[01:17] no, whats currently in svn tells me its beta3... ??? [05:04] meandtheshell (~markus@85.127.108.76) left irc: Quit: Leaving. [06:38] MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. [07:04] Nick change: \sh_away -> \sh [07:20] allee (~ach@dialin-212-144-131-228.pools.arcor-ip.net) joined #fai. [07:32] frale (~frale@tor-irc.dnsbl.oftc.net) joined #fai. [07:32] mfl (~mfl@apoll.fnb.maschinenbau.tu-darmstadt.de) joined #fai. [07:37] MT (~MT@dove.informatik.tu-muenchen.de) left irc: Remote host closed the connection [07:38] MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. [07:39] ErKa (keryell@193.50.97.167) joined #fai. [07:52] psibox (~dreger@smart.physik.fu-berlin.de) joined #fai. [08:12] MT (~MT@dove.informatik.tu-muenchen.de) left irc: Ping timeout: 480 seconds [08:12] MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. [08:39] hi, i have a patch now for #416633 against 3.1.8 (setup_hardisks, mount2dir) [08:39] for uuid/label in /etc/fstab [08:47] Mrfai (~lange@wireless-7-120.internal.dc7.debconf.org) joined #fai. [08:47] ah. [08:47] hi Mrfai [08:47] hi, good morning [08:47] i have a patch for the UUID/Label support against 3.1.8 now [08:48] in setup_harddisks [08:48] and mount2dir [08:48] there is code in mount2dir that explicitly calls fsck [08:48] this ist not used in default config, since fai-mount-disk calls it with fsckopt=0 [08:48] please send the patch to me or to the BTS [08:49] i have to discuss this fsck problem first [08:49] we don't use fsck [08:49] IF you use it, it will only work for ext2/3 [08:49] but this is not tested. [08:49] in the code [08:50] anyway, I can put in UUID/LABEL support even for fsck [08:50] but it will not work for anything except ext2/3 [08:50] fsck is not always called [08:50] well. [08:50] it looks like it's never called [08:52] [ "$fscheck" = 1 ] && fsck -nt $fstype $device [08:52] fscheck is always 0 or "" [08:52] fsck does only work for ext2/3. [08:53] this should at least check for the filesystem type and call the right fsck-program [08:53] but then i have the "problem" that not all fsck-programs unterstand UUID/LABEL [08:55] Mrfai: do you know any case, where mount2dir is called with fsck=1 ? i can't see any. [08:55] why is fscheck always 0 or ""? [08:55] Mrfai: solved the initrd problem yet? [08:55] maybe mount2dit is currently not calles with fsck enabled, but you can do it [08:56] i have a screenshot [08:56] stockholm: do you have a screenshot of the error message? [08:56] my point is: this breaks for anything except ext2/3 [08:56] in stock 3.1.8 [08:56] i can fix it for xfs/reiserfs [08:56] Mrfai: lange@debian.org? [08:56] but i can not support fsck for xfs/reisfers with UUID/LABEL [08:57] stockholm: yep [08:57] reiserfsck does not unterstand UUIDs [08:59] you got mail [08:59] psibox: ok. SO UUID support is only usefull when the user is only using ext2/3 filesystems [09:01] nono [09:01] UUID/LABEL work perfectly for reiserfs/xfs/ext2/3 [09:01] only problem is fsck [09:02] i have tested this with all three systems (we are actually using all three) [09:02] you can call fsck with UUID=1e95f00b-6408-4217-b69d-af6ab626eab0 instead of /dev/hdaX [09:03] this is not supported by xfs_check and reiserfsck [09:03] i could probably make it work using udev and /dev/disk/by-uuid/... but since it's probably not used by anyone, well... [09:04] proposal: i'll put in a check for ext2/3 befor calling fsck. [09:04] good idea, [09:04] fsck never worked for xfs/reisfers and won't with UUID/LABEL [09:04] so it's not really a change :) [09:04] I think there's also a wishlist bug for adding UUID support. [09:04] sure [09:04] that's the bug i'm looking at [09:05] we have been using this for years. our servers can't be installed without it. [09:05] no that you can't controll which sata-controllers are detected first... [09:06] another question is why anyone would want to call fsck on a filesystem, that was formatted 1 second ago... [09:09] psibox: send me the patch and will have a look at it [09:09] ok [09:09] it's quite small. [09:28] frale_ (~frale@tor-irc.dnsbl.oftc.net) joined #fai. [09:28] Mrfai: another thing that would be most useful for us would be if setup_harddisks would select the disk by serial-number [09:28] i tested this and it didn't work. i assume it's not (yet) supported ? [09:28] psibox: this is maybe a feature for the new partition tool [09:30] well, i looked at the new tool. [09:30] frale (~frale@tor-irc.dnsbl.oftc.net) left irc: Remote host closed the connection [09:30] that really has nothing to do with the tool, just udev [09:31] you just have to make sure '/dev/disk/by-id/scsi-SATA_ST380817AS_4MR4PM9S' is in place before calling setup_harddisks [09:32] anyway, if it's not supported yet, i'll just try to make it work and let you know... [09:33] one _could_ put the serial-number code inside setup_harddisks, but it's probably better to let udev do that... [09:35] udev is already running when setup_harddisk is called [09:36] i know. didn't work in a test for me. [09:36] have to look into that [09:44] MT (~MT@dove.informatik.tu-muenchen.de) left irc: Ping timeout: 480 seconds [10:06] Mrfai_ (~lange@wired-251-252.internal.dc7.debconf.org) joined #fai. [10:08] Mrfai (~lange@wireless-7-120.internal.dc7.debconf.org) left irc: Ping timeout: 480 seconds [10:09] Mrfai_: do you have an idea about the initrd problem? [10:09] Mrfai_: it is unrelated to the tftp one afaik [10:12] stockholm: please email me your file 59916A4B or paste it to paste.debian.net [10:12] certainly [10:15] MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. [10:26] http://paste.debian.net/31090 [10:26] Mrfai_: see above [10:30] mmmm. weired. [10:31] can't see any problems there [10:31] me either [10:31] link works [10:31] ls -l /srv/tftp/fai [10:31] are you using symlinks there? [10:31] Mrfai_: a symlink to from linux to the kernel [10:32] /srv/tftp/fai$ l [10:32] total 7176 [10:32] -rw-r--r-- 1 root root 5792342 2007-06-21 22:19 initrd.img-2.6.18-4-amd64 [10:32] lrwxrwxrwx 1 root root 22 2007-06-21 21:25 linux -> vmlinuz-2.6.18-4-amd64 [10:32] -rw-r--r-- 1 root root 13480 2007-06-21 22:19 pxelinux.0 [10:32] drwxr-xr-x 2 root root 4096 2007-06-22 12:31 pxelinux.cfg [10:32] -rw-r--r-- 1 root root 1514181 2007-05-04 05:35 vmlinuz-2.6.18-4-amd64 [10:33] what about a debugging option in tftp? [10:34] the log in paste.debian.net is with -v -v -v i think [10:34] i could increase that i guess [10:34] are you using the hpa version of tftpd? [10:34] yes [10:34] see in the log [10:36] the log files only says in.tftpd, not that thius ixs the hpa version [10:36] tftp dgram udp wait root /usr/sbin/in.tftpd /usr/sbin/in.tftpd -v -v -v -v -v -s /srv/tftp/fai [10:36] lets see [10:37] ^dpkg --get-selections | grep tftp [10:37] tftpd-hpa install [10:38] so yes [10:38] which version [10:39] apt-cache policy tftpd-hpa [10:39] tftpd-hpa: Installed: 0.43-1.1 Candidate: 0.43-1.1 [10:39] from etch [10:39] same as mine [10:40] Mrfai_: i've sent the patch to #416633. that mail contains a disk_config example and the corresponding results [10:40] psibox: fine [10:42] stockholm: install a tftp client on a machine (may on the install server), then try to get te vmlinuz-... file using the tftp client [10:43] Mrfai_: now i added a couple of -v more and get http://paste.debian.net/31091 [10:43] it does not find the file [10:44] but that is the other file [10:44] not the kernle one [10:44] so scratch that [10:45] try the tftp client to see if the file gets received corretly [10:45] remove the command line [10:48] Mrfai_: i'd suggest that you print the sfdisk input if verbose=1 or debug=1 in setup_harddisks [10:50] #364763 can easily be solved by this [10:51] should i send a patch ? :) [10:51] psibox: sure [10:51] are there any more bug related to setup_harddisks ? [10:52] remove the command line? which one? [10:53] stockholm: 31090 on paste, line 18 [10:58] comment ahh [11:02] Mrfai_: same error [11:03] Mrfai_: loading the vmlinuz file directly worked from out of atftp client [11:03] my theory is that the problem is on the client side [11:03] as the server never sees anything from it [11:03] as in an attempt to download a file [11:04] how feasable is it to debug pxelinux.0 and its operation? [11:05] Mrfai_: but that is unrelated to my real problem (initrd and live and generating the right ramdisk image) [11:06] yep [11:07] using the symlink is a workaround [11:08] I found the problem [11:08] AND the solution [11:09] Let's check, if you already has signed the Enterprise Support Contract [11:11] did you find the signed support contract? [11:11] i have a copy of it here, i filed it under "social contract" [11:11] since i am so social [11:14] http://paste.debian.net/31094 [11:14] The default and label lines are missing [11:14] MT (~MT@dove.informatik.tu-muenchen.de) left irc: Ping timeout: 480 seconds [11:23] jerkeby (~jerkeby@wireless-7-169.internal.dc7.debconf.org) joined #fai. [11:24] hey guys [11:24] Nick change: jerkeby -> kugg [11:25] Mrfai_ (~lange@wired-251-252.internal.dc7.debconf.org) left irc: Ping timeout: 480 seconds [11:26] lazyb0y: wuzah? [12:21] kugg: what? [12:21] Mrfai (~lange@wireless-8-41.internal.dc7.debconf.org) joined #fai. [12:23] live-initramfs is not in the repository :-( [12:24] stockholm: try again. should work now. [12:30] lazyb0y: wats up, in a slangier term :) [12:30] kugg: ey yo, digga [12:30] :D [12:31] which guide did you guys want me to translate to swedish? [12:42] I see you lazyb0y [12:44] kugg: hrm, ah! [12:44] and now? [12:47] :) [12:47] Action: lazyb0y waves [12:49] MT (~MT@dove.informatik.tu-muenchen.de) joined #fai. [12:53] stockholm: klappt es nun? [12:57] Mrfai: ich bin gerade dabei das problem in name2id in fcopy zu fixen [12:57] waere es ok, wenn files-modes auch kommentarzeilen vertragen wuerde ? [12:57] geht naemlich derzeit nicht... [13:00] psibox: ja, das sollte eigentlich nur eine Zeile mehr sein. KAnnst du machen [13:01] h01ger: Zeit fuer ein kleines FAI Treffen. Jetzt? Im Cafe. [13:06] Action: kugg went to a bof [13:07] MT (~MT@dove.informatik.tu-muenchen.de) left irc: Ping timeout: 480 seconds [13:08] kugg: I was a bit shocked to see your laptop being stolen cause I didn't take care... [13:08] but obviously that hasn't happened, ot at leas the thief is a nice guy who keeps talking to us... [13:11] lazyb0y: habe auch kein highlighting im emacs. [13:11] haha [13:11] hmm, daja eklig ... [13:29] stockholm: wie heissen nochmals diue paketnamen fuer linux-image und unionfs fuer amd64? [13:29] paste die namen einfach hier hin [13:30] ja [13:31] kernel-image-2.6-amd64-k8-smp unionfs-modules-2.6-amd64 [13:31] glaub ich nicht. Der kernbel sollte linux-image... heissen [13:32] hm [13:36] linux-image-2.6-amd64 ist wohl besser. den anderen gibt es zwar auch aber ich bin nicht sicher warum [13:37] es sollte reichen dann mit dem neuen packages.gz file make-fai-nfsroot auszuführen, gell? [13:38] Mrfai: ? [13:40] Mrfai sip'ed [13:40] allee (~ach@dialin-212-144-131-228.pools.arcor-ip.net) left irc: Read error: Connection reset by peer [13:41] stockholm: ja [13:44] funktioniert nicht [13:46] du weist ja jetzt wie man es debuggt. Wie du das eben gemacht hast [13:46] genauso [13:47] genau [13:48] live gibts immer noch nicht in der initrd [13:48] fehlermeldung? [13:50] die gleiche wie gestern: cant open /scripts/live [13:50] ich schau mal ob das in packages.gz im changeroot ist [13:51] schau dir die logs von make-fai-nfsroot an. Da sollte drin stehen warum live-ini... nicht installiert wird. [13:51] wo sind die? [13:51] stockholm: bitte keine Anfaengerfragen. Das hast du doch schonmal gemacht. [13:53] du meinst das ding mit tee? [13:54] ja [13:54] Action: stockholm macht das [13:57] WARNING: These unknown packages are removed from the installation list: live-initramfs [13:58] Action: stockholm untersucht das näher. [13:59] stockholm: brav :-) [13:59] Mrfai: wenn du sagst "logfile" dann denkt man an etwas unter /var/log [14:00] Get:2 http://www.informatik.uni-koeln.de developers/koeln Packages [1654B] [14:02] Mrfai, seit ihr immer noch im cafe? [14:03] Mrfai: ok, jetzt bin ich ratlos. [14:04] Mrfai: das packages file in /var/lib/apt/list hat live-initibla aber apt-cache findet es nicht (beides im chroot) [14:05] h01ger: ja [14:06] ja sind im cafe [14:07] stockholm: diese frage macht mich allerdings auch ratlos. was machst du denn da? [14:08] lazyb0y: nix besonders. komm und guck. [14:09] hacklab1, fensterplatz [14:09] folge dem geruch von earl gray [14:10] kugg (~jerkeby@wireless-7-169.internal.dc7.debconf.org) left irc: Ping timeout: 480 seconds [14:17] stockholm: du kennst den spruch mit dem berg und dem propheten ;) [14:17] live-initramfs is im packages file aber nicht im cache file [14:18] und ich habe keine ahnung wie das sein kann [14:19] live-initramfs_1.91.1~2143_i386.deb [14:19] i386 und nicht amd64 [14:20] Mrfai: spricht irgendwas dagegen, dass man folgendes: [14:20] $uid = ($user =~ /^\d+$/) ? $user : `$ENV{ROOTCMD} perl -e '\$uid = getpwnam $user; print \$uid'`; [14:20] ersetzt durch: [14:20] $uid = ($user =~ /^\d+$/) ? $user : `$ENV{ROOTCMD} id -u "$user"`; [14:21] ? [14:21] psibox: HAbe jetzt keine Zeit fuer Details. Muss ich mir in ruhe ansehen. Bitte via mail oder ans BTS [14:22] Mrfai: so are you building one for amd64? [14:22] Mrfai: i could build one, too. i have a amd64 machine here [14:23] kugg (~jerkeby@wired-230-252.internal.dc7.debconf.org) joined #fai. [14:24] ErKa (keryell@193.50.97.167) left irc: Ping timeout: 480 seconds [14:34] allee (~ach@dialin-145-254-255-156.pools.arcor-ip.net) joined #fai. [14:37] Mrfai: du kannst das von paket von http://developer.skolelinux.no/~andreas/ ziehen [14:37] Mrfai: passt die grösse? [14:37] (habe ich nicht gescheckt) [14:38] ja, sieht gut aus [14:38] ist nun da [14:38] Mrfai: auch schon ein neues packages.zg? [14:45] frale_ (~frale@tor-irc.dnsbl.oftc.net) left irc: Quit: . [14:49] Mrfai: i've submitted a patch for the name2num problem with a patch attached. [14:52] meandtheshell (~markus@85.127.111.125) joined #fai. [14:58] Nick change: \sh -> \sh_away [15:03] Mrfai: http://backports.org/debian/pool/main/l/live-initramfs/live-initramfs_1.91.1-1~bpo.1_amd64.deb [15:06] is now available in developers/.. But this version is slightly older that that before [15:53] now i boot again and it hangs after booting the root fs [15:54] no error message this time [15:59] Mrfai: setup_harddisks can currently not use disks by serial-number [16:00] ich würde gerne nochmal etwas hilfe beim debuggen in anspruch nehmen... Mrfai ? [16:00] Mrfai: ich bin wieder an meinem fensterplatz [16:00] Action: Mrfai langweilt sich im Grid Talk [16:01] ich habe jetzt das live-intibla paket nochmal im pbuilder gebaut und lade ihn nach developer hoch... in5min [16:01] what about: 'disk_config serial=4MR0ADP5' in disk_config file ? [16:02] psibox: write a short proposal and post it to linux-fai-devel. We should diskcuss that there. [16:07] Mrfai: jetzt habe ich ein neu paket frisch gebaut hochgeladen [16:07] neues [16:44] Mrfai (~lange@wireless-8-41.internal.dc7.debconf.org) left irc: Ping timeout: 480 seconds [16:46] Mrfai (~lange@wireless-8-41.internal.dc7.debconf.org) joined #fai. [16:51] what is the usual diretory people use for tftpd in fai ? [16:51] /boot/fai ? [16:51] MT (~MT@M309P023.adsl.highway.telekom.at) joined #fai. [16:52] Nick change: MT -> Guest90 [16:55] Guest90 (~MT@M309P023.adsl.highway.telekom.at) left irc: [17:03] stockholm: danke, neues amd paket liegt jetzt bei mir im repository [17:03] psibox: /srv/tftp/fai for FAI 3 [17:03] ah, ok. [17:06] Mrfai: ich kann aus dem rsync-support fuer savelog auch noch'n patch machen [17:06] aber ich muss halt immer wissen, wie man das aktivieren soll [17:06] habe es derzeit bei uns als hook drin [17:06] bin jetzt auf linux-fai-devel [17:07] warum werden die subjects eigentlich nicht mit [fai] oder sowas getaggt ? [17:07] faend ich besser ;) [17:08] weil dann jede Mail mit [fai[ getaggt wuerde. IMO eine ueberfluessige Info. [17:09] naja, das haengt so'n bisschen davon ab, ob man seine mails wegsortiert oder alle in die inbox kriegt. [17:09] rsync fuer savelog? Wozu? WArum nicht direkt rcp/scp? [17:09] aber ich weiss natuerlich, dass ich das auch selbst mit procmail/formail taggen kann... [17:09] scp kann keine links. [17:10] welche links? [17:10] rcp ist nicht gerade sicher, das haben wir nirgends. [17:10] wir haben im logdir links [17:10] ich nicht IIRC [17:10] z.B. zeigt last-install immer auf die letzte installation [17:10] MT_ (~MT@M309P023.adsl.highway.telekom.at) joined #fai. [17:10] jedenfalls funktioniert das mit rsync sehr schoen. [17:11] damit kann man z.B. das logdir auf dem server write only fuer die clients machen [17:11] fai macht das doch schon. auch ohne rsync [17:11] achso, scp braucht einen user, richtig ? [17:11] genau, ich wollte den LOGUSER weg haben [17:11] auch du meinst rsync nicht via ssh sondern auf ein NFS gemountetes dir. [17:11] mit rsyncd [17:12] habe ich mal ins fai wiki geschrieben vor einer weile [17:12] aber wenn das keiner nutzen will, soll mir das auch recht sein [17:12] ich finde, das funktioniert besser als alles andere ;) [17:12] und? Gabe es feedback das andere es auch wollten? [17:13] ja, irgendwer hat das mal genutzt. [17:13] aber egal. [17:13] ich kann damit leben, wenn's nicht in fai ist. [17:13] geht ja mit einen hook. [17:13] einem. [17:13] und es ist da sehr gut im Wiki dokumentiert. Danke nochmals dafuer. So sollte es sein [17:26] MT_ (~MT@M309P023.adsl.highway.telekom.at) left irc: Read error: Connection reset by peer [17:26] MT_ (~MT@M309P023.adsl.highway.telekom.at) joined #fai. [17:33] Lin (~igor@200.179.57.57) joined #fai. [17:35] MT_ (~MT@M309P023.adsl.highway.telekom.at) left irc: Quit: MT_ [17:40] kugg (~jerkeby@wired-230-252.internal.dc7.debconf.org) left irc: Ping timeout: 480 seconds [18:05] allee (~ach@dialin-145-254-255-156.pools.arcor-ip.net) left irc: Read error: Connection reset by peer [18:11] Lin (~igor@200.179.57.57) left irc: Quit: Ex-Chat [20:34] ErKa (keryell@keryell.pck.nerim.net) joined #fai. [21:41] Mrfai (~lange@wireless-8-41.internal.dc7.debconf.org) left irc: Quit: leaving [22:06] MT (~MT@M391P026.adsl.highway.telekom.at) joined #fai. [22:07] Nick change: MT -> Guest105 [22:09] ErKa (keryell@keryell.pck.nerim.net) left irc: Ping timeout: 480 seconds [22:20] Guest105 (~MT@M391P026.adsl.highway.telekom.at) left irc: Quit: Guest105 [00:00] --- Sat Jun 23 2007