[00:05] fai: 3 mt committed revision 6279 to 7 fai: Updated experimental to match current trunk, improved mdadm error message to fix [00:05] fai: recent bug report [00:15] msantana (msantana@SDF.ORG) left irc: Quit: Coyote finally caught me [00:16] msantana (msantana@SDF.ORG) joined #fai. [01:50] _julian (~quassel@hmbg-5f76600b.pool.mediaWays.net) joined #fai. [01:58] Nick change: pav5088_ -> pav5088 [01:58] _julian_ (~quassel@hmbg-4d06a2c7.pool.mediaWays.net) left irc: Ping timeout: 480 seconds [03:01] msantana (msantana@SDF.ORG) left irc: Quit: Coyote finally caught me [06:46] \sh (~shermann@nat.de.netviewer.com) joined #fai. [07:03] Action: \sh should think about changing his job [07:11] canonical became so bad? [07:20] <\sh> No I'm too old for this admin business [07:22] <\sh> well, I think in Egypt there will be a job position free in a couple of weeks...would be more fun "\sh - Dictator of the Pharaos" [07:25] I'm looking for someone who wants the job of "cheif investor". got $250K lying about? ;) [07:26] <\sh> lol [07:31] \sh: ha! [07:31] can't be [07:32] <\sh> what? too old for the admin business, or "Dictator of the Pharaos"? ;) [07:34] too old for admin business [07:34] just swing the big bat of endless experience [07:35] I know, this can be difficult in these modern times [07:44] but becoming pharao...sounds tempting ;) [07:45] ^Miaou^ (~AntoineR@82.230.94.236) left irc: [07:48] <\sh> hehe [07:48] <\sh> btw..anyone know about this bug https://bugs.launchpad.net/ubuntu/+source/fai/+bug/710311 ? I'm trying to find a reference on dbts, but failed [07:48] sitting on the top of a pyramid could be tiresome though... [07:49] torkel: when you get too tired, relax in the sarkopharg ;) [07:49] true :-) [08:03] ErKa (keryell@keryell.pck.nerim.net) joined #fai. [08:45] ErKa (keryell@keryell.pck.nerim.net) left irc: Ping timeout: 480 seconds [09:06] ErKa (keryell@m3.wifi.enstb.org) joined #fai. [09:25] Mrfai (~lange@kueppers.informatik.uni-koeln.de) joined #fai. [09:43] psibox (~dreger@login.physik.fu-berlin.de) joined #fai. [09:44] hi [09:44] does anyone remember why fai diverts /etc/init.d/nis during installs ? [09:44] this leads to problems with insserv on squeeze. dependencies are not correct on first boot [09:45] autofs does not work [09:45] i just can't see, why some would want to divert /etc/init.d/nis ... [09:49] no. but about nis+autofs+squeeze you want to read http://www.earth.li/~noodles/blog/2011/02/my-squeeze-upgrade-notes.html [09:55] h01ger: thanks. but there is still a fai problem [09:55] directly after install: [09:55] root@squeeze32-test:~> grep ^autofs /etc/init.d/.depend.start [09:55] autofs: rsyslog [09:55] running insserv (nothing else changed) [09:55] root@squeeze32-test:~> grep ^autofs /etc/init.d/.depend.start [09:55] autofs: rsyslog nis [09:55] and this comes from the diversion of /etc/init.d/nis in fai [09:56] grep divert $(dpkg -L fai-client) [09:56] + other fai packages [09:56] + in your configspace [09:56] though i dont get why init.d/nis is diverted [09:58] /usr/lib/fai/updatebase: fai-divert -a /etc/init.d/nis /sbin/start-stop-daemon /usr/sbin/policy-rc.d [10:00] maybe this comes from a time when /etc/init.d/nis did not use start-stop-daemon ? [10:01] guess so [10:01] +please file a bug [10:02] will do once I made sure it works with /etc/init.d/nis removed from that line [10:02] autofs already has a 'Should-Start: ypbind' [10:03] should be fine. the problem during installs is, that the file which fai-divert puts at /etc/init.d/nis does not contain any valid LSB headers for insserv [10:03] therefore all insserv calls during install log tons of errors [10:04] eek [10:08] Mrfai (~lange@kueppers.informatik.uni-koeln.de) left irc: Quit: leaving [10:08] Mrfai (~lange@malzmuehle.informatik.uni-koeln.de) joined #fai. [10:30] justL1 (~leen@rev-131-102.virtu.nl) joined #fai. [10:32] Hello All [10:33] can I ask for some help again? SInce I fixed the bootin on the BL360c G7, I've noticed that the G1's dont boot anymore... [10:33] firmware for the bnx2 gets loaded fine, but the boot hangs [10:34] last message is eth0: no ipv6 routers present [10:34] I dont see any weird errors [10:34] but im at a loss what might be the cause.. [10:46] siggg (~jean@tra78-3-82-244-85-196.fbx.proxad.net) joined #fai. [10:53] msantana (msantana@SDF.ORG) joined #fai. [11:08] <\sh> Mrfai: ping can you check on https://bugs.launchpad.net/ubuntu/+source/fai/+bug/710311 <- fai-cd bug..I fixed it temporarily for Ubuntu...I don't know if debian is/will be bugged too [11:13] yes, in squeeze grub-mkimage is also moved to grub-common [11:14] if we only change grub-pc to grub-common, fai-cd will not work with lenny any more (in certain configurations). [11:14] for fai stable branch we should install grub-pc _and_ grub-common [11:16] I'm not sure if grub-pc is also needed, or if only grub-common will fine [11:30] Mrfai, can you comment on the diverting /etc/init.d/nis issue discussed this morning between 10:44 and 11:04 ? [11:38] This file was diverted, because it causes problems in the past. Maybe because of not using start-stop-daemon. But maybe because of other problems. I'm not sure. Diverting the nis script works for me on lenny and squeeze. [11:43] the diversion of /etc/init.d/nis is from 2002 or earlier... [11:55] Mrfai, i think it only works on squeeze because you dont use/install autofs [11:55] Mrfai: are you using nis based automount maps ? [11:55] Action: h01ger believes its a serious bug even [11:55] if you don't use nis based automount maps, you can never see any effect of the bug ;) [11:55] (and definitly qualifiying for fixing in 6.0.1, expected in a month) [11:56] ich installed without the diversion and autofs works upon first boot [11:56] i'll check, if the nis script can do anything at all with start-stop-daemon diverted [11:56] don't think it can. [11:56] what does the diversion do? [11:56] it disabled /etc/init.d/nis [11:56] it does nothing. [11:57] but the problem is, that the diversion creates a file with content like "this is a diversion... blah" [11:57] missing all LSB headers for insserv [11:58] yeah [11:58] insserv is enabled on default, this breaks other software (autofs) -> serious bug in fai [11:58] every call to insserv leads to somthing like: [11:58] software.log:insserv: warning: script 'nis' missing LSB tags and overrides [11:59] ah [11:59] anyway, i'll just remove the diversion of /etc/init.d/nis for now, it works. [11:59] it doesnt fail? [11:59] then its not serious [11:59] autofs fails [11:59] after reboot [11:59] start order of nis and autofs is wrong [12:00] we have a lot of nis maps in autofs [12:00] so for us it is serious. [12:00] calling insserv once after reboot fixes it [12:00] is there a hook in fai that i can use AFTER diversions are taken back ? [12:01] stop chatting, file a bug [12:01] this needs to be fixed in debian anyway [12:01] using reportbug ? [12:02] and until it is i'd suggest to use a custom fai package with a lower version then the next in squeeze will have. so you can get upgrades.. [12:02] yes [12:02] severity: important [12:02] ok, will do after lunch. [12:02] thanks. [12:03] (important because fai is still usable for most cases and it also doesnt break autofs completly, there is a easy workaround) [12:03] thank you too [12:03] (important bugfixes also usually qualify for stable updates if the fix is sensible+targeted) [12:15] ErKa (keryell@m3.wifi.enstb.org) left irc: Ping timeout: 480 seconds [12:41] verwilst (~verwilst@router.begen1.office.netnoc.eu) joined #fai. [12:42] <\sh> Mrfai: I fixed it like that: http://paste.ubuntu.com/564406/ [12:57] <\sh> mgoetze: you did a centos5 rinse install, right? how do you fix that inside the chroot that yum only downloads from a local mirror and from an external mirror? [12:58] \sh: hm? mir ist nicht klar was du meinst [13:01] <\sh> mgoetze: nachdem rinse das grundsystem installiert, geht er ja in das chroot rein, und versucht durch das script "/usr/lib/rinse/centos-5/post-install.sh" den rest zu installieren [13:02] <\sh> durch das package centos-release*.rpm werden aber yum repos.d dateien mitinstalliert, die gemeinhin nach aussen zeigen und nicht auf ein lokales yum repos zeigen [13:02] <\sh> frage ist, sollte ich noch im post-install.sh script meine eigenen repos dateien verwenden, und die schon installierten ueberschreiben? [13:03] <\sh> sodass yum update nicht alles von aussen zieht sondern von lokalen yum repo mirrors? [13:03] \sh: guck mal in http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=612164 [13:04] \sh: wenn du dir ein skript schreibst welches das alles macht was ich versucht habe da reinzupacken kannst du es mit --before-post-install einbinden [13:05] <\sh> mgoetze: ok...lass mich mal basteln ;) [13:05] <\sh> erstmal yum repos bauen ;) [13:05] schreib dafuer am besten direkt ein howto ins fai wiki ;) [13:07] I'm using NIS based automount maps, but not using autofs but am-utils [13:31] <\sh> mgoetze: ok, I created a CentOS-Base.repo file like this: http://paste.ubuntu.com/564436/ [13:32] <\sh> mgoetze: and created a script like this: http://paste.ubuntu.com/564437/ [13:33] <\sh> mgoetze: and executed rinse like this: sudo rinse --distribution centos-5 --arch amd64 --directory /home/shermann/centos5 --before-post-install /usr/lib/rinse/centos-5/pre-install.sh [13:33] <\sh> mgoetze: and it worked [13:33] great [13:34] <\sh> mgoetze: I think you can just push your script inside the pre-install.sh and prefix is your chroot directory, and you can just replace ${prefix}/etc/yum.repos.d/* with your magic ;)( [13:34] <\sh> mgoetze: ii rinse 1.7-1 RPM installation environment [13:34] right, i was hoping to have everything configurable in one place, debian mirror, centos mirror, ubuntu mirror etc. - that's why i filed that wishlist bug [13:35] yeah i don't think there have been any new versions for 3 years or more :( [13:40] <\sh> mgoetze: you want to create the chroots dynamically? [13:41] <\sh> mgoetze: anyways I think your example in bug #612164 won't work 'cause "system( $CONFIG{ 'before-post-install' }, $prefix );" in /usr/sbin/rinse [13:41] \sh: well i wrote a generalized Makefile to allow any fai user to create whatever chroot he likes without thinking too much about how to do it [13:42] it fails even before with the -x check :) [13:42] <\sh> yes [13:48] <\sh> mgoetze: much better...you don't need the the before-post-install stuff- [13:49] <\sh> mgoetze: just put everything you need to do inside the post-install.sh script and you are done...I think some parts you need to hardwire (like the local repository server) [13:50] sure but this would defeat the purpose of what i want to achieve... that some person who has never been on this irc channel can do aptitude install fai-quickstart rinse; cd /srv/fai/config/basefiles; make CENTOS5_64.tar.xz [13:50] there are many ways to fix it for just me but i want to fix it for everyone ;) [13:55] siggg (~jean@tra78-3-82-244-85-196.fbx.proxad.net) left irc: Ping timeout: 480 seconds [13:57] <\sh> mgoetze: well, but wouldn't it be easier to create a patch for the post-install.sh scripts for rinse, to achieve that? [13:58] sure but what do you propose... nmu? or have fai-server divert it? or...? [14:02] <\sh> mgoetze: well, reading again and again your bugreport, when you want to have something like "make centos5_64.tar.gz" then rinse does the job without a problem, your replacement for the yum mirrors needs a bit more knowledge ;) like knowing a centos mirror and exporting it in MIRROR_CENTOS [14:04] right, so my idea is that the user can configure these variables at the top of Makefile [14:04] then in the rules of the makefile, which the user doesn't want to touch because after all makefiles are ugly, it would pass this off to a script called by rinse [14:05] that's why i would need different semantics for the rinse invocation [14:08] <\sh> mgoetze: so you need to patch rinse with some more intelligence ;) [14:08] siggg (~jean@tra78-3-82-244-85-196.fbx.proxad.net) joined #fai. [14:11] ErKa (keryell@keryell.pck.nerim.net) joined #fai. [14:20] <\sh> mgoetze: and your perl snippet has a bug [14:20] hm? [14:20] <\sh> oh no [14:20] <\sh> I misread it [14:20] didn't you know perl is a write-only language :) [14:22] <\sh> mgoetze: when I give you a patch, for rinse that is, and you test it that it works for you, will you push it to steve kemp ? [14:23] <\sh> MIRROR_CENTOS=http://172.24.24.15/centos sudo rinse --before-post-install "/usr/bin/perl -pie 's,mirrorlist=,\#mirrorlist=,;s,\#baseurl=http://mirror.centos.org,baseurl=${MIRROR_CENTOS},;' ${TMPDIR}/etc/yum.repos.d/CentOS-Base.repo" --distribution centos-5 --arch amd64 --directory /home/shermann/centos5/ <- this is what you want, right? [14:24] <\sh> if so, then I can tell you, that it works ;) [14:24] <\sh> I'll have to refine the patch, but in general it works... [14:26] i've never heard back from steve kemp, i write bug reports and he never reacts :) [14:26] but i can try [14:27] <\sh> mgoetze: this is opensource, you can always fork ;) [14:28] nmu is a bit difficult as non-dd :) [14:29] <\sh> you don't want to do an nmu...what you want is intrusive, and can break other peoples setup...so I introduced actually a new cli switch (--before-post-install-command, --after-post-install-command) [14:29] <\sh> so it doesn't break eventually existing scripts and gives new features ;) [14:30] well... it breaks the setup of people who have spaces in the file name of their scripts. they surely deserve it :) [14:31] <\sh> when I'm really adventurous, I go and patch Ubuntus version and push it to ubuntu ;) therefore mr. steve kemp needs to react somehow...and eventually he will [14:36] <\sh> hmmm..is he on irc somehow? [14:41] dunno [14:41] <\sh> ok...I'll make a patch and reply to your bug report with the patch as attachment [15:14] Mrfai: in fai4's make-fai-nfsroot you check [ -d $NFSROOT/usr/share/live-initramfs ], but this is no longer correct with the squeeze version of live-initramfs [15:15] what is it instead? [15:16] Mrfai: hm, maybe [ -x $NFSROOT/usr/share/initramfs-tools/hooks/live ] ? depends what exactly you want to check i guess [15:17] I think I wanna check if live-initramfs is installed [15:18] it's a transitional package, in squeeze you want live-boot-initramfs-tools [15:18] anyway how about $ROOTCMD dpkg -l live-boot-initramfs-tools ? [15:21] It's faster just to check the existance of an file or directory. [15:21] ok well i would go with that hooks/live one then, but probably mikap has a more informed opinion [15:21] mikap: ^^ [15:22] cheing for hook/live is fine IMO [15:23] i guess fai4 is targeted at squeeze+ ? or do we have to worry about lenny compatibility? [15:24] Yep, fai 4 ist squeeze+. We can drop lenny support in this version. [15:25] ok i will test again with the hooks/live test and commit to trunk if it works [15:26] i just successfully installed a kvm guest with squeeze nfsroot and fai4 btw :) [15:32] Mrfai: do you mind if i also update the dependencies of fai-nfsroot? [15:35] Which dependency? because of the aptitude problem in Ubuntu? [15:36] Mrfai: no i mean the dependency on live-initramfs, which is only a transitional package in squeeze [15:37] mgoetze: what's the problem? [15:37] please file a bug report for this. Even we can drop lenny support in 4.x I will try to drop it later, not yet. [15:38] It's more a cosmetic thing, not a real bug IMO. [15:38] mikap: what's a good quick test to see whether live-boot was installed successfully in the nfsroot? [15:39] mgoetze: for which suite? [15:39] mgoetze: dpkg --list live-boot for >= squeeze [15:39] squeeze i guess, but if it also works with (future) grml versions... [15:40] fallback to dpkg --list live-initramfs for anything older [15:40] dpkg -l too slow for thomas it seems :) [15:40] no garuantee for anything else [15:40] are you transitioning to live-boot in grml as well? [15:40] checking for according /usr/share/live*$FOO [15:41] mgoetze: we already did [15:41] check the bugfixes that squeeze got from grml... [15:46] FBI (~FBI@fbi.bot.oftc.net) left irc: Server closed connection [15:46] FBI (~FBI@fbi.bot.oftc.net) joined #fai. [15:47] eh apparently this issue was already fixed but hasn't made it to the beta packages on fai-project.org :) [15:48] maybe i should start building my own from trunk... [15:53] Mrfai: trunk/debian/changelog is at 4.0~beta2 but fai-project.org fai4/koeln repo has ~beta0, are you uploading them somewhere else? [15:55] <\sh> Mrfai: the aptitude bug is fixed in ubuntu...the same for fai-cd [15:56] <\sh> Mrfai: uploaded the fixes this morning... [16:02] im really really starting to dislike fai.. [16:02] why does it break every time you look at it? [16:02] bah [16:02] justL1 (~leen@rev-131-102.virtu.nl) left irc: Quit: Crap! [16:10] famous last words? [16:12] mgoetze: I do not upload very often the fai trunk version. Therefore only beta0 is available [16:12] s/upload/build/ [16:18] ErKa (keryell@keryell.pck.nerim.net) left irc: Ping timeout: 480 seconds [17:00] \sh (~shermann@nat.de.netviewer.com) left irc: Remote host closed the connection [17:00] fai: 3 mgoetze-guest committed revision 6280 to 7 fai: Deduplicate scripts/{DEBIAN,FAIBASE}/40-misc in simple-examples [17:06] hm someone remind me... how do i use unpartitioned disks with setup-storage? [17:06] flag_initial=1 in vlass/*.var [17:06] class/*.var [17:07] i mean i don't want setup-storage to create a partition table [17:10] verwilst (~verwilst@router.begen1.office.netnoc.eu) left irc: Quit: Ex-Chat [17:13] Mrfai (~lange@malzmuehle.informatik.uni-koeln.de) left irc: Quit: leaving [17:22] siggg (~jean@tra78-3-82-244-85-196.fbx.proxad.net) left irc: Quit: Leaving. [17:40] vicky (~luser@dhcp-83-219-97-225.customers.tvtnet.ch) left irc: Server closed connection [17:41] vicky (~luser@dhcp-83-219-97-225.customers.tvtnet.ch) joined #fai. [18:03] aptituz (~schoenfel@thehappy.de) left irc: Quit: switch! [18:14] tokkee (tokkee@osprey.tokkee.org) left irc: Server closed connection [18:14] tokkee (tokkee@osprey.tokkee.org) joined #fai. [18:54] ErKa (~keryell@keryell.pck.nerim.net) joined #fai. [19:17] allee (~quassel@allee.mpe.mpg.de) left irc: Server closed connection [19:17] allee (~quassel@allee.mpe.mpg.de) joined #fai. [20:01] alekibango (~alekibang@ip-94-113-217-204.net.upcbroadband.cz) left irc: Server closed connection [20:01] alekibango (~alekibang@ip-94-113-217-204.net.upcbroadband.cz) joined #fai. [20:08] how do i get the newly bootet system to handle device changes? it pxe-boots of a nic that turns out to become eth2 afterwards. [20:08] do i need to do ip=:::hostname:eth3:dhcp in the pxe config file that fai-chboot generates? [20:13] siggg (~jean@tra78-5-88-172-128-32.fbx.proxad.net) joined #fai. [20:15] thats the only way i have bin able to do it... [20:34] msantana (msantana@SDF.ORG) left irc: Quit: Coyote finally caught me [20:34] ErKa (~keryell@keryell.pck.nerim.net) left irc: Read error: Connection reset by peer [20:45] ErKa (keryell@keryell.pck.nerim.net) joined #fai. [21:22] Guest616 (52aa2d71@ircip2.mibbit.com) left irc: Quit: http://www.mibbit.com ajax IRC Client [21:31] _julian (~quassel@hmbg-5f76600b.pool.mediaWays.net) left irc: Server closed connection [21:31] _julian (~quassel@hmbg-5f76600b.pool.mediaWays.net) joined #fai. [21:53] ErKa (keryell@keryell.pck.nerim.net) left irc: Ping timeout: 480 seconds [22:00] pav5088 (~Mark@ppp118-208-33-206.lns20.bne1.internode.on.net) left irc: Remote host closed the connection [22:03] pav5088 (~Mark@ppp118-208-33-206.lns20.bne1.internode.on.net) joined #fai. [22:09] siggg (~jean@tra78-5-88-172-128-32.fbx.proxad.net) left irc: Ping timeout: 480 seconds [23:31] msantana (msantana@SDF.ORG) joined #fai. [00:00] --- Wed Feb 9 2011