[08:09] hot-line (~root@89.189.71.105) joined #fai. [08:12] hi i have problem in my wifi can any body help me ? [08:13] hm. I think you're wrong here. [08:13] ? [08:16] is you question FAI related? do you want to install a host via wifi? [08:18] no i wont join [08:19] want join [08:19] sory [08:19] can you help me [08:19] you're wrong here, believe me. :) [08:22] where can i find the answer [08:24] I don't know :) depends on your OS and wifi card. [08:26] my os is backtrack 4 [08:26] its linux [08:29] you might want to try #backtrack-linux [08:29] i have it now [08:34] hot-line (~root@89.189.71.105) left irc: Remote host closed the connection [09:07] ErKa (keryell@keryell.pck.nerim.net) joined #fai. [09:20] alekibango (~alekibang@ip-94-113-35-25.net.upcbroadband.cz) left irc: Remote host closed the connection [09:41] ErKa (keryell@keryell.pck.nerim.net) left irc: Ping timeout: 480 seconds [09:52] fai: 3 mt committed revision 6228 to 7 fai: Added (modified) fcopy patch as posted on linux-fai-devel by Sabine Schmidt. [10:27] alekibango (~alekibang@ip-94-113-35-25.net.upcbroadband.cz) joined #fai. [10:40] hi, i am trying to do simple disk partitioning with raid, but it fails.... my config: http://pastebin.com/8g2UiUFn [10:40] can someone please give hint or example for raid config? [11:09] Hi alekibango! [11:10] alekibango, could you give a more detailed description of what you mean by "it fails"? [11:57] alekibango (~alekibang@ip-94-113-35-25.net.upcbroadband.cz) left irc: Ping timeout: 480 seconds [12:05] pav5088 (~Mark@120.152.48.8) joined #fai. [12:12] pav5088 (~Mark@120.152.48.8) left irc: Read error: Connection reset by peer [12:12] pav5088 (~Mark@120.152.48.8) joined #fai. [12:14] pav5088 (~Mark@120.152.48.8) left irc: [12:31] alekibango (~alekibang@ip-94-113-35-25.net.upcbroadband.cz) joined #fai. [12:44] mt, will, just a moment, i had power failure here... [13:02] is there picture alternative to pastebin ? :) [13:12] mt: here is log http://pastebin.com/T42nPNZ1 [13:13] i am using debian squeeze with fai 3.4.5 [13:13] (and squeeze) [13:27] sepski (~sep@87.238.45.202) joined #fai. [13:27] Nick change: sepski -> sep_oslo [13:51] sep_oslo (~sep@87.238.45.202) left irc: Ping timeout: 480 seconds [13:58] sepski (~sep@87.238.46.238) joined #fai. [14:02] Nick change: sepski -> sep_oslo [14:54] alekibango, sorry for responding late, I'll take a look at your logs [14:54] np, i can understand :) [14:54] thanks [14:55] had a power outage here at home :-/ [14:55] the thing is, i am not very skilled using mdadm, so the error does nto make much sense for me [14:55] it's not an mdadm error actually [14:55] well i was not sure what it is [14:56] hmm, why are you using virtual for your configs? [14:56] is that indeed such a setup? [14:56] not sure, copied it from some example [14:56] mt: if you can provide some example for raid, it would be best [14:56] reading gramma from man pages didnt really help me muc [14:56] much* [14:57] well, your configuration is correct [14:57] examples are missing all over FAI :) [14:57] I just wonder whether you are intentionally using "virtual" [14:57] no i do not know what that even means :) [14:57] seems like there is a bug with virtual+raid [14:57] is it some lvm thingy? [14:58] so how could i make it working? :) [14:58] ok, you should only be using "virtual" when /dev/sda1, /dev/sda2, etc. already exist, e.g., when using a XEN domU [14:58] ah, no, they do not exist [14:58] then you should remove the virtual in your sda and sdb configs [14:58] and give it another try :-) [14:59] btw in example i have seen 5 primary partitions [14:59] i wanter what that would do :) [15:00] I don't quite understand?! [15:01] RAID_XEN_VIRTUAL [15:01] (simple) [15:01] contains 5 primary partitions for each drive [15:01] thats what i used to copy and paste to make my own [15:02] yes - so you wonder why more than 4 can be used? [15:02] yes [15:02] with "virtual" there's no disk label (like the MSDOS disklabel, which only permits 4 primary partitions) [15:02] "virtual" means that setup-storage assumes all these partitions already exist [15:02] ah :) i never liked xen :) [15:03] but thanks [15:03] setup-storage will only create file systems and RAID [15:03] so the "EXAMPLES" section of the setup-storage man page isn't yet sufficient? [15:04] will look again [15:05] (I'm right now just working on the man page, so any input is very welcome!) [15:06] problem SYNTAX is too dense for most people. i studied computer science, so i can read it, still tis too hard to decipher... [15:06] its* [15:06] simple example is great [15:07] Create a softRAID -- example looks incomplete [15:07] should be whole config imho [15:07] (where did sda1,sdd1 come from?) [15:08] other than that, it looks fine [15:09] having more examples always helps a ton to newbies [15:11] As first step I moved the Examples section to the front, before this large grammar section. [15:12] nice... [15:12] thanks for the work... i will contribute too, if i will use fai in production [15:14] may I use your configuration as example? [15:14] (without the "virtual" though :-) ) [15:15] yes, if that will work [15:15] i am now having funny problem with booting from eth1 [15:15] ip=:::::eth1:dhcp worked few hours ago [15:15] multiple interfaces? [15:15] yes [15:15] :)) [15:15] now it again tries to boot from eth0 [15:16] you can't rely on eth1/eth0 being consistent interface names [15:16] thankfully i have ipmi interface... [15:16] see a recent mailinglist post for that [15:16] names are ok [15:16] BOOTIF might help [15:16] bootif -- will check ty [15:19] looks like BOOTIF=eth1 helped... thansk! [15:19] disks partitioned :) [15:19] works like charm now [15:20] Action: alekibango is sending 2 beers to mt UP UP [15:20] cool :-) [15:21] hehe, but worked only on 1 server, 2 failed to boot lol [15:21] what's going wrong? [15:21] linux kernel is using eth0 [15:22] and will not ask eth1 for 'next server' [15:22] failing to connect 192.168.0.1 using eth0 [15:23] so, the issue is kernel is using dhcp on eth0 instead of eth1 [15:25] maybe you could briefly review https://lists.uni-koeln.de/pipermail/linux-fai/2010-December/008539.html [15:25] maybe there's something in there that helps [15:39] not really, but interesting still.. reading kernel docs for ip= now [16:35] justleen (~leen@188.206.129.195) joined #fai. [16:45] justleen1 (~leen@92.69.88.14) joined #fai. [16:51] justleen (~leen@188.206.129.195) left irc: Ping timeout: 480 seconds [17:11] justleen1 (~leen@92.69.88.14) left irc: Ping timeout: 480 seconds [17:26] mt: imho having some ready for usage fai examples (whole config sets) would be very helpfull [17:27] well, at least your configuration is now in the man page [17:27] :) [17:27] I'll check the others ... [17:31] few simple examples == like pictures, worth 1000 words in man [18:04] fai: 3 mt committed revision 6229 to 7 fai: Reworked setup-storage man page [18:38] msantana (msantana@SDF.ORG) left irc: Read error: Connection reset by peer [18:48] another problem is, when i use raid, it installs well, but after reboot there is only dead cursor there. grub install probably failed somehow on /dev/md0 [18:48] found the proof: GRUB/10-setup FAILED with exit code 1. [18:49] aha, its trying to copy menu.lst :) [18:51] well but in squeeze there is grub2 [18:52] ehm, grub-pc vers. 1.98+xxxx [18:52] strange thing is, without raid it worked well [19:30] alekibango, which version of FAI are you using? [19:30] you need to use a proper script for grub-pc [19:30] in the simple examples there's a GRUB_PC/10-setup [20:00] 3.4.5 [20:01] ha, my examples might be from oldero ne [20:01] older one :) [20:03] /usr/share/doc/fai-doc/examples/simple/scripts/GRUB -- [20:03] there is fcopy -v /boot/grub/menu.lst [20:03] do you have the GRUB_PC directory? [20:03] (in the examples) [20:03] ha, yes i do,. i guess i used grub dependency hehe [20:04] will try [20:04] thanks for tip [20:04] well, I think you should just copy the 10-setup script from examples /GRUB_PC to your config as GRUB/whatever [20:04] that is, just replace the script that was used before [20:04] (ifclass I386 || ifclass AMD64) && echo GRUB_PC [20:04] exit 0 [20:05] ah, ok [20:05] thats at the end of 50-host-classes [20:05] so probably GRUB_PC would be fine :-) [20:05] and i have GRUB there [20:05] :) [20:05] thats it [20:06] i hope squeeze will become stable as promised :) even with some bugs [20:06] just use some resolution that will execute the new script [20:06] as promised? [20:06] i mean planned [20:07] btw, no guarantees that things will work right away after fixing this GRUB_PC/GRUB mess [20:07] but let's see, together we'll make it work :-) [20:07] ha, on debian there is no release date has been set [20:07] but i have seen its expected to come this year [20:08] just before total crash of the american economy [20:11] its ready when its ready [20:12] i know what you mean, and you know what i want :) [20:12] or you are talking about the economy, heh [20:13] fai-economy? ;) [20:13] (see gerald celende, founder of trendsresearch.com [20:13] that guy is interesting [20:14] Gerald Celente [20:14] yes, my fingres are somehawt lazy today [20:23] btw i cant find documentation for FAI_FLAGS [20:28] oh yes, I think you're right [20:28] I'll ask thomas if this is really missing, if so [20:29] we'll file a doc bug [20:29] if i am not right, its not obvious where to look for it [20:29] http://fai-project.org/fai-guide/ [20:33] should i paste 8 lines, or use pastebin? [20:33] http://pastebin.com/uW2dw195 [20:33] grub --> error: no such disk. [20:33] (md0) [20:34] oz: i didnt found all FAI_FLAGS listed/explained in guide [20:35] only mentioned few times [20:36] oh yes, they are there... sorry -- http://fai-project.org/fai-guide/ar01s07.html [20:38] but as 7.3.1 is not mentioned in contents (and has space between fai and flags), its not that easy to find there if you are curious... 1 html page manual might make sense for some people :) [20:45] so, how i installed grub by hand: [20:45] mkdir /dev/md ; ln -s /dev/md0 /dev/md/0 [20:45] grub-install --recheck --root-directory=/target /dev/md0 [20:46] # -there are some bugs in grub-install or related scripts [20:54] alekibango, could you do another round of installs? [20:54] with an additional hack [20:54] sure, i did many and i will do another 100 for sure [20:55] luckily i have impi interface :) [20:55] which makes it easier a bit when servers are 100+km away [20:55] please add to your 10-setup script, before any grub-* calls [20:55] mount -o bind /dev $FAI_ROOT/dev [20:55] mount -o bind /proc $FAI_ROOT/proc [20:55] that's a cruel hack [20:55] ah, that might work :) [20:57] well, problem with grub is something around md/0 instead of md0 [20:58] hmm, ok, I just read that this was fixed in some version, but I don't know whether this holds for your version [20:58] it imho is, its mounted already [20:59] dev was --binded and proc is mounted 2 times [20:59] and sysfs etc [20:59] mt: its more like grub problem [20:59] so try another hack: [20:59] rewrite /dev/md/0 to /dev/md0 [20:59] via sed magic [20:59] i just will make the directory and link for me :) [21:00] I think /dev/md0 should exist as well [21:00] writing it atm [21:00] simple change and might work for me, but grub still has some bug [21:01] which grub version is that? [21:02] squeeze - grub-pc 1.98+20100804 [21:03] ehm 1.98+20100804-8 [21:03] hmm, that one was fixed long ago: [21:03] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=475585 [21:08] you know, bugs are really hard to kill [21:08] thats why insecticides are so dangerous [21:09] :-)= [21:34] alekibango: I think thomas will be happy to include a little patch in the doku ;) [21:34] (and I really this this would be a real improvement with almost zero work) [21:35] oz_: my bug is still alive, even when i used insecticides [21:35] ? [21:36] well, grub_pc fails to install on raid [21:36] for me -- on squeeze [21:36] squeeze the bug, then. ;) [21:36] trying [21:36] it moves lol [21:36] this talk reminds me starship troopers [21:36] dunno, did you google for it? [21:37] iirc, there were a lot of mails about it [21:37] and a lot of changes recently [21:37] ok, reading now [21:38] what $ROOT_COMMAND does in scripts? is it only chrooting to /target ? [21:39] yes, & set the vars accordingly. [21:40] but I don't know too much about recent FAI [21:40] will get into it again around january, I guess. [21:40] my last big fai job was in 2005 ;) [21:40] oh no 2007 [21:40] :) [21:40] but that was just a weekend to setup a linux pool [21:41] nothing big [21:41] $ROOTCMD expands to chroot /target [21:41] that's all [21:41] but that's already quite a lot :-) [21:41] i am trying to install openstack cluster using fai... [21:42] later bigger cloud... [21:42] alekibango: I hope you do more than just trying to do it! [21:42] for now its few machines [21:42] well. openstack is still young, would be scared to use it on customers [21:43] but its very promising project [21:43] Action: oz_ want to do an opensource cash point soultion \w FAI [21:43] cash point ? troll on a bridge? [21:43] :D [21:44] no for supermarkets and such [21:44] also, nothing really big [21:44] ic [21:44] I have a friend running a supermarket, and he has some visualbasic coded crap [21:44] oh, that sounds evil as google [21:45] it's very small, I think I can do it. [21:45] oz_: there si some floss software for this [21:45] not sure about name, but i have seen some [21:45] ? [21:45] floss? [21:45] free/libre/opensource software [21:46] yea, I'll look when it'S time. I think so, too. [21:46] but these merchants are special [21:46] they all want some specialities :) [21:47] but we're getting off-topic. [23:12] DavidDreezer (~Adium@c-71-227-156-146.hsd1.wa.comcast.net) joined #fai. [23:13] does anyone know of there has been some kind of regression in setting HOSTNAME in recent 4.0 experimental versions? [23:14] hmm, no, not that I would know of [23:15] what's happening? [23:15] I have tried everything Micheal [23:15] with everything being what? [23:15] i put HOSTNAME=myname in pxelinux.cfg script [23:15] no go [23:15] it should be lowercase hostname= [23:15] tried a /$FAI/config/class script [23:15] as in something such as 45-hostnanme [23:16] which read hostname myname [23:16] no go [23:16] no matter how I try the 50 hostname script finds no hostname [23:16] however older versions of 40 experimental this worked [23:16] i am baffled [23:16] which version of FAI *and* which distro? [23:16] debin [23:16] lenny [23:17] and today i updated to 40 45, which is the latest [23:17] I was trying the new /GRUB_PC 10 script for you [23:17] or, planning to [23:18] root@faiserver:/lib# dpkg --list | grep fai [23:18] ii fai-client 4.0~beta2+experimental45 Fully Automatic Installation client package [23:18] ii fai-doc 4.0~beta2+experimental45 Documentation for FAI [23:18] ii fai-server 4.0~beta2+experimental45 Fully Automatic Installation server package [23:18] ii fai-setup-storage 4.0~beta2+experimental45 automatically prepare storage devices [23:18] it is on a lenny server, installing lenny [23:18] was working fine, now not so fine. [23:19] let me check my pxelinux.cfg file again [23:19] would you know which version you were using before? [23:20] i can check. in my pxelinux.cfg file i have HOSTNAME in the kernel vars, not lower case [23:20] logs in /var/log/dpkg.log might help, if necessary [23:20] it should be lower? [23:20] yes iirc [23:20] yes, should always be lowercase [23:20] for that matter. why does it need to be there? [23:21] if i have a /classes/45-hostname script [23:21] i echo hostname myname, i echo it into additional.var and simply echo HOSTNAME-myname [23:21] why is it that 50 script cannot understand that? [23:22] well, the question is how your system (upon installation) finds out what its hostname is [23:22] you could get that via DHCP [23:22] if that isn't set up or doesn't work, you can use the kernel command line [23:23] i was using 41 previous to today, just checked. [23:23] kernel command line should really be a fallback as you need a separate cfg file for each system [23:24] i agree, i'd much rather not put it into kernel or even dhcp at that [23:24] I was successfully using a file in /classes [23:24] I wrote a script 45-mac-vars [23:24] hmm, ok, so that seems like a different issue then!? [23:25] well, that stopped working [23:25] so i fell back to kernel [23:25] and found it not working for me either [23:25] however maybe the uppercase was the issue on that [23:26] anyway my 45 script read the mac, and found a file named by mac (i perfer that over hostname) which contained speciual vars for the individual host [23:26] such as ip address, hostname, [23:26] was working. stopped working [23:27] sorry, I fail to find a change that could actually be causing this [23:27] i was relying on this [23:27] http://www.mail-archive.com/linux-fai@uni-koeln.de/msg02860.html [23:27] which no longer works [23:27] But as it stands this seems to be all your code that is being executed!? [23:27] the code is executing yes [23:28] but the 50 script fails to detect the hostname i set in 45 [23:28] fails to detect it from any way I can figure out how to set it [23:28] did you verify that at the very end of 45 hostname is still set to what you expected? [23:29] yes, at the end of 45 i echo $HOSTNAME >> addition.var [23:29] btw, just to make one thing sure: your NFSROOT also has the experimental packages? [23:30] yes, I am sure. [23:30] so, i look in addtrional.var and the correct hostname is written [23:30] yet 50 fails to find it [23:30] what is the exact name of your scripts? [23:30] 01-set-date 10-base-classes 20-hwdetect.source 45-mac-vars 50-host-classes [23:31] and which mode of communication are you expecting [23:31] just via variables or should the system hostname be set [23:31] could you maybe paste(bin) those two scripts? [23:32] what I am expecting is that 50-host-class understands the hostname I am trying to set, and applies the right classes [23:32] paste(bin)? [23:32] http://paste.debian.net/ [23:32] (and paste the link here) [23:34] ah. ok. the 45 script is now stripped to but a line or two from what it was. I've been working with it all morning trying to find what happened. let me restore it and I shall paste it [23:34] I doublechecked the nfsroot, it has the new files [23:34] I rebuilt it after I updated my mirror [23:35] mirror ... using fai-mirror? [23:36] that is, your mirror might be broken if you used fai-mirror at some point [23:37] using mkedebmirror [23:38] i added the experimentals to it though [23:38] seems to work, so why not use the tools you provide? :) [23:38] ok, I have my scripts ready to paste [23:38] ok, let's see :-) [23:38] there are a few debug echoes in there, as you shall see. [23:40] ok, 45-mac-vars [23:40] http://paste.debian.net/101531/ [23:41] 50-host-classes http://paste.debian.net/101532/ [23:41] would you also have the (currently broken) additional.var at hand? [23:42] i do [23:42] also one of the files in /MAC.vars [23:42] http://paste.debian.net/101533/ [23:42] do things get any better if you use `hostname` instead of $HOSTNAME in 50-host-classes? [23:42] the name of the file is, as you would expect, a mac address [23:42] i can try that, cannot recall if I have [23:44] additional http://paste.debian.net/101534/ [23:44] honestly, I'm not really sure why this interaction of 45-mac-vars and 50-host-classes would work [23:44] FAIBASE.var makes up the rest of the info that is needed, it is the default file [23:44] would, or would not? [23:45] I'm trying to figure out :-) [23:45] sorry. [23:45] i do appreciate all fo the help you've provided as I go through this project [23:46] ok, it would work if $HOSTNAME is set properly by a new shell spawned to execute 50-host-classes [23:46] you're welcome :-) [23:46] but I don't know whether this new shell gets spawned at all [23:46] so I cannot update the %HOSTNAME variable within my existing shell? [23:47] (which si what I'd assumed I had been doing :) ) [23:47] well, you can do that, but it won't have effects outside your shell [23:47] changing a variable in a subshell should never affect its parents... [23:47] but running "hostname" should have those effects [23:49] I seem to recall someone having an issue similar in the 3.x family, but I cannot find the post about it now. [23:50] so, if you could paste an "additional.var" file that would be nice [23:51] the way i know for sure that this is all failing is that after 50 has run $TYPE is set to SW_RAID and classes are FAIBASE DEBIAN AND DHCPC [23:51] i did, did I fail to put it here? [23:51] you might also want to try adding calls to "hostname" in both of your scripts and logging its results [23:51] oops, sorry [23:51] http://paste.debian.net/101534 [23:52] huh, that looks fine!? [23:52] exactly! [23:52] HOSTNAME and HOSTNAME2 seem to have the same value [23:52] yes! [23:52] so what is it that is going wrong? [23:52] hence my amazement/lack of understanding [23:53] do you have 50 in front of you? [23:53] yep [23:53] with the hostnamer db099.db01.grouipee-inc [23:53] the type is secified as SW_RAID in the file [23:53] would you allow me a break? [23:53] telephone ... [23:53] however 50 assigns it NOT_SW [23:53] sure [23:54] and it gives the * classes [23:54] not the db classes [00:00] --- Sun Dec 5 2010