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


Channel log from 15 October 2010   (all times are UTC)

00:13M4gic5t0rM has quit IRC
01:22alkisg has joined #ltsp
01:23feinbein_fred has joined #ltsp
02:04map7_ has quit IRC
02:57maslak has quit IRC
03:01Da-Geek has joined #ltsp
03:30biomorph_ has joined #ltsp
03:36biomorph_ has quit IRC
03:36biomorph_ has joined #ltsp
03:38biomorph_ has quit IRC
04:55
<zamba>
anyone hear about itcan?
04:55
http://resources.adellgroup.net/itcan/
04:56
heard*
04:59alkisg has quit IRC
04:59alexqwesa has quit IRC
05:29alkisg has joined #ltsp
06:06alkisg has quit IRC
06:11ogra_ac has quit IRC
06:12ogra_ac has joined #ltsp
06:16alkisg has joined #ltsp
06:23pmatulis has joined #ltsp
06:28ogra_ac has quit IRC
06:28ogra_ac has joined #ltsp
07:10MorningSon has joined #ltsp
07:39mgariepy has joined #ltsp
07:44
<shamino>
ok, what am i doing wrong now.. i am trying to build kiosk-enviroment with command ltsp-build-client --chroot i386kiosk-new --kiosk and it builds normal desktop environment.
07:53mikkel has joined #ltsp
07:54wwx has quit IRC
07:58tomak has joined #ltsp
07:58
<tomak>
Hi to everyone
07:59
does anyone use K12 Fedora ?
08:01* Appiah watches tumbleweed fly across the #ltsp channel
08:07Gadi has joined #ltsp
08:20wwx has joined #ltsp
08:29pthsWork has quit IRC
08:31Da-Geek has quit IRC
08:31ogra_ac has quit IRC
08:35ogra_ac has joined #ltsp
08:37CrypTom has joined #ltsp
08:38
<CrypTom>
Hi all, I've installed a kubuntu maverick server, installed the ltsp-server package (not standalone as we already have a dhcp server), i built the client, but when I boot the client, it always hangs
08:39
I changed /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default (I removed quiet and splash) and the client hangs after
08:39
* Starting LTSP client......
08:39
any ideas?
08:39alexqwesa has joined #ltsp
08:40
<CrypTom>
after two minutes or so I get further lines telling me that "task lsb_release:xyz blocked for more than 120 seconds"
08:40komunista has joined #ltsp
08:41
<CrypTom>
a bit higher up, I see a line: "[ 38.359132] nbd0:Hangup"
08:41
<alkisg>
Ugh, sounds like nbd-proxy...
08:41
https://bugs.launchpad.net/ltsp/+bug/589034
08:41
<CrypTom>
and later "unknown partition table"
08:42
alkisg: ok, I'm reading, thanks
08:44
alkisg: I've got no file named /usr/share/initramfs-tools/scripts/ltsp_nbd
08:44
on the server
08:44
<alkisg>
CrypTom: that's in the chroot
08:44
<CrypTom>
ok
08:45
<alkisg>
After you apply the patch, run sudo chroot /opt/ltsp/i386 update-initramfs -u && sudo ltsp-update-kernels
08:45
<mgariepy>
just add the option nbd_proxy=false in /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default
08:45
<alkisg>
Ah my code is in maverick, right :D
08:45
<mgariepy>
yeah :)
08:45
<alkisg>
Forgot about that, still in 10.04... ;)
08:45
<ogra_ac>
pfft so outdated
08:45
<alkisg>
CrypTom: ^^^
08:45
ogra!!!
08:46
Long time no ...read
08:46
<ogra_ac>
hey :)
08:46
<alkisg>
How are your ARMs?
08:46
<ogra_ac>
heavy :)
08:46
pandaboard.org is up ...
08:46
and i just registered an ltsp spec for UDS
08:47
(i hope i dont collide with anything stgraber was planning, i searched but couldnt find anything)
08:47
<mgariepy>
alkisg, i'm currently trying to debug nbd-proxy, but with the debugging, it doesn't fall into the same raise condition, my thin client has been rebooting for like 12 hours successfully...
08:47
<CrypTom>
ok thanks, so I'm not using the patch, just the nbd_proxy=false
08:48
<alkisg>
CrypTom: right
08:48
mgariepy: yeah I know I haven't been able to reproduce that here either, but I've heard many teachers having the same problem so we completely disabled nbd_proxy in Greece so I don't have any feedback about it anymore...
08:49
<mgariepy>
yeah
08:49
i know, in the office, we have like 20 or so thin client and on a mass reboot 2 of them failed to boot.
08:49
<alkisg>
I've seen "negotiation errors", "squashfs errors" etc, all solved after removing nbd-proxy
08:49
<ogra_ac>
alkisg, also http://ograblog.wordpress.com/2010/09/30/ubuntu-on-arm-the-best-since-sliced-bread/ :)
08:50
<alkisg>
ogra_ac, any netbooks we can buy with ARM? Or is it only for thin clients / embedded stuff etc?
08:50
<mgariepy>
with the version of nbd-proxy i have for debugging, it gives me like 800k of data per boot but doens't fail at all.
08:50
<alkisg>
Oooh funky blog theme :P
08:50
<ogra_ac>
alkisg, you can buy the ac100 and install ubuntu on it
08:50
<alkisg>
mgariepy: try rsync -an /rofs /tmp after you boot a client
08:50
<ogra_ac>
(thats what i'm typing on atm)
08:50
<alkisg>
And check syslog for errors
08:50
(fat clients work better for this, bigger images)
08:51
-an ==> read but not write
08:52
<mgariepy>
ok i see
08:53
doing it right now :)
08:53
<CrypTom>
thanks, it worked!!!
08:53
<alkisg>
mgariepy: also try re-enabling compression, I think it makes the problem more obvious (compression doesn't give problems to clients without nbd-proxy, in my experience, only to those with nbd-proxy)
08:54
CrypTom: if you please, report the nbd_proxy=false workaround for maverick+
08:54
(in the bug report)
08:54
<mgariepy>
i don't use compression here.
08:55
<alkisg>
mgariepy: yes, that makes problems *less* frequent, that's why I'm suggesting to reenable it...
09:01
<CrypTom>
alkisg: done
09:02
<alkisg>
nice :)
09:07
<CrypTom>
bye, thanks!
09:07CrypTom has quit IRC
09:11vongrippen has joined #ltsp
09:18ogra_ac has quit IRC
09:19ogra_ac has joined #ltsp
09:19Uzzi has joined #ltsp
09:19
<Uzzi>
hi
09:26
<mgariepy>
alkisg, ok i'm starting to tests with the compression on.
09:28
<alkisg>
mgariepy: isn't it possible to send a patch to upstream nbd-client for whatever feature nbd-proxy adds to it?
09:28
I think the maintainer is very cooperable, he already implemented the "multiple roots" option we asked him...
09:32
<mgariepy>
i will discuss that with julien an stgraber on monday, but anyway atm the code is not working relaibly enough
09:34
<alkisg>
mgariepy: what's the main feature of nbd-proxy? Reconnect on disconnect? Multiple trasmits?
09:34
(i mean if the first transmission fails)
09:35
Or changing servers for ltsp-cluster?
09:35* alkisg has no clue about what nbd-proxy actually does..
09:35
<mgariepy>
nbd-proxy implement, reconnection on disconnect after several minutes, and put the FS in iowait when disconnected
09:35
;)
09:36
<alkisg>
Doesn't nbd-client -persist support the reconnection part?
09:36
<mgariepy>
and resume transfert of fs access on reconnexion
09:36
no
09:36
not correctly
09:36
if you pkill the process it will reconnect
09:36
but if it take 20 second it will fail
09:36
<alkisg>
OK, both sound like they would be welcomed upstream, as much as I can tell... at least with a --root-file-system switch or something like that
09:37
And maybe the developer there could help with any current issues nbd-proxy has
09:38
<mgariepy>
and every file that is accessed on the thin client when disconnected will have a squashfs error and you need to clear the cache for those files.
09:42
we used nbd-proxy because nbd is a kernel module and the time would have been too long to have it upstream
09:43
<alkisg>
Understandable, but now that there's no time pressure, it'd be good to have 1 tool for nbd connection instead of needing 2 of them
09:43
Better speed (no localhost transfers), lower memory usage, maybe even easier bug fixing because it would be 1 app instead of 2 etc...
09:44
<mgariepy>
yeah but there is always time presure ;) we do not want to maintain a kernel module for 2 years
09:46
<alkisg>
There is an nbd maintainer, you would just send a feature request / patch... imho...
09:47
Anyway, bye all :)
09:47alkisg has quit IRC
09:49ogra_ac has quit IRC
09:53ogra_ac has joined #ltsp
09:53Sorinan has joined #ltsp
09:55
<Sorinan>
Why an app like "VirtualBox Manager" running on server would crash only on one terminal, and not on all others? There's no difference in the config...
09:55jhutchins_lt has joined #ltsp
10:06jhutchins_lt has quit IRC
10:09jhutchins_lt has joined #ltsp
10:13Sorinan has quit IRC
10:18alkisg has joined #ltsp
10:45[GuS] has joined #ltsp
10:45[GuS] has joined #ltsp
10:46Uzzi has quit IRC
10:47vongrippen has quit IRC
10:52vongrippen has joined #ltsp
11:04
<alkisg>
mgariepy: are the nbd-server/nbd-client in the same source tree as the nbd kernel module? I think they're separate, userland tools... http://nbd.git.sourceforge.net/git/gitweb.cgi?p=nbd/nbd;a=shortlog
11:04
So even if the nbd-server/client maintainer doesn't accept the reconnection patches (I don't see why he wouldn't), couldn't those be maintained as ubuntu specific patches in a debian/patch directory, for cpu / speed / memory / stability reasons, instead of maintaining a wrapper tool?
11:05kusanagi has quit IRC
11:06
<alkisg>
(another handy patch would be to have a named "swap" export, so that we could merge nbdswapd and nbdrootd and use the IANA assigned port for nbd-server)
11:08sweetpi has joined #ltsp
11:29alkisg has quit IRC
11:40Mava has joined #ltsp
11:57
<Kyle__>
Ahh drat, alk just left.
11:58
Anyone here using NFS4 on fat clients?
12:06feinbein_fred has quit IRC
12:38wwx has quit IRC
12:38wwx has joined #ltsp
12:38[GuS] has quit IRC
12:51sweetpi has quit IRC
12:59vagrantc has joined #ltsp
13:30vagrantc has quit IRC
13:31heru has joined #ltsp
13:31
<heru>
anybody home....
13:32
i was wondering if anyone was familiar with a kernal panic issue on the Dell Optiplex 755 and 760
13:33
we are actually using 4.2 so forgice us
13:37
forgive us
13:39
<atkuepker>
we run optiplex 2x0, but we're on Ubuntu w/ LTSP5.x
13:44
<johnny>
heru, forgive is one thing.. but getting help for that is nearly impossible
13:44
few people around here have ever used ltsp 4.2
13:44
might wanna post on the mailing list instead
13:45
<heru>
yeah i figured.... have you had any driver issues with the network card?
13:45
on the optiplex 755
13:52vongrippen has quit IRC
14:23
<abeehc>
still out of scope for this chan
14:23
not gonna guess the nic chipset; at the least it would make sense to post what lspci says
14:25
<Kyle__>
heru: what type of kernel panaick? Can you give us a snippet of the log?
14:26
heru: I had an issue with my fat clients on optiplex gx620s.
14:26bobby_C has joined #ltsp
14:31
<heru>
okay
14:32
it posted that it could not determine the type of nic
14:32
it advises us to add the info "option-129" in the dhcpd.conf
14:33
<Kyle__>
heru: Oh. Well if it can't find the type of nic, then it sounds more like a kernel/driver issue than anything else, but I'm not sure really.
14:34
<heru>
that is what i thought.... so we need to determine the driver/chipset for this machine
14:34
<Kyle__>
heru: what does lspci show you?
14:38heru has quit IRC
14:39
<Kyle__>
Humm, it appears to have shown him the futility of his ways....
14:41
<abeehc>
:\
14:42
<Kyle__>
abeehc: You're right, it was out of scope, but still may have been fun to dig into/help fix.
14:43* Kyle__ has a perverse sense of fun.
14:44* Gadi puts money on intel nic with e1000e driver that never existed back in LTSP 4.2 days
14:45pmatulis has quit IRC
14:45mistik1 has quit IRC
14:45
<Kyle__>
Oooh.
14:46
I never had the money, or worked with places that would spend the money, for a good enough card to be on the unsupported list in the days 4.2 would have been used.
14:48
<Gadi>
well, it wasn't an unsupported list - 4.2 had a niclist of what it could figure out on its own
14:48
<atkuepker>
Kyle__, no kidding. back then, was happy to be able to buy $15 3com SOHO NICs and not make do with used NICs.
14:49
<Gadi>
4.2 was very minimal
14:49
it had a "niclist" in the initramfs that served as its lookup table
14:49
(4.2 predated udev by and large)
14:49* Kyle__ still remembers being happy about finnaly getting rid of the last ISA nics he had to deal with.
14:50mistik1 has joined #ltsp
14:53mgariepy has quit IRC
15:05
<Kyle__>
Is there an easy or at least streightforward way to enable an autologut/timeout for clients?
15:08
<johnny>
read lts.conf manpage
15:08mgariepy has joined #ltsp
15:08
<johnny>
some variables there
15:11
<atkuepker>
assuming you've installed the ltsp-docs package =)
15:12
<abeehc>
i've not seen anything in lts.conf that will acheive that, Kyle__
15:12
sounds like it should be a function of gnome-session i dunno
15:15
<Kyle__>
I must have missed them.
15:19
johnny: My version only lists some things abut an LDM timeout, to auto-login if no one has logged in yet.
15:20
<johnny>
oh.. logout .. i think alkisg had something for that
15:20
i only saw "autolog" hah
15:21rad4christ has joined #ltsp
15:21
<rad4christ>
Hey guys, hope I can get a really quick bit of help
15:21
<Kyle__>
Well, honestly I have two logout issues. one I need something that will stop/kill a session that's gone on too long, and two, I need it to run a specific script as part of the logout (to cleanly shutdown a running virtualbox session(.
15:22
<rad4christ>
I have 10.10 installed iwth an LTSP server, all is working, except I've installed Firefox and flash locally. when I try to run it, I get an error that Firefox is already running
15:22
And BTW stgraber, very awesome work on this!
15:24
<Kyle__>
rad4christ: Well is the same user already running firefox on a different station? Either that, or firefox crashed leaving aroudn it's lock file.
15:24
<rad4christ>
Kyle__: this is the first the users have been logged in.
15:24
And each station (2 at the moment) are under different logons
15:25
<Kyle__>
rad4christ: If (big if), there is nothing important in that users firefox config, blow it away, and re-run firefox. See how it goes.
15:25
<rad4christ>
IT's a test box, I'm looking to roll it tonight, so there's no info I'm worried to lose.
15:25
<Kyle__>
rad4christ: If your users will be logging into multiple sessions at once, know that this could be an issue. google-chrome has a --temp-profile option, which is useful for simultanious login issues.
15:26
rad4christ: test box? Then rm -rf ~/.mozilla and try and launch it again :)
15:26
<Gadi>
Kyle__: autologout would be up to the desktop environment to handle - so, I would have a look into gnome, if that is what you use
15:26
<Kyle__>
Gadi: OK, I'll dig through it.
15:26
Gadi: I'm not a huge fan of how gnome tries to take over the system, but I guess it makes sense in this one situation :)
15:27
<Gadi>
hehe
15:29
<atkuepker>
Kyle__, don't get me started on the whole black-box DBUS mess.
15:29
<rad4christ>
Kyle__: Worked! Thanks!
15:29
<Kyle__>
NP.
15:29* Kyle__ hates dbus
15:29
<rad4christ>
Flash isn't working, though, doggonit.
15:29* Kyle__ loves linux, and hates ubuntu& gnome...but ends up using them anyway.
15:40rad4christ has quit IRC
15:59Gadi has left #ltsp
16:21gentgeen__ has quit IRC
16:21gentgeen__ has joined #ltsp
16:22mgariepy has quit IRC
16:27tomak has left #ltsp
16:49mikkel has quit IRC
17:05M4gic5t0rM has joined #ltsp
17:27jhutchins_lt has quit IRC
17:30komunista has quit IRC
17:31ogra_ac has quit IRC
17:44ogra_ac has joined #ltsp
17:46johnny has left #ltsp
17:48bobby_C has quit IRC
17:56alkisg has joined #ltsp
18:04
<alkisg>
Kyle__: for cleaning up after logout, see the xexit program that sbalneav made, it's much better than gnome-watchdog
18:05
(in his ppa)
18:06cliebow has joined #ltsp
18:06cliebow_ has joined #ltsp
18:14ltspbot` has joined #ltsp
18:15jhutchins_kc has quit IRC
18:15jhutchins_kc has joined #ltsp
18:28mistik1 has quit IRC
18:39mistik1 has joined #ltsp
18:54alkisg has quit IRC
19:08staffencasa has quit IRC
19:14cliebow has quit IRC
19:27Lumiere has quit IRC
19:35Lumiere has joined #ltsp
19:39pmatulis has joined #ltsp
20:18jhutchins_lt has joined #ltsp
20:24litlebuda has joined #ltsp
20:40jhutchins_lt has quit IRC
20:43litlebuda has quit IRC
21:18heru has joined #ltsp
21:19
<heru>
peace Fam
21:22leio_ has joined #ltsp
21:25vvinet_ has joined #ltsp
21:25wwx has quit IRC
21:25vvinet has quit IRC
21:25leio has quit IRC
21:26wwx has joined #ltsp
21:34
<heru>
has anyone had problems with network cards dropping configuration
21:34sweetpi has joined #ltsp
21:59M4gic5t0rM has quit IRC
22:03M4gic5t0rM has joined #ltsp
22:15ogra has quit IRC
22:23vagrantc has joined #ltsp
22:29ogra has joined #ltsp
22:31vagrantc has quit IRC
22:37heru has quit IRC
22:52MorningSon has quit IRC
23:56cfkane has quit IRC