| juri_ (~demo@ip70-178-33-171.ks.ks.cox.net) left irc: Remote host closed the connection |
| juri_ (~juri@ip70-178-33-171.ks.ks.cox.net) joined #fai. |
| siggg (~jean@tra78-5-88-172-128-32.fbx.proxad.net) joined #fai. |
| siggg (~jean@tra78-5-88-172-128-32.fbx.proxad.net) left irc: Ping timeout: 480 seconds |
| MT (~MT@s0478.vpn.hrz.tu-darmstadt.de) joined #fai. |
| MT_ (~MT@vpn.forsyte.informatik.tu-darmstadt.de) joined #fai. |
| Nick change: MT -> Guest1338 |
| Nick change: MT_ -> MT |
| Guest1338 (~MT@s0478.vpn.hrz.tu-darmstadt.de) left irc: Ping timeout: 480 seconds |
| MT (~MT@vpn.forsyte.informatik.tu-darmstadt.de) left irc: Read error: Connection reset by peer |
| MT (~MT@vpn.forsyte.informatik.tu-darmstadt.de) joined #fai. |
| MT (~MT@vpn.forsyte.informatik.tu-darmstadt.de) left irc: Ping timeout: 480 seconds |
| mludi42 (~or4000@213-205-85-140.net.novis.pt) joined #fai. |
| siggg (~jean@dyn-83-156-148-130.ppp.tiscali.fr) joined #fai. |
| Mrfai (~lange@suenner.informatik.uni-koeln.de) joined #fai. |
| MT (~MT@vpn.forsyte.informatik.tu-darmstadt.de) joined #fai. |
| Nick change: MT -> Guest1349 |
| Guest1349 (~MT@vpn.forsyte.informatik.tu-darmstadt.de) left irc: Read error: Connection reset by peer |
| MT_ (~MT@vpn.forsyte.informatik.tu-darmstadt.de) joined #fai. |
| Nick change: MT_ -> MT |
| MT_ (~MT@130.83.161.205) joined #fai. |
| Nick change: MT -> Guest1350 |
| Guest1350 (~MT@vpn.forsyte.informatik.tu-darmstadt.de) left irc: Read error: Connection reset by peer |
| Nick change: MT_ -> MT |
| MT_ (~MT@vpn.forsyte.informatik.tu-darmstadt.de) joined #fai. |
| MT (~MT@130.83.161.205) left irc: Read error: Connection reset by peer |
| MT (~MT@130.83.161.205) joined #fai. |
| MT_ (~MT@vpn.forsyte.informatik.tu-darmstadt.de) left irc: Read error: Connection reset by peer |
| MT_ (~MT@vpn.forsyte.informatik.tu-darmstadt.de) joined #fai. |
| MT (~MT@130.83.161.205) left irc: Read error: Connection reset by peer |
| MT_ (~MT@vpn.forsyte.informatik.tu-darmstadt.de) left irc: Remote host closed the connection |
| MT (~MT@vpn.forsyte.informatik.tu-darmstadt.de) joined #fai. |
| siretart (siretart@tauware.de) left irc: Ping timeout: 480 seconds |
| siretart (siretart@tauware.de) joined #fai. |
| MT_ (~MT@vpn.forsyte.informatik.tu-darmstadt.de) joined #fai. |
| Nick change: MT -> Guest1357 |
| Guest1357 (~MT@vpn.forsyte.informatik.tu-darmstadt.de) left irc: Read error: Connection reset by peer |
| Nick change: MT_ -> MT |
| 10:37 mludi42 | I am just thinking, i have written a ln -s in my script files to link to the ldap.conf. this works fine in a softupdate. but i am not sure if this links correctly when installed because I used the ${target} variable. |
| MT_ (~MT@vpn.forsyte.informatik.tu-darmstadt.de) joined #fai. |
| Nick change: MT -> Guest1358 |
| Guest1358 (~MT@vpn.forsyte.informatik.tu-darmstadt.de) left irc: Read error: Connection reset by peer |
| Nick change: MT_ -> MT |
| 10:47 mludi42 | ah, I guess something like: ln -s ldap.conf ${target}/etc/libnss-ldap-conf should work |
| MT_ (~MT@vpn.forsyte.informatik.tu-darmstadt.de) joined #fai. |
| Nick change: MT -> Guest1359 |
| Guest1359 (~MT@vpn.forsyte.informatik.tu-darmstadt.de) left irc: Read error: Connection reset by peer |
| Nick change: MT_ -> MT |
| MT (~MT@vpn.forsyte.informatik.tu-darmstadt.de) left irc: |
| MT (~MT@vpn.forsyte.informatik.tu-darmstadt.de) joined #fai. |
| MT (~MT@vpn.forsyte.informatik.tu-darmstadt.de) left irc: Quit: Leaving |
| MT (~MT@vpn.forsyte.informatik.tu-darmstadt.de) joined #fai. |
| jo (~joerg@ppc214.mipool.uni-jena.de) joined #fai. |
| 11:51 jo | Does FAI support incremental updates of the clients, i.e. run apt-get upgrade on each client? How are the clients maintained with FAI? What happens after the installation? |
| 11:59 Mrfai | lFAI softupda is a incremental update. If you only like to run apt-get update; ... upgrade, you can do it with a small script on your own. You also have to decide by yourself when to run such a script. Via cron, or only on a reboot, or oush from a central server? It's your decission how to do this, FAI will not force you to use a certain methed. |
| d_v_l (~dvl@193.72.17.15) joined #fai. |
| 13:19 d_v_l | Hi folks. I have two NICs on my fai-clients. Both NICs are served by DHCP. Unfortunately when starting the fai-boot the wrong interface gets dhcp-ed. Since there are a lot of different servers, I cannot realy control, which interface asks at first for DHCP. Is there any workaround or solution for controlling this? |
| jo (~joerg@ppc214.mipool.uni-jena.de) left irc: Quit: leaving |
| siggg (~jean@dyn-83-156-148-130.ppp.tiscali.fr) left irc: Ping timeout: 480 seconds |
| 13:49 mcarvalh | In live.log upon booting my FAI client, I get the message "mount: Mounting unionfs on /root failed: No such device", even after compiling in UFS support |
| 13:49 mcarvalh | this is with FAI version 3.2.4 on a Lenny server |
| 13:51 mcarvalh | Does ufs support need to be built into the kernel that exists within my NFS root, or just in the kernel which I use to boot my client via PXE? |
| 13:52 Mrfai | AFAIK ufs is not unionfs |
| 13:53 mcarvalh | really...ok |
| 13:54 mcarvalh | that would explain it then, looks like i need to patch my kernel. I will try that. Thanks |
| siggg (~jean@dyn-83-156-148-130.ppp.tiscali.fr) joined #fai. |
| fai-guy (~christian@16.121.113.82.net.de.o2.com) joined #fai. |
| 15:16 mcarvalh | yeah! that did it...im finally booting a Lenny client from FAI 3.2.4 |
| 15:17 mcarvalh | so things I've learned that bogged me down...the boot kernel needs root_nfs support and unionfs support built into it, and in my test case, and pcnet32 network driver for my vmware clients |
| stockholm (~andreas@marta.schuldei.org) joined #fai. |
| 16:13 stockholm | eartoast: where is IPADDR set running softupdates? |
| siggg (~jean@dyn-83-156-148-130.ppp.tiscali.fr) left irc: Ping timeout: 480 seconds |
| mludi42 (or4000@213-205-85-140.net.novis.pt) left #fai. |
| 16:34 Mrfai | maybe nowhere, because normally IPADDR is set by DHCP, which may not be called during softupdate |
| Mrfai (~lange@suenner.informatik.uni-koeln.de) left irc: Quit: leaving |
| 17:21 d_v_l | Nachdem heute nachmittag keiner geantwortet hat, traue ich mich einfach mal, meine Frage nochmal zu stellen... |
| fai-gu1 (~christian@16.121.113.82.net.de.o2.com) joined #fai. |
| 17:22 d_v_l | Ich habe auf den Clients zwei NICs, die via DHCP versorgt werden. Gibt es eine Möglichkeit, zu steuern, welche NIC als erstes via DHCP versorgt wird? |
| 17:23 d_v_l | Sobald die sog. zweite die IP bekommt, mach FAI Probleme (Config-Dir und NFSROOT) |
| fai-guy (~christian@16.121.113.82.net.de.o2.com) left irc: Ping timeout: 480 seconds |
| d_v_l (~dvl@193.72.17.15) left irc: Quit: KVIrc 3.2.0 'Realia' |
| Nick change: fai-gu1 -> fai-gui |
| Nick change: fai-gui -> fai-guy |
| mcarvalh (~mcarvalh@144-212-103-20.mathworks.com) left irc: Quit: Leaving |
| mcarvalh (~mcarvalh@144-212-103-20.mathworks.com) joined #fai. |
| MT (~MT@vpn.forsyte.informatik.tu-darmstadt.de) left irc: Ping timeout: 480 seconds |
| MT (~MT@flag.forsyte.informatik.tu-darmstadt.de) joined #fai. |
| fai-guy (~christian@16.121.113.82.net.de.o2.com) left irc: Quit: Leaving. |
| fai-guy (~christian@16.121.113.82.net.de.o2.com) joined #fai. |
| fai-guy (~christian@16.121.113.82.net.de.o2.com) left irc: |
| torkel (torkel@ip64.degernas.se) left irc: Quit: leaving |
| mac6cast (~mac6cast@marxbook.stattfernsehen.com) joined #fai. |
| 19:07 mcarvalh | I'm getting the following error when my Lenny client attempts to mount its mirror |
| mac6cast (mac6cast@marxbook.stattfernsehen.com) left #fai. |
| 19:07 mcarvalh | Can't set permissions on mtab: Operation not permitted |
| 19:08 MT | are you sure that unionfs is working? |
| 19:08 mcarvalh | However, I can manually mount it if i call mount -n -o nolock...is this a bug in FAI 3.2.4? |
| 19:08 MT | not a known one, at least |
| 19:08 mcarvalh | i was looking at http://layer-acht.org/fai-irc/fai.log.20070925.html, which looks like a similar problem mounting the config dir |
| 19:08 MT | does something like touch /bla work? |
| 19:09 mcarvalh | i can make directories files etc and mount my debian mirror from the fai client after it drops me to the shell, yes |
| siggg (~jean@tra78-5-88-172-128-32.fbx.proxad.net) joined #fai. |
| 19:11 mcarvalh | i imagine i can fix it by setting my mount options for task_mirror, but im not sure where to do so |
| 19:12 MT | it might be worth bringing this up on the fai mailinglist |
| 19:12 MT | just in case you don't get immediate help here |
| 19:12 MT | (I'm on the leave, will be back online in 30mins or so) |
| 19:12 mcarvalh | ok, not a problem, i might post tomorrow if Mrfai is on |
| 19:19 mcarvalh | i also signed up on the mailing list. thanks for the suggestion |
| MT (~MT@flag.forsyte.informatik.tu-darmstadt.de) left irc: Ping timeout: 480 seconds |
| MT (~MT@drms-4d0067b1.pool.mediaWays.net) joined #fai. |
| sepski (~seppity@217.17.211.51) joined #fai. |
| MT_ (~MT@vpn.forsyte.informatik.tu-darmstadt.de) joined #fai. |
| MT (~MT@drms-4d0067b1.pool.mediaWays.net) left irc: Ping timeout: 480 seconds |
| 19:58 mcarvalh | fwiw, i had to add the option -n to romountopt in my nfsroot |
| 19:59 mcarvalh | so no /srv/fai/nfsroot/live/filesystem.dir/etc/init.d/rcS has the following line changed: |
| 19:59 mcarvalh | so now* |
| 19:59 mcarvalh | export romountopt="-n -o async,noatime,nolock,ro,actimeo=1800" |
| 20:01 MT_ | hmmm |
| 20:01 MT_ | -n is already used for getting the config space |
| 20:01 MT_ | so only the mirror is still affected |
| 20:02 MT_ | ok, this is a bug |
| 20:02 MT_ | mcarvalh, would you mind filing a bug report? |
| 20:03 mcarvalh | ok, this will be my first :) |
| 20:04 MT_ | say that -n must be added to the mount command in task_mirror |
| 20:04 MT_ | just as is done in get-config-dir-nfs |
| 20:04 MT_ | and tag it "patch" |
| 20:04 mcarvalh | is that the correct place i should add this option, or is there a more approprate place for it, (e.g something equivalent to get_config_dir_nfs) |
| 20:04 mcarvalh | where is task_mirror located? |
| 20:04 MT_ | to me it seems to be most appropriate |
| 20:04 MT_ | in subroutines-linux |
| 20:05 MT_ | that is /usr/lib/fai/subroutines-linux |
| 20:05 mcarvalh | ok, thanks |
| 20:05 MT_ | in get-config-dir-nfs it says something like mount -n $romountopt |
| 20:05 MT_ | in task_mirror that -n is simply missing :-) |
| 20:06 MT_ | it was added to get-config-dir-nfs only very recently |
| 20:06 MT_ | so just nobody had stumbled upon it yet ... |
| 20:07 mcarvalh | heh, glad to be the first |
| BarryAllen (~marx@200.195.196.16) joined #fai. |
| 20:08 MT_ | but even cooler that you already have the fix ready :-) |
| 20:08 BarryAllen | hi all |
| 20:08 MT_ | hi :-) |
| 20:08 BarryAllen | im installing 3.2.35... but im missing the KERNELPACKAGE variable in the config files... is that deprecated ? |
| 20:08 MT_ | .35 doesn't exist (yet) ... |
| 20:09 BarryAllen | (im using the cvs fai-source branch) |
| 20:09 MT_ | :-) |
| 20:09 BarryAllen | ops, sorry :) |
| 20:09 MT_ | yes, it is no longer there |
| 20:09 BarryAllen | MT_, where is it now? |
| 20:09 MT_ | the kernel is now simply installed via apt |
| 20:09 MT_ | thus you need to modify /etc/fai/NFSROOT |
| 20:09 BarryAllen | nice :) |
| 20:10 BarryAllen | tks MT_ ! |
| 20:10 MT_ | in case you need to do so, at all |
| 20:27 mcarvalh | hmm, so if i submit a bug report from my older etch box, it autofills the version of FAI as 3.1.8 in my report...can i just edit that line? |
| 20:27 MT_ | yes, just do so |
| 20:27 MT_ | The only essential lines are |
| 20:27 stockholm | MT_: you work on setup-storage again? |
| 20:27 MT_ | Package: fai |
| 20:27 MT_ | Version: 3.2.4 |
| 20:27 stockholm | MT_: i was installing the newest version the other day |
| 20:27 MT_ | Tags: patch |
| 20:27 MT_ | ( :-) ) |
| 20:27 stockholm | and it did not work |
| 20:28 MT_ | well, first I started using it myself |
| 20:28 stockholm | i fixed some obvious bugs on the way to get it to compile but there came more and more |
| 20:28 MT_ | then I did some hacking |
| 20:28 MT_ | which SVN version? |
| 20:28 stockholm | i am using it too |
| 20:28 stockholm | that was head yesterday or the day before |
| 20:28 MT_ | I fixed some bugs in 4871 just yesterday |
| 20:28 MT_ | ah, ok :-) |
| 20:28 MT_ | I had integrated the encryption thingy |
| 20:29 MT_ | I think r4872 should be ok again |
| 20:29 stockholm | neat |
| 20:29 MT_ | but in case there are still any errors, I'm ready to fix them here and now :-) |
| 20:30 stockholm | i could not reach you when i had the problems so i went back to a known good version |
| 20:30 stockholm | i will come back and try it again soon |
| 20:31 MT_ | I've been mostly off IRC as I was pretty busy setting up the new system |
| 20:32 MT_ | in case I don't respond more or less immediately, just drop me an email, be it PM or via the list |
| 20:34 stockholm | great, will do |
| 20:34 stockholm | "the new system" sounds like i could know what you mean. what is it? |
| 20:35 MT_ | our group has moved to TU Darmstadt |
| 20:35 MT_ | packed my SVN from Munich |
| 20:35 MT_ | reworked the necessary bits |
| 20:35 MT_ | upgrade to latest SVN-FAI |
| 20:35 MT_ | and reinstalled :-) |
| 20:35 stockholm | ok |
| 20:35 stockholm | :-) |
| 20:36 stockholm | so you moved house from munich to darmstadt, too? |
| 20:36 MT_ | well, I had been sharing a flat back in in MUC, moved into a new flat here |
| 20:37 MT_ | but moving those things took even longer than moving the system :-) |
| 20:39 stockholm | sure... then finding friends and stuff takes years |
| 20:39 MT_ | well, currently I'm doing hardly anything else than working |
| 20:41 stockholm | gee. :-) |
| 20:41 stockholm | how can i express negative preferencs regarding packages for a class? |
| 20:41 MT_ | what exactly do you mean by negative preferences? |
| 20:42 stockholm | how can i say "if CLASS is defined, dont install PACKAGE"? |
| 20:42 MT_ | at least with PACKAGES install |
| 20:42 MT_ | that does work |
| 20:42 MT_ | as long as the entries end up in the same list |
| 20:42 MT_ | (install packages splits a some point) |
| 20:42 MT_ | just say |
| 20:42 stockholm | like a line PACKAGE- in the CLASS file in package_config ? |
| 20:43 MT_ | right |
| torkel (torkel@ip64.degernas.se) joined #fai. |
| 20:43 MT_ | so |
| 20:43 stockholm | MT_: hm? |
| 20:43 MT_ | PACKAGES install THE_CLASS_WHERE_YOU_DONT_WANT_IT |
| 20:43 MT_ | package-goes-away- |
| 20:43 MT_ | if you have some |
| 20:44 MT_ | PACKAGES install |
| 20:44 MT_ | package-goes-away |
| 20:44 MT_ | say the package-goes-away- |
| 20:44 MT_ | close to the other one |
| 20:45 MT_ | otherwise you could end up in an install - remove - install loop |
| 20:45 MT_ | if I'm not mistaken |
| 20:45 MT_ | once I think about it again, I might well be mistaken |
| 20:45 MT_ | and you can put it anywhere |
| 20:46 stockholm | where is that documented? |
| 20:46 MT_ | that - what? |
| 20:46 MT_ | removing packages? |
| 20:46 stockholm | that "PACKAGES install THE_CLASS_WHERE_YOU_DONT_WANT_IT" line |
| 20:46 MT_ | hm!? |
| 20:47 MT_ | meaning that you can put a classname after the command? |
| 20:47 MT_ | I'm just not sure what you're unclear about |
| 20:47 stockholm | i did not know that was possible |
| 20:48 MT_ | you get the magic <class_of_the_filename> && THE_OTHER_CLASS |
| 20:49 stockholm | hm, do you have a real world example? |
| 20:49 stockholm | just to be sure? :-) |
| 20:49 MT_ | yes, just a sec |
| 20:50 MT_ | it's even in the fai-guide |
| 20:50 MT_ | Software package configuration |
| 20:50 MT_ | everything's there, including the - thing |
| 20:50 stockholm | wow |
| 20:50 stockholm | ok, i look it up there |
| 20:51 MT_ | well, for the real world: |
| 20:51 MT_ | consider my DEFAULT |
| 20:51 MT_ | PACKAGES install |
| 20:51 MT_ | acl |
| 20:51 MT_ | acpid apmd- |
| 20:51 MT_ | apt-show-versions |
| 20:51 MT_ | attr |
| 20:51 MT_ | bash |
| 20:51 MT_ | bzip2 |
| 20:51 MT_ | PACKAGES install |
| 20:51 MT_ | acl |
| 20:51 MT_ | acpid apmd- |
| 20:52 MT_ | apt-show-versions |
| 20:52 MT_ | attr |
| 20:52 MT_ | bash |
| 20:52 MT_ | bzip2 |
| 20:52 MT_ | PACKAGES install |
| 20:52 MT_ | acl |
| 20:52 MT_ | acpid apmd- |
| 20:52 MT_ | apt-show-versions |
| 20:52 MT_ | sorry, that was twice |
| 20:52 MT_ | but that continues ... down to ... |
| 20:52 MT_ | PACKAGES aptitude mole |
| 20:52 MT_ | # using e1000-source, there is no need for these |
| 20:52 MT_ | # linux-image-2.6-amd64/etch-backports |
| 20:52 MT_ | # linux-headers-2.6-amd64/etch-backports |
| 20:52 MT_ | mpt-status/etch-backports |
| 20:52 MT_ | PACKAGES install AMD64 |
| 20:52 MT_ | linux-image-2.6-amd64 |
| 20:52 MT_ | linux-headers-2.6-amd64 |
| 20:52 MT_ | PACKAGES install AMD |
| 20:52 MT_ | linux-image-2.6-k7 |
| 20:52 MT_ | linux-headers-2.6-k7 |
| 20:52 MT_ | PACKAGES install IA32 |
| 20:52 MT_ | linux-image-2.6-686 |
| 20:53 MT_ | linux-headers-2.6-686 |
| 20:53 MT_ | PACKAGES install GRUB |
| 20:53 MT_ | grub |
| 20:53 MT_ | so I'm using the DEFAULT and specify all the specific kernel things |
| 20:53 MT_ | instead of having a bunch of separate files |
| 20:53 stockholm | why dont you use aptitiude all the time? then you would not bloat your system if dependencies change |
| 20:54 MT_ | oh, well, that's right but |
| 20:54 MT_ | - I was too lazy to change |
| 20:54 stockholm | :-) |
| 20:54 stockholm | ok |
| 20:54 MT_ | - there is no checking for existing packages with aptitude |
| 20:54 MT_ | I think it will just fail ... :-( |
| 20:58 mcarvalh | <MT_> I just finished submitting that bug report FYI |
| 21:20 mcarvalh | http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=477564 |
| 21:21 mcarvalh | hmm, but if i make that change in /usr/lib/... and then re-make my nfsroot, it doesnt seem to apply in there |
| 21:21 MT_ | nop |
| 21:22 mcarvalh | how do i get it in there without manually chroot'ing into it and making the change? |
| 21:23 MT_ | either, build new packages and install them manually |
| 21:23 MT_ | or make a script to do the patching |
| 21:24 mcarvalh | k, i think ill leave that up to the maintainers :) |
| 21:26 MT_ | so you'd better ping Mrfai tomorrow to get it done soon :-) |
| 21:26 mcarvalh | at least i can wedge the fix into my nfsroot to continue my testing |
| 21:44 mcarvalh | does the kernel name of the PXE boot kernel need to match the name of the kernel within the nfsroot? |
| 21:45 MT_ | not the name, but the version |
| 21:45 MT_ | you'll need to have the appropriate modules ready |
| 21:45 mcarvalh | for example, my boot kernel (with unionfs and root_nfs compiled in) is named 2.6.22-fai, but the one in the nfsroot is 2.6.22-3-486... |
| 21:46 mcarvalh | and it seems to have a hard time finding modules since it refers to the wrong path now |
| 21:46 MT_ | ok, yes |
| 21:46 MT_ | you'll need to get your modules installed into the nfsroot |
| 21:46 mcarvalh | ahh..do i just add the to /etc/fai/NFSROOT? |
| 21:47 MT_ | if you have a proper Debian package, which is available via some mirror, yes |
| 21:47 mcarvalh | i made the package myself |
| sepski (~seppity@217.17.211.51) left irc: Quit: Leaving |
| 21:53 mcarvalh | i give up for the evening...thanks for your help today! |
| siggg (~jean@tra78-5-88-172-128-32.fbx.proxad.net) left irc: Ping timeout: 480 seconds |
| --- Thu Apr 24 2008 |