><20051118144319.GA3947@titan.runlevelzero.net> Message-ID: <004001c5ec7b$cc5da970$6800a8c0@astevens> I can verify it runs great on the ----- Original Message ----- From: "Stefano Bridi" To: "Community Assembled OS" Sent: Friday, November 18, 2005 6:58 AM Subject: Re: [cAos] Quick question... > 2005/11/18, Greg M. Kurtzer : > >> > AFAIK cAos work well with multiple cpu, but i have tried only machine >> > with 2 Xeon (with HT enabled) that appears as 4 cpu. >> > The only limitation i have found in the cAos kernel is the support to >> > only 4GB of memory. >> > I have recompiled it with the 64GB support and all works fine. >> >> IIRC, the 32 bit kernels are limited, but the x86_64 kernels are not. > > Many thanks for the info! I suppose that but an official voice is better > :-) > > bye > stef > _______________________________________________ > cAos mailing list > cAos at caosity.org > http://lists.caosity.org/mailman/listinfo/caos > From astevens at gravitypark.com Fri Nov 18 12:24:30 2005 From: astevens at gravitypark.com (Arthur Stevens) Date: Fri, 18 Nov 2005 12:24:30 -0800 Subject: [cAos] Quick question... References: <32d4d46a605edfe9ee175af2c98c28da@www.pythagoras.no-ip.org><20051118144319.GA3947@titan.runlevelzero.net> <004001c5ec7b$cc5da970$6800a8c0@astevens> Message-ID: <002e01c5ec7e$147dcb70$6800a8c0@astevens> Sorry about the last mail, lets try that again :) Hi Benjamin, It runs very well on the xeon. The EMT64 equipped xeons will greatly benefit from the 64 bit version. Basically if it's a brand new system, it should have EMT64, HT and use the 64 bit version of cAos; while if it is an older xeon system (like my dual 2.0 ghz xeon), the 32 bit version is what you will want. Feel free to mail me if you need any help setting it up. Thanks for giving cAos a run, we hope to see you around. Arthur Intel Link to see if your processor has EMT64 http://indigo.intel.com/compare_cpu/default.aspx?familyID=5&culture=en-US ----- Original Message ----- From: "Benjamin Donnachie" <> To: "Community Assembled OS" Sent: Friday, November 18, 2005 4:19 AM Subject: [cAos] Quick question... > Will cAos work with multiple processors? I'm about to take delivery of a > quad Xeon box and would like to give cAos a go! > > Many thanks, > > -- > Benjamin > benjamin at pythagoras.no-ip.org > > > _______________________________________________ > cAos mailing list > cAos at caosity.org > http://lists.caosity.org/mailman/listinfo/caos > From astevens at gravitypark.com Fri Nov 18 12:22:36 2005 From: astevens at gravitypark.com (Arthur Stevens) Date: Fri, 18 Nov 2005 12:22:36 -0800 Subject: [cAos] Quick question... References: <32d4d46a605edfe9ee175af2c98c28da@www.pythagoras.no-ip.org> Message-ID: <004501c5ec7d$d0b35d10$6800a8c0@astevens> Hi Benjamin, It runs very well on the xeon. The EMT64 equipped xeons will greatly benefit from the 64 bit version. Basically if it's a brand new system, it should have EMT64, HT and use the 64 bit version of cAos; while if it is an older xeon system (like my dual 2.0 ghz xeon), the 32 bit version is what you will want. Feel free to mail me if you need any help setting it up. Thanks for giving cAos a run, we hope to see you around. Arthur Intel Link to see if your processor has EMT64 http://indigo.intel.com/compare_cpu/default.aspx?familyID=5&culture=en-US ----- Original Message ----- From: "Benjamin Donnachie" <> To: "Community Assembled OS" Sent: Friday, November 18, 2005 4:19 AM Subject: [cAos] Quick question... > Will cAos work with multiple processors? I'm about to take delivery of a > quad Xeon box and would like to give cAos a go! > > Many thanks, > > -- > Benjamin > benjamin at pythagoras.no-ip.org > > > _______________________________________________ > cAos mailing list > cAos at caosity.org > http://lists.caosity.org/mailman/listinfo/caos > From benjamin at pythagoras.no-ip.org Fri Nov 18 12:47:44 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Fri, 18 Nov 2005 20:47:44 +0000 Subject: [cAos] Quick question... Message-ID: Benjamin Donnachie wrote: >Will cAos work with multiple processors? I'm about to take delivery of a >quad Xeon box and would like to give cAos a go! Many thanks for all the replies! I thought that you needed a completely separate download for smp support and was worried when I couldn't find one for cAos! The box is a bit old with each processor rated at about 500MHz, but it should certainly be better than the 1.2GHz Celeron that I've got currently doing the job! :-) I'm currently using Fedora Core 3, so I expect I'll be asking lots of daft questions soon! Take care, -- Benjamin benjamin at pythagoras.no-ip.org From benjamin at pythagoras.no-ip.org Sat Nov 19 07:07:52 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Sat, 19 Nov 2005 15:07:52 +0000 Subject: [cAos] Quick question... Message-ID: Benjamin Donnachie wrote: >I'm currently using Fedora Core 3, so I expect I'll be asking lots of daft >questions soon! I've already run into a spot of bother! The server is a Compaq Proliant 6400 quad Pentium III Xeon server and, unfortunately, the cAos install disk does not recognise the Compaq SCSI RAID Array. I gave FC4 a go and noted what module it used, but cAos didn't have exactly the same one and the one I tried didn't seem to work. Also, FC4 only recognised one processor... So I'm expecting cAos to do the same... :-( I can't be more specific as I had to dash off to work. I shall have another look at it tonight and see how far I get! ! -- Benjamin benjamin at pythagoras.no-ip.org From gmk at runlevelzero.net Sat Nov 19 07:10:40 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Sat, 19 Nov 2005 07:10:40 -0800 Subject: [cAos] Quick question... In-Reply-To: References: Message-ID: <20051119151040.GA6844@titan.runlevelzero.net> When you get back, let us know what module FC was using for that SCSI device. Also, if you can send the output of: 'lspci; lspci -n' when run from Fedora. Thanks and good luck! On Sat, Nov 19, 2005 at 03:07:52PM +0000, Benjamin Donnachie wrote: > Benjamin Donnachie wrote: > >I'm currently using Fedora Core 3, so I expect I'll be asking lots of daft > >questions soon! > > I've already run into a spot of bother! > > The server is a Compaq Proliant 6400 quad Pentium III Xeon server and, > unfortunately, the cAos install disk does not recognise the Compaq SCSI > RAID Array. I gave FC4 a go and noted what module it used, but cAos didn't > have exactly the same one and the one I tried didn't seem to work. > > Also, FC4 only recognised one processor... So I'm expecting cAos to do the > same... :-( > > I can't be more specific as I had to dash off to work. I shall have > another look at it tonight and see how far I get! > > ! > > -- > Benjamin > benjamin at pythagoras.no-ip.org > > > _______________________________________________ > cAos mailing list > cAos at caosity.org > http://lists.caosity.org/mailman/listinfo/caos -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From benjamin at pythagoras.no-ip.org Sun Nov 20 05:04:09 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Sun, 20 Nov 2005 13:04:09 +0000 Subject: [cAos] Quick question... Message-ID: "Greg M. Kurtzer" wrote: >When you get back, let us know what module FC was using for that SCSI >device. Problem solved! cAos was actually loading the right drivers (I'm at work with the details at home again!) it was just missing the device nodes for the Compaq Smart Array controller - see http://www.cpqlinux.com/mkdev_ida.txt Unfortunately, the install script wouldn't recognise /dev/ida/c0d0 as a valid disk drive and as I was feeling lazy I've decided to stick with FC4 at the moment - particularly as it offers an easy way to set up LVM and I'm looking to add further disks soon. I also solved the problem of not detecting all the processors - I set the BIOS Operating System option to OTHER as Linux wasn't listed. It turns out that Linux is actually a sub-option off Unix! I've no doubt that I'll be giving cAos another go soon, but perhaps on a different machine. Take care, -- Benjamin benjamin at pythagoras.no-ip.org From gmk at runlevelzero.net Sun Nov 20 06:53:06 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Sun, 20 Nov 2005 06:53:06 -0800 Subject: [cAos] Quick question... In-Reply-To: References: Message-ID: <20051120145306.GA26597@titan.runlevelzero.net> On Sun, Nov 20, 2005 at 01:04:09PM +0000, Benjamin Donnachie wrote: > "Greg M. Kurtzer" wrote: > >When you get back, let us know what module FC was using for that SCSI > >device. > > Problem solved! cAos was actually loading the right drivers (I'm at work > with the details at home again!) it was just missing the device nodes for > the Compaq Smart Array controller - see > http://www.cpqlinux.com/mkdev_ida.txt Will do. This will be fixed in the next version. Thanks for bringing it to my attention! > Unfortunately, the install script wouldn't recognise /dev/ida/c0d0 as a > valid disk drive and as I was feeling lazy I've decided to stick with FC4 > at the moment - particularly as it offers an easy way to set up LVM and I'm > looking to add further disks soon. My testing version already does LVM auto configuration and manual. > I also solved the problem of not detecting all the processors - I set the > BIOS Operating System option to OTHER as Linux wasn't listed. It turns out > that Linux is actually a sub-option off Unix! Very interesting. :) > I've no doubt that I'll be giving cAos another go soon, but perhaps on a > different machine. You can also consider it in 6 or so months when the Fedora install has been EOL'ed, and no longer supported. ;) Seriously.. Glad you are giving it a try, and thanks for the information. I will be sure that the next version of cinch works properly for these devices. Best of luck, and keep us informed! -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From benjamin at pythagoras.no-ip.org Sun Nov 20 10:14:28 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Sun, 20 Nov 2005 18:14:28 +0000 Subject: [cAos] Quick question... In-Reply-To: <20051120145306.GA26597@titan.runlevelzero.net> References: <20051120145306.GA26597@titan.runlevelzero.net> Message-ID: <4380BD04.4060104@pythagoras.no-ip.org> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Greg M. Kurtzer wrote: > My testing version already does LVM auto configuration and manual. Is it available for download? Pretty please? :-) >>I also solved the problem of not detecting all the processors - I set the >>BIOS Operating System option to OTHER as Linux wasn't listed. It turns out >>that Linux is actually a sub-option off Unix! > Very interesting. :) Given that Compaq don't exist anymore, I think I can forgive them! This server is very well put together, but a little noisy! :-) >>I've no doubt that I'll be giving cAos another go soon, but perhaps on a >>different machine. > You can also consider it in 6 or so months when the Fedora install has > been EOL'ed, and no longer supported. ;) If I can persuade you to let me try the test version, I'd happily give cAos another go tonight or tomorrow! :-) > Seriously.. Glad you are giving it a try, and thanks for the > information. I will be sure that the next version of cinch works > properly for these devices. ... I might even be persuaded to set up the drives manually... :-) Ben -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQ4C9A+gNmph0Y1E2AQKGKxAApCbhMJH2T+Gv8tmpfawCI23wQhOQA18Y POEaDFCQ3vxBx9uOXPdSi045goqclSHFqY9Il4zetEfFaeLkqacxb29NrpmY9Cvb yPVfIaStRJxWqKXbCKiBwJ8WJJ9VXnz8LLcY9LJIRYL6tqBeom/yfWcVYyYgcEw/ rRwpr0HAK1UqAFVvL9vtjze+p21zAzrbbgplPhMUYC+dimkLzoF78wO44OcOWpKj umvoOTFob9tHJbcmoWDtOXnBq0ZRe9yF6vuhu6+jhdxX4XQ2yfIt05U6AYIsZEu/ aC3dZTC77WkPMrBHdYktQGOuUxk+pEeSuNnBWPXWCYd8B/InIQgfRXVkSNS0IlPS D0iFt3blUs5bSsJ1TrqzoVK94SdrZA16zaaJw4xlTXF6or2ST7t0hCbg5w96oXjh s3i6s0H2e+jMmIxzCGC7qoQIHZDD62lwjA4XWTml3hrZ8GGhEuli3Xy7Zr2y2wv3 /JH7fAohQn1bQR0WbugdN0GtGBC6O2lmft85lSwTIyYg93d0CzBdSIiZJFbZ9xa5 HMU/zrkxrvNBrjwVAF0akLTooDfDdYecVHbWSjmIEmSCzFfyAflw+0v00WgufeGk 8WgFqpl+6OXqP/OHEs3tHCeLI4b+eB+J1Ga8K3ju4ycUFfyNfkZGBGUJHyGJNssY 758A3+Jvp/8= =VyY+ -----END PGP SIGNATURE----- From benjamin at pythagoras.no-ip.org Sun Nov 20 10:29:42 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Sun, 20 Nov 2005 18:29:42 +0000 Subject: [cAos] Quick question... In-Reply-To: <4380BD04.4060104@pythagoras.no-ip.org> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> Message-ID: <4380C096.8070306@pythagoras.no-ip.org> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Benjamin Donnachie wrote: >>>My testing version already does LVM auto configuration and manual. > Is it available for download? Pretty please? :-) Currently working my way through the cvs repository... Ben -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQ4DAlOgNmph0Y1E2AQJrOQ//eL9vU+kLm18B4gQ/zyyNAOqEulKox49Q UTfVwCB9lW+ScCZPDV/i91JIntfDbYWuXA9v08o9wOG8Kj2gGt4+kNeXm2l1HRjf lak8qK5opegeNLfLBQZTjL+g+/XyCI1LSQVZDTg3LcP6WeiNqXy69p42uN3oOpbP yeaER1umfMOVL/JL6Dhe4fI1LYe+2c+ZfGaNo2W1ctfO63FkTR/pcOKyfuil/+hk yBoq+gObJ1shK1P8xvJqwAf7jdaYc4uD8Z9d9hl+XNF+qWTBFbmM3g6ZIpo0AS3j 4v/3FHTETt57Ha8BjrLYqfBMKlhHSg9RwVDuRezUNpR0YiLc0cCoW4Mh3FYsrbSb 1Y230S8ZOmWFjYbECWBwEZq3X3+1AbrknvcNUGH9lOFfIRSoQkh1aXfP9uejyCW4 cqiAc1qEiFLKU1j1MTq4Cuu80c4lNm119Xl6AGzfugZdjGuw+Ok7yoCTMX/M1rKf HgipRiiXW+XcIBCcToiQz5gHiDgzzKpePLU7w5Zdc/qUiQhXlruQXyhE9QInp/yP PQCxaqNsnqsdj2zNSdJAqJjd00JQsdJxlIBRDT08GV4Z6V6UKrsTvNcnUsu2F7RR Yu4HujW1i/Lc/TTOgB4FYeoF2bhd4SGZPebttd0yT2KGPkxvmTd30ffd7zz2PFRH A7UueZDnjVw= =pNMF -----END PGP SIGNATURE----- From gmk at runlevelzero.net Sun Nov 20 10:33:05 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Sun, 20 Nov 2005 10:33:05 -0800 Subject: [cAos] Quick question... In-Reply-To: <4380BD04.4060104@pythagoras.no-ip.org> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> Message-ID: <20051120183305.GB26597@titan.runlevelzero.net> On Sun, Nov 20, 2005 at 06:14:28PM +0000, Benjamin Donnachie wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Greg M. Kurtzer wrote: > > My testing version already does LVM auto configuration and manual. > > Is it available for download? Pretty please? :-) Sure, but keep in mind that it is only in testing and I haven't had a chance yet to add default support for those devices. http://runlevelzero.net/static/caos/caos-2-2.3test.i386.iso > >>I also solved the problem of not detecting all the processors - I set the > >>BIOS Operating System option to OTHER as Linux wasn't listed. It turns out > >>that Linux is actually a sub-option off Unix! > > Very interesting. :) > > Given that Compaq don't exist anymore, I think I can forgive them! This > server is very well put together, but a little noisy! :-) LOL! > >>I've no doubt that I'll be giving cAos another go soon, but perhaps on a > >>different machine. > > You can also consider it in 6 or so months when the Fedora install has > > been EOL'ed, and no longer supported. ;) > > If I can persuade you to let me try the test version, I'd happily give > cAos another go tonight or tomorrow! :-) Haha,... no need to persuade me its yours. But I would defiantly appreciate feedback! :) I can also work on creating the devices this evening and get you a new ISO (granted it will probably be late tonight PST). > > Seriously.. Glad you are giving it a try, and thanks for the > > information. I will be sure that the next version of cinch works > > properly for these devices. > > ... I might even be persuaded to set up the drives manually... :-) I will try to do the ISO work a bit earlier and send you the link for testing. It would be nice for you to test it this evening. :) Thanks! -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From gmk at runlevelzero.net Sun Nov 20 10:37:39 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Sun, 20 Nov 2005 10:37:39 -0800 Subject: [cAos] Quick question... In-Reply-To: <4380C096.8070306@pythagoras.no-ip.org> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> <4380C096.8070306@pythagoras.no-ip.org> Message-ID: <20051120183739.GC26597@titan.runlevelzero.net> On Sun, Nov 20, 2005 at 06:29:42PM +0000, Benjamin Donnachie wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Benjamin Donnachie wrote: > >>>My testing version already does LVM auto configuration and manual. > > Is it available for download? Pretty please? :-) > > Currently working my way through the cvs repository... The CVS is current, but it wont contain the core file system that is installed. It will need to be created. I wrote a HOWTO somewhere... I will write a new one and put on our wiki(.caosity.org). -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From benjamin at pythagoras.no-ip.org Sun Nov 20 11:49:28 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Sun, 20 Nov 2005 19:49:28 +0000 Subject: [cAos] Quick question... In-Reply-To: <20051120183305.GB26597@titan.runlevelzero.net> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> <20051120183305.GB26597@titan.runlevelzero.net> Message-ID: <4380D348.30201@pythagoras.no-ip.org> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Greg M. Kurtzer wrote: > Sure, but keep in mind that it is only in testing and I haven't had a > chance yet to add default support for those devices. > > http://runlevelzero.net/static/caos/caos-2-2.3test.i386.iso Drat - I've made the daft mistake of downloading it over a wireless link! Not to worry about the lack of device nodes - I can create them myself. I'm about to take a look at cinch - any tips on getting it to recognise the array as a valid drive? If all else fails I could always make it appear as /dev/hdb I suppose... > I can also work on creating the devices this evening and get you a new > ISO (granted it will probably be late tonight PST). I'm in the UK, so I hope to be fast asleep by then! :-) Take care, Ben -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQ4DTRugNmph0Y1E2AQLg0w/+LjhIAeJg3PWNZFXEZpSULKVgAbc25zEp IC2mQjWBmuDSJKwX3itJ5LhJbqihgOwctopv0uPAH8VqvPJOr5g+2V4AEB4ZzAVZ bKIVPMQZWvJTeDHAQ9f3HEqvL1PGgcHPGxBhMCCR52pQgaT7WnprGao6GsaS5bXx F1QqWsFo0gtHenxtoHag1ZXNABEimkmKw2QvbfIp6VRGFXxodhRboYxJIYELMv+v ydbDG3LS0eET11ALgPUJ1hLWS/L7t6qekVG63tgwQNFQ/gm32OozrlcaDK9TicIr BEZTJ/hr62sH4rfEBbm3urdGg0UahXSb50D4ItXrIcMRdWwpleuR9YXm7qkoHics gyl/4o+tY+q1zO3Fxvgv7JbYR6WOk/Qg3Q6aT5wVEA76pblwA4lUWXQPwW0HbpZl qdQW/DvbZbgXIXTnhApg6vNPT9Z6snz0NcSA9UZKy7EH0TivxlvwEPSOXhKtKRQE MUSbTSSMCplWpxUYiKtiVyzqsYL7TgV1VVnlLlmKpQk2V8zPwP5jXwiGTssEO2yB Z//8KFUtyS+3PkzdCkYFtqMv2Jf29cfdqdqEzdjyyLuMziA3XJ7P12QTbfrQ9a0b 5LL9O/vCIBNK58Mx96leZo42BSOpTcGLmk4h3cEYYupeYW41/odrHNx38LSiCqbV BYY8s1DOwPk= =OJOL -----END PGP SIGNATURE----- From benjamin at pythagoras.no-ip.org Sun Nov 20 14:08:27 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Sun, 20 Nov 2005 22:08:27 +0000 Subject: [cAos] Quick question... In-Reply-To: <4380D348.30201@pythagoras.no-ip.org> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> <20051120183305.GB26597@titan.runlevelzero.net> <4380D348.30201@pythagoras.no-ip.org> Message-ID: <4380F3DB.3060700@pythagoras.no-ip.org> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Benjamin Donnachie wrote: > Not to worry about the lack of device nodes - I can create them myself. > I'm about to take a look at cinch - any tips on getting it to recognise > the array as a valid drive? If all else fails I could always make it > appear as /dev/hdb I suppose... A few problems so far... With cinch I just deleted the sda* devices and used mknod to make sda* reference the array instead. Using the expert set up for LVM etc - it then complains that /cinch/tmp/partitions is empty. No matter what I put in there by dropping to a shell from the main menu, the file it empty again by the time I need to select what each partition is to do. I thought this might be linked to the crafty sda* business, but fdisk displays the partitions fine - sda1 /boot and sda2 LVM. When I've entered the main setup, I've noticed a problem with the network settings - if you select static it doesn't give you the option to specify a DNS server and everything then fails. I'm not very fond of manual LVM setups, but I'm going to give it a go so I can get on with setting up the server. Take care, Ben -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQ4Dz2+gNmph0Y1E2AQK/NxAAgCcK12ALj4O6xTB/LZY1WYVY2K+fLc1Y dZDh4ESTP3tZJi6uv463Sdkzs9KsVcFVd23MEMkyo1JTTAmVKB5C3B0Q8QiS+E4x JRVv6x1UjT94iazVrt6hNEsy9hm1Vf71l97hrXup2bdfj6sgxIVk5ZuqC6mZw3hk 0KoR9OifHPfY/1arVj05Q15JPCjec6z7NzlEStt+ihtTUW+xb3OFA7ThnVNnIMV/ xTJZ/5lnc9k5F2DJXmmIMDSvLdYsQYWeZ/W0+my7mQr1NpCUTUkX+b1dw0fIRcSy doWWhkQ4LdWQhvf1koXeY8u18VwCtMyJ9z6iWH2cLr2BVXXCEH62rRpPRCX+ovda wEYA4vj5ekBdHoE50zSwOQ3XY+G+cHtntd4PoWNQSrjUYB0CyAB+JyKHlXxZqpUE a3I+17vjB4Z0nxae5DQpUZMrew/vR01tE9aqA1muDKXnVLL1sOj5+IISjz78GaY/ gMbDZqdnPczCl2KTrSOQ6jNAFNNW5MmKG3PRpxj8KcCj0gpOmWyUjvHPr5yWk2wg EnCsp7auPyvMhJhp+OHMPBIgL2TOeVsWTMcd43ZWqlrjlATA/gEbMpF7Mi9Qi4HZ 0ytGtDaV/rhwpkhT5B7osbZt7y+LdJYPOW1Ck0bVRLkiEeTC++AToidw0botqm2m rNGf3mFF7KQ= =5uq9 -----END PGP SIGNATURE----- From benjamin at pythagoras.no-ip.org Sun Nov 20 14:45:26 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Sun, 20 Nov 2005 22:45:26 +0000 Subject: [cAos] Quick question... In-Reply-To: <4380F3DB.3060700@pythagoras.no-ip.org> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> <20051120183305.GB26597@titan.runlevelzero.net> <4380D348.30201@pythagoras.no-ip.org> <4380F3DB.3060700@pythagoras.no-ip.org> Message-ID: <4380FC86.4000002@pythagoras.no-ip.org> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Benjamin Donnachie wrote: > A few problems so far... Just given the Assisted Base LVM configuration a go under QEMU on my Mac and it doesn't appear to be creating any of the LVM mapper files - ie no /dev/mapper or any subdirectories. Ben -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQ4D8hegNmph0Y1E2AQJK4g//Tave7ZBum5B0TQWGz0B7q+pKsnqU5pu7 jND8eDyy90znWaLdPVqc1NRsLEMHww23IuVZswB3d5C7pVd49dqK4L4accTXrFY0 6fkGSmPpp2AOVVqCtpT45Z2yPqP6sCvd8p2UWo22KyQ4+lPwsSLKFrr4DPzWDugd 25Cb2T9/w4DXHZTznY6nNPjPFUNCdk0N7V2l89I7FBmvN5ru433cS/ClW74EbnWW WyegsVVnq6XD4Z4PHKUDdXGRW6T4+nK42BcQzxrAXvsZBcwxg3EOLdV9S7t5s3H1 Sj9GpgvsOBfkVjcUoiRQ8SGq7BeZlaiNEb8isnSGfw6EMvT6ibEW0ImJwuIq0dX8 j4LpFRc3JxXK15lzBV2+PScEIzT5iMPcEO8XWnPcoPtmS/KUtgbXvqQnpHO/ozDx FKSf1V7PcU83Q/GzrsYPPsoR+lJZEaLExEkoP45YjkXOCgTaeP/4hqzF1QZcds1s R/iT10gV4BQUzi/knpzIWAxPZ4C2jIpTGMdPaagoWeN08c4dZSRGJMLdxrJQ9NtP 3DQthTDLFqkA8z4w9e0edDfMGQ3hHTlO07aXHjVyEoQHJ+akpNrlqjd2Ei31HVbj +hPlw6SI0RqCOFmlA3GJskUX+kfMOFc3Gpi27+ibomJ2Yik+Sp4b9oBYo+DX7PK/ yvNTxj0sTfY= =esFf -----END PGP SIGNATURE----- From benjamin at pythagoras.no-ip.org Sun Nov 20 15:58:54 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Sun, 20 Nov 2005 23:58:54 +0000 Subject: [cAos] Quick question... In-Reply-To: <4380FC86.4000002@pythagoras.no-ip.org> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> <20051120183305.GB26597@titan.runlevelzero.net> <4380D348.30201@pythagoras.no-ip.org> <4380F3DB.3060700@pythagoras.no-ip.org> <4380FC86.4000002@pythagoras.no-ip.org> Message-ID: <43810DBE.8040805@pythagoras.no-ip.org> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Benjamin Donnachie wrote: > Just given the Assisted Base LVM configuration a go under QEMU on my Mac > and it doesn't appear to be creating any of the LVM mapper files - ie no > /dev/mapper or any subdirectories. Same on server. Cheated and used the FC4 install disks to create the partitions. However, device nodes under /dev/mapper were still missing so I created them manually. Using lvm.static vgscan successfully found "VolGroup00" but vgchange reports: /proc/misc: No entry for device-mapper found Is device-mapper driver missing from kernel? /dev/mapper/control: open failed: No such device Failure to communicate with kernel device-mapper driver. 0 logical volume(s) in volume group "VolGroup00" now active. I've looked in /lib/modules and can't find anything that seems relevant. Suggests to me that the kernel needs to be compiled with LVM support and that I shouldn't be manually creating device nodes for LVM LVs! Ben -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQ4ENvegNmph0Y1E2AQLDKA/8DG52CNSIbEKoYF5dbfuZr5RBde5pSvvC j6F977IKY6L+jT9T9TtuC4JssIP4OtM3J344ZY6igjAyqQ0hGmsG7mAwvuglndlR DM5296juM+U/u1/5w4u7anLp2pJ9+P02LT+DhysTKtLsc+018ltfyxiul17WVFVh LhbfS6/PH47wuyn/jCJK2p5yxnQrstweSspHEi5YXXN3wdwp6PJDTt9Zr2/yi6jp F3apa8N4+/SeUkFn9g5n1oWKNHfPH57RxU6p1hkOwsTcgYrWJMtrZHnNJn1qgrSI 35dmonqiIk+z/kQRY8uqD0QGba7IKrsb0AqLyoSroVKhcIK2zzmzx8ZiVvPatDxD DdDNm0pbwFVASDo7a9BSKIvultPf+0eTAaMAp034iyu/0+Q+3+/ICaX7bNwGtsHY U19g0fghJvFZ7bpCzNSQZ4zQQzuG0kRYdpcN/X4ZYhvaj2E9ZNyyAOf1KrF68obT wy3DkXsgzDTi3c0+mtBBHvsW6cL3JX4lIbFWnagAJ1XvC4ocdn8AzDPCwU3MTeXC jQ5o44hxs2qT6TpSYfBpuk5W8WcP51unHDPghYIYPXvVkzSZn8t5xNiaWYVH43zt ZH6XBWKuAT/cXeAJ5HaAD10IggXwvE87cpP3c17OfwPyMhjb1kDUhoyGsvWOTQRG VLgVEqz1WQo= =IKQh -----END PGP SIGNATURE----- From benjamin at pythagoras.no-ip.org Sun Nov 20 16:03:44 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Mon, 21 Nov 2005 00:03:44 +0000 Subject: [cAos] Quick question... In-Reply-To: <43810DBE.8040805@pythagoras.no-ip.org> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> <20051120183305.GB26597@titan.runlevelzero.net> <4380D348.30201@pythagoras.no-ip.org> <4380F3DB.3060700@pythagoras.no-ip.org> <4380FC86.4000002@pythagoras.no-ip.org> <43810DBE.8040805@pythagoras.no-ip.org> Message-ID: <43810EE0.6030100@pythagoras.no-ip.org> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Benjamin Donnachie wrote: > I've looked in /lib/modules and can't find anything that seems relevant. > Suggests to me that the kernel needs to be compiled with LVM support > and that I shouldn't be manually creating device nodes for LVM LVs! Ignore that - I just found dm_mod.ko! :-) Ben -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQ4EO3+gNmph0Y1E2AQLhkRAAhKJl9av30QMh/R9dRHSwRb4JHyzlb5nx IDu6mBGjQHC9lcGhj43a9xVlAmRQAU1+GQPrZx3BI6xzktTHbnM5JX4Eig6dn3zv staVDIDkAfohWrjvEx5yxU43tql7Zmswi24gFKJEAA1B/CRU6WGAdQ7WFt2394Df vjHhgIcjeVkdkJlueCZJmR0Tdf6hKEaGS5rMMEi2FODonzM7q7IiPM4RUlJ76/8a TnZMSc5oCqtc/gt9RJSexUTLxRdbANwJAzBnyHRLEPZoHCT0l5CYOXSt+F3OMdnB LhcS+m7tEujJiHQLoN7uyiCUkNlOIgVMpf0MSq4Se6M6QS0O+w60LOXNN+JWMOc1 SkUPiXGScpXIxS2Sa/ysUFjsGqzy/OWCYJvBItS8qItfwdm/kJh8Ub485z6SkV+P 7A2UktKtWa0GrrDD/aCQn/jMlkIKbVmHouOCxbXrFfHmDlheWhvcWK9EuRGgcYDg 9BhOc0QfJdP96HvBsdx6NMiRQO1KWIEI8tSelThCEjJHy3abhaf4yrMhCN1U3M8J +E5iVjGA6Bf9WM7dkMjf1uihxQzALev2rbLduX1T+S0hntSE/BsgxU2FwvaRvC6O RtVc5fVxxiUNFlKZ0CI0cY6y0UioX1FIhnsaIWJR5gEsXenfZ1cx8QerCs7WfZQU Ew8qvaYrWFs= =B6xQ -----END PGP SIGNATURE----- From benjamin at pythagoras.no-ip.org Sun Nov 20 17:31:59 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Mon, 21 Nov 2005 01:31:59 +0000 Subject: [cAos] Installed at last! :-) In-Reply-To: <43810EE0.6030100@pythagoras.no-ip.org> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> <20051120183305.GB26597@titan.runlevelzero.net> <4380D348.30201@pythagoras.no-ip.org> <4380F3DB.3060700@pythagoras.no-ip.org> <4380FC86.4000002@pythagoras.no-ip.org> <43810DBE.8040805@pythagoras.no-ip.org> <43810EE0.6030100@pythagoras.no-ip.org> Message-ID: <4381238F.3070209@pythagoras.no-ip.org> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Benjamin Donnachie wrote: > Ignore that - I just found dm_mod.ko! :-) Had some problems with grub refusing to install - so I created a boot floopy and installed it natively... Got there in the end, and sidekick is now busy installing everything... Phew! :-) Right... where's my bed? G'night, Ben -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQ4EjjOgNmph0Y1E2AQIYCQ/+KAbtVkPmBmDZPF0laur/HTnqOrJHEZSN Qm8u6Wdyw6oalEM00cJaizzmsqwU9zJjb8s+msjtOm0moNEjYcwumVXSb9U/0DOn 2Nh+2b1BkDVyQEvBeCP73/86uZDeFnQJq/u4VuC4ndDMP9N3KY4SR16K8OnzpddA tctOW6v623O9XGXMOmZIOEe5k8SGvrW3TtJf586xk2hBc9bYSzQ22gaT0L4sHrhR 69gf6cA0lkfXh3bnUhQJcR5uFnUdyO+V1qh7nVHmjPwLlN/K6G9vZKac2IQx7UAs zt+k3gc7EiZf84VUderI2zD0Y7weZ4fY8N4uKUkeMQhMqrbsT9t+3XY3jq1bqhRK 0r4BBS4Do6Q9JBwEX8GHfnF0ud2G8vWI01x/an8SJi/9/sjP6QEUy3IG1xaI6SdN /VvQbZ94IM78qXuVvoISt3a8eiLGQZkEXbahcviim9UhuqognaOGSMBpyrbh5Gtb BTWwaExfKiOT/TWEsaXkRVDuur/z8VpR3x0pMynAQbQPVU7bVKyzUWJm7Vuqp9GD +8Cu4ZGYRcBPQZLq2/FHEtI/ZQ1rueX6yiFCYP093u/D4GZSF4N1Io6Xhb71Afzx 0ura7gHKB+6DWcCngFAWuoR7RDkkUaE5insnU/yvcAAKGFM+a+jmFqlGt7pRQpkc 0xHbHjixeQE= =Z5+a -----END PGP SIGNATURE----- From benjamin at pythagoras.no-ip.org Sun Nov 20 17:35:03 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Mon, 21 Nov 2005 01:35:03 +0000 Subject: [cAos] Quick question... In-Reply-To: <4380F3DB.3060700@pythagoras.no-ip.org> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> <20051120183305.GB26597@titan.runlevelzero.net> <4380D348.30201@pythagoras.no-ip.org> <4380F3DB.3060700@pythagoras.no-ip.org> Message-ID: <43812447.4060306@pythagoras.no-ip.org> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Benjamin Donnachie wrote: > With cinch I just deleted the sda* devices and used mknod to make sda* > reference the array instead. For info - I ended up using the "proper" device files created in /dev/ida, instead of messing about trying to trick the installer. Ben -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQ4EkRegNmph0Y1E2AQKJKQ/9EuPABZ9PNll6H3Hd/NKpjt6UfLFlw73Z vn3NZPXWfV1wnhWgvfQkbUxsn+/BXMmO8vVa/+B8d0qFUuRlf+x2mCHD+fuBd0rV 2eeF16zAacufj0OPk46bkIqSlKCNVVT9/eceY8/uUUUJ3JmzT+OwedViwzBCZZnL xgMZpQYrDJiihS5lhFILX49YlAiz0MbhGg6i3G6hMejm7mH0EAQBzt7se9NxBrO7 vzrG0F03lrbgAV4Vgg2Uo4FoYvuf6PHHuakfvEBkzZ9rxvMSiWmKNod9p04aXuXh BHYSxsxOSqnGnZh6uawcqVpOpO0v3OOXSzTgTnTRJDxPniHwxww+rYNFI0uWqjpz MCDIaQTl8nlo4ZI74lToRtb08v8HyB90sq3dC16CvfVagfup0hVideHUkLjjdCTG hPk/kwdVcPXgryFJRuXx9Q60cwj1AhNHNIr6uMdlxwg+mTD1G02MPcE8Ycd2BOMy QTCy3Qu/peFJEwh4yYmTYHI+/DyLMaYqbrjAYbXGNnJcMaXxrp5FnSBnXPWQOZ/G GgYzTinUgmbHuET/xbYCxCSd38jqP5str2XpcJKgPc9N3ZCsM0vdWdoyGZg87hZp bMtDYyjATw6PaOkUkJNNa8fClNYxm69sl2Igr6aEF+jOubY3tLsukyLhBlzB26zz qhcaDkwr3Zs= =Zke4 -----END PGP SIGNATURE----- From benjamin at pythagoras.no-ip.org Sun Nov 20 18:50:17 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Mon, 21 Nov 2005 02:50:17 +0000 Subject: [cAos] Installed at last! :-) In-Reply-To: <4381238F.3070209@pythagoras.no-ip.org> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> <20051120183305.GB26597@titan.runlevelzero.net> <4380D348.30201@pythagoras.no-ip.org> <4380F3DB.3060700@pythagoras.no-ip.org> <4380FC86.4000002@pythagoras.no-ip.org> <43810DBE.8040805@pythagoras.no-ip.org> <43810EE0.6030100@pythagoras.no-ip.org> <4381238F.3070209@pythagoras.no-ip.org> Message-ID: <438135E9.9000507@pythagoras.no-ip.org> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Benjamin Donnachie wrote: > Had some problems with grub refusing to install - so I created a boot > floppy and installed it natively... Drat! I used the version in /usr/lib/grub/i386-caos but it doesn't support splashimages... Oh well! With any luck this machine won't be reset very often anyway! Right... I really must crash into bed now... Ben -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQ4E16egNmph0Y1E2AQIUmRAApGVCnCqXhm2mvzOCz7wrra6RSg3UxGtj R9v30MKnecpjIR88XkWZ+fkJHkeZY89vfhsVKlNM4pA89sOf7TQ3MKiRIgiMCHxI XjRIqovGxvawraTbOZ6n59k/brHqtPvXzXPk0ud3u8MUVdiYoySI5+QvMDsdfhjI tjIwo8Aitb2s4hIKxtHh7pOA9xbYpy/PD+0XqXWOM2uKnxO/sCymiVtFW46EQyYP KDKTLQUnKa/buRXDsP5CzmH7GlhP77Ilf1YJ6GF1da2vVcQ8u22eXQ3T/3An57Sn ays1goz1o8JkPHthTtrrLaihU99HAKdILCbR1Y74TPo4I/SckLp1q9XrS3K3ynfL amgo77M7VTEh269/TU68eg8CRzBoqenSBm9yaM1sPsNOlBcX4p/ffDc/Y4YunXDO O1T16FwZV7drj0xbUxUmIBNlWp5xGLVLzOu9kWX+3SV8MW/8lpJzuO2p+Z56nHHe E62XP9Zfl/GPIF6wMKzsLfGdjYwtAVlhADnWHYtgbQGwrJh+mVA0UKoUBJmhzAi+ XEr0YmnbBV3q2wyNNj1Ur/z8Apujnp/rq8uEm3R54UAHjMWYl2pKj6S4XY/1U87w iet79XDQzMTvqcvBW6goIAJ4PUSe9Ce7nU95odadJIi/qYXWvHlbTQNaLao+ZQY2 sxGZy3JIN1I= =ayWA -----END PGP SIGNATURE----- From benjamin at pythagoras.no-ip.org Mon Nov 21 02:52:19 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Mon, 21 Nov 2005 10:52:19 +0000 Subject: [cAos] Quick question... In-Reply-To: <4380F3DB.3060700@pythagoras.no-ip.org> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> <20051120183305.GB26597@titan.runlevelzero.net> <4380D348.30201@pythagoras.no-ip.org> <4380F3DB.3060700@pythagoras.no-ip.org> Message-ID: <4381A6E3.7020608@pythagoras.no-ip.org> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Benjamin Donnachie wrote: > When I've entered the main setup, I've noticed a problem with the > network settings - if you select static it doesn't give you the option > to specify a DNS server and everything then fails. It also doesn't allow you to specific a default gateway. However, cAos is running nicely on my server and I now need to start setting up all the services :-) Ben -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQ4Gm4ugNmph0Y1E2AQLC6hAAjW/WICiXPEchbxNlKoCIHqYR/bmGC6kU 1x/FsJ360RKg78Ofbf3X+JKf6X+uNxNYuB7vbOwtXxsjirSEwW7vq1mf57KoCYYv X6z06OpX8aMNSy4EpfCrog39jZtGQQRIL+mgeRt0lfp+/WfrFYRa0SKWnZzuQSsC 6Ytdz/lI+ePzlwpwPKYDPvb+9YiLusQdtId6TE9qUp5AeqLi2yLj1kMpSoXYd+27 M5OXFiWKNXxdccm6/yOcOO/Umde52RiQg+nGQwaRaiWX5BavvrU8ZkjmdcQmEefh AtxbmqgqAzQ6vXulv2lCvqHGauHgsZhHzMlnqLJteMuSB2GtpuqYKyk39ypAhs3h gUhlpwWNGBMFdKCUPvsECcwlD1/yG/SIdw+XWcyGdTDUAul+uTuGzuIckmMkx36K HqwludkpCnLnjATYQlcqNe0oK5rJOnmHc4ZdaqRtRtABHTWg1FXolbfiyuTAVqs+ O6xveetrfrQNHmV/Y8OvqeK2Rb0i77jwIH5Enw+VPGNZT9xN/LkDrQYBudqrCL28 yZZtyGfGcfPiHv3G3BCY9cQvxhoWFvfpLHcmQ3eXC4vyqn7o8d/8YbjDAS+zXx6d N1ejqtnV2cPHoATEUj8bd0Xz28SUDdAsholhY5S8nA5hf3CK6RlKfYqa3aUaSFXI qNmBVN58kR0= =pKT8 -----END PGP SIGNATURE----- From gmk at runlevelzero.net Mon Nov 21 09:53:44 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Mon, 21 Nov 2005 09:53:44 -0800 Subject: [cAos] Quick question... In-Reply-To: <4381A6E3.7020608@pythagoras.no-ip.org> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> <20051120183305.GB26597@titan.runlevelzero.net> <4380D348.30201@pythagoras.no-ip.org> <4380F3DB.3060700@pythagoras.no-ip.org> <4381A6E3.7020608@pythagoras.no-ip.org> Message-ID: <20051121175344.GA30443@titan.runlevelzero.net> My apologies. I meant to do cinch work all night, but then had some problems with VMWare. Thus I was using qemu which works but was very slow. Then I spent much of the evening bringing the accelerated qemu into our ports tree. It is actually very sweet... kqemu is not freely distributable. It is freely available but can't be redistributed and has a module which is not OSS. Using the ports system, it is downloaded and built directly from the upstream maintainer which enables the distribution from within caos and doesn't violate any licensing terms. Also it is much faster then the completely OSS qemu. So in a nutshell, I got the accelerated qemu going and it is almost as fast as VMWare but didn't get as much work on cinch done as I hoped. I will take your notes anyway and make sure all issues are addressed. If you can send me a consolidated list of issues and comments that would be very helpful. Thanks again for testing and your efforts with it. :) About the cAos Ports system: The cAos ports system adds enhanced functionality without getting in the way or polluting the base system which is managed by the RPM database. Its goal is to leverage the best features of package distribution in FreeBSD and Linux. We obviously think that binary package distribution has a huge advantage for reproducibility, consistency as well as ease of use, but there are also advantages to a source based package system for certain packages outside the base distribution. While the ports tree is not fully featured or meant to be redundant with the packages distributed via RPM, we are bringing in lots of development, HPC and various services that many people like to compile from scratch. It is also integrated with environment modules so it allows for multiple versions to be installed and easily used by various users. If anyone is not familiar with this sorta thing, I encourage you to play with it and/or ask questions. It is a very powerful balance between a binary replicated and managed distro and a source based packaging system. While qemu is distributed in our binary RPM tree, the accelerated version is not freely distributable (it is $free$ but not open source or distributable). The ports system allows it to be retrieved via the upstream maintainer's site directly on a per user/install basis and compiles and installs it from scratch (thus not violating any license restrictions). Here is the howto for installing the accelerated qemu (once we release the updated ports tree): $ su - # cd /usr/src/ports/programs/qemu/0.7.2/ # make install # exit $ module avail $ module load qemu/0.7.2 $ qemu ... $ module unload qemu note: accelerated qemu (kqemu) also builds and installs a kernel module into /lib/modules/`uname -r`/... That should be modprob'ed at system boot or before you use the accelerated qemu. On Mon, Nov 21, 2005 at 10:52:19AM +0000, Benjamin Donnachie wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Benjamin Donnachie wrote: > > When I've entered the main setup, I've noticed a problem with the > > network settings - if you select static it doesn't give you the option > > to specify a DNS server and everything then fails. > > It also doesn't allow you to specific a default gateway. > > However, cAos is running nicely on my server and I now need to start > setting up all the services :-) > > Ben > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.4.1 (Darwin) > Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org > > iQIVAwUBQ4Gm4ugNmph0Y1E2AQLC6hAAjW/WICiXPEchbxNlKoCIHqYR/bmGC6kU > 1x/FsJ360RKg78Ofbf3X+JKf6X+uNxNYuB7vbOwtXxsjirSEwW7vq1mf57KoCYYv > X6z06OpX8aMNSy4EpfCrog39jZtGQQRIL+mgeRt0lfp+/WfrFYRa0SKWnZzuQSsC > 6Ytdz/lI+ePzlwpwPKYDPvb+9YiLusQdtId6TE9qUp5AeqLi2yLj1kMpSoXYd+27 > M5OXFiWKNXxdccm6/yOcOO/Umde52RiQg+nGQwaRaiWX5BavvrU8ZkjmdcQmEefh > AtxbmqgqAzQ6vXulv2lCvqHGauHgsZhHzMlnqLJteMuSB2GtpuqYKyk39ypAhs3h > gUhlpwWNGBMFdKCUPvsECcwlD1/yG/SIdw+XWcyGdTDUAul+uTuGzuIckmMkx36K > HqwludkpCnLnjATYQlcqNe0oK5rJOnmHc4ZdaqRtRtABHTWg1FXolbfiyuTAVqs+ > O6xveetrfrQNHmV/Y8OvqeK2Rb0i77jwIH5Enw+VPGNZT9xN/LkDrQYBudqrCL28 > yZZtyGfGcfPiHv3G3BCY9cQvxhoWFvfpLHcmQ3eXC4vyqn7o8d/8YbjDAS+zXx6d > N1ejqtnV2cPHoATEUj8bd0Xz28SUDdAsholhY5S8nA5hf3CK6RlKfYqa3aUaSFXI > qNmBVN58kR0= > =pKT8 > -----END PGP SIGNATURE----- > _______________________________________________ > cAos mailing list > cAos at caosity.org > http://lists.caosity.org/mailman/listinfo/caos -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From benjamin at pythagoras.no-ip.org Tue Nov 22 10:14:39 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Tue, 22 Nov 2005 18:14:39 +0000 Subject: [cAos] Quick question... In-Reply-To: <20051121175344.GA30443@titan.runlevelzero.net> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> <20051120183305.GB26597@titan.runlevelzero.net> <4380D348.30201@pythagoras.no-ip.org> <4380F3DB.3060700@pythagoras.no-ip.org> <4381A6E3.7020608@pythagoras.no-ip.org> <20051121175344.GA30443@titan.runlevelzero.net> Message-ID: <4383600F.8070601@pythagoras.no-ip.org> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Greg M. Kurtzer wrote: > So in a nutshell, I got the accelerated qemu going and it is almost as > fast as VMWare but didn't get as much work on cinch done as I hoped. Not to worry - I got there in the end! :-) > I will take your notes anyway and make sure all issues are addressed. If > you can send me a consolidated list of issues and comments that would be > very helpful. Firstly, what's the difference between the stable and current repos? Are there any issues with using RPMs for RH distros? I'm happy to install from source, but like the easy upgrade path offered by rpms... Alternatively, I might look into how to create rpm files... The issues I had with installing using test3 were: Device nodes not created for RAID array - ie /dev/ida/c0d0 etc Device mapper module not loaded - needed for LVM - modprobe dm_mod Despite using "install smp", it only installs single processor kernel. Bootloader did not install correctly - probably because I bypassed the disk setup part of install. When configuring static network - no option for DNS or gateway. Grub does not support splash images. Here's what I did to install it with LVM: Use the Fedora Core install disk to recreate LVMs. Resulting in harddisk with two partitions - first one the boot partition and the second the LVM partition. Boot cAos install. Drop to shell and then create /dev/ida directory and then create device nodes - ie mknod c0d0 b 72 0, mknod c0d0p1 b 72 1 and mknod c0d0p2 b 72 2 Involke lvm.static and then execute commands vgscan and then vgchange -ay VERY IMPORTANT (As I eventually discovered!) if you used Fedora to create ext3 paritions you must now reformat them! For me this was /dev/ida/c0d0p1 and all the logical volumes under /dev/LogGroup00/ (except the swap partition) and as I couldn't find mkfs.ext3 I used mkfs.ext2 followed by tune2fs -j (Not sure whether you can do the latter this at this stage - if not, treat as ext2 and see later) NB If you fail to reformat cAos won't load properly on reboot as Red Hat have "kindly" modified their versions of ext3!!! (Sounds a lot like Microsoft's "embrace and enhance" which I detest so much!) For good measure, execute mkswap on your swap partition too. Create the directory /newroot, then mount all your new partitions under here and then involve swapon on your swap partition. Create the file /cinch/etc/fstab listing all your new partitions and how they should be mounted. For good measure, I copied this to /newroot/etc/fstab too. Exit the shell and install core file system. Drop to shell and edit the grub device.map file - in my case adding the line (hd0) /dev/ida/c0d0 Edit grub.conf so it correctly references hd0,0 (and add a password if you want). NB During one install it failed to copy the stage1 and 2 loaders from /lib/grub/i386-caos to /newroot/boot/grub and I had to do this manually. We will need to boot from floppy to install the bootloader - so grab a spare floppy and execute the following: # cd /usr/lib/grub/i386-caos # dd if=stage1 of=/dev/fd0 bs=512 count=1 # dd if=stage2 of=/dev/fd0 bs=512 seek=1 Keep the floppy to hand for later. Then execute yum install linux-smp followed by yum remove linux.i386 (assuming you want smp, of course!) Exit and then continue install. When prompted for network configuration enter dhcpd. Continue build. When installation complete and presented with login screen, login and edit /etc/sysconfig/network-scripts/ifcfg-eth0 if necessary - in my case changing to static and adding entry for GATEWAY. If you were unable to invoke tune2fs earlier, now invoke tune2fs -j on all ext2 filesystems that you wish to add journalising (ie ext3) to. Reboot with floppy in drive. Grub should load then type the following to install bootloader: grub> root (hd0,0) grub> setup (hd0) Then reboot without the floppy. All being well, grub should start and then boot cAos! :-) > Thanks again for testing and your efforts with it. :) No worries! I'm just glad to have it up and running! :-) > If anyone is not familiar with this sorta thing, I encourage you to play > with it and/or ask questions. It is a very powerful balance between a > binary replicated and managed distro and a source based packaging system. Is this the answer to my "I need to install stuff that's only available as source but want it easy to upgrade" problem? Take care, Ben -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQ4NgCugNmph0Y1E2AQKASw//e08lar9CY5HyhbZSCN74O6rPXqHC9M9N VBZZYTtcGyE0SZykVQhjE7ySqIzZpvDLAWunnm+e7sApw2tC6RLFc1JsUWchAbWL juQl+PUHtykvaETvGfB8K5Be4C6GBnMPLeeiffDaAMdsIVEUAKYQQlOldFx/FYDW VO4mljY95fSrRY4F7+74WMnd6HeBzbLUEeBcMvCObX20UCUsQ+N0q65d3p8SMlXF rtFdP8f4LErdqG05KOHgQ3yIYwOrqhGO9ddJ/rqUYUn32UfgyS0sR8qN4ukHL/qY LcChuIVsKpmJc6BHkfyGy33Kos13wJYUdVMBwjEVq6xjZu0wltBAZuPpOkSuxOHE xNg7HaFuQ5CYNDc8UcohhP+tbexsfDLBaceHEfHKICeM0kwwgKOIMlImqOYvqXdD 8EO/ES+8iQv3adC42Cu8sqx1HvCdhBHOyl8v/zM7DOtGMTfuXiTQavEvFxCTZnBP zExoRCSQMJflxpFVLi35ZUKawZAD7j4uakGeMHCdSDaG9AfR8IcLoS3kc+1v3E2X gNxxm3mu5m02xBSJ0aANyLurWXy/gzsw4e8/J0J30V/V8LNnUqRaoyqO4lOgvC6t 4LQ/aS5OnFwq0de7KGBWF5HcVAl94kXgaz5t3UEVMZkdBMIHO8PgIFlbYC+Wml4D GFXdDiO1mlQ= =8XZS -----END PGP SIGNATURE----- From gmk at runlevelzero.net Wed Nov 23 10:05:58 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Wed, 23 Nov 2005 10:05:58 -0800 Subject: [cAos] Quick question... In-Reply-To: <4383600F.8070601@pythagoras.no-ip.org> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> <20051120183305.GB26597@titan.runlevelzero.net> <4380D348.30201@pythagoras.no-ip.org> <4380F3DB.3060700@pythagoras.no-ip.org> <4381A6E3.7020608@pythagoras.no-ip.org> <20051121175344.GA30443@titan.runlevelzero.net> <4383600F.8070601@pythagoras.no-ip.org> Message-ID: <20051123180558.GA4799@titan.runlevelzero.net> On Tue, Nov 22, 2005 at 06:14:39PM +0000, Benjamin Donnachie wrote: > > I will take your notes anyway and make sure all issues are addressed. If > > you can send me a consolidated list of issues and comments that would be > > very helpful. > > Firstly, what's the difference between the stable and current repos? > Are there any issues with using RPMs for RH distros? I'm happy to > install from source, but like the easy upgrade path offered by rpms... > Alternatively, I might look into how to create rpm files... Primary difference is that current will have all currently known good packages. This may mean that there are several versions of a particular program there (eg. warewulf 2.4.4 and 2.4.5). as new packages are committed the autobuilder builds them and leaves them in the testing repository. Once they have undergone some human testing, they graduate to current. Stable now is the oldest known good package in current. So for example, warewulf will be 2.4.4 with the above example. When warewulf 2.4.4 is removed from current, stable will also be updated again to the oldest package in current (since current only has known good packages, it makes it the oldest known *GOOD* package). > The issues I had with installing using test3 were: > > Device nodes not created for RAID array - ie /dev/ida/c0d0 etc I am working on this now. I ended up rewriting some of the LVM stuff and how disk probing works to make this easier. > Device mapper module not loaded - needed for LVM - modprobe dm_mod What LVM partition mechanism did you use? If you did it by hand, the device module would not have been loaded. > Despite using "install smp", it only installs single processor kernel. ACK! I will check on that. > Bootloader did not install correctly - probably because I bypassed the > disk setup part of install. Probably. This just means you would have to configure grub by hand if not using the autopartitioning tools. > When configuring static network - no option for DNS or gateway. Actually, I know of this. The version I gave you actually is using an older core FS with an older sidekick. Current version doesn't do that. As a matter of fact the version that you have is arched... Do the following to upgrade: # rpm -e sidekick # yum install sidekick sidekick is now noarch.rpm and yum won't do an upgrade of differing arches. > Grub does not support splash images. It supports it, but that splash doesn't work with all video cards. This is on the TODO. Problem is that it works on all of our (developers) systems that do graphics... Thus nobody can debug. ;) If you can resolve, you will have some good friends! lol > Here's what I did to install it with LVM: > > Use the Fedora Core install disk to recreate LVMs. Resulting in > harddisk with two partitions - first one the boot partition and the > second the LVM partition. > > Boot cAos install. > > Drop to shell and then create /dev/ida directory and then create device > nodes - ie mknod c0d0 b 72 0, mknod c0d0p1 b 72 1 and mknod c0d0p2 b 72 2 There is that script that will do that but it creates a boatload of devices. Are they all needed (I am not familiar with this hardware)? > Involke lvm.static and then execute commands vgscan and then vgchange -ay > > VERY IMPORTANT (As I eventually discovered!) if you used Fedora to > create ext3 paritions you must now reformat them! For me this was > /dev/ida/c0d0p1 and all the logical volumes under /dev/LogGroup00/ > (except the swap partition) and as I couldn't find mkfs.ext3 I used > mkfs.ext2 followed by tune2fs -j (Not sure whether you can do the > latter this at this stage - if not, treat as ext2 and see later) NB If > you fail to reformat cAos won't load properly on reboot as Red Hat have > "kindly" modified their versions of ext3!!! (Sounds a lot like > Microsoft's "embrace and enhance" which I detest so much!) If the partition was created with a newer e2fsprogs then there could be a problem. We have a version that is static in our core (we won't upgrade it for caos2 because of our static core ABI policy). And yes... you are right. This can be a problem. > For good measure, execute mkswap on your swap partition too. > > Create the directory /newroot, then mount all your new partitions under > here and then involve swapon on your swap partition. > > Create the file /cinch/etc/fstab listing all your new partitions and how > they should be mounted. For good measure, I copied this to > /newroot/etc/fstab too. Nice catch. > Exit the shell and install core file system. > > Drop to shell and edit the grub device.map file - in my case adding the > line (hd0) /dev/ida/c0d0 > > Edit grub.conf so it correctly references hd0,0 (and add a password if > you want). NB During one install it failed to copy the stage1 and 2 > loaders from /lib/grub/i386-caos to /newroot/boot/grub and I had to do > this manually. > > We will need to boot from floppy to install the bootloader - so grab a > spare floppy and execute the following: > > # cd /usr/lib/grub/i386-caos > # dd if=stage1 of=/dev/fd0 bs=512 count=1 > # dd if=stage2 of=/dev/fd0 bs=512 seek=1 > > Keep the floppy to hand for later. > > Then execute yum install linux-smp followed by yum remove linux.i386 > (assuming you want smp, of course!) > > Exit and then continue install. > > When prompted for network configuration enter dhcpd. Continue build. > > When installation complete and presented with login screen, login and > edit /etc/sysconfig/network-scripts/ifcfg-eth0 if necessary - in my case > changing to static and adding entry for GATEWAY. > > If you were unable to invoke tune2fs earlier, now invoke tune2fs -j on > all ext2 filesystems that you wish to add journalising (ie ext3) to. > > Reboot with floppy in drive. Grub should load then type the following > to install bootloader: > > grub> root (hd0,0) > grub> setup (hd0) > > Then reboot without the floppy. > > All being well, grub should start and then boot cAos! :-) Very nice write up. Can you copy this to our wiki at http://wiki.caosity.org please? > > Thanks again for testing and your efforts with it. :) > > No worries! I'm just glad to have it up and running! :-) > > > If anyone is not familiar with this sorta thing, I encourage you to play > > with it and/or ask questions. It is a very powerful balance between a > > binary replicated and managed distro and a source based packaging system. > > Is this the answer to my "I need to install stuff that's only available > as source but want it easy to upgrade" problem? Yes, and more. It is also the answer to "User 1 needs version x user 2 version y" etc... Using the modules, every user can select on a per shell basis what versions of apps that they want to run. It also allows us to distribute things that are REdistributable (meaning they can only come from a single source perhaps in binary form only). An example of this is "real player". I have found this to be incredibly valuable for development and scientific needs. for example, on our clusters users can select which compiler they want, which mpi they want, which math libs, etc... they can even select specific versions (egg. gcc-4.0.2 and openmpi). Not to mention even valuable for desktops, etc... We are taking requests so post your thoughts (not saying we will accomodate, but all requests will be evaluated). -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From gmk at runlevelzero.net Wed Nov 23 10:21:23 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Wed, 23 Nov 2005 10:21:23 -0800 Subject: [cAos] Quick question... In-Reply-To: <20051123180558.GA4799@titan.runlevelzero.net> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> <20051120183305.GB26597@titan.runlevelzero.net> <4380D348.30201@pythagoras.no-ip.org> <4380F3DB.3060700@pythagoras.no-ip.org> <4381A6E3.7020608@pythagoras.no-ip.org> <20051121175344.GA30443@titan.runlevelzero.net> <4383600F.8070601@pythagoras.no-ip.org> <20051123180558.GA4799@titan.runlevelzero.net> Message-ID: <20051123182123.GB4799@titan.runlevelzero.net> On Wed, Nov 23, 2005 at 10:05:58AM -0800, Greg M. Kurtzer wrote: [.. snip ..] > Not to mention even valuable for desktops, etc... We are taking requests > so post your thoughts (not saying we will accomodate, but all requests > will be evaluated). I forgot about the link to the ports apps done/todo: https://wiki.caosity.org/tiki-index.php?page=Portstree+Programs&bl -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From benjamin at pythagoras.no-ip.org Sat Nov 26 05:17:41 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Sat, 26 Nov 2005 13:17:41 +0000 Subject: [cAos] Quick question... In-Reply-To: <20051123180558.GA4799@titan.runlevelzero.net> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> <20051120183305.GB26597@titan.runlevelzero.net> <4380D348.30201@pythagoras.no-ip.org> <4380F3DB.3060700@pythagoras.no-ip.org> <4381A6E3.7020608@pythagoras.no-ip.org> <20051121175344.GA30443@titan.runlevelzero.net> <4383600F.8070601@pythagoras.no-ip.org> <20051123180558.GA4799@titan.runlevelzero.net> Message-ID: <43886075.10504@pythagoras.no-ip.org> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Greg M. Kurtzer wrote: > Primary difference is that current will have all currently known > good packages. This may mean that there are several versions of a > particular program there (eg. warewulf 2.4.4 and 2.4.5). as new > packages are committed the autobuilder builds them and leaves them > in the testing repository. Once they have undergone some human > testing, they graduate to current. Stable now is the oldest known > good package in current. So for example, warewulf will be 2.4.4 > with the above example. So the current repo should be sufficient for a server? > > When warewulf 2.4.4 is removed from current, stable will also be > updated again to the oldest package in current (since current only > has known good packages, it makes it the oldest known *GOOD* > package). Would I be correct in thinking that the stable repo is very stable, current likely to be stable and testing, well, not likely to be stable? >> Device mapper module not loaded - needed for LVM - modprobe >> dm_mod >> > What LVM partition mechanism did you use? If you did it by hand, > the device module would not have been loaded. I ended up using the Fedora install disk to create the LVMs. However, I did try the assisted LVM install and it failed without any of the /dev/ LVM nodes being present - I assumed that it was due to the lack of dm_mod, but it I didn't run vgscan and vgcreate to check. >> Despite using "install smp", it only installs single processor >> kernel. >> >> > ACK! I will check on that. I seem to recall it installing hte smp sources okay, but just not linux.smp >> Bootloader did not install correctly - probably because I >> bypassed the disk setup part of install. >> >> > Probably. This just means you would have to configure grub by hand > if not using the autopartitioning tools. Done! :-) >> When configuring static network - no option for DNS or gateway. >> >> > Actually, I know of this. The version I gave you actually is using > an older core FS with an older sidekick. Current version doesn't do > that. > That's good - I initially thought that my network was playing up when I set it to static and it couldn't find any of the repos! Then I realised... :-) > As a matter of fact the version that you have is arched... Do the > following to upgrade: > > # rpm -e sidekick # yum install sidekick > > sidekick is now noarch.rpm and yum won't do an upgrade of differing > arches. Done! >> Grub does not support splash images. >> >> > It supports it, but that splash doesn't work with all video cards. > This is on the TODO. Problem is that it works on all of our > (developers) systems that do graphics... Thus nobody can debug. ;) > > If you can resolve, you will have some good friends! lol > I've got grub-0.97-12.caos installed and a grep of /boot/grub/stage2 fails to find any mention of splashimage, which suggests that it doesn't support it: # rpm -qa | grep grub grub-0.97-12.caos # grep splashimage /boot/grub/stage2 # When I enable the splashimage in grub.conf nothing at all gets displayed - not even a corrupt image. However, I don't plan to reboot this machine that often! :-) >> Drop to shell and then create /dev/ida directory and then create >> device nodes - ie mknod c0d0 b 72 0, mknod c0d0p1 b 72 1 and >> mknod c0d0p2 b 72 2 >> >> > There is that script that will do that but it creates a boatload of > devices. Are they all needed (I am not familiar with this > hardware)? It really depends upon your hardware - c0d0 refers to controller 0, disk 0, with c0d0p1 then refering to controller 0, disk 0, partition 1 and so on... Perhaps it would be easier just to include the script with instructions on how to run them? Or maybe even just have the script run during the installer's startup? >> Involke lvm.static and then execute commands vgscan and then >> vgchange -ay >> >> VERY IMPORTANT (As I eventually discovered!) if you used Fedora >> to create ext3 paritions you must now reformat them! For me this >> was /dev/ida/c0d0p1 and all the logical volumes under >> /dev/LogGroup00/ (except the swap partition) and as I couldn't >> find mkfs.ext3 I used mkfs.ext2 followed by tune2fs -j (Not >> sure whether you can do the latter this at this stage - if not, >> treat as ext2 and see later) NB If you fail to reformat cAos >> won't load properly on reboot as Red Hat have "kindly" modified >> their versions of ext3!!! (Sounds a lot like Microsoft's >> "embrace and enhance" which I detest so much!) >> > If the partition was created with a newer e2fsprogs then there > could be a problem. We have a version that is static in our core > (we won't upgrade it for caos2 because of our static core ABI > policy). > > And yes... you are right. This can be a problem. I used xfs first time round and didn't have any problems... with all the fuss with ext3 I wish that I'd stuck with xfs! > Very nice write up. Can you copy this to our wiki at > http://wiki.caosity.org please? Thank you! :-) I'll post it in just a second! :) >> Is this the answer to my "I need to install stuff that's only >> available as source but want it easy to upgrade" problem? >> > Yes, and more. It is also the answer to "User 1 needs version x > user 2 version y" etc... Using the modules, every user can select > on a per shell basis what versions of apps that they want to run. > It also allows us to distribute things that are REdistributable > (meaning they can only come from a single source perhaps in binary > form only). An example of this is "real player". > > I have found this to be incredibly valuable for development and > scientific needs. for example, on our clusters users can select > which compiler they want, which mpi they want, which math libs, > etc... they can even select specific versions (egg. gcc-4.0.2 and > openmpi). > > Not to mention even valuable for desktops, etc... We are taking > requests so post your thoughts (not saying we will accomodate, but > all requests will be evaluated). > I can definitely see the advantages on a desktop system. Would it also be suited to ensuring that only specified users have access to development tools on a server? Does it consume much disk space? Once I'm happy that the new server is running properly (I'd just like to state for the record that setting up cyrus-imap is a complete ?$%?@$%, but worth it in the end!) I plan to put cAos on my desktop machine and I'll give it a go then! :) Take care, Ben -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQ4hgc+gNmph0Y1E2AQJ9Ew//eZN/4Sgeq/EqoydI305vMWjJu5dP2GBX 0C9URC1dOBIwhr8MQYXoXCE3Ckbv22y3AJCEFbDzbtLstPAYfmphBRMf8MdA7zaO QxpYTdaUnrzBoOYq9lgfuOSlycqoWPgEDDLJmxivvFSg5IT9uURWvIzAIL9flkAs 6MUiz9PTjdGr4bM+HeuvypAefcj7AfEq6XuFiGWkQoEUpKIyXFVU+S+PjbhMf1Bj 9UyXMGvSRxLajcg3g9f7k5AlBC4Gu5DDdSCjo9JBCYAyg18Ayk8st4+a/oi5SAgR 6ZfmjQvhSUGuRiBUmDYJj3HIiyfdKcjiQn5k0MJlFFtJFbNBlKGycxFM0wbq1KIK IJ8pNsI92lT+4AU2u7ch6HL2qWkhV9j244U3JwcBCFUR1NXmy/dZysit8/RcI594 lRM27UOZBPjlc+XFb+E3cpuu8l4wVjwWKl1+H/GxD4gU2EOCXIab6R/8ei2WGsqD E2YtmNeDzXTtWCF+VsMfDL7jyaw+TFEzR8l7kIjyHwqX2Padzg3+Ccrv2vTtn7BA 6Qy31HeLZpcyajSyDTEPsofxmRL9wmk866Dv45fEx+nxCt0RmibEmM7BmxUidohH P2JCJ3ZDodwb7v3wks1nFSiv76KKVLw7jy6LLj4X4phe/0lG8RhzAXaJ3wGBc4bB rh63f1dK5nM= =X4kg -----END PGP SIGNATURE----- From gmk at runlevelzero.net Sat Nov 26 07:06:34 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Sat, 26 Nov 2005 07:06:34 -0800 Subject: [cAos] Quick question... In-Reply-To: <43886075.10504@pythagoras.no-ip.org> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> <20051120183305.GB26597@titan.runlevelzero.net> <4380D348.30201@pythagoras.no-ip.org> <4380F3DB.3060700@pythagoras.no-ip.org> <4381A6E3.7020608@pythagoras.no-ip.org> <20051121175344.GA30443@titan.runlevelzero.net> <4383600F.8070601@pythagoras.no-ip.org> <20051123180558.GA4799@titan.runlevelzero.net> <43886075.10504@pythagoras.no-ip.org> Message-ID: <20051126150634.GA19285@titan.runlevelzero.net> On Sat, Nov 26, 2005 at 01:17:41PM +0000, Benjamin Donnachie wrote: > So the current repo should be sufficient for a server? I typically use stable for servers and will tap into current for several things when needed: # yum --enablerepo current install openssh-server for example. > Would I be correct in thinking that the stable repo is very stable, > current likely to be stable and testing, well, not likely to be stable? The pathway is: autobuilder -> testing -> current -> stable There is only some basic sanity checking that goes on between autobuilder and testing. Packages must receive human testing to get to current. In cAos-1 this was done via a web page. I will probably make something like that again as I have time (any volunteers)? > >> Device mapper module not loaded - needed for LVM - modprobe > >> dm_mod > >> > > What LVM partition mechanism did you use? If you did it by hand, > > the device module would not have been loaded. > > I ended up using the Fedora install disk to create the LVMs. However, > I did try the assisted LVM install and it failed without any of the > /dev/ LVM nodes being present - I assumed that it was due to the lack > of dm_mod, but it I didn't run vgscan and vgcreate to check. Ahh, yeah. The autoLVM install had a bug. I don't think it will do it any more. ;) > >> Despite using "install smp", it only installs single processor > >> kernel. > >> > >> > > ACK! I will check on that. > > I seem to recall it installing hte smp sources okay, but just not > linux.smp I found this one too. :) > >> Bootloader did not install correctly - probably because I > >> bypassed the disk setup part of install. > >> > >> > > Probably. This just means you would have to configure grub by hand > > if not using the autopartitioning tools. > > Done! :-) I am modifying the manual partition steps so that it will autoprobe the partitions. So even if someone grabs a shell to partition by hand, it will help configure the mount points (and set things up for grub). > >> When configuring static network - no option for DNS or gateway. > >> > >> > > Actually, I know of this. The version I gave you actually is using > > an older core FS with an older sidekick. Current version doesn't do > > that. > > > That's good - I initially thought that my network was playing up when > I set it to static and it couldn't find any of the repos! Then I > realised... :-) Lol... sorry. > > As a matter of fact the version that you have is arched... Do the > > following to upgrade: > > > > # rpm -e sidekick # yum install sidekick > > > > sidekick is now noarch.rpm and yum won't do an upgrade of differing > > arches. > > Done! :) > >> Grub does not support splash images. > >> > >> > > It supports it, but that splash doesn't work with all video cards. > > This is on the TODO. Problem is that it works on all of our > > (developers) systems that do graphics... Thus nobody can debug. ;) > > > > If you can resolve, you will have some good friends! lol > > > I've got grub-0.97-12.caos installed and a grep of /boot/grub/stage2 > fails to find any mention of splashimage, which suggests that it > doesn't support it: Nice catch. I forgot that you had a different version of grub which fixed things like xfs on /boot... > # rpm -qa | grep grub > grub-0.97-12.caos > # grep splashimage /boot/grub/stage2 > # > > When I enable the splashimage in grub.conf nothing at all gets > displayed - not even a corrupt image. However, I don't plan to reboot > this machine that often! :-) I will check on the grub sources for this. > > There is that script that will do that but it creates a boatload of > > devices. Are they all needed (I am not familiar with this > > hardware)? > > It really depends upon your hardware - c0d0 refers to controller 0, > disk 0, with c0d0p1 then refering to controller 0, disk 0, partition 1 > and so on... Perhaps it would be easier just to include the script > with instructions on how to run them? Or maybe even just have the > script run during the installer's startup? Does udev create the devices properly? > > If the partition was created with a newer e2fsprogs then there > > could be a problem. We have a version that is static in our core > > (we won't upgrade it for caos2 because of our static core ABI > > policy). > > > > And yes... you are right. This can be a problem. > > I used xfs first time round and didn't have any problems... with all > the fuss with ext3 I wish that I'd stuck with xfs! Lol, that will teach ya! Seriously, I use xfs for just about everything. > > Very nice write up. Can you copy this to our wiki at > > http://wiki.caosity.org please? > > Thank you! :-) I'll post it in just a second! :) Cool. I may request that you update it once I get the new cinch released :) > > Not to mention even valuable for desktops, etc... We are taking > > requests so post your thoughts (not saying we will accomodate, but > > all requests will be evaluated). > > > I can definitely see the advantages on a desktop system. Would it > also be suited to ensuring that only specified users have access to > development tools on a server? Does it consume much disk space? All access can have access by default. But if you play with the permissions it would be easy to contain the users because all build packages are separated by [appname]/[version]. > Once I'm happy that the new server is running properly (I'd just like > to state for the record that setting up cyrus-imap is a complete > ?$%?@$%, but worth it in the end!) I plan to put cAos on my desktop > machine and I'll give it a go then! :) Sweet. I encourage you to check out the new enlightenment desktop (e17). It hasn't been formally released yet upstream but we have a snapshot that works very well. Glad things are now working for you, and thanks for all the feedback. Greg -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From benjamin at pythagoras.no-ip.org Sun Nov 27 14:52:35 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Sun, 27 Nov 2005 22:52:35 +0000 Subject: [cAos] Quick question... In-Reply-To: <20051126150634.GA19285@titan.runlevelzero.net> References: <20051120145306.GA26597@titan.runlevelzero.net> <4380BD04.4060104@pythagoras.no-ip.org> <20051120183305.GB26597@titan.runlevelzero.net> <4380D348.30201@pythagoras.no-ip.org> <4380F3DB.3060700@pythagoras.no-ip.org> <4381A6E3.7020608@pythagoras.no-ip.org> <20051121175344.GA30443@titan.runlevelzero.net> <4383600F.8070601@pythagoras.no-ip.org> <20051123180558.GA4799@titan.runlevelzero.net> <43886075.10504@pythagoras.no-ip.org> <20051126150634.GA19285@titan.runlevelzero.net> Message-ID: <438A38B3.5080800@pythagoras.no-ip.org> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Greg M. Kurtzer wrote: > I typically use stable for servers and will tap into current for > several things when needed: > > # yum --enablerepo current install openssh-server > > for example. I've been doing that - but quickly discovered, for the programs I wanted to update, the current and stable repos contain the same version. > The pathway is: > > autobuilder -> testing -> current -> stable > > There is only some basic sanity checking that goes on between > autobuilder and testing. Packages must receive human testing to get > to current. In cAos-1 this was done via a web page. I will probably > make something like that again as I have time (any volunteers)? /me raises hand! :-) Would you like a hand with packaging up some of the programs? For example, logwatch is quite out of date (Though that's distributed as a no-arch rpm so shouldn't be any bother), I've been trying to get the graphs in dspam to work but perl is having problems which under RH meant perl had to be recompiled (and if I'm going to do that I might as well package it up!) and I'm using dspam and cyrus-imap and it would be quite handy to have them available as rpms. Alternatively, I suppose I could just package them up and make them available on my server. However, I'm not sure how I'd stand with cross compiling. >> I ended up using the Fedora install disk to create the LVMs. >> However, I did try the assisted LVM install and it failed without >> any of the /dev/ LVM nodes being present - I assumed that it was >> due to the lack of dm_mod, but it I didn't run vgscan and >> vgcreate to check. > > Ahh, yeah. The autoLVM install had a bug. I don't think it will do > it any more. ;) I'll let you know! :-) > I am modifying the manual partition steps so that it will autoprobe > the partitions. So even if someone grabs a shell to partition by > hand, it will help configure the mount points (and set things up > for grub). That would be great! >> When I enable the splashimage in grub.conf nothing at all gets >> displayed - not even a corrupt image. However, I don't plan to >> reboot this machine that often! :-) > > I will check on the grub sources for this. Le me know and I'll give it a go on my other machine! :-) >> It really depends upon your hardware - c0d0 refers to controller >> 0, disk 0, with c0d0p1 then refering to controller 0, disk 0, >> partition 1 and so on... Perhaps it would be easier just to >> include the script with instructions on how to run them? Or >> maybe even just have the script run during the installer's >> startup? > > Does udev create the devices properly? /me slaps head! Why do I always miss the obvious? Yes - udev creates them perfectly! > Seriously, I use xfs for just about everything. As will I when repartition! But for now the server's staying as it is... unfortunately! >>> Very nice write up. Can you copy this to our wiki at >>> http://wiki.caosity.org please? >> >> Thank you! :-) I'll post it in just a second! :) > > Cool. I may request that you update it once I get the new cinch > released :) > Certainly! Though I won't be doing a complete install on my server, but I will certainly check that the relevant devices get created for the disk array. >> Once I'm happy that the new server is running properly (I'd just >> like to state for the record that setting up cyrus-imap is a >> complete ?$%?@$%, but worth it in the end!) I plan to put cAos on >> my desktop machine and I'll give it a go then! :) > > Sweet. I encourage you to check out the new enlightenment desktop > (e17). It hasn't been formally released yet upstream but we have a > snapshot that works very well. No worries - let me know when the next install CD is ready for testing and I'll give it a go then. > Glad things are now working for you, and thanks for all the > feedback. Please let me know if there's anything else I can do to help. Take care, Ben -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQ4o4sugNmph0Y1E2AQK+HxAAq/HHSMuWu1ryF00YKzRI6Jg8jbEo9edF 1vy1hcjPCosWLB6TqG9OjE6Tgr9LE7hM67z42iUkoDu/Fxc4xXaTHUpxu63GhLeE 6eUYXsWEpAibHUz2rkoisFdQ8EFlkNfSI6rAxh3mcRZgDeC0oD5vhMKFWmY3mgn9 q9MS05Gix3dSU6lD/BtbBaEDxX2TSNgAdtG3uxoWIMPIumIZPEkgckbFpeuori2F Muzey9SQqEQYTY5HHK6QDmkHS2NnRMbbTs5eeTpvM91SwO0qFjdsvYXnlZeNBNUE alTDFyuWQ2fjkO1Ou+CFZGl7RI+PgK50f4+pa2kXEpjrs5ermTIv8Tq3R3y/3C6F G1i/oC5IOFHLgSZUUY0p90GEaAPsIpuhRavFBzfLmv+XvnEhBsWf/46tu5MiXa40 uFFt/qINefzabEr7+hwZ6sWN6jJDkKHDLtpr0LVy1JsVpfCW5FhpEAekE+6A/uhq o3tElXHO2tqwD5k8VAlMJsvLA4YofDJ1Xzjgu+FFuefgcz5C0Fiy8/dExm3TEIAx 2Gznnf0KBBXtSR6/NYdGVBEqAuBVJ869Vnl/qvqnwuKWyd25YgfBQJ+g1tanQDnn q2TwZlnAvQThPpADmu4FxNijlNDCPM5bfsb9yrCAISBDw6XgIc1VbAtJMnXyU3xz UHnrO5c8KRY= =/zFs -----END PGP SIGNATURE----- From kulick at gmail.com Mon Nov 28 07:33:12 2005 From: kulick at gmail.com (Matt Kulick) Date: Mon, 28 Nov 2005 10:33:12 -0500 Subject: [cAos] Failure of 2x SATA2 drives under cAos-2 linux --- driver issue? Message-ID: <650e5e400511280733l24c1ded1sc9d56c2d8e2ef75d@mail.gmail.com> Dear fellow cAosers, After installing a new 500GB Seagate SATA2 drive into our cAos-2 powered server, we've experienced several problems that indicate that driver issues may be at fault. I am curious if (1) there are any known SATA2 compatability issues, (2) if anyone else has had similar experiences with this or other SATA2 drives, under load, on recent cAos kernels, and (3) why a 'hung' disk would prevent halt/reboot/shutdown commands from working? Below is a description of what has happened. We bought 2 new drives, qualified them with badblocks runs (destructive read/write tests), and ran one for over a week in light-use with no problems. Then, after a single day of heavy load, it started generating errors. Up until ths point, we had been using a 400gb SATA1 Seagate disk with no problems at all. After the failure, we ran Seagate diagnostics, which only seemed to run a SMART long test, but it still passed. After repeated badblocks 7+ hour runs on the second 500gb sata2 disk, we also encountered the same type of errors the first 500gb sata2 was generating. Here are some details: We are running: Linux voom.astro.cornell.edu 2.6.12-76.caoscustom x86_64 x86_64 x86_64 GNU/Linux The custom kernel just has an FNN patch. The main hardware is a Tyan Tomcat K8S S2850G2NR (SATA1) motherboard with an Opteron 148 CPU. The system included a 400GB Seagate SATA1 disk: Barracuda 7200.8 Seagate 400 Gbytes Model: ST3400832AS P/N: 9y7385-501 Firmware: 3.01 This was stable for many months. Then we installed a new Seagate 500GB SATA2 disk: Barracuda 7200.9 Seagate 500 Gbytes Model: ST3500641AS P/N: 9BD148-301 Firmware: 2.AAA I tested the new disk using: badblocks /dev/sdd -o whatever.bb -w -t random -s -v -c 32768 It tested fine after one pass, and ran fine for a few weeks of light use. Then it started throwing errors a day or so after we started heavy use: Nov 19 04:02:33 voom kernel: ata4: command 0x25 timeout, stat 0xd9 host_stat 0x1Nov 19 04:02:33 voom kernel: ata4: status=0xd9 { Busy } Nov 19 04:02:33 voom kernel: SCSI error : <3 0 0 0> return code = 0x8000002 Nov 19 04:02:33 voom kernel: sdd: Current: sense key: Aborted Command Nov 19 04:02:33 voom kernel: Additional sense: Scsi parity error Nov 19 04:02:33 voom kernel: end_request: I/O error, dev sdd, sector 825000119 Nov 19 04:02:33 voom kernel: ATA: abnormal status 0xD9 on port 0xFFFFC20000004EC7 Most processes that accessed the disk after this point went into diskwait, and the system would not reboot from software. After a hardware reboot and fsck, the same thing happened: a day or so of good running, then the same errors and behavior. I then installed a second, identical SATA2 disk, and tested both disks repeatedly with the badblocks command above. After several successful 7+ hour passes of badblocks, the second SATA2 disk failed in (almost) exactly the same way: Nov 24 12:12:19 voom kernel: ata3: command 0x25 timeout, stat 0xd8 host_stat 0x1Nov 24 12:12:19 voom kernel: ata3: status=0xd8 { Busy } Nov 24 12:12:19 voom kernel: SCSI error : <2 0 0 0> return code = 0x8000002 Nov 24 12:12:19 voom kernel: sdc: Current: sense key: Aborted Command Nov 24 12:12:19 voom kernel: Additional sense: Scsi parity error Nov 24 12:12:19 voom kernel: end_request: I/O error, dev sdc, sector 212810760 Nov 24 12:12:19 voom kernel: ATA: abnormal status 0xD8 on port 0xFFFFC20000004E87 The sdd disk passed Seagate's diagnostics, which I think is just the SMART long test (which is not a read-write test). It took a day or so of tests before the errors started flowing. Since two disks are failing the same way, I suspect a driver problem (though it could be a bad batch at Seagate). Any ideas? Thanks for the input! -matt From tmattox at gmail.com Mon Nov 28 08:23:49 2005 From: tmattox at gmail.com (Tim Mattox) Date: Mon, 28 Nov 2005 11:23:49 -0500 Subject: [cAos] Failure of 2x SATA2 drives under cAos-2 linux --- driver issue? In-Reply-To: <650e5e400511280733l24c1ded1sc9d56c2d8e2ef75d@mail.gmail.com> References: <650e5e400511280733l24c1ded1sc9d56c2d8e2ef75d@mail.gmail.com> Message-ID: Hi Matt, SATA2 is rather new... you might need a newer kernel version with a more recent libata driver. Can you test the drives on a machine without the FNN? If so, you can try with a 2.6.14 kernel to see if the problems go away. If they do, I'll see what I can do about quickly porting my FNN patches to the 2.6.14 kernel. I had hoped to avoid that, but hey, It shouldn't be too bad to update them... I hope... I spent a few minutes trying to find any relevant SATA patches between the 2.6.12 and 2.6.14 kernels, but there are so many patches that I wasn't able to quickly find anything relevant. You might try using the kernel.org's online GIT patch system to look yourself if you feel up to it. ;-) See: http://kernel.org/git/?p=linux/kernel/git/jgarzik/libata-dev.git;a=summary and http://kernel.org/git/?p=linux/kernel/git/bart/ide-2.6.git;a=summary Maybe you can search on your motherboard's SATA controller's name/part number to find any relevant patches that might fix things. On 11/28/05, Matt Kulick wrote: > Dear fellow cAosers, > > After installing a new 500GB Seagate SATA2 drive into our cAos-2 > powered server, we've experienced several problems that indicate that > driver issues may be at fault. I am curious if (1) there are any > known SATA2 compatability issues, (2) if anyone else has had similar > experiences with this or other SATA2 drives, under load, on recent > cAos kernels, and (3) why a 'hung' disk would prevent > halt/reboot/shutdown commands from working? > > Below is a description of what has happened. We bought 2 new drives, > qualified them with badblocks runs (destructive read/write tests), and > ran one for over a week in light-use with no problems. Then, after a > single day of heavy load, it started generating errors. Up until ths > point, we had been using a 400gb SATA1 Seagate disk with no problems > at all. After the failure, we ran Seagate diagnostics, which only > seemed to run a SMART long test, but it still passed. After repeated > badblocks 7+ hour runs on the second 500gb sata2 disk, we also > encountered the same type of errors the first 500gb sata2 was > generating. > > Here are some details: > > We are running: > > Linux voom.astro.cornell.edu 2.6.12-76.caoscustom x86_64 x86_64 x86_64 GNU/Linux > > The custom kernel just has an FNN patch. > The main hardware is a Tyan Tomcat K8S S2850G2NR (SATA1) motherboard > with an Opteron 148 CPU. > The system included a 400GB Seagate SATA1 disk: > > Barracuda 7200.8 Seagate > 400 Gbytes > Model: ST3400832AS > P/N: 9y7385-501 > Firmware: 3.01 > > This was stable for many months. Then we installed a new Seagate > 500GB SATA2 disk: > > Barracuda 7200.9 Seagate > 500 Gbytes > Model: ST3500641AS > P/N: 9BD148-301 > Firmware: 2.AAA > > I tested the new disk using: > > badblocks /dev/sdd -o whatever.bb -w -t random -s -v -c 32768 > > It tested fine after one pass, and ran fine for a few weeks of light > use. Then it started throwing errors a day or so after we started > heavy use: > > Nov 19 04:02:33 voom kernel: ata4: command 0x25 timeout, stat 0xd9 > host_stat 0x1Nov 19 04:02:33 voom kernel: ata4: status=0xd9 { Busy } > Nov 19 04:02:33 voom kernel: SCSI error : <3 0 0 0> return code = 0x8000002 > Nov 19 04:02:33 voom kernel: sdd: Current: sense key: Aborted Command > Nov 19 04:02:33 voom kernel: Additional sense: Scsi parity error > Nov 19 04:02:33 voom kernel: end_request: I/O error, dev sdd, sector 825000119 > Nov 19 04:02:33 voom kernel: ATA: abnormal status 0xD9 on port > 0xFFFFC20000004EC7 > > Most processes that accessed the disk after this point went into > diskwait, and the system would not reboot from software. After a > hardware reboot and fsck, the same thing happened: a day or so of good > running, then the same errors and behavior. > > I then installed a second, identical SATA2 disk, and tested both disks > repeatedly with the badblocks command above. After several successful > 7+ hour passes of badblocks, the second SATA2 disk failed in (almost) > exactly the same way: > > Nov 24 12:12:19 voom kernel: ata3: command 0x25 timeout, stat 0xd8 > host_stat 0x1Nov 24 12:12:19 voom kernel: ata3: status=0xd8 { Busy } > Nov 24 12:12:19 voom kernel: SCSI error : <2 0 0 0> return code = 0x8000002 > Nov 24 12:12:19 voom kernel: sdc: Current: sense key: Aborted Command > Nov 24 12:12:19 voom kernel: Additional sense: Scsi parity error > Nov 24 12:12:19 voom kernel: end_request: I/O error, dev sdc, sector 212810760 > Nov 24 12:12:19 voom kernel: ATA: abnormal status 0xD8 on port > 0xFFFFC20000004E87 > > The sdd disk passed Seagate's diagnostics, which I think is just the > SMART long test (which is not a read-write test). It took a day or so > of tests before the errors started flowing. Since two disks are > failing the same way, I suspect a driver problem (though it could be a > bad batch at Seagate). > > > Any ideas? Thanks for the input! > > -matt > _______________________________________________ > cAos mailing list > cAos at caosity.org > http://lists.caosity.org/mailman/listinfo/caos > -- Tim Mattox - tmattox at gmail.com http://homepage.mac.com/tmattox/ I'm a bright... http://www.the-brights.net/ From gmk at runlevelzero.net Mon Nov 28 10:54:21 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Mon, 28 Nov 2005 10:54:21 -0800 Subject: [cAos] Quick question... In-Reply-To: <438A38B3.5080800@pythagoras.no-ip.org> References: <20051120183305.GB26597@titan.runlevelzero.net> <4380D348.30201@pythagoras.no-ip.org> <4380F3DB.3060700@pythagoras.no-ip.org> <4381A6E3.7020608@pythagoras.no-ip.org> <20051121175344.GA30443@titan.runlevelzero.net> <4383600F.8070601@pythagoras.no-ip.org> <20051123180558.GA4799@titan.runlevelzero.net> <43886075.10504@pythagoras.no-ip.org> <20051126150634.GA19285@titan.runlevelzero.net> <438A38B3.5080800@pythagoras.no-ip.org> Message-ID: <20051128185421.GA14322@titan.runlevelzero.net> On Sun, Nov 27, 2005 at 10:52:35PM +0000, Benjamin Donnachie wrote: > > # yum --enablerepo current install openssh-server > > > > for example. > > I've been doing that - but quickly discovered, for the programs I > wanted to update, the current and stable repos contain the same version. openssh < 4.2 has been removed from current and was synced to stable. If you are running stable now... do the update, and then enablerepo current and do another update or list updates to see the differences. > > The pathway is: > > > > autobuilder -> testing -> current -> stable > > > > There is only some basic sanity checking that goes on between > > autobuilder and testing. Packages must receive human testing to get > > to current. In cAos-1 this was done via a web page. I will probably > > make something like that again as I have time (any volunteers)? > > /me raises hand! :-) Would you like a hand with packaging up some of > the programs? For example, logwatch is quite out of date (Though > that's distributed as a no-arch rpm so shouldn't be any bother), I've > been trying to get the graphs in dspam to work but perl is having > problems which under RH meant perl had to be recompiled (and if I'm > going to do that I might as well package it up!) and I'm using dspam > and cyrus-imap and it would be quite handy to have them available as > rpms. Alternatively, I suppose I could just package them up and make > them available on my server. However, I'm not sure how I'd stand with > cross compiling. Yes... you in the front! Initially if you want to help out with packaging, check out the mezzanine howtos at: http://www.kainx.org/mezzanine/mezz-howto.txt http://troy.jdmz.net/mezz/mezz-cvs.txt Then build the packages and send a link to the SRPMS to the caos-devel list. If you find yourself wanting access to our CVS tree, we can do that with the proper handshakes. ;) You might also find it worth while to join up to our IRC channel at irc.freenode.net, #caos and talk to us real time (same for others interested in real time communication). > > Ahh, yeah. The autoLVM install had a bug. I don't think it will do > > it any more. ;) > > I'll let you know! :-) OK. I think I got it. http://runlevelzero.net/static/caos/caos-2-2.2a.i386.iso http://runlevelzero.net/static/caos/cinch-2.2-raid_lvm-notes.txt > > I am modifying the manual partition steps so that it will autoprobe > > the partitions. So even if someone grabs a shell to partition by > > hand, it will help configure the mount points (and set things up > > for grub). > > That would be great! This has been implemetned. :) > >> When I enable the splashimage in grub.conf nothing at all gets > >> displayed - not even a corrupt image. However, I don't plan to > >> reboot this machine that often! :-) > > > > I will check on the grub sources for this. > > Le me know and I'll give it a go on my other machine! :-) The new iso is running an updated core file system and the same grub as you have now. I will integrate the spashscreen stuff after. > > Does udev create the devices properly? > > /me slaps head! Why do I always miss the obvious? Yes - udev creates > them perfectly! :) > > Seriously, I use xfs for just about everything. > > As will I when repartition! But for now the server's staying as it > is... unfortunately! :) > > Cool. I may request that you update it once I get the new cinch > > released :) > > > Certainly! Though I won't be doing a complete install on my server, > but I will certainly check that the relevant devices get created for > the disk array. Excellent. Thanks! > >> Once I'm happy that the new server is running properly (I'd just > >> like to state for the record that setting up cyrus-imap is a > >> complete ?$%?@$%, but worth it in the end!) I plan to put cAos on > >> my desktop machine and I'll give it a go then! :) > > > > Sweet. I encourage you to check out the new enlightenment desktop > > (e17). It hasn't been formally released yet upstream but we have a > > snapshot that works very well. > > No worries - let me know when the next install CD is ready for testing > and I'll give it a go then. Let me know how it works out for ya. > > Glad things are now working for you, and thanks for all the > > feedback. > > Please let me know if there's anything else I can do to help. /me pulls out the book... Hrmm... probably better to let me know that you want to help, and XYZ interests you. :) No doubt we can work you into a happy position! -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From mej at caosity.org Mon Nov 28 11:34:54 2005 From: mej at caosity.org (Michael Jennings) Date: Mon, 28 Nov 2005 14:34:54 -0500 Subject: [cAos] Quick question... In-Reply-To: <20051128185421.GA14322@titan.runlevelzero.net> References: <4380D348.30201@pythagoras.no-ip.org> <4380F3DB.3060700@pythagoras.no-ip.org> <4381A6E3.7020608@pythagoras.no-ip.org> <20051121175344.GA30443@titan.runlevelzero.net> <4383600F.8070601@pythagoras.no-ip.org> <20051123180558.GA4799@titan.runlevelzero.net> <43886075.10504@pythagoras.no-ip.org> <20051126150634.GA19285@titan.runlevelzero.net> <438A38B3.5080800@pythagoras.no-ip.org> <20051128185421.GA14322@titan.runlevelzero.net> Message-ID: <20051128193454.GA16826@kainx.org> On Monday, 28 November 2005, at 10:54:21 (-0800), Greg M. Kurtzer wrote: > Initially if you want to help out with packaging, check out the > mezzanine howtos at: > > http://www.kainx.org/mezzanine/mezz-howto.txt > http://troy.jdmz.net/mezz/mezz-cvs.txt Actually, this one is probably the best and most up-to-date, although it's still not 100% as our docs guy hasn't been able to update it in quite some time.... :( Anyway, here: https://wiki.caosity.org/tiki-index.php?page=Package+Maintenance&bl > Then build the packages and send a link to the SRPMS to the > caos-devel list. If you find yourself wanting access to our CVS > tree, we can do that with the proper handshakes. ;) > > You might also find it worth while to join up to our IRC channel at > irc.freenode.net, #caos and talk to us real time (same for others > interested in real time communication). Yup, I'd say that's the best place to find experienced Mezzanine users and packagers. Michael -- Michael Jennings (a.k.a. KainX) http://www.kainx.org/ n + 1, Inc., http://www.nplus1.net/ Author, Eterm (www.eterm.org) ----------------------------------------------------------------------- "I don't know what will be used in the next world war, but the fourth will be fought with stones." -- Albert Einstein From tmattox at gmail.com Mon Nov 28 12:44:38 2005 From: tmattox at gmail.com (Tim Mattox) Date: Mon, 28 Nov 2005 15:44:38 -0500 Subject: [cAos] Installing the Intel compilers on cAos-2? Message-ID: Hello, Is there anyone on the list who has successfully installed the Intel compilers on a cAos-2 system? How about on an Athlon 64 system even? I tried this past spring and was unsuccessful. I don't recall the error, or even the version of the compiler... yeah, I'm being so helpful on debugging this... ;-) Anyway, I will be trying again next week, and was just hoping to hear from anyone who might have already gotten it to work. Thanks! -- Tim Mattox - tmattox at gmail.com http://homepage.mac.com/tmattox/ I'm a bright... http://www.the-brights.net/ From astevens at gravitypark.com Mon Nov 28 12:34:52 2005 From: astevens at gravitypark.com (Arthur Stevens) Date: Mon, 28 Nov 2005 12:34:52 -0800 Subject: [cAos] Failure of 2x SATA2 drives under cAos-2 linux --- driverissue? References: <650e5e400511280733l24c1ded1sc9d56c2d8e2ef75d@mail.gmail.com> Message-ID: <001001c5f45b$339e70b0$6800a8c0@astevens> Hi Matt, interesting issues. As a side note, I had a bad 500gb seagate from the manufacturer. The seagate software said all was well, but we were loosing a lot of data. Seagates just dont seem like the drives they use to be when they were sub 100gb. When I took it back, my hardware guy gave me a 400 and said the 500's are dieing after a week of heavy load. Not sure if that is what is causing your issues or not. I have since killed the 400 as well. My 160's have been running for months without any issues, and i'm hearing not so good stuff about the larger drives right now. I have been runnin 4 160gb SATA2 drives here just fine, sometimes at a very heavy load. I am using kernel 2.6.14 currently with no issues. I had a SuperMicro server doing this a while back, and it was fixed by adding a little silicon glue to the cable so the vibration of hevy drive use would not wiggle it, as I hear this is a common issue, that is why most server palces that use sata glue the cables on this way. Not sure if this helps at all. Good luck, we all know how frustrating drive issues can be. Arthur ----- Original Message ----- From: "Matt Kulick" To: Sent: Monday, November 28, 2005 7:33 AM Subject: [cAos] Failure of 2x SATA2 drives under cAos-2 linux --- driverissue? > Dear fellow cAosers, > > After installing a new 500GB Seagate SATA2 drive into our cAos-2 > powered server, we've experienced several problems that indicate that > driver issues may be at fault. I am curious if (1) there are any > known SATA2 compatability issues, (2) if anyone else has had similar > experiences with this or other SATA2 drives, under load, on recent > cAos kernels, and (3) why a 'hung' disk would prevent > halt/reboot/shutdown commands from working? > > Below is a description of what has happened. We bought 2 new drives, > qualified them with badblocks runs (destructive read/write tests), and > ran one for over a week in light-use with no problems. Then, after a > single day of heavy load, it started generating errors. Up until ths > point, we had been using a 400gb SATA1 Seagate disk with no problems > at all. After the failure, we ran Seagate diagnostics, which only > seemed to run a SMART long test, but it still passed. After repeated > badblocks 7+ hour runs on the second 500gb sata2 disk, we also > encountered the same type of errors the first 500gb sata2 was > generating. > > Here are some details: > > We are running: > > Linux voom.astro.cornell.edu 2.6.12-76.caoscustom x86_64 x86_64 x86_64 > GNU/Linux > > The custom kernel just has an FNN patch. > The main hardware is a Tyan Tomcat K8S S2850G2NR (SATA1) motherboard > with an Opteron 148 CPU. > The system included a 400GB Seagate SATA1 disk: > > Barracuda 7200.8 Seagate > 400 Gbytes > Model: ST3400832AS > P/N: 9y7385-501 > Firmware: 3.01 > > This was stable for many months. Then we installed a new Seagate > 500GB SATA2 disk: > > Barracuda 7200.9 Seagate > 500 Gbytes > Model: ST3500641AS > P/N: 9BD148-301 > Firmware: 2.AAA > > I tested the new disk using: > > badblocks /dev/sdd -o whatever.bb -w -t random -s -v -c 32768 > > It tested fine after one pass, and ran fine for a few weeks of light > use. Then it started throwing errors a day or so after we started > heavy use: > > Nov 19 04:02:33 voom kernel: ata4: command 0x25 timeout, stat 0xd9 > host_stat 0x1Nov 19 04:02:33 voom kernel: ata4: status=0xd9 { Busy } > Nov 19 04:02:33 voom kernel: SCSI error : <3 0 0 0> return code = > 0x8000002 > Nov 19 04:02:33 voom kernel: sdd: Current: sense key: Aborted Command > Nov 19 04:02:33 voom kernel: Additional sense: Scsi parity error > Nov 19 04:02:33 voom kernel: end_request: I/O error, dev sdd, sector > 825000119 > Nov 19 04:02:33 voom kernel: ATA: abnormal status 0xD9 on port > 0xFFFFC20000004EC7 > > Most processes that accessed the disk after this point went into > diskwait, and the system would not reboot from software. After a > hardware reboot and fsck, the same thing happened: a day or so of good > running, then the same errors and behavior. > > I then installed a second, identical SATA2 disk, and tested both disks > repeatedly with the badblocks command above. After several successful > 7+ hour passes of badblocks, the second SATA2 disk failed in (almost) > exactly the same way: > > Nov 24 12:12:19 voom kernel: ata3: command 0x25 timeout, stat 0xd8 > host_stat 0x1Nov 24 12:12:19 voom kernel: ata3: status=0xd8 { Busy } > Nov 24 12:12:19 voom kernel: SCSI error : <2 0 0 0> return code = > 0x8000002 > Nov 24 12:12:19 voom kernel: sdc: Current: sense key: Aborted Command > Nov 24 12:12:19 voom kernel: Additional sense: Scsi parity error > Nov 24 12:12:19 voom kernel: end_request: I/O error, dev sdc, sector > 212810760 > Nov 24 12:12:19 voom kernel: ATA: abnormal status 0xD8 on port > 0xFFFFC20000004E87 > > The sdd disk passed Seagate's diagnostics, which I think is just the > SMART long test (which is not a read-write test). It took a day or so > of tests before the errors started flowing. Since two disks are > failing the same way, I suspect a driver problem (though it could be a > bad batch at Seagate). > > > Any ideas? Thanks for the input! > > -matt > _______________________________________________ > cAos mailing list > cAos at caosity.org > http://lists.caosity.org/mailman/listinfo/caos > From dgs at gs.washington.edu Mon Nov 28 13:13:39 2005 From: dgs at gs.washington.edu (DGS) Date: Mon, 28 Nov 2005 13:13:39 -0800 Subject: [cAos] Installing the Intel compilers on cAos-2? In-Reply-To: References: Message-ID: <20051128211338.GA1867@zambezi.gs.washington.edu> > Is there anyone on the list who has successfully installed the Intel compilers > on a cAos-2 system? How about on an Athlon 64 system even? > I tried this past spring and was unsuccessful. > I don't recall the error, or even the version of the compiler... yeah, > I'm being so > helpful on debugging this... ;-) I've got version 8.1 of 'icc' and friends running on a Xeon x86 cAos-2 system. I don't recall any problems with installing it. As for Athlon systems, you might look at http://www.swallowtail.org/naughty-intel.html David S. > From gmk at runlevelzero.net Tue Nov 29 22:45:10 2005 From: gmk at runlevelzero.net (Greg M. Kurtzer) Date: Tue, 29 Nov 2005 22:45:10 -0800 Subject: [cAos] caos2 repository update Message-ID: <20051130064510.GA11531@titan.runlevelzero.net> current: + SRPMS/awstats-6.4-2.caos.src.rpm + SRPMS/cmt-1.15-2.caos.src.rpm + SRPMS/DirectFB-0.9.24-1.caos.src.rpm + SRPMS/drupal-4.6.3-1.caos.src.rpm + SRPMS/e3-2.6.3-1.caos.src.rpm + SRPMS/gimp-2.2.9-0.caos.src.rpm + SRPMS/gstreamer-ffmpeg-0.8.7-1.caos.src.rpm + SRPMS/ladspa-1.12-3.caos.src.rpm + SRPMS/lftp-3.3.3-1.src.rpm + SRPMS/libmp3splt-0.2-1.caos.src.rpm + SRPMS/linux-2.6.14.2-95.caos.src.rpm + SRPMS/mambo-4.5.2.3-1.caos.src.rpm + SRPMS/mp3splt-2.1-1.src.rpm + SRPMS/mp3wrap-0.5-1.caos.src.rpm + SRPMS/MySQL-4.1.15-3.src.rpm + SRPMS/nagios-1.2-2.caos.src.rpm + SRPMS/perl-DBD-MySQL-3.0002-8.caos.src.rpm + SRPMS/perl-DBD-Pg-1.43-1.caos.src.rpm + SRPMS/perl-File-Copy-Link-0.061-1.caos.src.rpm + SRPMS/perl-File-Copy-Recursive-0.16-1.caos.src.rpm + SRPMS/perl-Glib-1.101-3.caos.src.rpm + SRPMS/perl-Gtk2-1.101-2.caos.src.rpm + SRPMS/perl-Module-Info-0.28-1.caos.src.rpm + SRPMS/perl-Module-ScanDeps-0.51-1.caos.src.rpm + SRPMS/perl-XML-TreeBuilder-3.09-1.caos.src.rpm + SRPMS/php-4.3.9-6.caos.src.rpm + SRPMS/portstree-0.2-1.caos.src.rpm + SRPMS/rat-4.2.25-2.caos.src.rpm + SRPMS/SDL_image-1.2.3-4.3.caos.src.rpm + SRPMS/SDL_mixer-1.2.5-5.caos.src.rpm + SRPMS/smart-0.40-6.caos.src.rpm + SRPMS/tikiwiki-1.9.1.1-3.caos.src.rpm + SRPMS/time-1.7-27.src.rpm + SRPMS/tvtime-1.0.1-12.caos.src.rpm + SRPMS/ucl-common-1.2.16-3.src.rpm + SRPMS/usermin-1.179-1.caos.src.rpm + SRPMS/webmin-1.249-1.caos.src.rpm - SRPMS/ecore-0.9.9.013-1.src.rpm - SRPMS/edje-0.5.0-1.20050224.src.rpm - SRPMS/embryo-0.9.1-1.20050224.src.rpm - SRPMS/embryo-0.9.1.013-1.src.rpm - SRPMS/emotion-0.0.1-1.20050224.src.rpm - SRPMS/engage-0.0.9-1.20050220.src.rpm - SRPMS/entice-0.9.3-1.20050224.src.rpm - SRPMS/epsilon-0.3.0-1.20050224.src.rpm - SRPMS/esmart-0.9.0-1.20050224.src.rpm - SRPMS/evas-0.9.9.013-1.src.rpm - SRPMS/ewl-0.0.4-0.20050224.src.rpm - SRPMS/linux-2.6.13-82.caos.src.rpm - SRPMS/linux-2.6.13.4-86.caos.src.rpm - SRPMS/php-4.3.9-5.caos.src.rpm - SRPMS/portstree-0.1-1.caos.src.rpm - SRPMS/SDL_image-1.2.3-4.2.caos.src.rpm - SRPMS/SDL_mixer-1.2.5-4.caos.src.rpm stable: + SRPMS/awstats-6.4-2.caos.src.rpm + SRPMS/drupal-4.6.3-1.caos.src.rpm + SRPMS/ecore-0.9.9.018-1.20051017.src.rpm + SRPMS/edje-0.5.0.018-1.20051017.src.rpm + SRPMS/embryo-0.9.1.018-1.20051017.src.rpm + SRPMS/emotion-0.0.1.004-1.20051017.src.rpm + SRPMS/engage-0.0.9-1.20051017.src.rpm + SRPMS/entice-0.9.3.004-1.20051017.src.rpm + SRPMS/epsilon-0.3.0.004-1.20051017.src.rpm + SRPMS/esmart-0.9.0.004-1.20051017.src.rpm + SRPMS/evas-0.9.9.019-1.20051017.src.rpm + SRPMS/ewl-0.0.4.004-1.20051017.src.rpm + SRPMS/ladspa-1.12-3.caos.src.rpm + SRPMS/libmp3splt-0.2-1.caos.src.rpm + SRPMS/mambo-4.5.2.3-1.caos.src.rpm + SRPMS/mp3splt-2.1-1.src.rpm + SRPMS/mp3wrap-0.5-1.caos.src.rpm + SRPMS/perl-File-Copy-Recursive-0.16-1.caos.src.rpm + SRPMS/perl-Gtk2-1.101-2.caos.src.rpm + SRPMS/perl-Module-Info-0.28-1.caos.src.rpm + SRPMS/perl-Module-ScanDeps-0.51-1.caos.src.rpm + SRPMS/perl-XML-TreeBuilder-3.09-1.caos.src.rpm + SRPMS/php-4.3.9-6.caos.src.rpm + SRPMS/portstree-0.2-1.caos.src.rpm + SRPMS/rat-4.2.25-2.caos.src.rpm + SRPMS/SDL_image-1.2.3-4.3.caos.src.rpm + SRPMS/SDL_mixer-1.2.5-5.caos.src.rpm + SRPMS/smart-0.40-6.caos.src.rpm + SRPMS/time-1.7-27.src.rpm + SRPMS/ucl-common-1.2.16-3.src.rpm - SRPMS/ecore-0.9.9.013-1.src.rpm - SRPMS/edje-0.5.0-1.20050224.src.rpm - SRPMS/embryo-0.9.1-1.20050224.src.rpm - SRPMS/emotion-0.0.1-1.20050224.src.rpm - SRPMS/engage-0.0.9-1.20050220.src.rpm - SRPMS/entice-0.9.3-1.20050224.src.rpm - SRPMS/epsilon-0.3.0-1.20050224.src.rpm - SRPMS/esmart-0.9.0-1.20050224.src.rpm - SRPMS/evas-0.9.9.013-1.src.rpm - SRPMS/ewl-0.0.4-0.20050224.src.rpm - SRPMS/php-4.3.9-5.caos.src.rpm - SRPMS/portstree-0.1-1.caos.src.rpm - SRPMS/SDL_image-1.2.3-4.2.caos.src.rpm - SRPMS/SDL_mixer-1.2.5-4.caos.src.rpm -- Greg M. Kurtzer http://runlevelzero.net/ http://caosity.org/ http://warewulf-cluster.org/ Do not look anywhere for truth, for all that is needed is to refrain from allowing concepts to arise. From benjamin at pythagoras.no-ip.org Wed Nov 30 09:08:48 2005 From: benjamin at pythagoras.no-ip.org (Benjamin Donnachie) Date: Wed, 30 Nov 2005 17:08:48 +0000 Subject: [cAos] Quick question... In-Reply-To: <20051128185421.GA14322@titan.runlevelzero.net> References: <20051120183305.GB26597@titan.runlevelzero.net> <4380D348.30201@pythagoras.no-ip.org> <4380F3DB.3060700@pythagoras.no-ip.org> <4381A6E3.7020608@pythagoras.no-ip.org> <20051121175344.GA30443@titan.runlevelzero.net> <4383600F.8070601@pythagoras.no-ip.org> <20051123180558.GA4799@titan.runlevelzero.net> <43886075.10504@pythagoras.no-ip.org> <20051126150634.GA19285@titan.runlevelzero.net> <438A38B3.5080800@pythagoras.no-ip.org> <20051128185421.GA14322@titan.runlevelzero.net> Message-ID: <438DDCA0.9030709@pythagoras.no-ip.org> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Greg M. Kurtzer wrote: > OK. I think I got it. > > > http://runlevelzero.net/static/caos/caos-2-2.2a.i386.iso > http://runlevelzero.net/static/caos/cinch-2.2-raid_lvm-notes.txt I tried to install the above on my desktop machine but, when prompted for the system roles, it states that it cannot contact the repo mirror and hence cannot download the package group information. I tried with both current and stable repos but still no luck - however, the network was definitely up as I could ping all over the place! Unfortunately, I haven't got time to isolate the fault as I am repairing a friends Windows (spit!) laptop and, consequently, I'm installing test3 instead... Which is working fine! Take care, Ben -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.1 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQIVAwUBQ43cn+gNmph0Y1E2AQKcpA//TcbAnUDtUxXEa87qgccAoTR9HxL6+5tV ptcyHW/4IZJIxjpL+mNxNS/n2YaGFDqTdD8Wx+fSSufizW6f+j5blHDcnYcfZSnV 8DmlAThWcULV642VRtzb8+yBqv8MfDmWTdv1cPdfSZFE9SH9D83JzYO2AjVxajZi Y45/fn/Lz00BgiLjX5qSGtEfW/u5NZySH2gA9taM/paHTtAPpxY9X9bY8YVvkRx2 mLgCLMHGRsfGbB+x5eTStpBP1wfjQTW8QN9oRY/h5zQRR7MF/LmjIkiE+NHuvKOB nA+sF9M45TTxJLLlsx57F9XBD6ZGmjq7LS9sWK0fiKFz78FTLgK0Pyb68KiiWycy 7E4gvDBR/xv7kYvSVesJGwQnl9r5OHK252zo4bEx4RMygjsDO9IDCNfnwcMwDfdC 6sMcf4WtgtGQy/rIRtPanGkujpDaAXJRlwp3w+NLfbtlDbPg4B3j9DWisOrXcZi4 Kc9ioJPQfIpfFZi7msurNz6Etq1ePC55Be0j5vHgK/zYJ2uUYGu5M8GAqqidDrP2 Zq5tmOVtmnYM5bMZ2LL7mcLpD/XuMS2IBkqRfu39d0TQmECe11Gu+Iop3YcfWlJC 9bqOUwtTADriS06wZWzE3r8iWpxVGzeAL2tD2W7I35/aOc0cIrkxM/8cpj5Otwy9 vuKZqwWxDZc= =gB4u -----END PGP SIGNATURE-----