[05:35] rcy (~rcy@209.90.177.236) joined #fai. [05:35] hello [06:13] Nick change: \sh_away -> \sh [07:19] Mrfai (~lange@kueppers.informatik.uni-koeln.de) joined #fai. [07:33] glance (glance@montezuma.acc.umu.se) left irc: Ping timeout: 480 seconds [07:56] Bluemoon (~mj@www.bfr.bund.de) joined #fai. [07:56] guten morgen [07:58] glance (glance@shaka.acc.umu.se) joined #fai. [08:11] I want to use FAI to do ubuntu installations. Should I run the server from a debian machine, or ubuntu, or does it matter? The demohost installation has some issue with grub not being available when working from a feisty server [08:12] rcy: in the end it doesnt matter, its just up to you. [08:13] ok... that's sorta what I suspected, I just wanted to make sure making the wrong choice wasnt going to cause me endless problems [08:13] its taking me a little while to get my head wrapped around how this all works [08:14] anyhow it whill end whith you needing to write/customize some scripts to get it all working as you want it to. [08:16] oh definitly [08:17] im looking to just have the demohost installed and booting as the first step [08:25] siretart: did you already tested 3.2betaxx with ubuntu? I like to release FAI 3.2 next week. [08:38] should /srv/fai/nfsroot/proc be an actual proc filesystem on my faiserver? [08:38] make-fai-nfsroot is failing because it cant remove that directory [09:14] meandtheshel1 (~markus@85.127.103.154) joined #fai. [09:15] h01ger (~holger@socket.layer-acht.org) left irc: Remote host closed the connection [09:16] h01ger (~holger@socket.layer-acht.org) joined #fai. [09:17] type mount and you will see that there are remaining mounts. These are from a previous make-fai-nfsroot call, that you've interrupted or suspended [09:18] Mrfai_ (~lange@suenner.informatik.uni-koeln.de) joined #fai. [09:18] yeah yer right... I cancelled one [09:25] Mrfai: yes, I was able to install it, but I think I had not dma on while installing, so it tool really ages [09:25] Mrfai: I didn't boot the system yet, but I'm confident [09:26] Mrfai: did you include the workaround for update-initramfs pointing to /bin/true? [09:34] yep [09:35] no dma is mostly cased by a missing kernel module inside the initramfs. If the generic ide driver is loaded before the special driver the the chipset, you do not have dma. [09:52] I'm not sure what script is running atm, but I see "Mounting /dev/hda8 to /tmp/target/usr" followed by a bunch of "[: 93: ==: unexpected operator" errors [09:52] put the fai.log to pste.debian.net. and also you disk_config file and format.log [09:53] ok, ill wait till the install finishes [10:06] well, after all that, it did work this time... i booted into the system! [10:08] it immediatly rebooted though... is the fai.log available somewhere? [10:11] on the install client in /var/log/fai, on the install server in homedir of $LOGUSER if you configured it to do so. Mmm, this maybe a FAQ for the wiki [10:11] i did see stuff in the config files about it... but didnt bother touching it yet [10:12] maybe it should write it to the target hd somewhere as well [10:13] weird that it just rebooted automatically though... had i not been around to pull the ethernet, it would have started the install again [10:15] it does write it into the target [10:15] oh... where? [10:15] i see /var/log/bootstrap.log, but thats not it [10:16] /tmp/tar/var/log/fai [10:16] sorry [10:16] /tmp/target/var/log/fai [10:17] excellent [10:23] fai.log: http://paste.debian.net/34670 [10:24] format.log: http://paste.debian.net/34671 [10:24] now it's a FAQ: http://faiwiki/index.php/User_FAQ#Where_do_I_find_the_log_files.3F [10:26] disk_config: http://paste.debian.net/34672 [10:28] where do i find the code that does this: "Calling task_mountdisks"? [10:30] /srv/fai/nfsroot/usr/lib/fai/subroutines and ...usr/lib/fai/mount2dir [10:34] it might be nice if the echo calls in the various scripts did a echo "$0: ...", so one could more easily track down where things are coming from [10:41] mount2dir is the only script with == in it that i can see... but it looks fine [10:46] Nick change: \sh -> \sh_away [10:46] add set -x to if, then you will get much debugging output [10:48] to if? I don't exactly know what you mean [10:58] Nick change: \sh_away -> \sh [11:01] Mrfai: do you happen to have a quick workaround for having dma in live-initramfs? [11:35] woffs (~woffs@osiris.taz.de) joined #fai. [11:45] s/if/it/ [11:46] siretart: which chipset? IDe or sata. Sata sould work IMO since it always enables dma. [11:55] Action: h01ger thinks sata might depend on the kernel version as some sata chipsets are not supported by 2.6.18 [12:01] lazyb0y, http://udrepper.livejournal.com/17577.html - a quote: "For sysadmin types this means: do what you have to do with Xen for now. But keep the investments small. For developers this means: don't let yourself be tied to a platform. Use an abstraction layer such as libvirt to bridge over the differences. For architects this means: don't looking to Xen for answers, base your new designs on KVM." [12:01] interesting read [12:02] Xen is to die? [12:02] Mrfai_: should be piix, but using libata [12:08] which kernel version? check if piix is included into the initramfs. [12:23] morning all (morning in brazil) [12:23] how can I define the computer hostname after a fai instalation [12:24] I have added lines like this in dhcp.conf [12:24] host cyber01.copacabana { hardware ethernet 00:19:d1:46:5e:7d; fixed-address cyber01.copacabana; } [12:24] and like this in /etc/hosts: [12:25] 192.168.1.101 cyber01.copacabana.ibeu.org.br cyber01.copacabana [12:25] but the hostname is just cyber01 not cyber01.copacabana (as I wish) [12:25] any tips? [12:30] h01ger: sorry for lameness. But what is the KVM model? [12:32] kernel based virtual machine (GIYF) [12:32] ;-) [12:34] :) [12:35] h01ger: I have some Hypervisors here.. [12:40] h01ger: as I could notice there is a lib (libvirt) to compile your codes with that enable the code run "natively" over KVM, something "like" java. [12:42] seems weird since it doesn't create a full environment for every application. [12:47] OMG! There is a pletora of VMs [12:48] <\sh> Mrfai, do you know any good nfs stress test tools? [12:52] no. [12:53] <\sh> bonnie++ running on an nfs share is breaking everything ,-) [12:55] \sh: maybe nfs does't like stress tests.. [12:56] nfs doesn't even like of file copies. ;-) [12:56] <\sh> Lin, it loves stress tests...because if not, netapp solutions wouldn't be there on the market ,-) [12:56] netapp... [12:57] IMHO nfs is good for nothing (except for remote rootfs) [12:58] netapp is a storage? [12:58] based only on NFS? [12:59] <\sh> netapp is a san solution...i think low- to midrange (not so expensive and big as symetrix/emc) [12:59] and the storage is shared using nfs? [13:00] if so Im REALLY amazed. [13:33] <\sh> Lin, there are several methods..but nfs is one of the prefered ones... [13:35] \sh: there are crazier people than I. [13:39] <\sh> Lin, why? it just works :) [13:39] <\sh> iozone is a good stress test tool for nfs , yeah [13:39] \sh: nfs is unreliable. [13:40] low security. [13:41] <\sh> Lin, this is not true :) [13:42] <\sh> well at least you shouldn't use nfs over internet...but inside a secure network it's quite good and a good solution for mass storage... [13:42] we cannot create redundant shares with nfs [13:43] nfs doesn't work in unreliable networs (TCP/IP) [13:43] <\sh> Lin, we can ;) [13:43] how? [13:43] you cannot just remount in another server or use a pool of servers [13:43] <\sh> Lin, just use heartbeat and use nfs3 over tcp [13:44] <\sh> you don't need...you just need several servers, run heartbeat on it, give him a single virtual ip and everything works [13:44] nfs doesn't work *WELL* in unreliable networs (TCP/IP) [13:44] \sh: have you tried this? [13:44] \sh: doesn't show any freezes? [13:45] \sh: I had bad times using NFS, mounts frozen, zombie processes, systems stoped. A complete chaos [13:46] <\sh> Lin, believe me it does...it does at lycos-europe, unity media (2nd biggest cable provider in germany) and it does at our company :) and I think in several more companies around the globe :) [13:57] \sh: just a simple question. How do you synchronize the repository content? [13:58] <\sh> Lin, repository content? [13:59] 2 nfs servers. [13:59] a lot of files, someone request a file /data/firefox/firefox2.exe [13:59] or someone write a new file on this repository. [14:00] how do you keep both them synchronized? [14:00] single rsync everyday? [14:00] <\sh> Lin, the sync is done via drbd [14:00] <\sh> Lin, it's something like raid1 over network [14:00] i know drbd [14:00] I use it here.. to syn my xen servers. [14:01] drbd do not allow access to both devices at same time. just one. [14:01] but Ok. I understood ;-) [14:02] <\sh> Lin, you just need one...push a MSA 1000 on your hp server and you are fine ;) [14:03] <\sh> or just (as we do) use 2 oem machines with 16x500GB sata hds in raid6 mode, which gives you around 7 TB storage... [14:04] hp? we use dell ;-) there is some incopatibility? :-P [14:59] Nick change: \sh -> \sh_away [16:34] Mrfai_ (~lange@suenner.informatik.uni-koeln.de) left irc: Quit: leaving [16:54] woffs (~woffs@osiris.taz.de) left irc: Quit: Leaving. [18:17] zirpu2 (~zirpu@nefud.org) joined #fai. [18:26] I'm using ubuntu 7.04 and the fai that comes with it to install some amd64 boxes with 3ware 9650se cards. [18:26] anyone have exerience with the 3ware 9650se cards? [18:26] Nick change: zirpu2 -> zirpu [18:32] Lin (~igor@200.179.57.57) left irc: Quit: Ex-Chat [18:47] Action: h01ger needs yet another fai config space... [18:47] hmm. this doesnt scale. [20:27] meandtheshel1 (~markus@85.127.103.154) left irc: Quit: Leaving. [21:43] Lin (~igor@200.179.57.57) joined #fai. [22:00] zirpu_ (~zirpu@nefud.org) joined #fai. [22:01] zirpu (~zirpu@nefud.org) left irc: Read error: Connection reset by peer [00:00] --- Thu Aug 16 2007