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


Channel log from 20 November 2007   (all times are UTC)

00:02
<vagrantc>
daduke: no idea. works for me.
00:02
can't test it again till mid december
00:08
<daduke>
vagrantc: you said earlier that it just installs ltsp-client and its deps. Could I add pulseaudio-esound-compat to this list?
00:08
<vagrantc>
daduke: you can try, but i can't see how that would change it
00:09
daduke: ltsp-build-client --early-packages
00:09
<daduke>
vagrantc: it's the best idea I have atm.
00:09
<vagrantc>
daduke: ltsp-build-client --early-packages "ltsp-client pulseaudio-esound-compat"
00:10
actually, i can test the build ... i just don't have a client to boot
00:10
<daduke>
vagrantc: running it now...
00:20
vagrantc: it worked! thanks, now I'm gonna test the build.
00:21
<vagrantc>
seems to be working for me without explicitly specifying pulseaudio-esound-compat, but whatever.
00:21
<daduke>
vagrantc: hmmmm
00:29
vagrantc: boots just fine, but X doesn't start, even with explicit SCREEN_01 = startx. The config's fine though, running X from the console works.
00:33
<vagrantc>
daduke: specifying SCREEN_ 01-06 will likely break
00:34
daduke: default is SCREEN_07=ldm if no screen sessions are defined
00:34
<daduke>
vagrantc: startx, ldm vt0 :0, I get the login window, but logging in takes forever and doesn't work - ldm quits. I've seen this will all new (non-stock etch) versions of ltsp/ldm
00:34
vagrantc: I
00:35
vagrantc: 'm trying 07=ldm now...
00:35
<vagrantc>
only problem i had with LDM with these backports was that ldm didn't display the background wallpaper
00:36
<daduke>
vagrantc: nope, no ldm. The logo is there if I start it like described above, but I cannot log in.
00:36
<vagrantc>
etch hasn't changed much
00:36
don't really understand why it breaks for you in this way.
00:37
<daduke>
vagrantc: also trying to change the session kills ldm: ldmgtkgreet:1906 glib-critical g_strsplit: assertion string != NULL failed
00:38
vagrantc: real bummer, we'd like to use it for our new student rooms. But we need USB.
00:40* vagrantc tries with qemu
00:40
<vagrantc>
i don't release software without testing on real hardware
00:41
unless it's really trivial fixes... but i definitely tested sound, usb cdrom and floppy support with the current etch backport.
00:41
<daduke>
vagrantc: we happily provide all the help we can.
00:41
<vagrantc>
daduke: you're using standard etch mirrors, plus the etch-ltsp-backport ?
00:42
<daya>
daduke, hi , have you tested the sound in ltsp client in debian
00:42
<daduke>
vagrantc: when I try to log in, it stays with 'Verifying pw, pls wait' for 2 mins, then crashes.
00:42
<vagrantc>
ok, so qemu test booted to ldm by default
00:42
daduke: have you removed /opt/ltsp after creating it?
00:43
<daduke>
vagrantc: assertion failed. Yes, we have our own mirror debian.ethz.ch, plus the alioth repo.
00:43
daya: yup, sound works fine, but local USB devices don
00:43
<vagrantc>
daduke: debian.ethz.ch is a full debian mirror?
00:44
<daya>
daduke, How, it works is it in etch?
00:44
daduke, I have tried using backport, as directed by vagranct But still not able to work
00:44
<daduke>
vagrantc: you mean in between the runs? yep, I cleaned it. debian.ethz.ch is a full mirror, we install/update 500+ machines from it.
00:45
<daya>
daduke, I don't know where I have committed mistake,
00:45
<daduke>
daya: this is what we're talking about, there still seem to be some glitches in here...
00:47
vagrantc: where are you? Should we send you some hardware? :-)
00:48
<vagrantc>
daduke: i'm travelling
00:49
daduke: when i get back, i'll have all of freegeek at my disposal (http://freegeek.org) as well as at least one model from disklessworkstations.com
00:49
<daduke>
vagrantc: ok, a second laptop then.
00:50
<vagrantc>
definitely looking for a tiny PXE capable laptop
00:51
<daduke>
vagrantc: I have a 12" Asus, but I never tried PXE booting. Are you in the US?
00:52
<vagrantc>
daduke: mostly
00:52
daduke: whereabouts are you?
00:52
<daduke>
vagrantc: Switzerland (still), but I could send it to you next time I'm in CA..
00:53
<vagrantc>
well, the md5sum on debian.ethz.ch's Packages file is identical to the mirror i've been using ...
00:53
so i don't understand why it works so differently for you
00:53
<daduke>
vagrantc: *relieved*
00:54
<vagrantc>
daduke: CA as in canada, or california?
00:54
<daduke>
vagrantc: neither do I. But our friend daya seems to have some probs too... let's figure it out, shall we?
00:54basanta has joined #ltsp
00:54
<daduke>
vagrantc: Bay Area, CA
00:54
<vagrantc>
daduke: i'll probably mostly be in portland, oregon next year
00:54
<daya>
daduke, `why not, I am seeeking it since some months
00:55
<vagrantc>
i've tried to help daya for quite some time, but i am stumped
00:55
<daduke>
vagrantc: I'm pretty sure USPS can handle that...
00:55
<vagrantc>
and not really able to do proper testing at the moment
00:55
<daya>
daduke, do u remeber have u made some configuration or install pkgs, other that is provided in the readme file,
00:55
<daduke>
vagrantc: let me be your workbench
00:56
daya: nope, I start from scratch every time.
00:56
<vagrantc>
ltsp-build-client --mirror http://debian.ethz.ch/debian/ --apt-keys /etc/apt/trusted.gpg --copy-package-lists --copy-package-cache -
00:56
-extra-mirror "http://pkg-ltsp.alioth.debian.org/debian etch-ltsp-backports main"
00:56
seems to be working fine for me
00:57
<daduke>
daya: if you don't need local USB, you can use plain vanilla etch, w/o alioth. Sound is working.
00:57
<vagrantc>
oh, --copy-package-lists is a custom plugin from 5.0.39debian1
00:57
sound works with esd with the older versions, but not alsa ...
00:58
<daduke>
vagrantc: coming from debian.ethz.ch? A had a colleague compile a backport, but it shouldn't be in there
00:58
<vagrantc>
daduke: that's the exact commandline i ran
00:58
<daduke>
vagrantc: and it boots into X and ldm works (or can't you test?) ?
00:59
<vagrantc>
ltsp-server 5.0.31debian2~0.etch.1 on an etch system
00:59
daduke: it boots into X, and ldm starts up. i can't really test if it logs in.
01:00
<daya>
daduke, plain vanilla etch ,
01:00
?
01:01
daduke, can I download it , and start from the scratch to build ltsp I didn't get it
01:01
<daduke>
vagrantc: trying it here.... says unrecognized --copy-package-lists
01:01
<vagrantc>
daduke: yeah, it's a plugin i grabbed from the newer packages
01:02
daduke: that's the only non-standard thing i have
01:02
<daduke>
vagrantc: right, sorry, forgot.
01:02
<vagrantc>
it basically just copies /var/lib/apt/lists/*Packages into the chroot ...
01:02
should save some downloading time
01:07
<daduke>
vagrantc: the above line (sans --copy-package-lists) is missing pulseaudio again.
01:12
<vagrantc>
daduke: not for me
01:16
<daduke>
vagrantc: I don't know. Another question: in /etc/ltsp.conf on the TS it still says PKG_SOURCE=http://ltsp.mirrors.tds.net/pub/ltsp/ltsp-4.2/, shouldn't this be 5.0 now? I changed it to 5, but I don't see a difference.
01:18
<vagrantc>
daduke: TS is .. the server ?
01:18
<daduke>
vagrantc: terminal server, yes.
01:18
<vagrantc>
daduke: /etc/ltsp.conf isn't used at all.
01:18
daduke: do you have ltsp-utils installed?
01:18
<daduke>
vagrantc: I do.
01:19
<vagrantc>
daduke: ltsp-utils is incompatible with ltsp5
01:19
<daduke>
vagrantc: does it mess things up?
01:19
<vagrantc>
daduke: it can
01:19
<daduke>
vagrantc: purging it RIGHT NOW
01:20
<vagrantc>
daduke: dpkg -l ltsp-server
01:20
<daduke>
vagrantc: ii ltsp-server 0.99debian11 Basic LTSP server environment
01:20
<vagrantc>
daduke: ah.
01:20
this is getting a little clearer.
01:21
<daduke>
vagrantc: ah?
01:21
<vagrantc>
daduke: you should install the backported ltsp-server package
01:21
<daduke>
vagrantc: I thought I did
01:21
<vagrantc>
clearly not :P
01:21
<daduke>
vagrantc: I see that, but why not?
01:22
<vagrantc>
daduke: because you didn't install it?
01:22
<daduke>
vagrantc: I added alioth to the sources.list... ok, let me retry.
01:22Egyptian[Home] has quit IRC
01:23
<vagrantc>
daduke: you followed the instructions on: http://wiki.debian.org/LTSP/Howto
01:23
?
01:23
<daduke>
vagrantc: yep.
01:24Egyptian[Home] has joined #ltsp
01:24
<vagrantc>
daduke: apt-key list | egrep 13D61A99
01:25
<daduke>
vagrantc: I have a suspicion... I added the sources last night, but it's missing now. Our automated distributed configuration system might've overwritten it. Sorry, my bad!
01:26
<vagrantc>
i see. :P
01:26
<daya>
daduke, what u have kept int sources.lst file,
01:28
<daduke>
vagrantc: now we're talking: ii ltsp-server 5.0.31debian2~0.etch.1
01:29
<vagrantc>
daduke: now i know why you were getting the pulseaudio problems ...
01:29
daduke: 0.99debian11 adds esound as an explicit dependency ...
01:29
er, not a dependency ... it includes it in the list of packages to install
01:30
which conflicts with some pulseaudio-esound-compat
01:30
<daduke>
vagrantc: I see. Rebooting now. stand by...
01:30
<vagrantc>
shouldn't need to reboot ... but sure
01:30
<daduke>
vagrantc: the thin client.... I think I should ;)
01:31
<vagrantc>
ah, yes.
01:31
<daduke>
vagrantc: hmm, still no X
01:31F-GT has quit IRC
01:32
<vagrantc>
daduke: what's in /opt/ltsp/i386/etc/lts.conf ?
01:32
daduke: note the comment at the bottom of the Howto
01:32
<daduke>
vagrantc: [Default]
01:32
LOCALDEV = True
01:32
LOCAL_STORAGE = True
01:32
SOUND = True
01:32
XF86CONFIG_FILE = /etc/X11/xorg.alix.conf
01:32
<vagrantc>
LOCAL_STORAGE is unused if LOCALDEV is set.
01:33
<daduke>
vagrantc: well, the cfg in my /etc/X11/xorg.alix.conf works fine, it just doesn't start
01:33
<vagrantc>
SOUND is enabled if the sound stuff is installed
01:35
daduke: how does it work fine if it doesn't start?
01:35
<daduke>
vagrantc: startx on console. Started ldm again manually, logging in still doesn't work.
01:36
vagrantc: I think I'll reinstall everything from scratch, who knows what's still lurking around in my system now....
01:49
<daya>
vagrantc, which mirror it to use that provided in wiki or that you mention earlier,
01:49
ltsp-build-client --mirror http://debian.ethz.ch/debian/ --apt-keys /etc/apt/trusted.gpg --copy-package-lists --copy-package-cache -
01:49
-extra-mirror "http://pkg-ltsp.alioth.debian.org/debian etch-ltsp-backports main"
01:50
<daduke>
daya: I'm not always sure what you're trying to ask, but the above is one line, that's what the \ is for.
01:52
daya: and debian.ethz.ch was just mentioned because it's our mirror. Use the one from the wiki
01:53
<daya>
daduke, :), sorry for misunderstanding, I mean this link is not exact that is provided in the http://wiki.debian.org/LTSP/Howto,
01:53
daduke, I mean is there any difference in using any one of them,
01:53
<daduke>
daya: use the one from the wiki, the mirrors are identical (vagrantc just checked)
01:54
<vagrantc>
daya: --copy-package-cache and --copy-package-lists just save download time
01:55
daya: and using a different --mirror might save download time if the mirror is near you
01:55
<daya>
vagrantc, daduke ok thanks
01:55
<vagrantc>
daya: but it might also be out of date
01:56
daya: though etch doesn't change much, so most mirrors are probably up to date
02:04
<daya>
vagrantc, I have used installed pulseaudio and managed to use pulseaudio in xmms as output sound plugins, when I run pulseaudio in client in tty1 it says main.c:"Daemon startup fails"
02:05
vagrantc, But sound in server is working with pulseaudio
02:05
<vagrantc>
daya: i have only used pulseaudio with LTSP as described in the README.Debian.sound
02:10
<daya>
vagrantc, I am a bit confused what libasound2-plugins does, Is it only the plugin for alsa to use pulse,could you please clarify it
02:14cyberorg has quit IRC
02:16
<vagrantc>
daya: it is the plugin for alsa to use pulseaudio, yes.
02:25cyberorg has joined #ltsp
02:32
<daya>
vagrantc,But we don't install pulseaudio and don't run pulseaudio process in server,
02:51
<vagrantc>
daya: yes, it's not necessary if you use libasound2-plugins
02:51
the pulseaudio libraries are installed
02:51
the binaries aren't needed
02:52
<daya>
vagrantc,ye
02:52
vagrantc, But I am too tired why it don't work in case of mine :(
02:58
<daduke>
vagrantc: yihaaaaa! I can log in!
02:58
vagrantc: even the background image works :)
02:58
<vagrantc>
daduke: really?
02:59
the background image never worked for me
02:59
<daduke>
vagrantc: yup. sound too.
02:59
<vagrantc>
daduke: now that you're an expert, maybe you can help daya :)
03:00
<daduke>
vagrantc: drumroll..... plugging in the USB stick now....
03:01
<daya>
daduke, vagrantc yes great !
03:01
<daduke>
vagrantc: WORKING!
03:01
vagrantc: thanks a lot!
03:02
<vagrantc>
daduke: glad we finally figured it out :)
03:03
<daduke>
vagrantc: me too, believe me.
03:03
<vagrantc>
hmmm... i guess 5.0.39debian1 doesn't really introduce many significant changes
03:05
<daya>
daduke, vagrantc can we now once again debug my prblem,
03:06
<daduke>
daya: what did you do, and what is your problem?
03:06
<daya>
daduke, sound problem in client,
03:07
<daduke>
daya: this only answers the second part of my question.
03:07
<daya>
daduke, I have installed LTSP in etch by follwing instrction in http://wiki.debian.org/LTSP/Howto, the lower one
03:08
daduke, I have configure the server as told in README.Debian.sound
03:09
daduke, and I managed the the player (xmms ) to use alsa as output plugin in client, But still sound don't work
03:11
<daduke>
daya: I did not configure anything, xmms is using esound. It's a bit choppy from time to time, but working. will try README.Debian.sound now...
03:11
<daya>
daduke, hm,
03:20
<daduke>
daya: I followed README.Debian.sound now, configured xmms for alsa, and it works w/o any probs. Not even choppy!
03:20
<daya>
daduke, but it didn't on me, no any sound,
03:21
<daduke>
daya: and you're using the right jack, the alsa devices are not muted?
03:22
<daya>
daduke, yes, I have tested the jack, and no mute , It is fine in ubuntu in the same machine
03:26
<daduke>
daya: check alsamixer again, I've had it before that the standard sound output was muted. The setting does not survice a reboot, hence ubuntu is not proof here.
03:26
<vagrantc>
uploaded some new backported packages: deb http://llama.freegeek.org/~vagrant/debian etch-backports/
03:26
will move them to the pkg-ltsp project on alioth when i get a chance to test them
03:27
<daduke>
vagrantc: .. what's the difference? I'm so happy everything is working now (I can even watch fullscreen video on the client!) that I don't want to break things again. But I'll test them of course.
03:28
<vagrantc>
daduke: mostly just trying to keep it in sync with unstable ... not many differences
03:29
maybe won't even bother to upload a new official backport until i've gotten more bugs fixed in unstable
03:29
<daduke>
vagrantc: ok, I'm being brave...
03:30
<vagrantc>
updated translations for many languages :)
03:35
well, my initial build at least partially worked
03:37
klausade: new backports for etch, if you care to test them
03:37
<daduke>
vagrantc: ok, updated everything, rebooting now....
03:39
<klausade>
vagrantc: do they work for you?
03:39
<vagrantc>
klausade: i'm not fully able to test them
03:39
klausade: not for a few weeks at least ... testing with qemu right now ...
03:40
<daduke>
vagrantc: sound is working, but for local usb I get 'X11 connection rejected because of wrong authentication' in the TS log.
03:40
<klausade>
vagrantc: I'll test them. Is there a changelog, what's new, right now i'm happily using 5.0.8debian3.
03:40
<vagrantc>
klausade: ah, you're still on that!
03:40
klausade: quite a long list of changelog entries, then.
03:41
klausade: i think you had tried one of the newer backports since, but didn't have much luck with it ... ?
03:41
hmmm... i'm getting no X ...
03:42
worked with 5.0.31debian2~0.etch.1 ... but 5.0.39debian1~40.etch.0 seems borked
03:42
maybe qemu is just being very slow ...
03:47
<klausade>
vagrantc: yes, I once tried a newer backport, but got no X.
03:48basanta has quit IRC
03:50
<daduke>
vagrantc: X is working for me, and the local usb probs can be resolved by # env['DISPLAY'] = 'localhost:10.0' in ltspfsmounter and ssh -X -S ${LDM_SOCKET} in S10-delayed-mounter
03:51
<vagrantc>
daduke: this is good news!
03:51
<daduke>
vagrantc: happy to be able to provide some at least...
03:53daya has quit IRC
03:53
<daduke>
vagrantc: I'm _very_ happy now, everything's working fine, and it's even integrated into our LDAP setup.
03:53
vagrantc: thanks again for your help.
03:54
<vagrantc>
daduke: very glad to hear
03:57rafaelcardoso has joined #ltsp
04:06
<klausade>
daduke: what inm ltsp (besides auth) do you have integrated into ldap?
04:08
<daduke>
klausade: well I installed ltsp-server on one of our machines which is integrated into our 500+ machines LDAP setup. Hence my LTSP uses our full system. Very nice.
04:09
<klausade>
daduke: and you have etch installed?
04:10
<daduke>
klausade: yup.
04:12
<klausade>
daduke: i'm also using backported xorg 7.3 in the chroot, gives me all the cool new features from xorg, like mirrored display with intel cards.
04:12
<daduke>
klausade: out for lunch, back in 20 min...
04:12daya has joined #ltsp
04:13
<vagrantc>
klausade: i would expect the newer backports to work decently with newer xorg
04:13
<klausade>
daduke: same here (free lunch with a customer having a POS machines from us)
04:13
free lunch .....?
04:15meduxa has joined #ltsp
04:15
<daya>
vagrantc, Sound is now working , !
04:15
<vagrantc>
daya: excellent!
04:16
daya: what was the fix?
04:17
<daya>
vagrantc, I don't know how it works
04:17
<vagrantc>
ah, well, you'll have fun trying to figure it out next time :)
04:21
<daya>
vagrantc, I have tried it in a new laptop,
04:22
vagrantc, Many many thanks for your cogent help and support,
04:23
vagrantc, I think My clients machine were the main reason
04:23
<vagrantc>
daya: of course :)
04:23meduxa is now known as toscalix
04:23
<daya>
vagrantc, It works on esd, pulseaudio,and alsa, but not in oss,
04:23
vagrantc, when I choose oss, its get played on server
04:23
<vagrantc>
ah, yes.
04:24
daya: that's to be expected
04:24
daya: if it only supports oss, you can use the padsp wrapper to get it to work sometimes
04:25
<daya>
vagrantc, padsp wrapper,is it a package or what ?
04:26
vagrantc, I mean which pacakge should I install to use it,
04:27
<vagrantc>
daya: i don't remember which package it is in ...
04:27
daya: but it's really a last resort
04:28
<daya>
vagrantc, ok
04:28
vagrantc, Now I am curious about usb,
04:28
vagrantc, It is detected in clinet, as I can see from tty1, dmesg, and fdisk -l,
04:29
vagrantc, But how can I use it in GUI client.
04:29
<vagrantc>
daya: pulseaudio-utils is the package that has padsp
04:29
daya: it should get mounted in /media/USER/DEVICE
04:35
<daya>
vagrantc, in gui or tty
04:35
vagrantc, it is not automounted
04:35
<vagrantc>
daya: on the server
04:36
i must get some sleep before the sun comes up
04:36
<daya>
vagrantc, in chrooted env
04:36
vagrantc, just some min,
04:37
<vagrantc>
daya: on the server
04:37
/media/USER on the server
04:37
on the server
04:37
also you can see if it's getting mounted: mount | grep ltspfs
04:37
<daya>
vagrantc, I don't get any /media/USER on the server,
04:37
<vagrantc>
then it's broken
04:38
i was hoping to get to sleep about 5 hours ago.
04:38
good luck
04:38vagrantc has quit IRC
04:44rafaelcardoso has quit IRC
04:46F-GT has joined #ltsp
04:48daya has quit IRC
05:00mcfloppy_ has joined #ltsp
05:03rafaelcardoso has joined #ltsp
05:05
<XTaran>
Hi daduke! *g*
05:06oldwolf has quit IRC
05:06
<XTaran>
.oO( The [IRC] world is small. :)
05:21tsurc has joined #ltsp
05:22
<tsurc>
does anyone know the port numbers need between client and server on an edubuntu system.
05:22
needed*
05:23Guaraldo has joined #ltsp
05:25
<tsurc>
I wrote instructions for a transparent proxy but to make it work I've had to use iptables, thing is locally it works but it stops remote logins from thin clients
05:28toscalix has quit IRC
05:31meduxa has joined #ltsp
05:37tsurc has left #ltsp
05:42cyberorg has quit IRC
05:45K_O-Gnom has joined #ltsp
06:03Q-FUNK has joined #ltsp
06:06subir has quit IRC
06:09rafaelcardoso has quit IRC
06:18rafaelcardoso has joined #ltsp
06:35mcfloppy_ has quit IRC
06:53cliebow_ has joined #ltsp
06:55mcfloppy_ has joined #ltsp
06:55elisboa has joined #ltsp
07:10rafaelcardoso has quit IRC
07:17K_O-Gnom has quit IRC
07:19rafaelcardoso has joined #ltsp
08:02K_O-Gnom has joined #ltsp
08:05daya has joined #ltsp
08:10plamengr has joined #ltsp
08:22ogra has quit IRC
08:24ogra has joined #ltsp
08:30Q-FUNK has quit IRC
08:34daya has quit IRC
08:35daya has joined #ltsp
08:45prpplague has joined #ltsp
08:52tsurc has joined #ltsp
08:54meduxa has quit IRC
08:58sbalneav has joined #ltsp
08:58
<sbalneav>
Morning all
08:58
<ogra>
!s
08:58
<ltspbot>
ogra: "s" is Scotty!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
08:58
<ogra>
he's alive !
08:59
<sbalneav>
heh
08:59
yeah, busy few days.
09:00
Lost a raid unit yesterday, weekend we had a huge snowstorm, so had to get snowtires on, and get the snowblower going
09:00
etc etc
09:03plamengr has left #ltsp
09:03
<ogra>
can you merge my tree ?
09:04
i think warren would like ot have something he can base on
09:05
<warren>
yes please
09:12sonjag has joined #ltsp
09:16
<ogra>
warren, when we taled about polkit you said fedora only uses it hand in hand with selinux ... is that to overcome any kind of weakness in policykit ?
09:16
*talked
09:16
(we're just discussing inclusion in ubuntu again)
09:17
apparently there are some issues with ptrace()
09:17
<warren>
ogra, that doesn't sound right.
09:17
<ogra>
thats what i understood from our conversation
09:17
(but there was sam adams involved ... so dont count on my mind :) )
09:17
<warren>
ogra, we're pretty paranoid about security and don't allow lots of stuff that is unsound by design. SELinux is not meant to be the first line of defense ever.
09:18
<ogra>
yeah, i understood that, but it sounded like you need that second line of defense with polkit
09:18
but i get thats not the case
09:19
<warren>
I'm trying to find somebody with a clue who can answer this.
09:20
Hmm... 2.6.21 seems to never crash the T170
09:20
this might be a newer kernel bug
09:22
<ogra>
my t170 runs fine with 2.6.22
09:22
(thats what we use in the current release)
09:28
where does it die ?
09:28
on kernel unpacking ? in the initramfs ? during mount of / ?
09:28
<warren>
ogra, really early kernel
09:28
ogra, hpet_late_init
09:28
ogra, intermittent kernel memory corruption
09:28
<ogra>
hmm, i dont see that here ...
09:28
<warren>
testing 2.6.22 now
09:28
<ogra>
culd it be your RAM =
09:28
?
09:29
<warren>
ogra, I suspect it is new in 2.6.23
09:29
ogra, yeah I should test that to be sure.
09:29
<ogra>
ah, we dont have any built of .23 yet
09:29
*build
09:29
and i wont compile my own :)
09:31
<warren>
ogra, I asked a few of our desktop people about PolicyKit. They suggested if there are concerns about the security architecture it should be posted on upstream discussion lists. But there are a few general comments...
09:31
<ogra>
i'm following the hal list ...
09:31
<warren>
Like... Red Hat wouldn't work on it if it were not a sound architecture, especially given the entire point of it is to be a security framework it BETTER be secure, even without SELinux.
09:32
<ogra>
but i'm no security expert to judge ... we have people for that i like to delegate such stuff to :)
09:32
<warren>
And I personally would add, it is easier to audit the narrow set of what PolicyKit does, than to audit all the setuid crap that it is meant to replace.
09:33
<ogra>
well, the ubuntu way of using pmount with a ten line-of-code suid sub binary was actually pretty appealing
09:34* ogra likes it if the security relevant pieces are separate and as trivial as possible
09:34
<warren>
ogra, PolicyKit is meant for other scenarios of desktop automation that are more complicated than those simple cases.
09:34
<ogra>
polkit is a beast compared to that ... but then it offers more than mounting
09:34
right
09:35
<warren>
ultimately we need something like this
09:35
perhaps this or another implementation
09:35
but I"m glad somebody is thinking about it
09:36
<ogra>
indeed, it surely will improve
09:37
<tsurc>
I hate to jump in but what ports are/could be used between cleint and server on an edubuntu gutsy server. I'm playing tith a transparent proxy that with an iptables rule stops users from loggin in from a client.
09:37
with*
09:37
<ogra>
tsurc, look in /etc/inetd.conf
09:37
<tsurc>
thanks ;-)
09:38
<ogra>
usually its only 2000 for the nbd image and tftp nowadays, unless you have multiple images
09:38
i wonder if we could ssh-tunnel nbd somehow
09:41
<tsurc>
my rule is: iptables -t nat -A OUTPUT -p tcp --match owner --gid-owner pupils -j REDIRECT --to-ports 8080
09:41
<stgraber>
hmm, you would need to have ssh in the initrd and a way to login on the server without login/pass, so tunneling is possible but far from easy :)
09:42
<tsurc>
but I guess that will stop all tcp traffic to and from the terminal
09:42
<ogra>
stgraber, well, my main concern with the ssh client side stuff is that you can just read the keys
09:43
<stgraber>
yes, if you block port 2000 you'll block all disk access which will basically make your clients to crash, if you just want to block login from running clients blocking ssh (tcp 22) should do it
09:46putty_thing has quit IRC
09:50
<tsurc>
well accepting all traffic to and from port 2000 didn't work. The client gets a login screen and can authenticate but the screen goes black with a mouse cursor after they have logged in, I guess its stopping the desktop session from starting up properly
09:53
Ah... that rule will redirect all traffic on "lo" as well as actual "outbound onto the internet" (eth0) traffic. Am I right?
09:58
thanks all, I'll have a think tonight and update the edubuntuProxy wiki page accordingly
09:59tsurc has left #ltsp
10:02
<ogra>
hmm
10:03
i wonder if he used LDM_DIRECTX :)
10:03
he would have to let X traffic through then
10:09elisboa has quit IRC
10:12putty_thing has joined #ltsp
10:16
<rafaelcardoso>
via unichrome vga hava any configuration to make more faster o
10:18sonjag has quit IRC
10:40daduke_ has joined #ltsp
10:42daya has quit IRC
10:42
<warren>
*sweet*
10:42daya has joined #ltsp
10:42
<warren>
really close to making LTSP install and update work with selinux enabled
10:43
<ogra>
heh
10:43
<warren>
ogra, if we get all aspects of the distro working by default with selinux enabled, then people aren't tempted to turn it off.
10:43
<ogra>
we dont have any users in a choot and root is locked :)
10:43
<warren>
ogra, selinux denials are bugs.
10:44
ogra, well I'm trying to make this in a general way, so the same chroot can be used for netbooted managed workstations.
10:44
<ogra>
in the nfs implementation there are only a handfull of files writeable additionally
10:44
<warren>
ogra, the nearer term issue however was selinux was denying the install/update itself
10:44
<ogra>
ah
10:44
yeah, we have an ongoing request for fat clients
10:45
<warren>
Fedora 8's selinux is really getting good, the default policies are very fine grained, and we keep adding more every day.
10:45
<ogra>
the thing is that you need a network auth mechanism (at least the client side) which makes you depend on stuff like ldap etc ... something we liked to avoid unpstream yet
10:45
<warren>
More developers and users know how to submit new rules for default inclusion.
10:45Q-FUNK has joined #ltsp
10:46
<warren>
ogra, right, I wont have all aspects of net managed workstations for quite a while.
10:46
<ogra>
right
10:46
<warren>
ogra, but I am getting the chroot install part correct now, so it isn't a problem later.
10:46
<ogra>
i'm pondering suchg a setup since a while ... with one "golden workstation" that has full rw access to the chroot
10:46
so you can maintain the whole lot from that machine
10:47
<warren>
Our workstation idea would be readonly filesystem, completely centrally managed, with data dirs mounted over the network. On top of that centrally managed selinux for another layer of security.
10:47
It is completely doable today to have your "golden workstation"
10:47
that is easy even if it is full rw
10:48
<ogra>
right
10:48
but i havent come around to find the time to work out a proper default setup that works out of the box
10:48
we dont have a redhat directory server in ubuntu :) (yet)
10:49
so i'D have to tackle that part as well, i dont want to provide half solutions
10:49staffencasa has joined #ltsp
10:49
<warren>
Somebody added debootstrap to Fedora yesterday
10:49
<ogra>
cool !
10:50
<warren>
It would be cool if the equivalent were in Ubuntu.
10:50
<ogra>
botstrapping fedora chroots ?
10:50
<warren>
ogra, no, it is to install debian chroots on top of fedora.
10:50jammcq has joined #ltsp
10:50
<ogra>
ah
10:50
what is the equivalent to debootstrap in fedora ?
10:51
<warren>
ogra, we have too many options, unfortunately.
10:51
<ogra>
ah
10:51
<warren>
ogra, all of our tools assume you already have a fedora system in /. It might be simpler to write a tool that downloads the base RPMS and create a new chroot, then use yum from that chroot to install the rest.
10:52
<ogra>
it would be cool to hackup debootstrap to build fedora chroots though ...
10:52
<warren>
not sure how debootstrap works
10:52
i'll look into it
10:53
<ogra>
it makes a lot use of dpkg ... not sure rpm can offer all features here ... beyond that it uses apt-get to install additional stuff after the base system is in place, i assume that part can easily be done by yum
10:55rafaelcardoso has quit IRC
11:01
<ogra>
and they refuse to look at our tools ... not written here ... pfft
11:03
<johnny>
:(
11:03
kill RPM
11:03
all over the place
11:03spectra has joined #ltsp
11:03
<ogra>
whoops, that was from a different conversation ...
11:03ace_suares has quit IRC
11:03
<johnny>
hehe
11:04ace_suares has joined #ltsp
11:05daya has quit IRC
11:05
<ogra>
rpm has its advantages and drawbacks, as dpkg has ...
11:05
<johnny>
hmm it would be nice to see an up to date comparison these days
11:05
i have't seen any of the good stuff in RPM since like redhat 5 :)
11:06
before the yum era
11:06
<ogra>
well, its rather the thing on top that counts ...
11:06
apt, yum, yast ...
11:06
<johnny>
yast ... haven't used suse since 8 or something..
11:06
i had it on multi disk dvd back in the day
11:07
or hmm.. really it musta been before 8..
11:08
ogra, is anybody familiar with ldm2 other than scott?
11:08
trying to figure out what my next step to help resolve this bug could be
11:08
i've looked in the code even, but nothin is jumping it, maybe cuz i rarely code in C
11:09
<ogra>
johnny, me a bit ... but i havent had any time yet to look at anything ...
11:09
<johnny>
he suggested i try ssh -X ltspuser@ltspserver
11:09
<ogra>
i'm pretty busy with job stuff atm
11:09
<johnny>
and that worked just fine
11:19rjune__ has joined #ltsp
11:19rjune_ has quit IRC
11:32sepski has joined #ltsp
11:42
<ogra>
sbalneav, https://code.launchpad.net/~ogra/ltsp/ltsp-build-client-gtk
11:42
feel free to play with it :)
11:43
<sbalneav>
Awesome
11:52
<stgraber>
ogra: Looks really good, too bad I don't have a LTSP server around (I get an ugly error as I don't have /opt/ltsp :)). I played a bit with the glade file and think it looks better if you set the mirror URLs, chroot name and base dir as filling horizontal space
11:53
as their initial width was really really too small here (1680x1050 display so they had plenty of free space around them)
11:53
<ogra>
feel free to patch it :)
11:53
i'll happily merge
11:54
<stgraber>
ok, will send you the patch
12:04vagrantc has joined #ltsp
12:06ace_suares has quit IRC
12:06
<stgraber>
ogra: http://www.stgraber.org/download/ltsp-build-client.diff, updated .glade + check if /opt/ltsp exists, if not show an error message (otherwise it's gtk who whines :))
12:06
<vagrantc>
!serversize
12:06
<ltspbot>
vagrantc: "serversize" is http://wiki.ltsp.org/twiki/bin/view/Ltsp/ServerSizing
12:09sgonzalez has quit IRC
12:10sgonzalez has joined #ltsp
12:11ogra1 has joined #ltsp
12:13
<cliebow_>
sbalneav: know anyone that misplaced by-the-sea 2005 sweatshirt?
12:23ogra has quit IRC
12:36Guaraldo has quit IRC
12:40
<vagrantc>
wahoo. finally figured out how to get qemu bridged networking at least sort of working
12:42
ironically, it was docs on getting bridged networking for virtualbox that helped me figure it out :)
12:42
<ogra1>
heh
12:42
why dont you use vitrualbox ?
12:42
i find it way more comfortable
12:42
<vagrantc>
just installed the etch backport today :)
12:43
but i'm not eagre to use something that requires kernel modules ...
12:43
ogra1: or are the kernel modules optional?
12:43
<ogra1>
i dont think so
12:44
<vagrantc>
and there's both host and guest modules?
12:44
<ogra1>
but virtualbox on my lappie is faster than the ebox 1000 :)
12:44
you make it what you want it
12:44
<vagrantc>
?
12:44
<ogra1>
its just a vm ...
12:45
i have one wiht a full edubuntu install and one that netboots from the other one and actes as client
12:45
<vagrantc>
well, if the vm requires installing kernel modules inside the vm ...
12:46
<ogra1>
i never found that necessary
12:46
indeed you can
12:46
<vagrantc>
ok, that's what i was trying to find out :)
12:47
<ogra1>
btw, in case you missed it: https://code.launchpad.net/~ogra/ltsp/ltsp-build-client-gtk
12:48
will need some tweaks for debian i guess
12:48prpplague has left #ltsp
12:48
<ogra1>
(i.e. a switch to skip the image creation at the end etc)
12:48
<vagrantc>
i gotta start testing the NBD root stuff again ...
12:49
ogra1: why doesn't it just use the normal plugins?
12:49
<ogra1>
it calls ltsp-build-client
12:49
<vagrantc>
which doesn't create an image on debian
12:49
<ogra1>
but with --skipimage ... and then runs ltsp-update-image separately
12:49
<vagrantc>
at least, not yet
12:49
ah.
12:50
<ogra1>
it currently emulates the ubuntu default setu but split into parts to get proper progressbar output
12:51
<vagrantc>
right
12:52
<dan__t>
hi
12:52* ogra1 currently pokes around on ltsp-image-shell, a tool that builds a unionfs from image and tmpfs and chroots into it ... on exit irt rebuilds the image
12:52
<dan__t>
sorry, didn't have a lot of time to dive into that last night, vagrantc.
12:52
Had to replace the power steering pump in my truck heh
12:52
<vagrantc>
no problem
12:53
ogra1: a utility like that could really alliviate one of my biggest annoyances with the whole NBD image stuff
12:53Guaraldo has joined #ltsp
12:53
<ogra1>
i also want to have a small status panel icon for ltsp sanity checks ;)
12:54
<vagrantc>
i'd also like to see what the performance of a plain ext2 image or whatever ... how much is squashfs and how much is NBD
12:54
<ogra1>
nbd is surely more involved here
12:54
but squashfs reduces the bandwith to 1/3 of a unsquashed fs
12:55
<vagrantc>
that big of an improvement, eh?
12:55
<ogra1>
well, unsquashing happens on the client
12:55
<vagrantc>
right
12:55
<ogra1>
so you trade bandwith against cpu power
12:55
<vagrantc>
right ... slower clients might boot slower
12:55
<ogra1>
exactly
12:56
still faster than nfs mounted ones though
12:57
shriek
12:57* ogra1 was brave today and thougt he'd upgrade to hardy ....
12:58
<vagrantc>
ogra1: y'all pulled in debian's ltsp
12:58
<ogra1>
Get:534 http://archive.ubuntu.com hardy/main ltsp-server-standalone 5.0.39debian1 [37,9kB]
12:58* ogra1 cries
12:58
<vagrantc>
ogra1: i've been telling you that for weeks
12:58
it's probably the first time debian had a newer version
12:58
<ogra1>
yeah, i didnt bother since i ran gutsy and scott didnt merge my branch yet to have something to package
12:59
no, there was a time around edgy ...
12:59
but we didnt sync back then
12:59
<vagrantc>
the whole source split before merging debian changes makes for conflict hell.
12:59
20-50 conflicts on any given merge.
13:00
<ogra1>
well, ugly but solveable
13:00
cant you try to make the same split first ?
13:00
that should reduce it significantly
13:00meduxa has joined #ltsp
13:00
<vagrantc>
i've solved them once, and then will have to solve them again... due to the "dists" shuffling
13:01
basically, i'm waiting to see a "stable" upstream before i try to merge again
13:01
<ogra1>
well, i'D like to see released whats there now
13:01
we can fix the issues on the go
13:01meduxa has quit IRC
13:01
<vagrantc>
i would *really* prefer to get debian merged before a release
13:02
<ogra1>
thats what i mean with on the go ...
13:02
either push your merge now or wait for a release
13:02
<vagrantc>
where do i push it?
13:02
<ogra1>
i dont mind either
13:03
i just want code to package
13:03
<vagrantc>
i've published my changes and asked for them to be merged
13:03
<ogra1>
ah
13:03
<vagrantc>
ogra1: your getting rid of the "dists" dir hasn't been merged yet either
13:04
<ogra1>
sbalneav, ??
13:04* ogra1 only has until dec 13th to get a stable source in
13:04
<vagrantc>
or wait ... maybe i'm wrong here
13:05
nope. still not merged
13:05* ogra1 missed the LP codebrowse functions
13:05
<ogra1>
*misses
13:05
<vagrantc>
i'm hesitant to work on LTSP until i see where LTSP is going ...
13:06
<ogra1>
well, i thought we were pretty much done with the changes to the source so far
13:06
<vagrantc>
because resolving 20-50 conflicts every merge is not my idea of fun.
13:06
<ogra1>
right
13:07
<vagrantc>
i think there were two changes you won't want for ubuntu ... dhcpd.conf and ... something else
13:07
ah yes, i only include the ltsp and Debian themes for ldm
13:08
<ogra1>
well, thats easily changed :)
13:08* vagrantc pushes again for a separate source package for additional themes
13:09* ogra1 isnt opposed ... but would like to see a proper ltsp theme first
13:09
<vagrantc>
LDM was something like 2MB with all the themes, and about 150k with just the "ltsp" and "Debian" themes
13:09
<ogra1>
68 lurkers in this channel
13:09
<vagrantc>
ogra1: have you seen my ltsp theme?
13:09
<ogra1>
IS THERE NO ARTWORKER ?? !!
13:10
vagrantc, got a screenshot
13:10
?
13:10
<vagrantc>
it's nothing amazing, but i think it'll ddo
13:10
<Guaraldo>
jammcq: pastebot id down on ltsp-br... :-/ Can you help us?
13:10
<vagrantc>
ogra1: no screenshot ... but it's just the LTSP.org logo from the website on a blue background
13:11
<ogra1>
uuh
13:11
<vagrantc>
ogra1: it's really small, and i would say passable
13:11* ogra1 is no friend of the blue :)
13:12
<vagrantc>
ogra1: well, for the "ltsp" theme, i tried to emulate the ltsp website :P
13:12
<ogra1>
right
13:12
<vagrantc>
i suspect most distros will probably make their theme default anyways
13:12
<ogra1>
yep
13:12
<vagrantc>
it's kind of just an example
13:13
and for that, i think a small example theme is perfect
13:13
<ogra1>
indeed
13:14
actually i'd just like to have LTSP overall a more modern look
13:15
<vagrantc>
ogra1: well, if it's any consolation, the background image isn't just straight blue ... it fades at a diagonal :)
13:15
<ogra1>
there are so many good artworkers out there
13:16
oooh, sexy :)
13:19* vagrantc is happy to have qemu doing etherboot/pxe boot
13:19
<vagrantc>
might be good to update the qemu-ltsp script
13:20
although maybe i should just document it
13:20
since the user-mode networking requires no special privledges... even if it doesn't fully test the entire setup
13:21
<warren>
user mode networking can't test PXE
13:22
or any of the netboot
13:23
<vagrantc>
right, that's the downside
13:24
hence, "doesn't fully test the entire setup"
13:25
hrm. last night i was getting X with ltsp in qemu ... today ... just a black screen
13:25lns has quit IRC
13:38
<vagrantc>
user registration with virtualbox? no thanks.
13:47nubae has quit IRC
13:47
<stgraber>
vagrantc: are you running the OSE or standard version ? IIRC only standard (the one with some proprietary modules) ask for registration
13:49Guaraldo has left #ltsp
13:53
<vagrantc>
stgraber: i'm running the ose version from backports.org (based on packages from debian.org)
13:54staffencasa has quit IRC
13:55
<stgraber>
hmm, so it should be OSE ...
14:00Guaraldo has joined #ltsp
14:00Guaraldo has left #ltsp
14:01joebake1 has joined #ltsp
14:02moquist_ has joined #ltsp
14:16joebaker has quit IRC
14:31moquist has quit IRC
14:33Q-FUNK has quit IRC
14:35shogunx has quit IRC
14:35mcfloppy_ has quit IRC
14:35mcfloppy_ has joined #ltsp
14:40daduke_ has quit IRC
14:58Drakeweb has left #ltsp
15:05
<cliebow_>
jammcq know anyone that misplaced by-the-sea 2005 sweatshirt?
15:07
<jammcq>
cliebow_: did you end up with the extra?
15:08
err, misplaced one?
15:08moquist_ is now known as moquist
15:08* jammcq was thinking it might belong to moquist
15:08
<moquist>
sbalneav: hey-a. didja get sshfsplus? any comments?
15:08
jammcq: I found mine at home.
15:08bobby_C has joined #ltsp
15:08
<jammcq>
hmm, maybe dtrask?
15:09
<moquist>
better chance of that.
15:09
<cliebow_>
it was an extra large
15:09
ill ask dtrask
15:09
<jammcq>
well, we know it's not mine
15:09
<cliebow_>
8~)
15:09
<moquist>
heh
15:11
<cliebow_>
i thought i was going crackerdog..id hang mine up..next thing i know..durn thing was downstairs again..
15:13edgarin has joined #ltsp
15:13
<cliebow_>
til finally they both appeared at once..reminded me of esiotrot..first it was huge..next thing i know it was little again..
15:14* moquist laughs out loud
15:16mcfloppy_ has quit IRC
15:17mcfloppy_ has joined #ltsp
15:23sepski has quit IRC
15:32cliebow_ has quit IRC
15:37Q-FUNK has joined #ltsp
15:39ogra1 is now known as ogra
15:45sonjag has joined #ltsp
15:56bobby_C has quit IRC
15:59
<sonjag>
Hello! Anyone out there ready/able to help me troubleshoot an issue? I think it's related to fuse.
16:00
<vagrantc>
sonjag: best to just ask your questions and people will respond if they can
16:00
ltspbot: factoids search --values
16:00
<ltspbot>
vagrantc: 'ltsp', 'sbalneav', 'icewm', 'frappr', 'wiki', 'debian', 'edubuntu', 'dhcpd', 'greyscreen', 'ltsp42', 'localdev', 'localdev', 'localdev', 'localdev', 'checklist', 'muekow', 'bestltspdistro', 'serversize', 'wireless', 'sound', 'topics', 'integration', 'lts.conf', 'pastebot', 'bootfloppy', 'ltsp5', 'tarball', 'download', 'monkeys', 'ogra', 'nfs', 'nfsnotresp', 'js', 's', 'troubleshooting', (1 more message)
16:01
<vagrantc>
!question
16:01
<ltspbot>
vagrantc: "question" is if you have a question about ltsp, please go ahead and ask it, and people will respond if they can.
16:01
<sonjag>
vagrantc, okay, here goes :)
16:02
I've been having occasional severe slow downs on my ltsp servers. If I restart, the issue goes away. I am not currently having the issue, but when I do, I can't really troubleshoot because I just need to get them running again.
16:02
So I looked around and saw this...
16:02
lsof | wc -l gave me this response
16:02
lsof: WARNING: can't stat() fuse file system /media/13tcair/usbdisk-sda1
16:02
Output information may be incomplete.
16:03
There are 5 similar responses, then lsof hangs. I have to ctl-c to get back to a prompt.
16:03
dmesg gives me this at the end
16:03
[98838.897576] end_request: I/O error, dev fd0, sector 0
16:03
<ogra>
fuse is owned by the user and not even readable by anyone else (not even root)
16:04
s/fuse/a fuse mount/
16:04
<sonjag>
I don't know if they're related, and I also don't know if they indicate that my server will start to slow to a crawl, but I'm just trying to understand and prevent the issue.
16:04
ogra, is that a command you are recommending?
16:05
<ogra>
there is no command
16:05
its a security feature of ltspfs ...
16:05
sadly it breaks some tools like lsof
16:05
<sonjag>
one last thing... when I run mount, I get quite a few of things similar to this
16:05
/tmp/.13tcair-ltspfs/usbdisk-sda1 on /media/13tcair/usbdisk-sda1 type none (rw,bind)
16:06
ogra, so are you saying that the fact that lsof hangs is not indicating problems? I don't understand what you meant by s/fuse/a fuse mount/
16:06
<ogra>
tmp/.13tcair-ltspfs/usbdisk-sda1 is actually mounted to the thin clients device
16:06
i replaced the first with the latter in my former sencence ;)
16:07
<sonjag>
ogra, it was. That is a removed usb drive. It is not still in the client machine.
16:07
<ogra>
sorry, i'm to geeky sometimes
16:07
hmm, then the mount should have gone away
16:07
<sonjag>
ogra, lol... it took me so long to get that.
16:08
<ogra>
can you find out how the users logged out for which the stale mounts are there ?
16:08
<sonjag>
ogra, what do you mean, how they logged out?
16:08
<ogra>
ldm has a command that forcefully cleans up fuse mounts after logout
16:09
if you dont properly log i.e. just hold down the power button until the client shuts down that cleanup wont hppen
16:10
<sonjag>
they rarely do that, but no guarantees (they're middle school students... everything is game :) I also have some of these lines
16:10
/tmp/.wdowning-ltspfs/floppy0 on /media/wdowning/floppy0 type none (rw,bind)
16:10
There are no floppy drives on the clients
16:11
<ogra>
hmm
16:11
thats weird
16:12
<sonjag>
Wait, that could be a lie. I have a laptop that the user listed there uses. It might have one. I would have to go look at it.
16:13
When my server ran slow on Monday morning, there were about 40 of these /tmp listings from mount. There was also swap space allocated that wasn't there on Friday when I left. No one uses these over the weekend, so I was surprised that the swap space got used.
16:17fernando1 has quit IRC
16:18atistirma has joined #ltsp
16:18
<moquist>
sonjag: do you have trouble with printing? e.g., whole TC session locking up when somebody tries to print?
16:19
<sonjag>
I just went to look a the laptop.... it does not have a floppy drive. It is unusual tho that the only user for whom the floppy error comes up uses a laptop not a thin client.
16:19atistirma has left #ltsp
16:19
<sonjag>
moquist, hi! No, not usually. When the server starts to slow then yes, almost everything seems to lock the client.
16:20
moquist, I was having an error like that due to some fonts that either weren't compatible with cups or my printer drivers.
16:20
I told the students to be boring and use Times :)
16:21
<moquist>
heh
16:21
sonjag: K. I'll keep that in mind when we try to figure out what's going on with ours (in one school).
16:21
I thought the problems had cleared up b/c I stopped hearing about them, but today I found out they haven't been using the TCs because of the above problem. 8-p
16:23
<sonjag>
moquist, to clear it, I had to restart the printer which clears its queue so I assume it's an issue with the printer driver. However, it did stall the terminals so I thought it might have to do with cups.
16:23Q-FUNK has quit IRC
16:23K_O-Gnom has quit IRC
16:25
<moquist>
sonjag: Nah, in our case they click File->Print in OO and the whole TC freezes instantly.
16:25
I haven't been able to reproduce it.
16:25* moquist hates not being able to reproduce this stuff
16:26
<moquist>
I'm scheduling a time when I can go to a classroom and watch it happen.
16:26
<sonjag>
ogra, others... does it seem like the fuse and /tmp results on dmesg and mount, and the mysterious swap allocation would cause a general slow down of server response?
16:27
moquist, definitely try using the fonts that start with ae. They seem to be the most problematic.
16:27
<vagrantc>
sonjag: what's your load average?
16:28
<sonjag>
Is that in iostat? I'm trying not to sound numb ;)
16:28
<vagrantc>
sonjag: cat /proc/loadavg
16:28
sonjag: or you can get it from top
16:29
<ogra>
sonjag, well, the swap files should just vanish after some time and the stale mounts will only have an effect if the user tries to mount the same device again neither should have any infulence on the system speed
16:29
<sonjag>
vagrantc, 2.03 2.24 2.30 3/447 813
16:29
<vagrantc>
not crazy bad ...
16:29
<sonjag>
vagrantc, the system is not loaded now... is this an average over time or right now?
16:30
vagrantc, by not loaded, I mean students are gone... no users but me.
16:30
<vagrantc>
sonjag: the three numbers are averaged over time, but i think only in a matter of minutes
16:30
sonjag: so you've got a load of 2 with no users?
16:30
that's not good
16:30
<ogra>
no, thats pretty bad actually
16:30
<vagrantc>
sonjag: look for runaway processes
16:30
<sonjag>
vagrantc, yikes... newbie time... how?
16:31
<ogra>
if you run top ... whats the top CPU consuming process ?
16:31
(top of the list)
16:31
<sonjag>
vagrantc, gnome-panel, and apparently I have a few users who didn't log out before they left so there are a few logged in but they're not doing anything
16:32
vagrantc, 24632 15jlamb 25 0 33132 12m 9548 R 100 0.2 216:19.21 gnome-panel
16:32
vagrantc, unless it's meaningful to have them there, I can kill their logins
16:33
<vagrantc>
i would suspect those processes can be killed
16:34fernando1 has joined #ltsp
16:34
<sonjag>
vagrantc, I ran the loadavg on another identical server with no users and got this
16:34
24632 15jlamb 25 0 33132 12m 9548 R 100 0.2 216:19.21 gnome-panel
16:34
or actually this...
16:34
0.07 0.02 0.00 1/448 18200
16:34
<moquist>
sonjag: I would totally kill that gnome-panel
16:34
<ogra>
thats rather normal
16:35
<sonjag>
ogra, what's normal?
16:35
<ogra>
the last values you pasted
16:35* moquist set the LTSP servers at one school to reboot at midnight every night...
16:35
<moquist>
I don't want it to be a long-term "solution", but hopefully it'll be a helpful troubleshooting step.
16:36
sonjag: The rebooting is, in part, to deal with such leftover processes as the ones you mention.
16:36
<sonjag>
moquist, I do that manually when I see swap used... it always dies when swap gets used. But I don't want to run it so that I never see the problem b/c then I can't troubleshoot.
16:36
<moquist>
I regularly have to kill more CPU-sucking processes than we have CPU cores, which means the entire server is busy running runaway processes until I log in and kill -9 them.
16:37
sonjag: Yes. You are right.
16:37
sonjag: I wasn't recommending anything here; I was confessing my inadequacy. :p
16:38
<sonjag>
Confessions of a Serial Rebooter by moquist.
16:38sgonzalez has quit IRC
16:38vagrantc has quit IRC
16:38jammcq has quit IRC
16:38ufo_ has quit IRC
16:39
<moquist>
oh, no.
16:39
I was going to publish under a sudo-nym. :(
16:40
Now sonjag gave it all away.
16:40vagrantc_ has joined #ltsp
16:40ufo_ has joined #ltsp
16:41
<warren>
ogra, sbalneav: did sbalneav pull in ogra's repo?
16:41sgonzalez has joined #ltsp
16:41vagrantc has joined #ltsp
16:41jammcq has joined #ltsp
16:42
<ogra>
warren, i pung him several times ... according to vagrantc it wasnt merged two hours ago though
16:43moquist_ has joined #ltsp
16:45
<sonjag>
Is it normal to have processes running after the xsession has ended? I have users that have gnome-panel, evolution-data, and bonobo-activati showing up in top that have no xsession running.
16:46jammcq has quit IRC
16:46sgonzalez has quit IRC
16:46vagrantc has quit IRC
16:46jammcq_ has joined #ltsp
16:46sgonzalez has joined #ltsp
16:48
<vagrantc_>
http://www.ltsp.org/~sbalneav/ltsp/ is missing the patches from http://www.ltsp.org/~ogra/ltsp/
16:49
<ogra>
yeah
16:50
<sonjag>
moquist, could this be hanging processes that cause problems with my servers? Didn't you once say that xterminator could be configured to kill processes on logout? Does it sound reasonable that changing to that might solve my issue?
16:50
<vagrantc_>
and everybody's missing the patches from debian :P
16:50
<ogra>
right
16:51Egyptian[Home] has quit IRC
16:51
<jammcq_>
=/quit
16:51
oops :)
16:51jammcq_ has quit IRC
16:54
<vagrantc_>
mostly just a few bugfixes and lots of translation updates
16:56henriquev has joined #LTSP
16:56* ogra tires a reboot ...
16:56
<ogra>
*tries too :)
16:56
<henriquev>
I've about 200+ users to add to LTSP, what can I use to automatize it?
16:56ogra has quit IRC
17:00F-GT has quit IRC
17:00daduke has quit IRC
17:00gonzaloaf_work has quit IRC
17:00oh207 has quit IRC
17:00rcy has quit IRC
17:00moquist_ has quit IRC
17:00joebake1 has quit IRC
17:00rjune__ has quit IRC
17:00spectra has quit IRC
17:00sutula has quit IRC
17:00Joris has quit IRC
17:00jbrett has quit IRC
17:00primeministerp has quit IRC
17:00warren has quit IRC
17:00Shingoshi has quit IRC
17:00klausade has quit IRC
17:00topslakr has quit IRC
17:00otavio has quit IRC
17:00ogra has joined #ltsp
17:00moquist_ has joined #ltsp
17:00joebake1 has joined #ltsp
17:00rjune__ has joined #ltsp
17:00spectra has joined #ltsp
17:00klausade has joined #ltsp
17:00otavio has joined #ltsp
17:00topslakr has joined #ltsp
17:00Joris has joined #ltsp
17:00warren has joined #ltsp
17:00sutula has joined #ltsp
17:00Shingoshi has joined #ltsp
17:00primeministerp has joined #ltsp
17:00jbrett has joined #ltsp
17:01F-GT has joined #ltsp
17:01rcy has joined #ltsp
17:01daduke has joined #ltsp
17:01oh207 has joined #ltsp
17:01gonzaloaf_work has joined #ltsp
17:02
<henriquev>
oh, as they're back from the split... => I've about 200+ users to add to LTSP, what can I use to automatize it?
17:03klausade has quit IRC
17:03topslakr has quit IRC
17:03otavio has quit IRC
17:04otavio has joined #ltsp
17:06topslakr has joined #ltsp
17:06klausade has joined #ltsp
17:06edgarin has quit IRC
17:09sonjag has quit IRC
17:11moquist has quit IRC
17:31shogunx has joined #ltsp
17:33henriquev has left #LTSP
17:49joebake1 has left #ltsp
18:05
<moquist_>
sbalneav: ping
18:37ltspbot has joined #ltsp
18:53staffencasa has quit IRC
18:59J45p3r has joined #ltsp
19:07sbalneav has joined #ltsp
19:07
<sbalneav>
Evening all
19:07* vagrantc_ waves
19:15ltspbot has joined #ltsp
19:17
<sbalneav>
hey vagrantc
19:50Amaranth has joined #ltsp
20:47Amaranth has quit IRC
20:48edgarin has joined #ltsp
20:53boog has joined #ltsp
20:53J45p3r has left #ltsp
20:57
<boog>
Is there a command for the lts.conf that will automatically log a user in at a given client?
21:03Artemka has joined #ltsp
21:08
<Artemka>
iocharset CP1251 on local device, how can i made this?
21:43
<vagrantc_>
boog: there's some autologin stuff, but it may be broken
21:45zodman has joined #ltsp
21:47
<boog>
vagrantc: can you point me in a direction?
21:48
I want to set up some kiosks at my community college campus. the kiosk admin tool seems to work well at locking them down, now i just need to figure out auto login
21:50
<rcy>
i have been trying to get autologin working in ltsp 5. all i get is a black screen with an X cursor
21:51
using ldm
21:51
<vagrantc_>
boog: supposedly LDM_USER and LDM_PASSWORD, but i think it's broken
21:53
<boog>
is it broken in ltsp 5?
21:54
<vagrantc_>
i think it's broken in ubuntu/gutsy, doesn't exist for ubuntu/feisty (although there are patches), and is broken in debian sid and lenny, and doesn't exist for etch, but there are patches for it.
21:55
boog: search for LDM_USERNAME and LDM_PASSWORD
21:55
<boog>
will do. i plan to use it on opensuse 10.3 under kdm
21:56
<vagrantc_>
ah, *no* idea how the suse support is
21:56
<boog>
i'll report back here after we try it next week
21:56
<vagrantc_>
actually, you might have better luck with gdm
21:56
i know there's support to use it with gdm
21:57
<boog>
can i force the login session to kde with gdm?
21:57
<vagrantc_>
probably
21:57
<boog>
the kiosk admin policies do not seem to apply to gnome sessions, but they do work if the user logs in via kde
22:00
originally tried kiosk-mode under edubuntu, but could not get printing to work
22:02
<Artemka>
anybody already made OO as local application7
22:03
<rjune__>
boog: are you using kdm or gdm for login manager?
22:03
IIRC sabyon is the GNOME tool to do Kiosk type stuff
22:06
<boog>
was using gdm
22:08
i hadn't heard of sabyon. does it do the same thing?
22:13
<rjune__>
kdm works better for autologin
22:13
at least it used to
22:13
and yah, if I remember correctly, sabyon is the gnome equiv to kiosktool
22:16
<boog>
i think i'll stick with kiosktool and try to accomplish the autologin with kdm. I am rebuilding this server with kde only.
22:22
thanks for your help and suggestions.
22:23boog has quit IRC
22:30zodman has quit IRC
22:42vagrantc_ has quit IRC
22:56Joris_ has joined #ltsp
23:00Artemka has quit IRC
23:01Artemka has joined #ltsp
23:03cyberorg has joined #ltsp
23:07subir has joined #ltsp
23:07edgarin has quit IRC
23:09Jach has joined #ltsp
23:13Joris has quit IRC
23:14
<warren>
*SWEET*
23:14
./mkdist --release can export Changelog directly from bzr log
23:14
so you don't have to maintain it in two separate places
23:14
unfortunately, Changelog originating from upstream's old old old bzr repos will contain lots of garbage
23:14
so this is only really usable on a fresh repo
23:18
<cyberorg>
.
23:19
<warren>
(for example, ltspfs will contain bzr log entries from everything else from the beginning of time instead of just talking about ltspfs. It is too confusing to use, unfortunately.)
23:19
<cyberorg>
warren, where is the new upstream repo now?
23:20
<warren>
cyberorg, sbalneav didn't import it yet
23:20
cyberorg, ogra has it in the new desired structure
23:20
I'll be applying fixes on top of ogra's structure
23:20
<cyberorg>
cool, looking forward to it :)
23:20
<warren>
this is the third version of new structure
23:21
Hmm...
23:32Egyptian[Home] has joined #ltsp
23:34johnny_ has joined #ltsp