IRC chat logs for #ltsp on irc.freenode.net (webchat)


Channel log from 13 February 2011   (all times are UTC)

00:00
<ball>
X11 for the graphics?
00:01
<alkisg>
Yup
00:01alkisg has quit IRC
00:03alkisg has joined #ltsp
00:05evil_root is now known as zz_evil_root
00:26ball has quit IRC
00:38alkisg has quit IRC
00:40alkisg has joined #ltsp
01:14alkisg has quit IRC
01:19chupacabra has joined #ltsp
01:42vvinet has quit IRC
01:54vvinet has joined #ltsp
01:58alkisg has joined #ltsp
02:04vvinet has quit IRC
02:06vvinet has joined #ltsp
02:34vvinet has quit IRC
02:38vvinet has joined #ltsp
03:13alkisg has quit IRC
03:17chupacabra has quit IRC
04:13vmlintu has joined #ltsp
04:14_UsUrPeR_ has quit IRC
04:23_UsUrPeR_ has joined #ltsp
05:15chupacabra has joined #ltsp
05:49komunista has joined #ltsp
07:59artista_frustrad has joined #ltsp
08:08Faithful has joined #ltsp
09:10mikkel has joined #ltsp
09:38artista_frustrad has quit IRC
09:53RiXtEr has quit IRC
09:54RiXtEr has joined #ltsp
10:16andygraybeal has joined #ltsp
10:37wwx has quit IRC
10:39wwx has joined #ltsp
10:47bobby_C has joined #ltsp
10:48xavierb has joined #ltsp
10:50Trixboxer has joined #ltsp
10:59
<muppis>
nouveau keeps loading even blacklisted. Any hint?
11:02MorningSon has joined #ltsp
11:15komunista has quit IRC
11:23alkisg has joined #ltsp
11:26alkisg has quit IRC
11:33xavierb has quit IRC
11:37alkisg has joined #ltsp
11:38mikkel has quit IRC
11:41xavierb has joined #ltsp
11:47
<alkisg>
muppis: XSERVER=nv ?
12:01xavier_brochard has joined #ltsp
12:05xavierb has quit IRC
12:10vagrantc has joined #ltsp
12:22xavier_brochard has quit IRC
12:27bobby_C has quit IRC
12:40vagrantc has quit IRC
13:19alkisg has quit IRC
13:21alkisg has joined #ltsp
13:31vagrantc has joined #ltsp
13:38alkisg has quit IRC
15:19Faithful has quit IRC
15:26xavier_brochard has joined #ltsp
15:27Trixboxer has quit IRC
15:28chupacabra has quit IRC
15:44xavier_brochard has quit IRC
16:10wwx has quit IRC
16:10wwx has joined #ltsp
16:15alkisg has joined #ltsp
17:02nutron has joined #ltsp
17:02
<nutron>
I have a server which has been running successfully for some time (since last July), as of last week, the system seems to freeze on the clients like crazy, every minute or so the client becomes unresponsive.
17:03
I've done what I could to check the pipes.
17:03
There doesn't seem to be a network issue.
17:03
All I'm left with is the ltsp server, and I can't for the life of me try to ascertain what the issue is.
17:03
Has anyone else run into something like this before?
17:04
System is currently only running 6 clients. It's an 8 core xeon with 48 gigs of memory. Running debian lenny. (afraid to go to squeeze just yet)
17:05
The home directories are remote mounted via nfs. No issue there.
17:05
I think I'm going to remove the NBD settings since they are thin clients.
17:05
(for swap)
17:05
Don't mind me, I'm just rambling. But this issue cropped up early last week, and the users are gettin' mighty pissed...
17:06
I even have suggestions to install windows, because windows wouldn't do things like this :(
17:11
<vagrantc>
nutron: you using NBD swap, or NBD root, or both?
17:23
<nutron>
vagrantc: nbd swap
17:23
Hmm after running iostat on the nfs server, it seems that I get spikes of 30%+ iowait, though I can't tie them to the lockups.
17:24
The nfs server runs the home directories.
17:24
<vagrantc>
seperate from the application server?
17:24
<nutron>
vagrantc: aye, seperate
17:25
but orbit, dbus and gconf all use the home directory... Though I'm not sure which would cause the locks if the nfs server wasn't responding fast enough.
17:25
<vagrantc>
nutron: could try taking the NFS server out of the picture and moving some home dirs directly onto the application server
17:25
nutron: are you running the ltsp backports?
17:25
<nutron>
vagrantc: aye, thought of that, but it's a bind mount and have to wait for the users to go home, so there's no instant love there.
17:26
vagrantc: yes I am.
17:26
<vagrantc>
nutron: they sshfs-mount the home dirs on the thin-clients in order to support localapps
17:26
<nutron>
oh..
17:26
<vagrantc>
well, if they're crashing every minute, i don't see what the difference is in intentionally taking them down
17:26
<HrdwrBoB>
nutron: can't runa rest user and change the homedir?
17:27
<nutron>
vagrantc: heh, they're not crashing.. they .. pause..
17:27
<vagrantc>
nutron: ah.
17:27
nutron: i *do* vaguely recall something like that...
17:27
<nutron>
like you click the application menu, then it appears but the highlight doesn't move with the mouse... it's unresponsive for a few seconds... then it seems to catch up again.
17:27
same for when typing an email etc.
17:27
you type.. it freezes but it retains everything you typed and the buffer clears and all your words appear.
17:28
<vagrantc>
think i just eventually replaced the server ... or ram, or something like that
17:29
<HrdwrBoB>
that sounds liek an I/O issue
17:29
<vagrantc>
nutron: any cron jobs?
17:29
running every minute
17:30
<nutron>
Yeah it is io... i see the nfs server having large io waits..
17:30
Hmm nope, I've been watching the logs for days... every five minutes I get the php poller from cacti running, but otherwise the app server is clear to do its thing.
17:30
<vagrantc>
nutron: are they running gnome? i seem to recall something that took heavy i/o toll by doing indexing of the homedir by default
17:31
<nutron>
yep gnome, heavy indexing?
17:31
<vagrantc>
some search indexing program...
17:31
<HrdwrBoB>
yeah
17:31
uses couchdb
17:31
<nutron>
I don't that would be it... here I'll paste a snapshot of iostat on the nfs machine
17:31
<HrdwrBoB>
it used to go crazy on my laptop
17:31
had to kill it all the time
17:32
<nutron>
ps doesn't show any instances of couchdb
17:32
this is lenny still, nothing got installed .. unless I don't know about it :P
17:33
<vagrantc>
one day it worked, the next it started acting up?
17:33
<nutron>
http://pastebin.com/MCgt7kZq
17:34
though the iowait is low, there are times when it jumps quite high
17:34
vagrantc: yeah, last tuesday people started complaining about the pauses.
17:34
and it's all about the writes
17:34
obviously.
17:34
<vagrantc>
bad disk in a raid array?
17:34
<HrdwrBoB>
nutron: have you dropped a disk in your raid array
17:34
and not noticed?
17:35
<nutron>
cat /proc/mdstat doesn't show me much
17:35
and dmesg again, doesn't show much
17:35
I'll check it again
17:36
wwwow.. cat /proc/mdstat took 3 seconds on the nfs server...
17:36
ok there's something wrong there for sure
17:36
<HrdwrBoB>
yeah
17:36
there's your problem
17:36
<nutron>
no errors though?
17:36
<HrdwrBoB>
should look something like this:
17:36
md1 : active raid5 sdj[3] sdc[4] sda[2] sdb[0] sdh[6] sdd[1] sdi[7] sdg[5] 10255969472 blocks level 5, 64k chunk, algorithm 2 [8/8] [UUUUUUUU]
17:36
<nutron>
ie. should it complain about _something_?
17:37
yeah, that's right, mine looks like that
17:37
<HrdwrBoB>
all Us
17:37
no underscores?
17:37
<nutron>
but running cat against it took a long time
17:37
yeah no underscores..
17:37
md0 : active raid1 sda1[0] sdd1[2] sdc1[3](S) sdb1[1] 1945310720 blocks [3/3] [UUU]
17:37
<HrdwrBoB>
anything suspeicious in top?
17:37
yeah that looks fine
17:38
<nutron>
mysql.. but.. it's always been a hog on the nfs machine
17:38
though.. wait i did have ata problems at boot on the nfs server... they magically went away after I forced a rebuild
17:38
hrrm
17:38
32116 mysql 19 -1 4237m 1.2g 6004 S 2 4.9 240:08.05 mysqld
17:39
1.2g resident and 4.2g virtual
17:39
<vagrantc>
maybe a little swappity
17:39
<HrdwrBoB>
ram is not your problem, but it may be indicative of a larger issue
17:39
yeah
17:39
what does free say
17:39
<nutron>
0k of swap used
17:39
on both boxes..
17:40
24 gigs on the nfs server
17:40
<vagrantc>
well, good luck sorting that out...
17:40* vagrantc heads out
17:40
<nutron>
vagrantc: heh, thanks
17:40vagrantc has quit IRC
17:43alkisg has quit IRC
17:57jhutchins_kc has joined #ltsp
17:58Gibby has left #ltsp
17:59jhutchins has quit IRC
19:22b3n6 has joined #ltsp
19:36b3n6 has quit IRC
19:55gentgeen__ has quit IRC
19:58gentgeen__ has joined #ltsp
20:40chupacabra has joined #ltsp
21:01tech_dvo has joined #ltsp
21:03
<tech_dvo>
good day!
21:03
Error: failed to connect to NBD server --- ?
21:09tech_dvo has quit IRC
21:10MorningSon has quit IRC
21:28tech_dvo has joined #ltsp
21:30nutron has quit IRC
21:40beginer has joined #ltsp
21:43tech_dvo has quit IRC
21:45vmlintu has quit IRC
21:59tech_dvo has joined #ltsp
22:13tech_dvo has quit IRC
22:26tech_dvo has joined #ltsp
22:29nutron has joined #ltsp
23:05mistik1 has quit IRC
23:06mistik1 has joined #ltsp
23:15tech_dv has joined #ltsp
23:18alexqwesa has quit IRC
23:19tech_dvo has quit IRC
23:25tech__dav has joined #ltsp
23:25alkisg has joined #ltsp
23:28tech_dv has quit IRC
23:31
<muppis>
alkisg, had to try it when I get back to home.
23:31
<alkisg>
np. Good morning all.
23:32
<muppis>
Good moorning.
23:41cyberorg has joined #ltsp
23:46alexqwesa has joined #ltsp
23:53
<tech__dav>
alkisg: how to bypass pxelinux?
23:53
<alkisg>
tech__dav: please ask on the channel, not specific persons. What do you mean? You want to load the kernel directly?
23:54
<tech__dav>
sorry, I was reading previous chats --
23:55
you taught me about gpxe -- ctrl+b ..... to load the kernel by passing the pxelinux before...
23:55alexqwesa_ has joined #ltsp
23:56tech_dv has joined #ltsp
23:57
<alkisg>
tech__dav: use google translate on that forum post: http://users.sch.gr/alkisg/tosteki/index.php?topic=1451.msg24373#msg24373
23:58alexqwesa has quit IRC