[02:43] istaz (~olivier@77.222-242-81.adsl-dyn.isp.belgacom.be) joined #fai. [02:50] istaz_ (~olivier@250.106-65-87.adsl-dyn.isp.belgacom.be) left irc: Ping timeout: 480 seconds [03:15] pjustice (~pjustice@yagi.h-net.msu.edu) left irc: Ping timeout: 480 seconds [04:14] pjustice (~pjustice@yagi.h-net.msu.edu) joined #fai. [07:13] fai-guy (~fai-guy@p4FC23452.dip0.t-ipconnect.de) joined #fai. [07:29] siggg (~jean@dyn-83-156-145-242.ppp.tiscali.fr) joined #fai. [08:13] ErKa (keryell@m2.wifi.enstb.org) joined #fai. [08:28] siggg (jean@dyn-83-156-145-242.ppp.tiscali.fr) left #fai. [10:04] h01ger: im hpc-cluster verkraftet das netz alle 56 maschinen gleichzeitig [10:05] ui...was für ein netz ist das? [10:05] 1GB ? [10:05] oz__: ein high performance nety [10:05] netz [10:05] daher der name hpc [10:05] Action: oz__ konnte maximal 10 [10:05] im "normalen" netz, also arbeitsplatzrechner und server, sind nicht mehr 20 drin, ohne dass die user leiden ;) [10:05] ab 15 wurde es...naja [10:06] aber das war ein extrem verswitchtes netz...nicht wirklich gut [10:06] oz__: hpc: tcp/ip laeuft ueber 1gb [10:06] der mpi-traffic der neuen maschinen wird ueber IB geregelt [10:08] die normalen maschinen sind teilweise ueber 1gb, teilweise ueber 100mbit angebunden. der erfahrungswert 20 stammt noch aus der 100mbit-only zeit ;) [10:08] eartoast: das war auch mein wert aus Freiburg [10:08] in Jena ging's auf 10-15 runter [10:09] welcher dienst gibt denn zuerst auf? [10:09] nfs? oder svn? [10:11] eartoast, hm, ich dachte eher von der gui. der screenshot sieht nicht so aus, als ob das für 56 oder 560 machinen scaliert [10:11] _das gui_ [10:11] weil soviel nicht aufn bildschirm passt [10:15] stockholm: cvs vergibt read-locks, die rechner werden an der stelle zunaechst einmal ausgebremst [10:18] ah, ich habt cvs. :-) [10:18] die 20-rechner-gleichzeitig-grenze war eigentlich das, was die benutzer hier im netz noch toleriert haben... wenn 20 rechner gleichzeitig von nem 1gb-angebundenen server openoffice per nfs updaten, war im netz selbst nicht mehr viel luft (ok, da war auch mehrere switches kaskadiert) [10:18] mittlerweile haben wir das ganze flacher... [10:19] keine ahnung, wie viel das jetzt verkraftet [10:19] h01ger: die spalten kann man skrollen [10:20] hmokay. gruppieren wäre noch schön, aber es ist ja noch nicht weihnachten :) [10:20] das layout war aber das effizienteste, was man mit 80x25 machen konnte [10:20] ja. jetzt versteh ich das layout auch besser. die in "success" sind ja quasi egal.. und da sollten die meisten sein :) [10:21] wenn man sehr viele rechner gleichzeitig im update-zustand sind, dauert das updaten des UI allerdings manchmal etwas [10:21] jo [11:49] ErKa (keryell@m2.wifi.enstb.org) left irc: Ping timeout: 480 seconds [12:06] b1gf1sh (~f1sh@193.16.163.152) joined #fai. [12:07] hey [12:07] I have a question concerning multiple network devices and FAI [12:08] thing is, we have HP DL380 servers here, which come with 2 network-cards with 2 ports each [12:09] and I need to have the FAI installation booted with eth2 (2nd nic, 1. port), but unfortunately, while booting via tftp it tries to boot from eth0 only (after dhcp on eth2 kicked in and started booting), so then the FAI installation stalls at the moment it would usually mount the live-system from server to client via nfs [12:10] obviously, since there is no connection between eth0 and the fai-server [12:18] ErKa (keryell@keryell.pck.nerim.net) joined #fai. [12:22] b1gf1sh: what kind of kernel are you using/ [12:22] initrd or monolithic? [12:23] im using 2.6.18-6 (debian etch default kernel) [12:23] initrd [12:23] network devices are both "bnx2" drivers [12:23] so thats pretty nasty, would have been easier to disable loading of one driver if the the nics were different brands [12:24] there must be something I could add to the file "fai-chboot" creates for the specific server [12:24] some kernel appened param [12:25] well, you could try to unpack the initrd and modify it; [12:26] kernel command line options only work reliably if you are using the kernel-level autoconfig code [12:27] kernel-level autoconfig code? never heard of it? can you please explain? [12:27] i was thinking to add: insmod=bnx2 netdevice=eth2 [12:27] to the append line [12:27] but it didnt work [12:29] append initrd=initrd.img-2.6.18-6-amd64 insmod=bnx2 netdevice=eth2 ip=dhcp root=/dev/nfs nfsroot=/srv/fai/nfsroot boot=live FAI_FLAGS=verbose,sshd,createvt,reboot FAI_ACTION=install [12:30] any hints? [12:32] what i meant with kernel-level autoconfig is the code within the kernel [12:33] which is used if you linked the network drivers statically into the kernel [12:33] and you don't have an initrd [12:33] ah ok [12:33] this code then parses the ip=$something line [12:34] makes sense, what would be the easiest work around then? [12:34] what you could try is to set dhcp only for eth2 and disable eth{0,1} detection. [12:34] yeah, when i disable eth0/1 in bios [12:34] all works fine [12:34] as soon as i reenable it [12:35] it will stall again during fai installation [12:35] i did not speak about the bios... i spoke about the ip=$something part [12:35] in the kernel command line [12:35] how would it look like then? [12:36] i have no idea how to disable eth0/1 in the command line while not disabling eth2/3 which are of same brand/same driver [12:36] dirty hack: unpack the initrd, and restrict the udhcpc-call to eth2 [12:37] ah alright [12:37] within the initrd, it is just a bunch of shell scripts which do the work [12:37] ;) [12:37] sounds cool [12:37] then repack it and reboot. [12:38] the initrd code is understandable, I remember doing a similar hack a year ago... [12:39] thanks a lot [12:40] you dont remember the cmds to unpack the initrd and mount it? [12:40] mount -o loop initrd.img-2.6.18-6-amd64 /test/ [12:40] mount: you must specify the filesystem type [12:41] ah figred it out, nm [12:44] problem is, the file called udhcpc-call is not in there? do I have to create it manually? [12:47] argh. [12:48] no. in some script, there is called the dhcp client [12:48] call in the sense of call of external program [12:49] try to do the shotgun-approach: user@host:~/unpacked-initrd> grep -ri 'dhcp' . [13:11] ok [13:11] ive tried to edit scripts/nfs [13:12] and changed DEVICE in the line: ipconfig -c ${IPOPTS} -d ${DEVICE} to [13:12] ipconfig -c ${IPOPTS} -d eth2 [13:16] no go unfortunately [13:16] anyway, thanks for help [13:16] i might be back tonight if I dont figure it out, cu [13:16] b1gf1sh (~f1sh@193.16.163.152) left irc: Quit: b1gf1sh [13:24] fai-guy_ (~fai-guy@p4FC23452.dip0.t-ipconnect.de) joined #fai. [13:53] fai-guy__ (~fai-guy@p4FC23452.dip0.t-ipconnect.de) joined #fai. [13:53] fai-guy__ (~fai-guy@p4FC23452.dip0.t-ipconnect.de) left irc: Read error: Connection reset by peer [13:55] fai-guy_ (~fai-guy@p4FC23452.dip0.t-ipconnect.de) left irc: Ping timeout: 480 seconds [13:58] fai-guy_ (~fai-guy@p4FC23452.dip0.t-ipconnect.de) joined #fai. [14:00] fai-guy_ (~fai-guy@p4FC23452.dip0.t-ipconnect.de) left irc: Read error: Connection reset by peer [14:02] juri_ (~juri@ip70-178-3-140.ks.ks.cox.net) left irc: Ping timeout: 480 seconds [14:06] juri_ (~juri@ip70-178-3-140.ks.ks.cox.net) joined #fai. [14:34] ErKa (keryell@keryell.pck.nerim.net) left irc: Ping timeout: 480 seconds [14:50] fai-guy_ (~fai-guy@p4FC23452.dip0.t-ipconnect.de) joined #fai. [14:55] siggg (~jean@dyn-83-156-145-242.ppp.tiscali.fr) joined #fai. [15:04] fai-guy (~fai-guy@p4FC23452.dip0.t-ipconnect.de) left irc: Quit: leaving [15:06] fai-guy_ (~fai-guy@p4FC23452.dip0.t-ipconnect.de) left irc: Quit: leaving [15:21] siggg (~jean@dyn-83-156-145-242.ppp.tiscali.fr) left irc: Remote host closed the connection [15:29] ErKa (~keryell@keryell.pck.nerim.net) joined #fai. [15:51] siggg (~jean@dyn-83-156-145-242.ppp.tiscali.fr) joined #fai. [16:27] siggg (~jean@dyn-83-156-145-242.ppp.tiscali.fr) left irc: Ping timeout: 480 seconds [16:43] allee (~ach@allee.mpe.mpg.de) left irc: Remote host closed the connection [17:46] juri_ (~juri@ip70-178-3-140.ks.ks.cox.net) left irc: Ping timeout: 480 seconds [19:51] siggg (~jean@tra78-5-88-172-128-32.fbx.proxad.net) joined #fai. [20:45] ErKa (~keryell@keryell.pck.nerim.net) left irc: Ping timeout: 480 seconds [20:57] ErKa (~keryell@keryell.pck.nerim.net) joined #fai. [21:54] ErKa (~keryell@keryell.pck.nerim.net) left irc: Remote host closed the connection [22:08] ErKa (keryell@keryell.pck.nerim.net) joined #fai. [22:13] siggg (~jean@tra78-5-88-172-128-32.fbx.proxad.net) left irc: Ping timeout: 480 seconds [22:28] ErKa (keryell@keryell.pck.nerim.net) left irc: Ping timeout: 480 seconds [00:00] --- Fri Jun 6 2008