[04:30] i-R2D247 (~deed@tor-irc.dnsbl.oftc.net) left irc: Remote host closed the connection [04:32] i-R2D247 (~deed@tor-irc.dnsbl.oftc.net) joined #fai. [05:36] moin [05:58] rcy (~rcy@209.90.176.14) left irc: Ping timeout: 480 seconds [06:30] MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. [07:23] Nick change: MT -> Guest1778 [07:23] Guest1778 (~MT@dove.informatik.tu-muenchen.de) left irc: Read error: Connection reset by peer [07:23] MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. [07:26] lazyb0y, nowadays http://layer-acht.org/debian etch fai is the same as thomas repository. this wasnt always the case and might not be [07:35] h01ger, but if it fails atleast i know who to blame now :) [08:04] Mrfai (~lange@kueppers.informatik.uni-koeln.de) joined #fai. [08:28] allee (~ach@allee.mpe.mpg.de) joined #fai. [08:39] http://paste.debian.net/39759 [08:39] MT (~MT@dove.informatik.tu-muenchen.de) left irc: Quit: MT [08:40] you scared him :) [08:40] seams that eh pv is not created during the install, but if i halt it there and use tty2 i can run pvcreate /dev/sda2 without problems [08:40] MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. [08:40] seams that way [08:40] and the install just happily continues and installs on the / raqmdisk. [08:40] it should perhaps do better error detection and not install [08:41] ErKa (~keryell@keryell.pck.nerim.net) joined #fai. [08:52] MT, seams that the pv is not created during the install, but if i halt it there and use tty2 i can run pvcreate /dev/sda2 without problems http://paste.debian.net/39759 [08:55] so, /dev/sda2 does exist? [08:58] yes [08:59] this used to work in vmware, i just copied /srv/fai/config to the real server [09:00] h01ger: also, ich denke nicht, dass ich es schaffe, die mirrors zu machen - hier ist grad "alles kaputt" [09:01] ich werde also eine leere 500GB platte mitbringen, die in den nslu kann, und eine mit einem etch mirror und den etch skolelinux oaketen (stable), die in meinem notebook steckt - die daten kann ich natuerlich umkopieren [09:01] sep, could you re-run the install [09:01] and interrupt it once again? [09:01] MTi just added a sleep after parted -s /dev/sda set 2 lvm on so i could see what was going on, and now it worked [09:01] i am restarting it [09:02] and could you then only run the pvdisplay /dev/sda2 command? [09:02] and an echo $? [09:02] afterwards? [09:02] I'd be interested in error messages and the exit code ... [09:02] MT,i tried the pvdisplay /dev/sda2 || pvcreate /dev/sda2 on tty2 and it succeeded [09:03] without any error message? [09:03] ah, with the sleep even the automatic thing worked out? [09:03] pvdisplay put out messages like [09:03] Failed to read physical volume "/dev/sda2" [09:03] Device /dev/sda2 not found (or ignored by filtering). [09:04] i think i have to remove the sleep again, since i can't make it fail now [09:04] well, in fact the sleep may be the solution [09:04] because it takes some time to rsync the tables [09:05] udev need to resettle after parted twiddle the partition ? [09:05] yes, I think so [09:05] probably worked in vmware since that's much slower then this server [09:07] is there no way for udev to signal when it's done ? [09:07] hmm, no idea [09:07] I just found a sleep 1 to be the simplest solution :-) [09:08] sep, I've just committed r4623 to SVN, which has two more sleeps in it [09:08] im svn up and restarting the install [09:09] MT this is basicaly exactly what i am doing http://wiki.debian.org/DebianEdu/HowTo/FaiInstallDebianEdu [09:10] sdd2 succeeded now [09:10] MT, thanks :) [09:11] MT, it might be a good idea to check that partitions get created and stop if something fails. now the install compleated and rebooted and i was in a grub meny with no remote access to the machine at all. [09:12] /target was not mouned but /target/boot was so the install was compleated to /target (ramdisk) [09:12] I've the problem, that a specific package installs just fine on unstable but is b0rken in etch; any ideas what I could do to resolve that issue so that I get the package (it's snort) in an sid-installation but not in an etch system without putting all the package list into a new class? [09:14] sep, was it really the case that the partition was not created, or was there some error in mounting it? [09:14] mikap: that is you are basically installing different distributions [09:14] (I think in debian they are even called so) [09:15] what I do when I have package differences between Debian and ubuntu is: [09:15] the basic usecase is handled with the class MYUSECASE [09:15] MT, vg_system was not created becouse /dev/sda2 was not a pv, [09:16] I mark that a system belongs to some distribution with classes like DIST-DEBIAN-ETCH and DIST-UBUNTU-DAPPER [09:16] then the lv's was not craeted becouse vg_system was missing, [09:16] using the latest SVN? [09:16] or was that the case with earlier versions where you reported the error? [09:16] earlier [09:16] ah, ok [09:16] now it workes fine [09:16] I added a note on self-tests to the wiki-page [09:17] #Implementation [09:17] and in the package list file for MYUSECASE I differentiate between packages that are installed for all systems with MYUSECASE, some that are only installed for DIST-DEBIAN-ETCH and some that are only installed for DIST-UBUNTU-DAPPER [09:17] (by adding the class name in an addtional section like "PACKAGES install DIST-DEBIAN-ETCH" [09:18] .oO( soo many highlights here...) [09:21] MT, the lv's was not mounted including /target, but /dev/sda1 was mounted as /target/boot and then the base was extracted and the install continued [09:22] lazyb0y: exactly; what I don't like in this variant is that the amount of classes is growing and growing :-/ [09:36] mikap: yes, that's the price you pay for fai's simplicity [09:37] the other way would be, have some kind of meta-information somewhere in package lists, and decide which package list to use for example depending on the value of the variable DISTRIBUTION (not used yet) [09:38] in any case, _somewhere_ you add complexity if you start installing more than one distribution(-version) [09:40] hmm, apart from that, if I look at my configspace - I have many different classes for many usecases, and only a few DIST-* classes - they don't blow it up that much [09:42] lazyb0y: ACK; main problem is that grml-live has a nice abstraction layer already and now I have to add some more complexity :-/ [09:45] lazyb0y: ich lege zB ein neues file an für eine partitionierung. FAIVSERVER zB [09:45] nun möchte ich sagen [09:45] ./installieen IP FAIVSERVER [09:49] baldy: ich verstehe nur bahnhof. Bitte schreibe doch mal dein anliegen in 3-5 vollstndigen, zusammenhngenden Stzen zusammen, und poste das an die Mailingliste(falls du das nicht schon getan hast, und ich das grad uebersehe [09:50] hm, would it be useful to just execute scripts that don't end with .dpkg-old? :) [09:52] mikap: I don't understand the last one? what and where are these scripts and how do they help with snort install problems on etch? [09:54] lazyb0y: has nothing to do with my snort-problem :) [09:54] lazyb0y: if you modified files inside /etc/path/to/fai and install an update (at least with grml-live where the files are marked as conffiles) you'll have .dpkg-new/.dpkg-old/... laying around and they are executed [09:55] though I'm wondering why nobody except me noticed that snort is uninstallable on Debian/etch when using FAI dirinstall [09:56] mikap: mikap dpkg-old: the fai configspace is "normally" not a conffiles dir - it's not even contained in any package at the location where it's actually used [09:57] it's only provided as example in the doc package [09:57] lazyb0y: I'm aware of that [09:57] lazyb0y: though people deploying their systems using FAI might keep such stuff inside a package (like what I'm doing with grml-live) [09:58] ah. but,. I guess nobody would mind if you provide a patch that make the old files not being executed :) [09:58] yes [09:58] sure [09:58] I just don't want to write patches nobody applies ;) [09:59] that's the decision of our source-dictator :) I vote for your proposal when asked [09:59] ok :) [09:59] Action: mikap has several new error-handling issues regarding http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=444210 as well [10:00] as for configspace in a package: yes! I also wrote up a proposal for a slightly different structure of the configspace - so each class has it's own directory, and therefore can be exchanged easier [10:01] cool :) [10:02] I'm pretty happy with current state of grml-live, though I have to handle all the errors of fai dirinstall run inside grml-live currently ;( [10:03] hopefully I'll have a few days time to sit down and write some patches against FAI which do all I need ;) [10:08] lazyb0y, ok. wir koennen da ja zu diesem edu-center gehen , da muesste es 34mbit oder sowas geben... ich nehm auf jeden fall das sata gehaeuse mit [10:16] mikap: I can confirm that installing snort wuring dirinstall is broken. But it's the fault of snort IMO. [10:18] moin [10:30] williamvergara (~chatzilla@nat-1.rz.uni-karlsruhe.de) joined #fai. [10:31] hi everybody, i have a fast question [10:32] how can i configure disk_config, so that it can always protect the third primary partition, regardless of its size [10:32] ?? [10:32] preserve3 [10:33] but setup_harddisks will complain, if you first and second partition are too big and would overlap into the third partition. [10:37] i get this :Partition is not DOS aligned [10:37] h01ger: achso - nee, das sata gehaeuse brauchst du nicht mitbringen - ich hab mich falsch ausgedrueckt. die sata platte kann ich nichts ausbauen aus dem rechner, aber ich hab noch eine ide platte gefunden, die man nehmen kann [10:37] reichen 80gb fuer diese mirrors? [10:38] first partition is 10 gb, second 1 gb [10:38] btw, thanks a lot for your help [10:38] williamvergara: what kind of partition is that? [10:38] is a ntfs partition [10:39] and i have it so in my disk_config: primary - preserve3 [10:40] I saw strange problems when trying to preserve thinkapd rescure partitions - FAI's setup harddsik wouldn't work with them for me, and the gentoo installer even decided to completely delete it without further notice (sinece then I am a big gentoo fan :) ) [10:41] williamvergara: ho does your whole disk config look like? [10:41] primary / 10000 ; format reiser [10:41] primary swap 1000 ; lazyformat -i 1024 [10:41] primary - preserve3 ; [10:41] lol, i also use gentoo for my personal computers [10:43] eeww wtf, stockholm's lvm issues is biting me on the real hardware, but not in vmware. [10:43] it's stuck during boot saying "waiting for root fs" [10:46] Mrfai: nope, fixed in unstable - see http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=443481 [10:47] stockholm, did you ever manage to boot on lvm ? [10:47] mikap: snort.preinst uses $LOGDIR, which is also used in FAI. therefore mkdir in this preinst will fail [10:48] yep, so snort in etch is still broken :-( and FAI has again to do the workaround :-(( [10:49] possible workaround in FAI: set LOGDIR in /etc/snort/snort.common.parameters before installing snort [10:49] use a hook for that [10:50] any advice for me, mrfai? [10:51] you can ignore the warning that a partition is not DOS aligned. Or is it an error and stops? [10:53] it stops. i think that disk_var file depends of that [10:54] setup_harddisks did not create /var/log/fai/current/disk_var.sh file. [10:54] you may add "sfdisk="--force" to a class/*.var file. [10:55] BTW, try without lazyformat [10:55] ok, ill try that out [10:55] thanks a lot [10:55] Mrfai: exactly; oh cool - I wasn't aware of this option as a workaround :) [10:55] and using format for the first primary is also not needed IIRC [10:56] mikap: I just looked into /var/lib/dpkg/info/snort.preinst and saw that they are reading this file. [10:56] Mrfai: ah :)) [10:57] is this right: [10:57] sfisk=--force [10:59] use: sfdisk="--firce" [10:59] s/i/o/ :-) [11:00] thanx, ;-) [11:04] Mrfai (~lange@kueppers.informatik.uni-koeln.de) left irc: Quit: leaving [11:04] Mrfai (~lange@suenner.informatik.uni-koeln.de) joined #fai. [11:04] lazyb0y, 80gb sind zu klein, glaub ich. [11:04] Mrfai, ist dein repo gesigned? [11:06] ja, mittlerweile schon [11:06] fain :) [11:07] mit deinem normalen key, nehm ich mal an? [11:11] yep [11:12] MT (~MT@dove.informatik.tu-muenchen.de) left irc: Ping timeout: 480 seconds [11:14] well, i made a new class in /class, ist name 001.var and in that file i wrote sfdisk="--force" [11:14] the result was the same, :-( [11:15] does your client belong to the class 001? [11:19] yes, i also put it in another one, default, just to be sure [11:22] Mmm, you have also specify another geometry using the sfdsik variable. Sometimes windo.. uses another disk geometry than linux. There's a logical and physical geometry. [11:28] so, it would be something like sfdisk="--forece something" [11:38] but really, i still can not figure out, why it just wont preserve the third partition, it shouldnt be that hard to get done [11:42] stockholm, MT, not sure if my lvm boot problem was the same as your boot problem, but if i added lvm2 as a package to be installed in the target system (i use FAIBASE) , then update-initramfs automagicaly added lvm support to the initrd [12:11] MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. [12:36] rcy (~rcy@209.90.176.221) joined #fai. [12:38] ErKa (~keryell@keryell.pck.nerim.net) left irc: Ping timeout: 480 seconds [12:43] MT, i seem to recall there is a better way to wait for udev than just sleeping some secs. but i forget which it was.. :) [12:43] forgto [12:43] Action: h01ger throws his hands up in disgust [12:43] s/forget/forgot/ [12:43] I'll try to look it up later on [12:44] in case you get an idea what we was, just ping me :-) [12:45] stockholm, MT, not sure if my lvm boot problem was the same as your boot problem, but if i added lvm2 as a package to be installed in the target system (i use FAIBASE) , then update-initramfs automagicaly added lvm support to the initrd [12:45] MT are your internet intermittant ? :) [12:46] MT might want to add a note on the wiki about lvm2 in the target as well as the nfsroot [12:46] no, not really, was just moving around :-) [12:46] yes, will do!! [12:46] thanks for the pointer!!! [12:46] hopefully it will work for stockholm too when he have a chance to test it [12:47] to some extent it should have been obvious that this is the cause of trouble :-) [12:47] sep, would you add that to http://wiki.debian.org/DebianEdu/HowTo/FaiInstallDebianEdu too? [12:48] MT sure [12:49] faiwiki updated ... [12:50] i am basicaly pointing to the faiwiki for that part tho. since it's a moving target i dont want to end up with too much outdated info in the debian wiki [12:50] have a few example configfiles. but the rest they must get from the faiwiki [13:14] MT (~MT@dove.informatik.tu-muenchen.de) left irc: Ping timeout: 480 seconds [13:18] MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. [13:32] ErKa (keryell@m5.wifi.enstb.org) joined #fai. [13:58] MT (~MT@dove.informatik.tu-muenchen.de) left irc: Ping timeout: 480 seconds [14:26] MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. [14:36] h01ger, there is a command udevsettle [14:36] probably we should use this one [14:37] sep, still there? [14:39] Robinayor_ (~robin@suw.rkara.de) left irc: Remote host closed the connection [14:39] Robinayor_ (~robin@suw.rkara.de) joined #fai. [14:43] MT (~MT@dove.informatik.tu-muenchen.de) left irc: Read error: Connection reset by peer [14:44] bye [14:44] williamvergara (~chatzilla@nat-1.rz.uni-karlsruhe.de) left irc: Quit: ChatZilla 0.9.78.1 [Firefox 2.0.0.7/2007091417] [14:45] udevsettle - wait until queued kernel/udev events are handled [14:45] MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. [14:45] looks promising [15:05] MT (~MT@dove.informatik.tu-muenchen.de) left irc: Ping timeout: 480 seconds [15:32] ErKa (keryell@m5.wifi.enstb.org) left irc: Ping timeout: 480 seconds [15:38] allee (~ach@allee.mpe.mpg.de) got netsplit. [15:38] oz_ (~oz@anubis.coresystems.de) got netsplit. [15:39] allee (~ach@allee.mpe.mpg.de) returned to #fai. [15:39] oz_ (~oz@anubis.coresystems.de) returned to #fai. [15:43] MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. [15:48] MT, udevsettle indeed sounds good [16:02] sepski (~sep@217.17.211.51) joined #fai. [16:03] MT (~MT@dove.informatik.tu-muenchen.de) left irc: Ping timeout: 480 seconds [16:10] woffs (~woffs@osiris.taz.de) joined #fai. [17:02] woffs (~woffs@osiris.taz.de) left irc: Quit: Leaving. [17:07] sepski (~sep@217.17.211.51) left irc: Remote host closed the connection [17:08] Mrfai (~lange@suenner.informatik.uni-koeln.de) left irc: Quit: Lost terminal [17:25] allee (~ach@allee.mpe.mpg.de) left irc: Remote host closed the connection [17:27] ErKa (~keryell@keryell.pck.nerim.net) joined #fai. [17:39] allee (~ach@allee.mpe.mpg.de) joined #fai. [18:24] jesper_ (~jesper@shrek.krogh.cc) left irc: Remote host closed the connection [18:46] h01ger: hmm, ich habe sonst nochmal 100... sonst nochmal gucken, ich kann noch eine 2,5 sata mit nochmal 80 oder 100 leermachen, vielleicht [18:48] dann muesstest du doch dein gehaeuse mal mitbringen... [19:03] MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. [19:26] ErKa (~keryell@keryell.pck.nerim.net) left irc: Ping timeout: 480 seconds [20:56] h01ger: apt-mirror sagt zu Beginn: downloading 56GB - ohne etch, weil ich das lokal kopiere [21:54] google knows what FAI is. :) [22:05] MT (~MT@dove.informatik.tu-muenchen.de) left irc: Ping timeout: 480 seconds [22:18] lazyb0y, also ich nehm das gehaeuse jetzt mit, correct? [22:19] http://en.wikipedia.org/wiki/Fully_Automatic_Installation is a stub.. [00:00] --- Tue Oct 16 2007