/home/eggdrop/logs/fai.log.20071009

rcy (~rcy@209.90.176.121) joined #fai. jair_ (~jair@adsl-67-36-185-84.dsl.chcgil.ameritech.net) joined #fai. jair (~jair@adsl-67-36-182-169.dsl.chcgil.ameritech.net) left irc: Ping timeout: 480 seconds MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. MT (~MT@dove.informatik.tu-muenchen.de) left irc: Ping timeout: 480 seconds MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. sep (~sep@40.211.jostedal.no) joined #fai. ErKa (keryell@m5.wifi.enstb.org) joined #fai. serieus (~serieus@office.nedstat.nl) joined #fai. 08:04 serieus I want help with the problem described on the following url: https://lists.uni-koeln.de/pipermail/linux-fai/2007-August/005283.html 08:05 serieus Does anybody know the solution for this problem ? The author of this thread won't respond, and i'm out of possible options :-) 08:07 MT so what exactly is your problem? 08:08 serieus My problem is that i'm getting the same error, but can't figure out which parameter in the fai.conf is causing or will solve this 08:08 serieus The log files come up with an error that it can't find the package linux-image-486 08:08 MT please paste your fai.conf to paste.debian.net 08:08 serieus one moment 08:10 serieus http://paste.debian.net/39174 08:11 serieus It's pretty standard 08:22 MT which architecture are you using? 08:23 serieus amd64 08:23 serieus On both, the install server and the client 08:30 MT hmm, then probably there is no such package as linux-image-486 08:31 MT (just a blind guess) 08:31 sep is this http://faiwiki.informatik.uni-koeln.de/index.php/Setup_harddisks_2 about the best docs ill find for setup_harddisk2 ? MT (~MT@dove.informatik.tu-muenchen.de) left irc: Remote host closed the connection MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. 08:34 MT sep, yes, actually it is 08:34 MT and probably this channel :-) 08:35 sep MT the examples in the svn helped a lot :) Mrfai (~lange@suenner.informatik.uni-koeln.de) joined #fai. Bokeh (~blaat@berchem.lorentz.leidenuniv.nl) joined #fai. 08:45 Bokeh morning 08:46 Bokeh just in case someone's awake at this early hour, but i'm having the following problem with fai: when i try to install ubuntu gutsy, the /var/lib/fai/config/class nfs mount won't work. 08:46 Bokeh i already installed nfs-common in the nfsroot, solving the first problem 08:46 Bokeh however, now nfsmount requires me to either mount with -o nolock or start rpc.statd 08:46 Mrfai which FAI version are you using? 08:47 Bokeh 1 sec 08:47 Bokeh 3.2.1 08:47 Bokeh is there a simple solution, is it a known bug or will i have to hack the -o nolock option into the nfsroot's init.d 08:49 Mrfai Have a look at http://faiwiki.informatik.uni-koeln.de/index.php/UbuntuTeam 08:49 Mrfai There's a ubuntu version of 3.2.1 08:49 Mrfai a gutsy ppa package 08:50 Bokeh cool 08:50 Bokeh let me see how that works out 08:51 Bokeh hmm..... that's the development tree though, do you know anything about how stable it is? 08:52 Mrfai it's the stable version for gutsy 09:00 stockholm MT: sollte es möglich sein auch lvm volumen als unveränderlich zu markieren? 09:02 MT meinst mit preserve? 09:12 Bokeh cheers 09:13 Bokeh the packages from launchpad worked 09:41 stockholm MT: ja, preserve 09:41 MT da sollte das dann halt einfach nur "preserve" heißen 09:42 MT also auch dort statt der size 09:42 stockholm MT: ich frage weil das auswirkungen auf die darunterliegenden platten hat 09:43 MT warum das? 09:43 stockholm MT: ein "preserve" auf ein raid oder lvm würde auch "preserv" für alle beteiligten partitionen und ev. platten bedeuten. 09:44 MT hmm, ja 09:44 stockholm ansonsten finde ich dein teil dass du da gebaut hast ziemlich beeindruckend MT (~MT@dove.informatik.tu-muenchen.de) left irc: Read error: Connection reset by peer 09:44 stockholm schade... :-) MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. 09:49 sep umm when i use vg vg_system sda2 in order to make a lvm it fail, but vg system sda2 works. is it prohibited to use vg_ in a volumegroup name ? vg_system is the default vg name in a debian-edu installation 09:51 stockholm MT: wie kann man denn lösen dass die bootreihenfolge der devices sich ändert? 09:51 sep also i use disk_config disk1 in order to partition the first disk. is there any way i can omit using sda2 to specify the pv's hda2 | sda2 | c0d0p2 09:51 stockholm also sda sdb sdc sdd das nächste mal nicht in dieser reihenfolge erkannt werden muss 09:57 sep does anyone know where Michael Tautschnig prefers comments or bugs on setup_harddrive2 ? 09:58 stockholm sep: you work on that one too? 09:58 stockholm sep: me too 09:58 sep stockholm, i only try to use it 09:58 stockholm sep: and how are you doing? 09:58 stockholm what kind of issue do you see? 09:58 sep i think i notice a oddity Action: stockholm noticed several :-) 10:01 MT sep, comments either via ICQ or probably better via the mailing list 10:01 MT sorry, meant IRC instead of ICQ meandtheshell (~markus@85.127.118.147) joined #fai. 10:05 stockholm MT: weisst du wie man disk-ids in das system reinbekommen könnte om eine reproduzierbare bootreihenfolge zu garantieren? 10:05 MT in das - welches - system? 10:06 MT später, booten? 10:06 MT oder in setup_harddisks? 10:07 sep stockholm, that i can not start the volumegroup name with vg_[name] 10:07 MT oh, does that confuse the parser? 10:08 sep i tried vg_system, and it said in the fai.log Volume group _system allready exsist. 10:08 stockholm MT: setup Hardsist muss das ja ermöglichen und einrichten! 10:08 sep using only system or any other name that dont start with vg is ok 10:09 MT also das sollte es schon können 10:09 oz_ stockholm: das muss man noch mit nem hook machen 10:09 oz_ geht derzeit so noch nicht 10:09 MT im neuen schon 10:09 oz_ o gott, echt? 10:09 MT mach mal disk_config disk1 fstabkey:uuid 10:09 oz_ dann muss ich ja meine config ändern! ;) 10:09 MT also setup_harddisks_2 kann es... 10:10 MT nur die grub-config dürfte noch ein Problem sein 10:10 oz_ hook. 10:10 MT da muss man AFAIK auch root=UUID=... haben 10:10 oz_ hook und initrd 10:10 MT ich hab da einfach ein preinst bei menu.lst dabei 10:11 MT aber die fstab sollte setup_harddisks_2 schon richtig mit UUIDs erzeugen 10:15 sep also http://faiwiki.informatik.uni-koeln.de/index.php/Setup_harddisks_2 might mention that you need dm_mod personally i added a modprobe in partition.<CLASS> but i'm sure there are more corerct ways 10:16 sep stockholm, other then that it seams to work for me atm 10:16 MT sep, could you change line 451 in shdd2-parser 10:17 MT to volume: /^vg\s+/ name devices 10:17 MT this should do the trick 10:17 sep MT sure, let me try 10:17 MT in fact, it would be cool if you could test whether this then works with vg_system 10:18 sep test is in progress 10:18 MT thanks, sep 10:18 MT I also updated the wiki page 10:19 MT added the modprobe dm_mod to the hook 10:19 sep very nice 10:21 sep from my previous test i notice that my lvm swap is correcty craeted as a lv swap , but no fstab entry is written. ? i use vg_system-swap - 512 swap - 10:21 MT hmm, probably a bug :-) 10:22 oz_ a bug! a bug! :) ErKa (keryell@m5.wifi.enstb.org) left irc: Ping timeout: 480 seconds 10:23 sep MT, test failed i used vg vg_system sda2 10:23 sep (CMD) vgdisplay -s 1> /tmp/zGDqmuYLjB 2> /tmp/f6mJjzOwY4 10:23 sep Unexpected vgdisplay output "system" 11.76 GB [10.93 GB used / 848.00 MB free] 10:26 MT ok, well :-) 10:26 MT it seems to accept your name vg_system, which is the good news 10:26 MT but the parser of vgdisplay output is broken 10:27 MT let me see whether I can fix that 10:27 MT mom ... 10:28 MT ok, sep, about the swap <-> fstab problem 10:32 MT its a bug in sshd2-fstab 10:32 MT it silently skips all partitions without a mountpoint 10:32 MT thus including swaps 10:32 MT I won't be able to fix this right now, I'll look into it this evening 10:33 MT so, let's go for the vgdisplay thing 10:34 sep sure 10:35 MT ok, its about the MB/GB 10:35 MT I was only expecting GB 10:35 MT not 848.00 MB 10:35 sep MT, what about TB ? 10:35 MT no idea whether lvm displays this 10:36 MT actually I had no idea whether vgdisplay was using GB only or MB, TB 10:37 MT ok, could you please change line 379 of shdd2-volumes 10:37 MT ( $line =~ /^\s*"(\S+)"\s+\d+\.\d+ [MGT]B\s+\[\d+\.\d+ [MGT]B\s+used \/ \d+\.\d+ [MGT]B\s+free\]$/ ) 10:38 MT hmm, shall I paste it to paste.debian.net? 10:39 MT I've committed the bugfixes to the SVN 10:40 sep MT is not not tricky to try to guess the output all the time ? vgdisplay -s --units m would make it more predictable 10:41 MT oh, cool 10:41 MT didn't know about that one 10:43 MT I'll do the respective changes this evening 10:48 sep MT, your test on line 379 is booting nowe 10:48 MT meaning, it did the trick? 10:48 sep MT, you can use whatever units you want thol 10:48 sep meanint i am booting the fai client now 10:48 MT for another install attempt? 10:49 sep yeah 10:50 MT stockholm, did you already try the fstabkey:uuid thing? 10:52 sep MT, it failed to add the partition into the vg vg_system becouse it was allready a member of vg system from my previous installation attempth 10:53 sep MT, then it failed to create lv's since vg vg_system does not exsists 10:53 MT hmm, could you manually remove that volume group? 10:54 sep yeah ill do that and try again 10:55 sep ill be back in 20 minutes, run to lunch before they remove the food. :) 10:56 MT I'll also be away for 40 minutes or so 10:57 MT sep, this is of course a bug, but manually removing it is the quick fix :-) 11:03 Bokeh ok, i'm a bit further now. i've installed the gutsy version of fai 2.3.1 from ppa.launchpad.net and the install procedure is working, until it gets to downloading the dkpg's with aptitude 11:03 Bokeh it just stops downloading at a random package 11:04 Bokeh we're running a local apt mirror here, it doesn't seem to have any performance problems. Anyone who could point me into the right direction or a logfile which would show more info? 11:06 oz_ Bokeh: what does software.log say? 11:06 oz_ there must be something in the logs 11:07 Bokeh the only warning software.log gives is about the fact that fai-client is untrusted (which is correct because it comes from another repository) and that it's ignoring it because Ignore-Trust-Violations is set 11:07 Bokeh the last line is Get:32 http://apt.lorentz.leidenuniv.nl gutsy/main librpcsecgss3 0.14-2ubuntu4 [35.4kB] 11:08 Bokeh the package name and number is usually different 11:08 stockholm MT: the fstabkey:uuid thing is a solution, but it would need to be automated in order to work 11:08 Bokeh dmesg is only giving me a message about unionfs (new lower inode mtime) 11:12 Bokeh status.log ends with instsoft.FAIBASE OK. in case that's interesting to know i-R2D2 (~deed@tor-irc.dnsbl.oftc.net) left irc: Remote host closed the connection MT (~MT@dove.informatik.tu-muenchen.de) left irc: Ping timeout: 480 seconds i-R2D2 (~deed@tor-irc.dnsbl.oftc.net) joined #fai. serieus (~serieus@office.nedstat.nl) left irc: Quit: My damn controlling terminal disappeared! MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. 11:37 Mrfai Bokeh: try using an official mirror. OR paste the whole fai.log to paste.debian.net 11:39 Bokeh yeah, was already working on seeing if it works with the regular ubuntu mirrors 11:39 Bokeh still weird though, never had problems with installing from the local repo ErKa (~keryell@keryell.pck.nerim.net) joined #fai. 11:56 Bokeh nope... still hanging while getting the dpkgs 11:56 Bokeh i'll put the log up in a second 11:59 Bokeh http://paste.debian.net/39195 12:00 Mrfai really strange. 12:01 Mrfai Log into your install client and use top and strace (maybe not installed into the nfsroot) to check what is going on. Maybe a partition is full? 12:02 Bokeh top isn't really showing anything, i'll put strace in the nfsroot now 12:02 Bokeh looks like aptitude is hanging for some reason 12:02 Bokeh it's still in pstree but not doing anything 12:03 Bokeh brb 12:29 Bokeh hmmm 12:29 Bokeh iostat isn't showing anything significant 12:30 Bokeh and the only thing that aptitude is doing according to strace is [pid 7302] futex(0xa4d67c, FUTEX_WAIT, 1, NULL <unfinished ...> 12:31 Mrfai Mmm, I never heard of this problem. Other people have success using ubuntu 12:31 Bokeh could be b0rked hardware, i'll give it a try on another machine later one 12:31 Bokeh on* 12:32 Bokeh yeah, i also didn't have any problems with installing dapper with fai 12:32 Bokeh though that was a while ago already on different machines with a different setup 12:33 Bokeh it seems all processes (install_packages, aptitude, the http processes spawned by aptitude are all in a wait state 12:34 Bokeh could be a messed up harddisk or ram, though it's weird dmesg doesn't show anything interesting Action: h01ger is hit by bashism in his own script :) ErKa (~keryell@keryell.pck.nerim.net) left irc: Ping timeout: 480 seconds 12:57 stockholm MT: gestern hast du gesagt ich solle auch die zeilen 95 und 96 in shdd2-pasrer ändern wenn ich die disk1.2 notation hinzufügen würde. 12:57 stockholm MT: aber in den zeilen ist doch garnix was damit zu tun hat, oder sehen ich das falsch? 12:58 stockholm Mrfai: http://www.zweinull.cc/spotify-neuer-musikdienst-stellt-alles-andere-in-den-schatten/ 12:58 MT also da sollte meines Wissens nach dieselbe regexp stehen wie im eigentlichen Code 12:58 MT ach, stockholm, was hast Du eigentlich vorhin gemeint mit 12:59 stockholm MT: der witz ist dass die regexp glaube ich garnicht geändert werden muss. 12:59 MT "it would need to be automated in order to work"? ErKa (keryell@keryell.pck.nerim.net) joined #fai. 12:59 MT hmm, stimmt 13:00 MT das nimmt ja alles 13:00 stockholm MT: ich meinte dass shdd das uuid labeling übernehem müsse und die richtigen werte in grubs boot.menu und fstab eintragen müsste 13:00 MT also in fstab sollte er das ja machen 13:00 stockholm und grub muss ja auch davon wissen, damit der kernel die devices richtig ordnen kann 13:00 MT ok, dann muss man also nur im parser das disk\d+\.\d+ besonders behandeln 13:01 MT ja, grub ist schon noch ein Problem 13:01 stockholm genau 13:01 stockholm das habe ich gerade gemacht 13:01 stockholm und checke es mal eben ein 13:01 MT ich schreib das mit grub mal auf die wiki-Seite 13:03 stockholm MT: schau mal ob das so passt in -parser 13:03 MT stockholm, vielleicht ist das mit dem grub gar nicht so kompliziert 13:03 stockholm ich teste es mal geschwind... 13:03 MT es gibt ja die Variable ROOT_PARTITION 13:04 stockholm und die muss man einfach exortieren? 13:04 MT hmm, mal kurz die Skripte ansehen 13:06 MT ja, ich glaub schon dass das funktionieren sollte 13:06 MT also man muss einfach nur in ROOT_PARTITION den richtigen Wert reinschreiben 13:06 Mrfai stockholm: danke. Interessanter Beitrag. Und meine FReundin ist immer noch begeistert von spotify. 13:07 MT stockholm, Zeile 124 in shdd2-fstab 13:09 stockholm MT: was sollte da rein, dann? 13:09 MT da muss man ROOT_PARTITION mit einem korrekten Wert belegen 13:09 stockholm MT: ja, das hatte ich verstanden. 13:09 MT im moment wird da halt /dev/sda oder so gebaut 13:10 stockholm ah 13:10 stockholm dann müsste man also die infrastruktur zum uuid-holen besorgen? 13:10 MT ich glaube, so ungefähr den Code ab Zeile 93 kopieren 13:10 MT ist eigentlich alles schon da 13:10 stockholm ohhh 13:11 sep MT, seams to work when the disk is clean 13:11 MT ok, thanks 13:11 MT just one thing 13:11 MT you had a volume group named system on disk and now have a volume group "vg_system", don't you? 13:12 MT stockholm, es ist noch viel einfacher 13:12 stockholm noch einfacher??? :-) 13:12 MT $FAI::disk_var{ "ROOT_PARTITION" } = $fstab_line[0] 13:12 sep well now it's called vg_system2, since i have been testing a few times. and i wanted to make sure the old one got erased and not reused, 13:13 MT in Zeile 124 statt dem was dort steht 13:13 MT das if in Zeile 125 muss bleiben 13:13 MT weil in fstab_line steht ja eh schon das richtige drin, muss man den device namen ja nicht nochmal bauen... 13:13 stockholm j 13:14 stockholm ja 13:14 stockholm ok, cool. 13:14 stockholm das kommitte ich dann auch noch eben. :-) 13:14 MT sep, ok, I just wonder why it failed to add a logical volume to vg_system which had previously existed only in system 13:14 stockholm dann sieht das im svn log danach aus als hätte ich richtig was drauf. :-) 13:15 sep MT, becouse when the md_mod is loaded the vg and lv's from the disks are 'valid' 13:15 sep so the disk is in the vg, and the vg is in use by lv's 13:16 sep partman in d-i have the same problem 13:16 MT ah, hmm, ok 13:16 MT this will require some further debugging/testing then 13:17 MT stockholm, kannst Du dann auch testen, ob das mit den UUIDs so auch wirklich klappt? 13:18 MT stockholm, noch eine Bitte 13:18 MT kannst Du TODO statt ToDo schreiben? 13:18 MT macht das grep einfacher :-) 13:18 oz_ MT: grep -i 13:19 MT sicher, aber wenn man das vergisst... 13:28 sep MT, you can check out dm_wipe_lvm() in partman http://svn.debian.org/wsvn/d-i/trunk/packages/partman/partman-auto/auto-shared.sh?op=file&rev=0&sc=0 13:29 stockholm MT: ja, klar 13:29 stockholm MT: jetzt setzt er ROOT_PARTITION=/dev/my_pv/_root 13:29 stockholm was ist mit boot? 13:29 stockholm er soll ja booten von /boot 13:29 stockholm und das ist auch das was grub liest 13:30 MT das sollte AFAIK kein Problem sein 13:30 MT also in /boot liegt grub selbst 13:30 MT bis dahin muss das BIOS kommen 13:30 MT und dann muss grub nur die richtige root=bla Kommandozeile an den kernel übergeben 13:31 MT und da muss dann eben root=/dev/my_pv/_root stehen 13:31 MT das sollte dann eigentlich schon klappen 13:31 MT falls die grub-tools das richtig reinschreiben 13:31 MT -> mal die erzeugte menu.lst prüfen 13:32 stockholm ja, ich boote jetzt gerade das installierte system 13:34 stockholm grup error 17 13:34 MT grub error? 13:34 stockholm ja 13:34 MT was findet er denn nicht? 13:34 MT ich mein, das hat dann ja mit root=bla noch nix zu tun ... 13:36 Mrfai vielleicht ist die device.map nicht OK? 13:37 Mrfai 17 : Cannot mount selected partition 13:37 Mrfai This error is returned if the partition requested exists, but the filesystem type cannot be recognized by GRUB. 13:38 stockholm 17 : Cannot mount selected partition This error is returned if the partition requested exists, but the filesystem type cannot be recognized by GRUB. 13:39 MT stockholm, wenn Du die grub config beim booten editierst, findet er da den kernel und so? 13:39 MT (einfach mal tab completion probieren...) 13:39 sep MT, any way one can avoid hardcoding device names in the line vg vg_system sda2 ? since i use the line disk_config disk1, 13:39 stockholm gerade macht grub garnix mehr. ich probiere gleich 13:40 MT sep, stockholm comitted a patch for this just a few minutes ago 13:40 MT if you grab shdd2-parser from SVN again, you can use 13:40 sep MT stockholm great 13:40 MT disk1.2 13:41 MT for partition 2 on sda, if sda is disk1 13:44 stockholm MT: grub kommt garnicht richtig hoch, error 17 kommt "sofort" bevor ich was manuell machen kann 13:44 stockholm MT: was macht dein patch? 13:45 MT Du meinst, die Sache mit ROOT_PARTITION? 13:45 stockholm ja 13:45 Mrfai ganz dumme frage. Kann grub denn lvm? 13:45 MT sollte eben eigentlich nur statt /dev/sda1 oder so evtl. eine UUID oder ein LABEL nehmen 13:45 stockholm Mrfai: das muss er nicht 13:46 stockholm Mrfai: erst bootet man /boot wo der kernel und die initrd drauf ist 13:46 stockholm Mrfai: aber ich frage mich warum die root partition dann ins spiel kommt 13:46 stockholm MT: zuerst muss doch erstmal /boot auf der richtigen hd gefunden werden. 13:47 stockholm wie funktioniert das denn wenn die festplatten in beliebiger reihenfolge auftauchen können? 13:52 MT stockholm, der letzte Patch sollte eigentlich nichts ändern bei Dir 13:52 MT weil die root partition ja sowieso im LVM liegt 13:52 MT hat es denn vorher funktioniert? 13:52 stockholm ah 13:53 stockholm MT: das letzte mal war das error 17 13:53 stockholm jetzt boote ich wieder 13:55 stockholm wieder error 17 13:55 stockholm und hier ist die fstab 13:55 stockholm # <file sys> <mount point> <type> <options> <dump> <pass> 13:55 stockholm /dev/sde1 none swap sw 0 2 13:55 stockholm /dev/sdb1 none swap sw 0 2 13:55 stockholm /dev/my_pv/_root / ext3 rw,errors=remount-ro 0 1 13:55 stockholm /dev/sdc1 none swap sw 0 2 13:55 Mrfai stockholm: stimmt /boot ist wohl nicht im LVM nur /. Verstehe. 13:55 stockholm /dev/sdb2 /boot ext3 rw,notail 0 2 13:55 stockholm /dev/my_pv/_var /var xfs rw,notail,noatime 0 2 13:55 stockholm /dev/sdd1 none swap sw 0 2 13:55 MT ich meine, hat es schon irgendwann mal mit setup_harddisks_2 geklappt? 13:55 stockholm MT: nee, vorher hatten wir ja andere probleme so dass es nie überhaupt gebootet hat 13:56 stockholm errr 13:56 stockholm ich meine: vorher ist shdd2 ja nie durchgelaufen 13:56 MT ja, verstehe 13:56 stockholm jetzt macht es das. 13:56 MT was steht denn in shell.log oder so drin? 13:56 stockholm wie markiert man die bootpartition denn? 13:57 MT irgendwelche errors, wenn er GRUB/10-setup ausführt? 13:57 stockholm MT: ich habe hier die fai.log, die ist fast nur shdd2 log :-) 13:57 stockholm ich schau mal 13:57 MT shell.log brauchst aber ... 13:57 MT hmm, boot partition 13:57 stockholm ja 13:58 stockholm Executing shell: GRUB/10-setup 13:58 stockholm GRUB/10-setup FAILED with exit code 1. 13:58 MT aber in shell.log sollte mehr drin stehen 13:58 stockholm ok 13:59 stockholm floood!!! 13:59 MT hmm, vielleicht brauchst du sowas 13:59 stockholm Report bugs to <bug-grub@gnu.org>. 13:59 stockholm Can't open /target/boot/grub/device.map 13:59 stockholm Searching for GRUB installation directory ... found: /boot/grub 13:59 stockholm Searching for default file ... Generating /boot/grub/default file and setting the default boot entry to 0 13:59 stockholm Searching for GRUB installation directory ... found: /boot/grub 13:59 stockholm Testing for an existing GRUB menu.lst file ... found: /boot/grub/menu.lst 13:59 stockholm Searching for splash image ... none found, skipping ... 13:59 stockholm Found kernel: /vmlinuz-2.6.18-5-amd64 13:59 stockholm Found kernel: /memtest86+.bin 13:59 stockholm Updating /boot/grub/menu.lst ... done 13:59 stockholm Grub installed on on 13:59 stockholm ERROR: postinst returned code 1 13:59 stockholm GRUB/10-setup FAILED with exit code 1. 13:59 MT disk_config disk1 fstabkey:uuid bootable:2 14:00 MT er braucht die Variablen BOOT_PARTITION und BOOT_DEVICE 14:00 sep i see this error too <stockholm> GRUB/10-setup FAILED with exit code 1 14:01 MT ok, switching to English 14:01 stockholm MT: ok, i test that 14:01 MT BOOT_PARTITION and BOOT_DEVICE seem not to be set 14:01 MT could either of you check that? 14:01 stockholm MT: i am reinstalling with that setting now 14:02 MT with bootable:2 ? 14:02 stockholm yes 14:02 MT I'd be interested in what happens if you don't have that 14:02 MT it should have properly set the values itself 14:02 MT I hope ... 14:03 MT paste the contents of disk_var.sh 14:03 MT without bootable:2 14:04 sep MT fai.log and disk_var.sh state that BOOT_DEVICE and BOOT_PARTITION is set 14:04 MT sep, do you have a separate /boot partition? 14:04 sep MT i have bootable:1 btw and a primary /boot 14:04 sep yes 14:05 MT but, did your grub report error 17 as well? 14:05 MT when booting the client for the first time? 14:05 stockholm again, grub exited with error 1 14:05 MT what about the log? 14:05 MT does it still say 14:06 MT Grub installed on on 14:06 MT or are there now sensible values? 14:06 MT after the on's ErKa (keryell@keryell.pck.nerim.net) left irc: Ping timeout: 480 seconds 14:07 stockholm http://paste.debian.net/39209 14:07 stockholm that is the grub part from shell.log 14:07 stockholm it looks quite unhappy 14:08 sep Grub installed on on 14:08 sep ERROR: postinst returned code 1 14:08 sep GRUB/10-setup FAILED with exit code 1. 14:08 oz_ Can't open /target/boot/grub/device.map 14:08 oz_ Thomas pointed in this direction already Action: Mrfai filed a bug report that the postinst of grub/menu.lst should check if those variables are defined 14:08 MT just a sec ... telephone ... 14:09 sep http://paste.debian.net/39211 14:09 sep it's identical 14:09 stockholm the uuid is not used in BOOT_PARTITION: 14:09 stockholm # <file sys> <mount point> <type> <options> <dump> <pass> 14:09 stockholm /dev/sde1 none swap sw 0 2 14:09 stockholm UUID=e6163ecf-98ff-41b8-ace6-a3ed7bd59ce3 none swap sw 0 2 14:09 stockholm /dev/my_pv/_root / ext3 rw,errors=remount-ro 0 1 14:09 stockholm /dev/sdc1 none swap sw 0 2 14:09 stockholm UUID=c3455830-36b4-45b2-aa7a-380c4c0ca295 /boot ext3 rw,notail 0 2 14:09 stockholm /dev/my_pv/_var /var xfs rw,notail,noatime 0 2 14:09 stockholm /dev/sdd1 none swap sw 0 2 14:09 stockholm SWAPLIST=/dev/sde1 /dev/sdb1 /dev/sdc1 /dev/sdd1 14:09 stockholm BOOT_DEVICE=/dev/sdb 14:09 stockholm ROOT_PARTITION=/dev/my_pv/_root 14:09 stockholm BOOT_PARTITION=/dev/sdb2 14:09 stockholm but it is used in the fstab 14:09 stockholm that is something already 14:10 stockholm Mrfai: can we work around those missing definitions? 14:10 stockholm can we define them on our own? 14:10 sep #groot= is blank in the menu.lst 14:14 Mrfai setup_harddisks2 should write those variables to diskvar.sh. Then they should be defined during postinst 14:15 MT I think the values written to disk_var.sh are ok 14:15 stockholm Mrfai: they are there: 14:15 stockholm SWAPLIST=/dev/sde1 /dev/sdb1 /dev/sdc1 /dev/sdd1 14:15 stockholm BOOT_DEVICE=/dev/sdb 14:15 stockholm ROOT_PARTITION=/dev/my_pv/_root 14:15 stockholm BOOT_PARTITION=/dev/sdb2 14:16 stockholm in diskvar_sh 14:16 MT which is pretty fine IMHO 14:16 stockholm jupp 14:16 stockholm except that we need them in UUID form 14:17 MT no, you don't 14:17 stockholm why not? 14:17 MT because nothing changes in that sense until you run grub-install 14:17 stockholm ah 14:18 stockholm and grub-install checks the fstab? 14:18 MT no, it is told 14:18 MT grub-install --no-floppy --root-directory=$target $BOOT_DEVICE 14:18 MT $BOOT_DEVICE is /dev/sdb 14:19 stockholm and that can change its name, but grub maps that to something stable internally? 14:19 Mrfai please add set -x into postinst, wo we can see if the variables are really defined in the postinst script 14:19 Mrfai s/wo/so/ 14:19 stockholm which postinst script? 14:19 MT yes, maybe even add some debugging output for now 14:19 stockholm the grub one? 14:19 MT files/boot/grub/menu.lst/postinst 14:19 MT in your config space 14:19 stockholm ah 14:19 Mrfai grub/menu.lst/postinst 14:20 sep in order to install grub manually i had to chroot /tartget ; cp /proc/mounts /etc/mtab ; grub-install /dev/sda, edit meny.lst with the #groot line and update-grub 14:20 sep my ride is here now so i have to leave. 14:21 stockholm MT, Mrfai do you have some suggestions regarding the additional debug output? 14:21 MT echo X${BOOT_DEVICE}X 14:21 MT something of that kind 14:22 MT and set -x would be pretty cool 14:22 Mrfai yep, that's perfect for debugging 14:23 MT stockholm, is /target/var/log/fai/disk_var.sh really in place? 14:23 stockholm i can check next time around 14:23 MT stockholm, where from did you read the variable settings you posted above 14:24 stockholm MT: from the install server 14:24 Mrfai /target/var/log/fai/disk_var.sh is only needed for softupdates 14:24 MT no, it's read in exactly this postinst script 14:25 MT ah, no, I get it 14:25 MT should these variables be exported? 14:25 stockholm # during softupdate use this file 14:25 stockholm [ -r $target/var/log/fai/disk_var.sh ] && . $target/var/log/fai/disk_var.sh 14:25 Mrfai the old (or current) setup_harddisks says this: 14:25 Mrfai Write FAI variables to file /var/log/fai/current/disk_var.sh 14:25 stockholm MT: dont you export them currently? 14:26 Mrfai the FAI automatically sources this file, so the variables are defined 14:26 Mrfai s/the/then/ 14:26 stockholm ok 14:29 stockholm echo "BOOT_DEVICE:${BOOT_DEVICE}" 14:29 stockholm + echo BOOT_DEVICE: 14:29 stockholm BOOT_DEVICE: 14:29 stockholm in shell.log 14:29 stockholm also sind die variablen nicht definiert. 14:30 stockholm MT: aber das sollte ja einfach zu fixen sein 14:30 MT existiert /tmp/fai/disk_var.sh 14:30 MT ? 14:30 stockholm ich schau gerad 14:30 stockholm fai-client:~# cat /tmp/fai/disk_var.sh 14:30 stockholm SWAPLIST=/dev/sde1 /dev/sdb1 /dev/sdc1 /dev/sdd1 14:30 stockholm BOOT_DEVICE=/dev/sdb 14:30 stockholm ROOT_PARTITION=/dev/my_pv/_root 14:30 stockholm BOOT_PARTITION=/dev/sdb2 14:31 stockholm ja 14:31 MT BINGO! 14:31 stockholm fai-client:~# ls -al /var/log/fai/current/disk_var.sh 14:31 stockholm -rw-r--r-- 1 root root 127 Oct 9 16:32 /var/log/fai/current/disk_var.sh 14:31 stockholm was bingo? 14:31 MT 112 # now define variable for root and boot partition and boot device 14:31 MT 113 . $LOGDIR/disk_var.sh 14:32 MT das passiert natürlich im task partition 14:32 MT der ja geskippt wird 14:32 stockholm ah, und das wird .... genau 14:32 MT Mrfai, hast Du eine gute Idee für eine saubere Lösung? 14:32 stockholm dann müssen wir das in den hook mit reintun, gelle? 14:32 MT naja, aber der hook wird nicht gesourct 14:32 MT Du musst dann den hook umbenennen 14:32 MT partition.BLA.source 14:32 MT dann klappt's 14:32 Mrfai der hook wird gesourced wenn er mit .source endet 14:33 MT ich ändere mal die wiki page 14:35 MT ok, wiki page ist aktualisiert meandtheshell (~markus@85.127.118.147) left irc: Quit: Leaving. 14:44 MT stockholm, konntest Du schon einen neuen Test machen? 14:45 stockholm ja, gerade eben: 14:45 stockholm immer noch grub error ... 14:45 stockholm grub-install --no-floppy --root-directory=$target $BOOT_DEVICE 14:45 stockholm + grub-install --no-floppy --root-directory=/target /dev/sdb 14:45 stockholm You shouldn't call /sbin/grub-install. Please call /usr/sbin/grub-install instead! 14:45 stockholm Probing devices to guess BIOS drives. This may take a long time. 14:45 stockholm /dev/mapper/my_pv-_root does not have any corresponding BIOS drive. 14:45 stockholm error=1 14:45 stockholm ++ error=1 14:46 stockholm jetzt gibt es also die umgebungsvariablen 14:47 stockholm aber bood device ist falsch gesetzt. 14:47 MT ? 14:47 stockholm nee 14:47 stockholm stimmt nict 14:47 stockholm er weiss nicht was /dev/mapper/my_pv-_root 14:47 stockholm ist 14:47 MT kannst Du mal nochmal den ganzen Teil aus shell.log -> paste.debian.net? 14:47 stockholm ja 14:47 MT thx 14:48 stockholm http://paste.debian.net/39216 14:49 MT blöde Frage: macht der error wirklich was aus? 14:50 MT oder tut's einfach 14:50 MT also bootet der client? 14:50 stockholm nee, error 17 14:50 stockholm wieder 14:52 Mrfai ok ich haetter gerne noch "echo $1" in dem postinst 14:53 Mrfai das ist wird als $ENV{$1} in dem perl einzeiler genutzt 14:54 Mrfai ist den der filetype fuer die /boot partition richtig gesetzt? 14:54 Mrfai was sagt fdisk oder file -s /dev/sdb... 14:55 stockholm das hier ist falsch: 14:55 stockholm + grub-install --no-floppy --root-directory=/target /dev/sdb 14:55 MT ? 14:55 MT warum falsch? 14:55 stockholm das muss von /target/boot booten, nicht von /target 14:56 MT aber das root-directory, sollte das nicht der mountpoint von / sein? 14:56 MT sodass er dann /boot finden kann? 14:56 MT was ich im Netz gefunden habe: 14:56 MT ist /boot wirklich gemountet? 14:56 MT also während grub-install aufgerufen wird 14:57 stockholm ah 14:57 stockholm gute frage. 14:57 Mrfai IMO ist /target richtig. 14:57 MT ich denk auch 14:57 Mrfai stockholm: paste mal das ganze fai.log 14:57 MT und bitte mal checken, ob alles gemountet ist ... 14:58 stockholm ich habe jetzt $1 und mount noch in postinst dazugeschrieben und boote nochmal, dann paste ich fai.log und shell.log nach paste.debian.net 14:59 MT super! 14:59 Mrfai echo $1 hoffentlich ;-) 15:08 stockholm http://paste.debian.net/39218 15:08 stockholm http://paste.debian.net/39219 15:10 MT also wenn ich's nicht übersehen habe, dann ist /boot nicht gemountet 15:10 stockholm ja, sehe ich auch nicht 15:10 stockholm das ist ja *schlecht* 15:10 MT # 15:10 MT Calling task_mountdisks 15:10 MT # 15:10 MT Mounting /dev/my_pv/_root to /target/ 15:10 MT # 15:10 MT Mounting /dev/my_pv/_var to /target/var 15:11 stockholm warum kein /boot? 15:11 Mrfai Physical volume '/dev/sdd2' is already in volume group 'my_pv' 15:11 Mrfai ist das OK? 15:11 Mrfai Und hier scheint ein syntax error drin zu sein: 15:11 Mrfai /var/log/fai/current/disk_var.sh: line 1: /dev/sda1: Permission denied 15:12 MT hmm, da brauchen wir wohl " 15:12 MT SWAPLIST hat vermutlich keine quotes 15:12 MT kannst Du mal die generierte fstab pasten? 15:13 stockholm ja 15:14 Mrfai und das fai.log war auch nicht ganz komplett. Nur als info. Weil am Ende von fai.log sieht man was alles gemountet ist. 15:15 Mrfai echo '$1' bringt nichts. $1 muss in "" stehen sonst wird nicht die variabel ausgegeben 15:15 MT ok, das mit SWAPLIST hab ich gefixt 15:15 MT r4612 15:15 stockholm # /etc/fstab: static file system information. 15:15 stockholm # 15:15 stockholm # <file sys> <mount point> <type> <options> <dump> <pass> 15:15 stockholm /dev/sdb1 none swap sw 0 2 15:15 stockholm UUID=621537db-f54f-4c57-bba7-c6ed2afb4176 none swap sw 0 2 15:15 Mrfai sorry vergiss mein $1 kommentar 15:15 stockholm /dev/my_pv/_root / ext3 rw,errors=remount-ro 0 1 15:15 stockholm /dev/sdc1 none swap sw 0 2 15:15 stockholm UUID=58e982cf-2632-45e1-9402-ffbff4c0183c /boot ext3 rw,notail 0 2 15:15 stockholm /dev/my_pv/_var /var xfs rw,noatime 0 2 15:15 stockholm /dev/sdd1 none swap sw 0 2 15:15 stockholm /dev/fd0 /floppy auto users,noauto 0 0 15:15 stockholm /dev/hda /media/cdrom0 udf,iso9660 ro,user,noauto 0 0 15:15 stockholm /dev/sr0 /media/cdrom1 udf,iso9660 ro,user,noauto 0 0 15:15 MT dacht ich mir doch 15:15 MT mount2dir kann vermutlich keine UUIDs 15:16 Mrfai ne woher auch ;-) 15:16 MT kann mount direkt UUIDs? 15:17 MT also kann man mount <UUID> /bla machen? 15:17 Mrfai stockholm: du hast vier swap partitionen? zwei nur mit UUID angegeben und zwei mit /dev/sd... ? STimmt das? 15:18 stockholm Mrfai: ja, jede festplatte hat eine swap partition 15:18 MT ja, mount kann UUIDs 15:18 Mrfai MT: mount -U kann das 15:18 MT also bei mir hat's jetzt auch ohne geklappt 15:18 MT also ohne -U 15:19 stockholm Mrfai: aber ich weiss nicht warum die eine in UUID und die andren drei in /dev/bla2 angegeben sind. 15:20 MT kannst Deine disk_config mal pasten? 15:20 stockholm die ist doch in fai.log mit drin 15:20 MT und Du wirst wohl vorerst mal $NFSROOT/usr/lib/fai/mount2dir hacken müssen 15:20 MT sorry, natürlich 15:21 MT fstabkey:uuid musst Du bei jeder disk angeben 15:21 stockholm MT: und mount2dir muss uuids verdauen können? wie sollen die denn *wirklich* aussehen? 15:21 MT das sollte nicht weiter schwierig sein 15:21 MT Zeile 76 von mount2dir 15:22 MT /dev/*|LABEL=*|UUID=*) 15:22 MT falls fsck und mount mit labels und UUIDs umgehen können 15:22 MT sonst muss man wohl mehr machen 15:23 Mrfai schau mal in /etc/init.d/checkroot.sh 15:24 stockholm ja 15:25 Mrfai patch fuer mount2dir bitte per email an mich. Ich bau ihn dann ein 15:26 MT ok, falls es nicht mit der obigen Änderung klappt muss man findfs bemühen 15:26 MT root@joey:~# findfs UUID=f23dffd4-1e71-4a36-976f-9fffa8b22695 15:26 MT /dev/sda1 15:26 stockholm testing... 15:32 stockholm fai-client:~# mount | grep boot 15:32 stockholm fai-client:~# 15:32 stockholm hm 15:32 stockholm immer noch nicht 15:32 stockholm mount | grep target 15:32 stockholm /dev/my_pv/_root on /target type ext3 (rw,noatime,data=ordered) 15:32 stockholm /dev/my_pv/_var on /target/var type xfs (rw,noatime) 15:32 stockholm proc on /target/proc type proc (rw) 15:32 stockholm sysfs on /target/sys type sysfs (rw) 15:32 stockholm udev on /target/dev type tmpfs (rw) 15:32 stockholm fai-client:~# 15:33 MT steht in fai.log irgendwas 15:33 MT irgendwelche Fehlermeldungen? 15:33 MT bzw. mount2dir nochmal manuell ausführen mfl (~mfl@adonis.fnb.maschinenbau.tu-darmstadt.de) joined #fai. 15:34 MT /usr/lib/fai/mount2dir /target /tmp/fai/fstab 15:34 MT das wär vielleicht eine gute Sache 15:34 stockholm http://paste.debian.net/39221 15:34 Mrfai und vielleicht sh -x mount2dir ..... 15:34 stockholm http://paste.debian.net/39222 15:36 MT ähm, kannst Du mount /dev/sda2 /target/boot machen? 15:36 MT # 15:36 MT Mounting UUID=51fe2312-d781-48c1-9287-ad41213c4357 to /target/boot 15:36 MT # 15:36 MT mount: wrong fs type, bad option, bad superblock on /dev/sda2, 15:36 MT # 15:36 MT missing codepage or other error 15:37 MT # 15:37 MT In some cases useful info is found in syslog - try 15:37 MT # 15:37 MT dmesg | tail or so 15:37 MT offensichtlich kann er also die UUID mappen 15:37 stockholm http://paste.debian.net/39223 15:37 stockholm da ist das nochmal der set -xv output von mount2dir 15:38 Mrfai UUID= muss wahrscheinlich weg beim mount aufruf 15:38 stockholm ja 15:39 MT sicher? 15:39 MT ich mein, probier einfach mal von Hand /boot zu mounten 15:39 MT mount -t ext3 -o noatime -o rw,notail UUID=51fe2312-d781-48c1-9287-ad41213c4357 /target/boot 15:39 MT solange anpassen, bis es klappt 15:40 stockholm mount /dev/sda2 /target/boot 15:40 stockholm das geht gut 15:41 stockholm fai-client:~# mount -t ext3 -o noatime -o rw,notail UUID=51fe2312-d781-48c1-9287-ad41213c4357 /target/boot 15:41 stockholm mount: wrong fs type, bad option, bad superblock on /dev/sda2, missing codepage or other error In some cases useful info is found in syslog - try dmesg | tail or so 15:41 MT mit -U, wie Mrfai gemeint hat, klappt das? 15:43 Mrfai mount -t ext3 -U 1234-123234234-234-2-34-234 /mnt klappt bei mir 15:44 Mrfai UUID=....... klappt aber auch 15:44 stockholm mount -t ext3 -U 51fe2312-d781-48c1-9287-ad41213c4357 /target/boot 15:44 stockholm geht 15:44 MT ok, dann musst wohl mount2dir doch etwas aufwändiger patchen 15:44 Mrfai stockholm: probier man UUID= 15:44 stockholm die mountoptions waren das problem, nicht die uuid oder -U 15:45 Mrfai notail gibt es IMO nur bei reiser 15:45 stockholm fai-client:~# mount -t ext3 UUID=51fe2312-d781-48c1-9287-ad41213c4357 /target/boot 15:45 stockholm das geht 15:45 MT und -o noatime -o rw 15:45 MT ? 15:45 stockholm ja, das war der notail der hankt 15:45 Mrfai yep, das notail bei ext3 erzeugt den FEhler 15:46 MT aber das steht in Diner disk_config 15:46 MT :-) 15:46 Mrfai stockholm: schick mir bitte den mount2dir patch 15:46 stockholm MT: ja das geht auch mit den -o options 15:47 MT Mrfai, ich glaub es war einfach /dev/*|LABEL=*|UUID=*) 15:47 MT in Zeile76 Action: Mrfai ist gespannt ob der Rechenr das naechste mal bootet 15:47 MT stockholm, pleas confirm 15:48 stockholm mt, Mrfai ja, bisher ist es nur das /dev/*|LABEL=*|UUID=*) 15:48 stockholm notail entfernt 15:48 Mrfai ok 15:48 stockholm neu gebootet 15:49 stockholm wenn der jetzt gleich bootet verschwinde ich ganz schnell nach hause... dann kriege ich nämlich onch meinen zug :-) Action: Mrfai drueckt die Daumen. 15:57 stockholm yes, grub OK! 15:59 stockholm YESSSSS! 15:59 stockholm tschüss! 15:59 stockholm weiter heute nacht zuhause. Action: Mrfai gratuliert MT und stockholm 15:59 Mrfai gute Arbeit! Action: MT dankt stockholm ... Mrfai (~lange@suenner.informatik.uni-koeln.de) left irc: Quit: leaving Mrfai (~lange@kueppers.informatik.uni-koeln.de) joined #fai. sepski (~sep@217.17.211.51) joined #fai. Mrfai (~lange@kueppers.informatik.uni-koeln.de) left irc: Quit: leaving Mrfai (~lange@xdsl-87-78-55-146.netcologne.de) joined #fai. 16:51 sepski hmm after svn up i cant get partitioning to work again, i did revert the changes MT instructed back to svn version of the files. is it broken atm or ? 16:52 MT just a sec .. MT (~MT@dove.informatik.tu-muenchen.de) left irc: Remote host closed the connection MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. 17:22 MT sorry sepski .. 17:23 MT no, the current state _should_ be fine 17:23 MT in what sense is it broken? 17:25 sepski does not create partitions, but it might be related to some dd commands i put into the hook to overwrite lvm info before loading the dm module 17:26 MT are there any error messages? 17:27 sepski ill clean up my hook and try, ill post a log if it still fail 17:28 sepski yeah it was my ugly dd if=/dev/zero of=/dev/sda that did it in the hook. 17:28 sepski nevermind the noise 17:29 MT hmm, was it really that line that caused the error? 17:29 MT why did it? 17:29 MT seems strange to me... 17:29 MT did you check the wiki page in the last hour or so? 17:29 MT there was a severe change to the hook 17:30 MT it must get sourced 17:30 MT otherwise we can't get the disk_var.sh contents along 17:30 sepski no ill check it out 17:35 sepski #dd if=/dev/zero of=/dev/sda bs=1k count=1 17:35 sepski #hdparm -z /dev/sda 17:35 sepski commented these and it worked again 17:36 MT I'd be interested in the error that setup_harddisks_2 reported (and that caused it to fail) 17:36 MT or does that simply mean that there was no error reported? 17:43 sepski ill uncomment and post a log 17:43 sepski after i see if this install ends with a successfull grub. 17:45 MT sepski, I'll be offline for little less than an hour 17:45 MT will you be online later on? 17:45 sepski most likely 17:45 MT I'd really be interested in your logs 17:45 MT would be cool if you could tell me later on 17:45 sepski perhpas not after 21:00 but ill be online again from tomorro 0800 17:45 MT I'll ask then... 17:45 MT ok, see you soon 17:46 sepski :) ErKa (keryell@keryell.pck.nerim.net) joined #fai. MT (~MT@dove.informatik.tu-muenchen.de) left irc: Ping timeout: 480 seconds ErKa (keryell@keryell.pck.nerim.net) left irc: Ping timeout: 480 seconds 18:11 stockholm now i finished the Giga/Tera/Petabyte patch. MT (~MT@ppp-82-135-65-206.dynamic.mnet-online.de) joined #fai. jesper_ (~jesper@shrek.krogh.cc) joined #fai. 18:31 jesper_ Hi. Nick change: MT -> Guest1344 MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. 18:31 jesper_ I get messages from the root-on-NFS like this: 18:31 jesper_ http://pastebin.com/m37ccfd00 18:31 sepski http://paste.debian.net/39237 18:32 jesper_ "Nfs server not responding, still trying". 18:32 jesper_ Any debugging suggestions. It is not a persistent error.. it only show up "sometimes" 18:32 MT sepski, was it you who had an issue with the swaps not entering fstab? 18:33 MT thanks sepski, I'll look into this 18:33 stockholm MT: hilfe! ich habe den giga/tera/petabyte patch fertig, lasse perltidy drueberlaufen und habe jetzt einen riesen diff! 18:34 stockholm svn diff | wc 2576 11495 90622 18:34 stockholm der patch selbst ist nicht SOOO gross 18:34 MT hmm, ja, ich hab irgendwann manuell weiterformatiert 18:34 MT könntest Du folgendes machen 18:34 MT Deine Änderungen reverten 18:34 MT dann perl tidy 18:35 MT und dann commit 18:35 stockholm ich habe es nicht eingecheckt 18:35 stockholm ah, ja 18:35 stockholm stimmt 18:35 MT und dann nur den eigentlichen patch 18:35 MT macht es einfacher, falls es mal Probleme gibt 18:35 MT wenn im commit log drin steht, dass das nur perl tidy war 18:35 MT (und sonst wirklich nichts!) Guest1344 (~MT@ppp-82-135-65-206.dynamic.mnet-online.de) left irc: Ping timeout: 480 seconds 18:42 stockholm MT: ja, so habe ich es jetzt gemacht 18:42 MT super! 18:44 sepski MT, yeah youll see the conffile in that paste actualy. 18:45 stockholm wie kriege ich denn das system dazu die initramdisk mit dm_mod zu bauen? 18:45 MT mom ... 18:46 MT stockholm: /etc/initramfs-tools/modules 18:47 stockholm MT: und wenn du mal den eigendlichen patch anschaust siehst du dass ich da auch kilobyte habe. aber es ist zweifelhaft dass das mit k und 1/1024 wirklich so funktioniert. weisst du man kommazahlen an parted geben kann? 18:47 MT nein, macht er nicht 18:47 MT rechnet alles in byte um 18:47 MT (und rundet wenn notwendig) 18:48 MT also vielleicht könnte parted das sogar 18:48 MT aber kleiner als byte wird eh nicht gehen :-) 18:48 MT bzw. man muss ja sogar zu cylinders runden 18:49 MT sepski, which paste do you mean? 18:49 stockholm MT: aber du wirfst ja Mbyte in range rein. rechnest du das später in byte um? 18:49 MT sepski, I only saw the one with the error 18:49 stockholm dann ist das mit kbyte kein problem so :-) 18:49 MT this has no fstab after all :-) 18:50 MT ja, stockholm - siehe z.b. Zeile 628 in shdd2-sizes 18:51 MT die perl-tidy version hast aber noch nicht eingecheckt, oder? 18:51 stockholm MT: doch 18:52 MT ah, ich hatte die schon 18:52 MT sorry 18:52 sepski MT, check line 94 in that one it's now i tried to do lvm swap 18:53 MT yeah, sure 18:53 MT but due to some bug setup_harddisks_2 failed entirely :-) 18:54 MT so I can't see anything about a missing entry in fstab :-) 18:54 stockholm uhh, did i break it??? 18:54 MT no 18:54 MT it is broken if you do hdparm -z before 18:54 MT for some strange reason 18:54 sepski well you wont find that in any log. i just ment to indicate how i tried towrite the swap into the config file 18:54 MT stockholm, see http://paste.debian.net/39237 18:55 MT ah, ok 18:55 MT ah, sepski 18:55 MT there's an error in that config 18:55 MT it should say 18:56 MT system2-swap swap 48 swap - 18:56 MT that is, the "mountpoint" is swap 18:56 MT not "-" 18:57 sepski MT, thanks 19:02 MT sepski, still there? 19:02 sepski yeah 19:02 MT did you fix the hook to use skiptask instead of skip at the end? 19:03 sepski yes 19:03 sepski this is what i get in fai.log now Calling hook: partition.EDUCONFIG 19:03 sepski Could not find method: Parse::RecDescent::convert_unit 19:03 sepski at shdd2-parser line 729 19:03 sepski disklist was: 19:04 sepski perhaps a new dependency 19:04 MT hmm, no 19:04 MT stockholm just added this method 19:04 sepski vn -u st 19:04 sepski Status against revision: 4616 19:05 MT 1 minute 19:06 sepski i use the etch package from http://layer-acht.org/debian and setup_hardrives from svn 19:06 h01ger thats 3.2.1 19:06 sepski yes Action: h01ger waves :) 19:06 sepski h01ger, :) 19:07 stockholm MT: taht was a typo, i just fixed it 19:07 MT but I think you need to say &FAI::convert_unit 19:08 MT when calling it 19:08 stockholm why? it is in the local scope only 19:08 stockholm it is just a local sub 19:09 stockholm but it still breaks 19:09 stockholm hu 19:09 MT I think it's because of the parser namespace 19:09 MT using &FAI::convert_unit should do 19:09 stockholm ok 19:09 stockholm i can fix that 19:12 stockholm checked in, testing it now 19:13 MT sepski, I guess it should do now... 19:13 sepski 4618 worked for me now 19:14 stockholm how fast are you guys booting?! 19:14 stockholm my box is till just mounting the nfs root 19:15 stockholm and did not get it 19:15 stockholm duh 19:15 sepski stockholm, im booting a virtual machine it's booting in a sec 19:15 stockholm a xen machine? 19:15 stockholm i need to do that too 19:15 stockholm later on :-) 19:15 sepski vmware, but virtaulbox is cheaper 19:16 stockholm sepski: are you doing this for debian-edu? 19:17 sepski ssh -X me@work vmware does not lag my machine here at home at all 19:18 sepski I'm trying to use it on a network i do consultant work for. but the idea is to at the very least make a howto for debian-edu and if it works very well it might become part of it. 19:19 juri_ qemu. ;) 19:19 stockholm juri_: is THAT fast?! 19:20 sepski juri_, qemu dies on me for some reason. i think i have messed toooo much with kqemu vbox-drv and kvm at the same time... 19:20 sepski stockholm, with either kqemu or kvm its fast. otherways it's syrup 19:21 sepski one think that vmware does much better then virtualbox and qemu is the networking between virtual machines. it's just there. without any hassle. 19:24 stockholm i am working with xen machines 19:24 juri_ stock: i work on reliable, not fast. :) 19:25 juri_ my fai server is a qemu box. 19:25 stockholm i had bad problems when trying to use it on my notebook 19:25 juri_ it takes 74 minutes to spin a FAI CD. ;P 19:25 juri_ (kqemu dosent work on my hardware/software) 19:25 sepski the production machines are real servers in my case this is just me trying things out, seeing if i can learn to use it or not. 19:25 stockholm my fai server ist a xen box :-) 19:25 MT sorry to interrupt, but sepski, does the whole thing now do it? 19:25 MT does it even work in case you add the hdparm and dd? 19:26 sepski MT, as long as i dont do any silly things in the hook yeah 19:26 sepski no that still fail 19:26 MT hmm, that's really strange 19:26 stockholm gah, my boot fails because it cant mount root 19:26 MT boot after installation? 19:26 stockholm lets see if it has dm_mod now 19:26 sepski and i had issues with grub eariler. going to be nice seeing how that went 19:27 stockholm MT: yes, boot after installation. the ramdisk is laoded and tries to mount / 19:27 MT I think we got things sorted out with grub 19:27 sepski MT, swap got into the fstab ysy 19:27 sepski yay 19:27 stockholm yes, this is no grub problem, it is a module/lvm problem 19:27 MT what is it trying to mount, i.e. what's there in menu.lst? 19:27 MT ah, no LVM support in the initrd? 19:28 stockholm MT: yes, that is why i added dm_mod to /srv/fai/config/files/etc/initramfs-tools/modules 19:29 stockholm but something must have gone wrong, dm_mod is not in the ramdisk 19:29 stockholm do i need more modules? 19:29 MT are you sure it isn't there, or is there some other error? 19:30 sepski stockholm, was that copied into the client, and is the initrd created _after_ ? 19:30 stockholm sepski: hu? 19:30 Mrfai stockholm: I would use ainsl to add a module to $target/etc/initramfs-tools/modules 19:30 MT jep, that's what I wanted to point out too 19:30 MT the entry must be there _before_ the initrd is built 19:31 MT alternatively, add a postinst script 19:31 stockholm ok. 19:31 MT that runs update-initramfs 19:31 stockholm Mrfai: where would the ainsl go? 19:31 Mrfai This is up to you. 19:31 stockholm when would the /srv/fai/config/files/etc/initramfs-tools/modules come into play? 19:33 stockholm i would have thought the /srv/fai/config/files/etc/initramfs-tools/modules was put into place right at the beginning 19:34 stockholm Mrfai: if the directory is not yet there, will ainsl create the missing directories? 19:35 sepski stockholm, the initrd is made during the base installation, and there is no /etc to put it into before that 19:36 sepski and if there was i guess baseinstall would overwrite them ? 19:36 Mrfai man ainsl or use ainsl -v and read the log files Action: stockholm shoots Mrfai and thanks the dead body 19:38 MT sepski, could you do another try with the hdparm and dd stuff in a minute? 19:38 MT I think I've found the bug 19:39 sepski i still had issue with grub too http://paste.debian.net/39248 19:39 sepski MT sure 19:39 stockholm Mrfai: and the docs dont answer my question either 19:39 stockholm Mrfai: i asked if the DIRECTORY would be created 19:39 stockholm not the file 19:39 stockholm i guess it is not 19:39 MT sepski, it's r4619 19:40 MT sepski, what is the name of your hook? 19:40 MT I guess it's still partition.EDUCONFIG 19:40 MT but it should be partition.EDUCONFIG.source 19:40 MT instead 19:40 sepski hooks/partition.EDUCONFIG 19:41 MT jep, rename it to hooks/partition.EDUCONFIG.source please 19:41 sepski darn missed that in the wiki 19:41 stockholm ok, newbi question: where should i put things that do do stuff like ainsl? in hooks? 19:41 MT yeah, added that one to the wiki as well 19:41 MT it is necessary to properly source disk_var.sh 19:41 MT use instsoft.FAIBASE 19:42 MT if you've got that one, stockholm 19:42 sepski MT is thatone run after everything else ? 19:42 MT then the line will be there before the kernel package is installed 19:42 MT that one - the instsoft.FAIBASE hook? 19:43 MT no, it is run before installing packages 19:43 MT I thought you were looking for some hook to work before packages are installed 19:43 MT but maybe I missed something 19:43 stockholm MT: jup 19:43 sepski that was him,. i am looking for a way to run a script just before the machine is rebooted 19:44 stockholm sepski: LAST? 19:45 sepski scripts/LAST ? 19:45 MT yes, you could put whatever you like to do in scripts/LAST/99-at-the-end 19:45 MT but there's some more going on 19:45 MT like saving logs 19:46 MT that happens even after scripts have been executed 19:46 sepski this is with dd and hdparm in the hook http://paste.debian.net/39250 19:46 sepski that's ok 19:47 stockholm sepski: what do you do with hdparm? 19:48 MT is that r4619? 19:48 MT I think sepski clears out any existing LVM 19:49 stockholm i see :-) 19:49 sepski the idea is to clear the drive with dd then run hdparm -z to reload the partitiontable (that just got cleared away) before losing dm_mod 19:49 sepski loading dm 19:50 MT did you really use r4619 for what you pasted in 39250? 19:50 sepski MT, AGGH /me hides 19:50 sepski forgot ! 19:50 MT huh 19:50 sepski sorry ! Action: MT hopes that it'll do in the next attempt :-) 19:52 sepski it seams to do it 19:52 MT cool :-) 19:52 MT thanks a lot for testing! 19:52 sepski altho it's a ugly workaround to dd the disk. :) 19:52 sepski MT, stockholm thanks a lot for developing code ! 19:53 sepski now lets see if grub likes it better 19:53 sepski :) 19:54 MT stockholm, any progress at your side? 19:55 stockholm MT: i have rebootet, the mounting of / hangs again and next i am going to see if dm_mod is in the initramfs 19:55 stockholm sepski: lol, not much from me:-) 19:56 sepski system2-usr /usr 2000 ext3 defaults 19:56 sepski where do the mkfs arguments go ? the wiki seam to indicate after the mount arguments 19:57 MT yes, they do 19:57 sepski system2-usr /usr 2000 ext3 defaults dir_index,resize_inode 19:57 MT hmm, there's only a single example 19:57 MT primary / 12000 ext3 rw -b 2048 19:57 MT # /dev/hda3 should be formatted using ext3 filesystem; when calling mkfs.ext3 19:57 MT # the option "-b 2048" is appended. 19:58 sepski ahh right there 19:58 sepski system2-usr /usr 2000 ext3 defaults -O dir_index,resize_inode 19:58 sepski it must be like this then 19:58 MT yes, correct 19:59 stockholm yes, now dm_mod is there 19:59 stockholm and it is loaded 20:00 sepski perhaps add -O dir_index,resize_inode to the lvm example since i think that's a pretty common wanted feature 20:01 MT sepski, will do ... 20:01 MT stockholm, does your comment mean that dm_mod is there but it still doesn't work? 20:01 sepski MT, i'v made a wiki user. so expect wiki spam later on :) 20:02 stockholm MT: yes, exactly 20:02 MT cool, feel free to edit whatever you like, sepski 20:02 MT I already added the _o dir_index,resize_inode to the last example 20:02 stockholm in /proc/cmdline it says that root=/dev/my_vp/_root 20:03 MT ah, are you in busybox? 20:03 stockholm /dev/my_pv/_root 20:03 stockholm yes 20:03 MT does /dev/my_pv exist? 20:04 MT or does /dev/mapper/ exist? 20:04 stockholm in busybox, on the commandline from hell, without the / key due to my bad terminal emulation 20:04 stockholm dev/mapper is there yes 20:04 MT but no volume groups below it? 20:04 stockholm there is only control 20:05 stockholm no volume groups 20:05 stockholm possibly the lvm creation did not work out? 20:05 stockholm no, we were intalling stuff into it after all 20:05 MT yep 20:05 sepski MT didnt' think reiserfs had the same commands like dir_index and resize_inode 20:06 MT oops, thanks, fixed 20:06 MT hmm, do you have any of the lvm tools inside the busybox? 20:07 MT or could you just manually mount the root fs 20:07 MT and then use the lvm tools to see what's going on? 20:08 stockholm how can i mount the root fs if the device is not there? 20:08 MT ah - well - 20:08 sepski MT, last try installed and booted OK Action: MT hides 20:08 MT including LVM? 20:08 sepski yes 20:08 sepski i have /boot as ext3 tho 20:08 sepski and bootable:1 20:09 stockholm yes, me too 20:09 stockholm not so root 20:09 sepski root on lvm 20:09 stockholm me too 20:09 stockholm why cant *I* boot /root and you can?! 20:09 sepski http://paste.debian.net/39252 this is the one i use 20:10 MT stockholm, could you try to run the dd and hdparm commands that sepski uses 20:10 MT sepski, could you paste these two lines? 20:10 stockholm why would that help? i dont want to get rid of anything. 20:11 sepski stockholm, are you preserving partitions between installs ? 20:11 stockholm no 20:11 stockholm no preserving at all 20:11 stockholm only ERASING all 20:11 sepski do you erase the lvm's manually between each install ? or does it simply work becouse the vg happen to be the same name ? 20:12 stockholm i dont know 20:12 sepski the hook file i use http://paste.debian.net/39253 20:12 stockholm i think mounting the thing should be straight forward... 20:13 sepski is it grub that fail to load or the kernel that fail to find root when booting ? 20:14 stockholm the kernel does not find a /root after booting the initramfs 20:14 sepski stockholm, ohhh ! i know i can boot becouse i install the usplash package later on, then the initrd is recreated !" 20:15 sepski so i get the dm_mod there probably 20:15 stockholm i have dm_mod in the initramfs and it is inserted in the kernel 20:15 oz_ let's see if it works without usplash... 20:15 sepski and use use ext3 ? 20:16 stockholm yes 20:16 sepski then i realy dont have any idea 20:17 sepski oz_, it's installing again now without usplash 20:17 stockholm i will try to add the lvm tools onto the initramdisk 20:20 sepski hmm why does the partition with lvm pv on it dont have type 8e (linux lvm) ? 20:20 stockholm (STDERR) Warning - device mapper device, but no dmsetup(8) found 20:20 stockholm what is dmsetup good for? 20:20 MT no idea... 20:20 MT (just on the phone, so I'm sort of slow...) 20:22 sepski stockholm, i dont have that package installed either 20:22 stockholm sepski: do you get any error like the above in fai.log? 20:22 stockholm i will try to add the package 20:23 stockholm where did i specify packages i want? 20:23 stockholm ah, i need it in the nfsroot 20:25 stockholm dmsetup manages logical devices that use the device-mapper driver. Devices are created by loading a table that specifies a target for each 20:25 stockholm sector (512 bytes) in the logical device. ErKa (keryell@keryell.pck.nerim.net) joined #fai. 20:30 sepski no not at all Mrfai (~lange@xdsl-87-78-55-146.netcologne.de) left irc: Quit: leaving 20:30 stockholm sepski: hu? 20:32 sepski but i have added libparse-recdescent-perl parted lvm2 to the nfsroot as explained in the wiki 20:35 sepski without the late recreation of the initrd by usplash it fails to boot 20:38 sepski i am trying now with script/LATE/99-update-initramfs containing $ROOTCMD update-initramfs 20:39 MT $ROOTCMD update-initramfs -k all -u 20:39 MT (-k all -u might be necessary) meandtheshell (~markus@85.127.117.208) joined #fai. 20:42 stockholm interesting.... 20:42 stockholm what is missing? 20:42 MT missing where? 20:43 stockholm earlier, when the initramfs is generated 20:43 stockholm the first time 20:44 sepski probably the mtab in the chroot or something 20:48 stockholm mtab is for raid, right? 20:48 stockholm i dont do any raid 20:49 sepski /etc/mtab is the mount list 20:50 stockholm ah, right, sorry, i mixed that up 20:54 MT hmm, stockholm, what is so fascinating about the point where the initramfs is generated? 20:56 sepski booted fine with script/LATE/99-update-initramfs 20:57 sepski dont know whats missing at the first kernel install time,. but i'd look at mtab or perhaps proc is not mounted before after base install is finished. 20:58 MT is the kernel already part of the base.tgz? 20:58 MT or is it installed later on, together with the other packages? 21:00 sepski later i think, 21:01 sepski but now it's night. I'll be back at 0800 for more fai fun :) 21:01 stockholm sepski: LATE was an existing class? 21:01 sepski script/LATE is a directory. i made a file called in there 21:02 stockholm yes 21:02 MT LAST 21:02 sepski LAST 21:02 sepski bah 21:02 sepski sorry 21:02 sepski night 21:02 MT good night! 21:02 stockholm right, i was confused abtou late vs last sepski (~sep@217.17.211.51) left irc: Quit: Leaving allee (~ach@allee.mpe.mpg.de) left irc: Quit: Konversation terminated! alvinc (~alvin.cur@office-ca.emaillabs.com) joined #fai. 21:54 alvinc by any chance, has anyone successfully FAI'd a Dell PowerEdge 2970? 21:55 MT what is it that troubles you? 21:55 alvinc i'm having that "cannot mount root problem" 21:55 MT not that I'd have such a machine, but maybe one can help anyway 21:55 alvinc i'm wondering if it needs a new network driver 21:56 alvinc it's getting PXE fine, kernel is loading, but mounting the nfsroot is being a pain 21:56 alvinc i'm guessing maybe a network driver, but i was curious if it had been seen befire 21:56 alvinc i'm loading feisty, btw 21:56 MT does it work on any other machine? 21:56 alvinc i'm building a 2.6.20 fai kernel now, just in case 21:56 alvinc but i was curious 21:56 alvinc it used to. :) 21:57 alvinc i mean, i FAI'd a Dell 850 and 750 a while back, before i started hacking. 21:57 alvinc and i can't re-FAI the 850 and 750 right now, because they are in use 21:57 alvinc hehehe 21:58 alvinc yeah. the feisty default is linux-image-2.6.17.13-ubuntu1-fai-kernels_1_i386.deb 21:58 alvinc so i'm hoping 2.6.20 might be the quick fix 21:58 MT well, with the Dells this may be true 21:59 MT you might want to ask on the mailing list though 21:59 alvinc make-kpkg is being slow though, and i was restless. so i thought i'd ask if anyone had already seen this one 21:59 alvinc yess, i might do that 21:59 MT I know of some guys having lots of Dell systems, but no idea whether they already have a 2970 21:59 alvinc nod 22:00 alvinc today is not a good day for me, it seems 22:00 alvinc i'm also having problems with a Realtek 8169 NIC 22:00 alvinc i've successfully used an etherboot floppy with it on other machines before 22:00 alvinc but this one is being a pain 22:01 alvinc haven't quite figured it out yet. i'm suspecting maybe spanning-tree on the cisco is not doing portfast correctly 22:01 alvinc even though i set it to portfast 22:01 alvinc dhcp is not happening quickly enough 22:01 alvinc guess it's just one of those days. :-) 22:01 MT hmm, cisco is popping up again and again in the context of FAI 22:02 alvinc nod 22:02 MT seems like people have quite some trouble with that sort of switches 22:02 alvinc it's not really FAI that is the issue, of course 22:02 alvinc it's that dhcp is not returning within an acceptable timeout 22:02 MT sure, but rather DHCP/NFS/PXE 22:02 alvinc so it's more to do with dhcp really 22:02 alvinc yessir, exactly 22:03 alvinc are you in Europe, MT? 22:03 MT yes, Germany 22:03 alvinc ah cool 22:03 alvinc so it's already late for you 22:03 MT thus I'm leaving for bed pretty soon 22:03 alvinc ja 22:03 MT yep :-) 22:03 alvinc :) 22:03 alvinc i'll try getting a dumber switch and see if it makes a difference 22:03 alvinc if i see you'll tomorrow, i'll let you know what happens 22:04 MT ok, cool, thanks 22:04 alvinc bitte 22:04 alvinc (that's all the german i know) 22:04 alvinc rofl 22:04 MT :-) 22:04 MT so, good luck 22:05 stockholm ok, still no booting install, despite the late intiramfs 22:06 stockholm i could try to unmount and remount it right after boot. 22:06 stockholm see if i get any errors or problems then 22:06 MT did you manage to install any of the LVM tools into the initramfs 22:06 MT ? 22:07 stockholm no, not yet, i still dont know much about lvm. 22:07 stockholm which ones would be good to have? 22:07 MT for debugging it might be useful to move your root to a standard partition 22:07 stockholm yes, that is true 22:07 MT and only keep some not-so-important volume in LVM 22:08 MT then debugging is probably a lot more fun 22:08 stockholm right 22:08 MT otherwise you'd somehow need all of lvm2 in your initramfs, including the libraries 22:09 MT I'd suggest to go for the only-swap-in-LVM way 22:09 MT this shouldn't hinder it from booting 22:09 MT then you can see what else you need to get it going 22:14 stockholm right. tomorrow :-) 22:14 MT jep, good night! meandtheshell (~markus@85.127.117.208) left irc: Quit: Leaving. MT (~MT@dove.informatik.tu-muenchen.de) left irc: Ping timeout: 480 seconds MT (~MT@ppp-62-245-162-223.dynamic.mnet-online.de) joined #fai. 22:30 alvinc hm. it looks indeed like it's the ethernet driver for Dell 2970 22:30 alvinc i've got my nForce ethernet up fine ErKa (keryell@keryell.pck.nerim.net) left irc: Ping timeout: 480 seconds MT (~MT@ppp-62-245-162-223.dynamic.mnet-online.de) left irc: Ping timeout: 480 seconds meandtheshell (~markus@85.127.117.208) joined #fai. --- Wed Oct 10 2007 Generated by irclog2html.pl 2.1_h01ger1 by Jeff Waugh - find it at freshmeat.net! Customized by Holger 'h01ger' Levsen