[00:07] |kardan| (~kardan@p54BE6C3B.dip.t-dialin.net) joined #fai. [00:15] kardan_ (~kardan@p54BE6880.dip.t-dialin.net) left irc: Ping timeout: 480 seconds [00:20] allee-k (~allee-k@dslb-092-074-034-235.pools.arcor-ip.net) left irc: Ping timeout: 480 seconds [00:57] amgarchIn9 (~amgarchin@p4FF036E4.dip0.t-ipconnect.de) left irc: Remote host closed the connection [01:07] ErKa (keryell@wireless-357.sc09.org) left irc: Ping timeout: 480 seconds [01:27] simNIX (~simNIX@156-60.bbned.dsl.internl.net) left irc: Ping timeout: 480 seconds [03:33] simNIX (~simNIX@156-60.bbned.dsl.internl.net) joined #fai. [04:08] ErKa (~keryell@72.11.113.122) joined #fai. [04:10] mikap (~mika@mail.michael-prokop.at) left irc: Ping timeout: 480 seconds [04:15] mikap (~mika@mail.michael-prokop.at) joined #fai. [04:20] ErKa (~keryell@72.11.113.122) left irc: Ping timeout: 480 seconds [07:39] ghislain (~Ghislain@LPuteaux-151-41-11-129.w217-128.abo.wanadoo.fr) joined #fai. [08:31] Mrfai (~lange@kueppers.informatik.uni-koeln.de) joined #fai. [09:05] ErKa (~keryell@72.11.113.122) joined #fai. [10:00] Action: allee stimmt \sh zu [10:01] <\sh> was? [10:12] siggg (~jean@tra78-5-88-172-128-32.fbx.proxad.net) left irc: Ping timeout: 480 seconds [10:13] cbmuser (~glaubitz@z6.physik.fu-berlin.de) left irc: Ping timeout: 480 seconds [10:27] Mrfai (~lange@kueppers.informatik.uni-koeln.de) left irc: Quit: leaving [10:28] Mrfai (~lange@suenner.informatik.uni-koeln.de) joined #fai. [10:30] cbmuser (~glaubitz@bender.physik.fu-berlin.de) joined #fai. [10:37] \sh: das'de was vernuenftiges geschafft hast :) [10:42] <\sh> ach so [11:18] siggg (~jean@AVelizy-151-1-20-46.w82-124.abo.wanadoo.fr) joined #fai. [12:27] cbmuser (~glaubitz@bender.physik.fu-berlin.de) left irc: Quit: leaving [12:27] cbmuser (~glaubitz@z6.physik.fu-berlin.de) joined #fai. [13:17] mfl (~mfl@adonis.fnb.maschinenbau.tu-darmstadt.de) left irc: Quit: leaving [13:19] ErKa (~keryell@72.11.113.122) left irc: Ping timeout: 480 seconds [13:37] mfl (~mfl@adonis.fnb.maschinenbau.tu-darmstadt.de) joined #fai. [14:55] hey, does setup_harddisks work in FAI 3.3 at all anymore? [14:55] or was it completely dumped now? [14:55] all of a sudden I cannot install machines with squeeze anymore :( [15:08] I really need help, FAI doesn't mount the newly partitioned disk anymore [15:08] so all packages end up being installed into RAM [15:08] Ramdisk [15:08] which is eventually full [15:17] I do not recommend to use setup_harddisks any more. It's still available in 3.3 but, it's untested an no more bugs will be fixed for it. [15:18] I myself do not yet use FAI with squeeze. [15:21] Mrfai: Thomas, nice to see you ;) [15:21] do you have any idea why FAI wouldn't mount the disks anymore all of a sudden? [15:22] can you point me to some resource explaining how to migrate from setup_harddisks to _storage? [15:23] I also want to use the latter now in order to use ext4 ;) [15:24] cbmuser: hav a log at the log files. Maybe your kernel cannot load the ext2/3 module [15:24] ah [15:25] right, it complained about some modules [15:25] thanks [15:25] I would be wondering why though [15:25] ah, I know [15:26] the kernel has been upgraded to 2.6.30-2 [15:28] ok, another problem, make-fai-nfsroot fails creating the initrd [15:30] <\sh> cbmuser, that problem is known for squeeze :) pav5088__ had the same issue [15:31] any known fix? [15:32] (using Lenny is not an option on a Dell OptiPlex 760) [15:34] .o0(trö-Dell...) [15:34] so? [15:35] so FAI is only supported on ancient hardware? [15:36] cbmuser, perhaps running update-initramfs -c -t -k 2.6.30-2-486 after chrooting into your nfsroot... this was a suggestion given to me... [15:36] ok, thanks, will give it a shot [15:36] ie. doing things manually for the time being. [15:37] thats good enough [15:37] in any case, I really can't go back to Lenny [15:37] siggg (~jean@AVelizy-151-1-20-46.w82-124.abo.wanadoo.fr) left irc: Ping timeout: 480 seconds [15:37] xorg doesn't support the intel graphics board [15:37] and the kernel the rest of the hardware [15:38] Nick change: pav5088__ -> pav5088 [15:42] cbmuser: you might be interested in http://faiwiki.informatik.uni-koeln.de/index.php/Grml_Kernel [15:42] cbmuser: (with an otherwise lenny nfsroot) [15:45] Is Squeeze still freezing in December 2009? [15:46] cbmuser: fai does support any hardware if your kernel supports it. [15:47] oz_, perhaps he's talking about the xserver in Lenny [15:47] pav5088: no. I think the plans say first quarter of 2010 [15:48] Mrfai, so they've changed the schedule? Or was the Dec freeze just for the installer with the other parts coming later? [15:50] pav5088: the dec freeze was a premature announcement without sufficient consultation... [15:50] I haven't really heard anything about it since midyear [15:50] i believe the current plan is to freeze in march [15:51] OK, that's good... :) [15:52] |kardan| (~kardan@p54BE6C3B.dip.t-dialin.net) left irc: Ping timeout: 480 seconds [15:52] http://lists.debian.org/debian-devel-announce/2009/10/msg00002.html [15:53] I was working under the assumption that the 2009 deadline was still on, and in that case I would be a bit behind schedule in my own project work. [15:55] oz_: I know, but when FAI supports Lenny only I can't use Squeeze and therefore not a current kernel without any trouble [15:57] cbmuser, have you tried backports for a more modern kernel? [15:58] ...but I guess that won't help if your xserver doesn't support your video anyway. [15:59] cbmuser: use FAI with squeeze and report problem to the bts. You can also help to make FAI work with squeeze if you send patches [15:59] |kardan| (~kardan@p54BE36AC.dip.t-dialin.net) joined #fai. [16:00] Mrfai: I will report back problems, I don't have time for patches unfortunately :( [16:01] I don't have the skills... ;) It has taken me a year to document GOsa as far as I have. [16:02] pav5088: well, Squeeze is already rising at the horizon [16:03] cbmuser: bug reports are fine. Most of the time it's easy for me to implement a fix, if you can tell me what's going wrong, and what's the new way to do it. [16:04] I know... some guy from Belgium asked me to join a documenting effort for GOsa, but if I'm one of the more skilled people then the GOsa world is in trouble. [16:05] My FAI+GOsa documentation SEEMS to work... but I have no idea how to use it. ;) I'm trying to work out Kolab+GOsa now... and it looks like I will have to do lots of LDAP mods etc... etc... and using Kolab in custom ways is not documented. :( [16:06] Mrfai: so far I've only had the isse with the initrd not being created by make-fai-nfsroot [16:06] <\sh> Mrfai, make-fai-nfsroot makes a divert of update-initramfs when upgrading the chroot, right? this is what I read from the source... [16:06] the rest works quite nicely, I had to update alot of stuff in our config though [16:06] \sh: yep [16:06] <\sh> Mrfai, so...it shouldn't happen when creating the chroot, or? [16:07] In the past, the initrd was build more than once, since the kernel itseld did this, and then also the live-initramfs and mdadm also recreated the initrd. There the code was added to skip this, and create the initrd at the end of make-fai-nfsroot [16:07] <\sh> pav5088, or cbmuser would you do me a favour, and paste the contents of /etc/lsb-release? [16:08] siggg (~jean@tra78-5-88-172-128-32.fbx.proxad.net) joined #fai. [16:08] therefore make-fai-fsnroot has the option -U [16:08] No such file or directory [16:08] is this in the NFSROOT? [16:08] cbmuser: can you please try it with -U [16:08] <\sh> Mrfai, well, strange thing is, that you set $divert to 0 for ubuntu [16:09] <\sh> Mrfai, could be that we need to set $divert to 0 for squeeze too, just a shot [16:09] Action: Mrfai .oO (man pages are really great) [16:09] /etc/lsb-release doesn't exist on debian, or does it? [16:09] <\sh> oz_, woot? [16:09] <\sh> it should have that [16:10] There's an lsb-base directory [16:10] package debian-edu-config has /etc/lsb-release [16:10] (with nothing in it) [16:10] http://packages.debian.org/search?searchon=contents&keywords=lsb-release&mode=path&suite=testing&arch=any [16:10] \sh: my lenny here doesn't have it. [16:10] lsb-release only exists if you have installed the package lsb-release. I also do not install it on my machines [16:11] <\sh> Mrfai, on ubuntu /etc/lsb-release belongs to base-files [16:11] \sh: vanilla debian is SLOW. :) [16:12] ubuntu is too fast [16:12] Mrfai: imho too, but $users want is like that... [16:13] <\sh> well, lsb-release is the only file which can really tell a app dev on what linux os the app is running, and if it would work...it's not that this file is useless or so [16:13] <\sh> even if you can fake it ;) [16:13] \sh: cat /etc/debian-release is the file for me. ;) [16:13] /etc/debian_version even [16:14] <\sh> imagine every linux distro has it's own -release file (imho redhat does have its own too) [16:14] Mrfai: make-nfs-root with -U? [16:14] <\sh> anyways, I wonder if we have to set divert to 0 for squeeze [16:14] I can't try right now because I have 3 clients being installed right now [16:14] \sh: why imagine? it's just like that [16:14] <\sh> cbmuser: make-fai-nfsroot -vU [16:15] <\sh> oz_, it will be our hell..if we don't get a standard on this, distro neutral standard that is [16:16] I'm doing the make-fai-nfsroot -vU ... although it's on a virtual machine running on a slow box [16:18] ie. my desktop. ;) [16:19] \sh: hm. I have been administering different UNIXes...HP/UX, SUNOS, AIX... [16:19] normal unix hell ,) [16:21] http://pastebin.com/d17d23f18 [16:22] <\sh> pav5088, thx...that's it [16:23] <\sh> oz_, can you paste your debian_release or whatever debian uses? ;) [16:23] <\sh> Mrfai, divert=0 for squeeze ;) [16:23] gold:~# cat /etc/debian_version [16:23] 5.0.2 [16:24] \sh: that's for lenny [16:24] \sh: ok, will try later, thanks [16:24] <\sh> how can someone tell that someone is running debian, programmtically? [16:24] <\sh> squeeze will be 6.0.0 then? [16:24] /etc/debian_version = sqeeze/sid [16:24] s/sqeeze/squeeze [16:24] I think make-fai-nfsroot must check etc/debian_version inside the nfsroot, not from the host ssytem [16:26] ow it's sqeeze/sid but I think it will be changed (to 6.0) if squeeze is released. So, not that easy to detect which version be have [16:26] Squeeze has changed over to dash as /bin/sh... I've been bitten by strange things like eg. echo -e will echo the -e (dash has an inbuilt echo command) [16:26] fai is only using bash scripts [16:26] OK [16:26] <\sh> pav5088, that's why all shell scripts are shebanging /bin/bash ;) [16:27] <\sh> we had this topic already for ubuntu in the past ;) [16:27] <\sh> oh well...shebang is somehow sexist ;) [16:27] \sh: i've found /etc/issue to be pretty good for identifying what distro you're on, generally [16:27] mgoetze: hehe [16:28] <\sh> hmm [16:29] pav5088: Using -U shows that squeeze will generate the initrd twice. But this is better than never. [16:30] i'd like to help test fai with squeeze, but i'm having such a hard time getting kvm working on squeeze :( [16:34] Mrfai, OK.... I'll tell this to the guys on #gosa on FreeNode so they can test [16:35] <\sh> there are more problems, if you check pav5088 paste, the MD stuff is failing [16:36] hmmmm.... OK [16:37] <\sh> # [16:37] <\sh> Setting up mdadm (3.0.3-2) ... [16:37] <\sh> # [16:37] <\sh> [16:37] <\sh> # [16:37] <\sh> W: mdadm: failed to load MD subsystem. [16:37] <\sh> # [16:37] <\sh> [16:37] <\sh> # [16:37] <\sh> Generating array device nodes... /var/lib/dpkg/info/mdadm.postinst: 158: /dev/MAKEDEV: not found [16:37] <\sh> # [16:37] <\sh> [16:37] <\sh> # [16:37] <\sh> failed. [16:37] <\sh> # [16:37] <\sh> [16:37] <\sh> argl [16:37] <\sh> damn [16:37] FAIL. :) [16:37] and we trust this guy with svn access? ;) [16:38] <\sh> I'm not pushing to trunk ;) and normally the paste works...*grmpf* [16:38] <\sh> *shameonme* [16:38] mgoetze: hm. I get headache when I think about our government and 'competence'.... [16:39] Further up : [16:39] update-initramfs: Generating /boot/initrd.img-2.6.30-2-486 [16:39] sed: can't read /etc/mdadm/mdadm.conf: No such file or directory [16:39] Examining /etc/kernel/postinst.d. [16:39] <\sh> squeeze should use upstart, right? [16:39] mgoetze: you can trust this guy, I'm sure. I've seen him IRL!!! [16:39] I think the mdadm stuff is not problem. It's similar to what I get on lenny. [16:39] anyway, i'm sure the MAKEDEV problem will be sorted out eventually in squeeze... it got moved from priority: required to priority: extra soon after the release of lenny [16:39] apt-file doesn't find a /etc/mdadm/mdadm.conf [16:39] <\sh> Mrfai, ok... [16:40] \sh: i haven't seen anyone seriously working on making upstart the default [16:40] <\sh> pav5088, looks like it's autogenrated [16:40] \sh: however, i can report that it works fine, so if you want to install it on your systems, go ahead :) [16:40] I just checked. The diversion of update-initramfs or mkinitramfs can also be remove for lenny. lenny also uysed trigger that do work, so the initrd is only created once. [16:40] <\sh> mgoetze, oh well...good to know...so you don't have a problem with bonds, bridges and vlans [16:40] <\sh> mgoetze, it doesn't work fine .. believe me [16:40] mdadm : tool to administrer Linux MD arrays [16:41] -r [16:41] <\sh> mgoetze, i had several discussions with scott about it...but it looks like he doesn't like server issues, and is only concentrating on desktop [16:41] \sh: well, i mean, it works for me (on my laptop) [16:41] what sorts of problems are you having? [16:42] <\sh> mgoetze, laptops + desktop installs are no problem...using auto eth0 etc. is also not a problem...problem are virtual devices...like bond, vlan, bridges [16:42] <\sh> there is a timing problem between /etc/init/network-interface.conf and /etc/init/networking [16:43] but is it a problem inherent in upstart or is it just a badly written configuration file? [16:43] <\sh> result: /etc/init/networking.conf will be triggered sometimes before /etc/init/network-interface.conf and then the whole /e/n/i handling failes, because /var/run/network/ is not avaiable, and ifup needs it [16:43] <\sh> it's a problem between udev, upstart and all this crap [16:44] <\sh> udev -> bus -> upstart -> for hardware , while upstart -> networking.conf -> ifup -a for /e/n/i handling [16:44] <\sh> minus all static hardware interfaces [16:44] i don't have any post-8.04 ubuntu systems available so i don't know what's in those files [16:44] <\sh> https://bugs.launchpad.net/ubuntu/+source/ifupdown/+bug/446031 [16:45] <\sh> there is a workaround from me...and this actually works for dist-upgrades from jaunty to karmic e.g. [16:47] <\sh> mgoetze, especially when you talk about something like this: http://www.sourcecode.de/content/network-setup-freaks-me-out <- if this fails after dist-upgrade you're fcked ;) [16:47] i don't have anything in /var/run/network, what does ubuntu do there? [16:47] <\sh> mgoetze, ifup needs it [16:47] just the directory to exist? [16:47] <\sh> the directory needs to exist [16:47] <\sh> yes [16:48] <\sh> and /var/run is tmpfs + /etc/init/network-interface.conf will create it normally, but /etc/init/networking.conf does not create it...so when it's executed before the network-interface, ifup -a fails totally, [16:48] so it's just a poorly written configuration file, and not a bug in upstart itself *shrug* [16:49] <\sh> mgoetze, no it's a timing problem in upstart + friends [16:49] <\sh> bonds need hardware interfaces...if they are not added from udev to the userspace, bonds are created without any hardware interface bound on it... [16:49] can you maybe just pastebin network-interface.conf and networking.conf so i can figure out what you're talking about? ;) [16:50] Mrfai (~lange@suenner.informatik.uni-koeln.de) left irc: Quit: leaving [16:50] <\sh> mgoetze, http://paste.ubuntu.com/321661/ <- network-interface.conf [16:50] <\sh> http://paste.ubuntu.com/321662/ <- networking.conf [16:51] thx [16:53] ErKa (keryell@wireless-357.sc09.org) joined #fai. [16:53] so net-device-added gets emitted for physical interfaces like eth*, but not for bond*, is that right? [16:53] <\sh> yes [16:53] <\sh> the net-device-added comes from udev [16:54] <\sh> and it's triggered for every physical NIC you have in your box...so it doesn't even make sense, to call "ifup -a" somehow in network-interfaces [16:56] so if you were just doing this for one specific configuration, you could replace networking.conf with stuff like bond0.conf, which has "start on (local-filesystems and stopped udevtrigger and net-device-added INTERFACE=eth0 and net-device-added INTERFACE=eth1) [16:56] ", right? [16:57] <\sh> mgoetze, well, that's what we tested already...this doesn't work ;) [16:58] <\sh> because when interface eth0 is added interface eth1 isn't ... so and somehow upstart forgets about old net-device-added interfaces... [16:58] <\sh> (actually we don't know if upstart just forgets about old net-device-added triggers, but it looks like it) [16:59] maybe the problem is that it's the same event name, just with different variables [17:00] <\sh> any combination we tried, didn't work..but just adding a script inside networking which creates /var/run/network helps..because then the configs for bonds are already loaded... and when the phys. interfaces are coming up, the bond is ready for usage...therefore all depend virt interfaces which are using the bond do come up as well [17:01] maybe it would work if you have a bond0.conf with "start on (local-filesystems and stopped udevtrigger and up-eth0 and up-eth1)", making sure of course to emit those events once the slave interfaces are up [17:02] <\sh> mgoetze, you really don't want to have 2 or three + more job/task configs for vlan interfaces e.g. in upstart... [17:02] you could autogenerate them via a script ;) [17:03] <\sh> mgoetze, well, upstart should work with /e/n/i [17:03] /e/n/i? [17:03] <\sh> /etc/network/interfaces [17:03] ah. I wonder why it doesn't [17:04] upstart isn't well tested, hm? ;) [17:04] <\sh> well, one of scotts great ideas was to just forget about /e/n/i handling and force people to use network-manager on servers ;() [17:04] oz_: we have set our debian alpha tester task force, aka ubuntu users, onto the job :) [17:04] <\sh> oz_, the problem between "desktop" and "server"...people running server are doing nasty things to the OS ;) [17:05] \sh: :) [17:05] <\sh> oz_, the real problem but: scott is desktop and upstart maintainer... and other people are server and don't have a clue how to debug upstart [17:05] <\sh> mgoetze, rotfl [17:06] \sh: :))) [17:06] <\sh> "how to debug upstart? Add --debug or --verbose to the kernel commandline" [17:06] once it's good it'll go into beta testing (sidux) [17:06] <\sh> "check the console and eventually you'll find some things in /var/log/syslog" [17:06] To be honest...I had some ubuntu servers here, but I replaced them all with vanilla debian... [17:06] as I have less hassle with debian. [17:08] <\sh> I always have troubles with linux in general, regarding server and regarding new hardware...but that's distro neutral [17:09] \sh: it's our job, I guess... [17:09] <\sh> yes...but making more troubles because of "we need 10seconds boottime" [17:09] <\sh> is not a good idea [17:09] whut? they switched to upstart to speed up booting? [17:10] <\sh> that was the only reason to use upstart [17:10] hm. I have a server here with 700days uptime. [17:10] jay. how clueless. [17:10] imho, of course :) [17:10] <\sh> well, actually I installed karmic on a vmware as server flavour and the boot time was less then 6 seconds ;) [17:11] i personally believe there are other, and better, reasons for using upstart ;) [17:11] <\sh> (no more services installed only openssh-server that was) [17:11] <\sh> mgoetze, which ones? apache2 will always wait for network [17:11] <\sh> mgoetze, while an Xserver can run with lo configured + mounted fs [17:11] upstart is better at things like respawning daemons... [17:13] <\sh> mgoetze, hmm...we had this already...was named daemontools, or? [17:13] did that ever make it into the debian archives? i believe there were some freeness issues :) [17:13] <\sh> and came from d.j. bernstein ;) qmail was also much better then sendmail and postfix or exim these days [17:13] jay, daemontools [17:14] $collegue installed it, and the apache could not be stopped anymore :) [17:14] anyway i think combining functionality of init + daemontools makes a lot of sense [17:14] <\sh> mgoetze, hehe...there was an issue with the license, yes [17:14] <\sh> mgoetze, but it worked like a charm... [17:15] <\sh> http://blog.blinkenlights.nl/2006/10/29/daemontools-with-upstart/ [17:17] <\sh> https://bugs.launchpad.net/ubuntu/+source/daemontools/+bug/253465 and this is a really nice one...check the dups [17:23] <\sh> ok going home for today... [17:23] <\sh> cu [17:44] tokkee_ (tokkee@osprey.tokkee.org) joined #fai. [17:44] tokkee_ (tokkee@osprey.tokkee.org) left irc: [18:19] ErKa (keryell@wireless-357.sc09.org) left irc: Ping timeout: 480 seconds [19:28] amgarchIn9 (~amgarchin@p4FF01909.dip0.t-ipconnect.de) joined #fai. [20:39] \sh: qmail better, you must be kidding [21:00] siggg (~jean@tra78-5-88-172-128-32.fbx.proxad.net) left irc: Ping timeout: 480 seconds [21:10] siggg (~jean@tra78-5-88-172-128-32.fbx.proxad.net) joined #fai. [21:38] mgoetze: http://faiwiki.informatik.uni-koeln.de/index.php/Grml_Kernel -> nice, thanks :) JFYI: kernel 2.6.31-grml[64] are available nowadays as well, maybe someone could mention that in the wiki? [21:38] siggg (~jean@tra78-5-88-172-128-32.fbx.proxad.net) left irc: Ping timeout: 480 seconds [21:54] mikap: done :) [21:55] mgoetze: thanks! :) [22:05] mgoetze: i'll send a note regarding this issue through our grml microblogging channel, want to "proof read" it? [22:05] "Did you know, that you can use #FAI (Fully Automatic Installation) with the #Grml kernel? http://bit.ly/1RSty4" [22:06] -> last chance for any modifications ;) [22:06] sure :) [22:07] mgoetze: so ok? then i'll send it ;) [22:10] ErKa (keryell@wireless-357.sc09.org) joined #fai. [22:11] mgoetze: http://twitter.com/grmlproject/status/5838877600 && http://identi.ca/notice/15042193 [22:45] so, gute nacht [22:56] allee-k (~allee-k@dslb-188-099-217-195.pools.arcor-ip.net) joined #fai. [23:16] ErKa (keryell@wireless-357.sc09.org) left irc: Ping timeout: 480 seconds [23:21] allee-k_ (~allee-k@dslb-188-099-217-195.pools.arcor-ip.net) joined #fai. [23:21] allee-k (~allee-k@dslb-188-099-217-195.pools.arcor-ip.net) left irc: Read error: Connection reset by peer [23:43] allee-k_ (~allee-k@dslb-188-099-217-195.pools.arcor-ip.net) left irc: Ping timeout: 480 seconds [23:53] ErKa (keryell@wireless-357.sc09.org) joined #fai. [00:00] --- Thu Nov 19 2009