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


Channel log from 17 January 2008   (all times are UTC)

00:05
<vagrantc>
lns: not that that fixes anything, really...
00:11
<lns>
vagrantc, ahh.. ok thx - that was just brought over from an upgraded server.
00:11
is pulse installed by default? I thought I had to install pulse by hand w/Ubuntu
00:11
(gutsy i386)
00:37F-GT has joined #ltsp
00:40
<vagrantc>
lns: well, if you're using ltsp-server-standalone on the server and ltsp-client in the chroot
00:40
it should all be there.
00:41
<lns>
vagrantc, ltsp-client in the chroot... ? is that not set up by ltsp-server[-standalone]?
00:41
err
00:41
ltsp-build-client ?
00:42
<vagrantc>
lns: yes, unless you changed something
00:42
it is possible to have an ltsp install without it, but you have to really try.
00:42
<lns>
i was wondering wtf that package was for... heh
00:42
wow
00:43
so chroot /blahblah apt-get install ltsp-client ?
00:43
<vagrantc>
lns: you should *never* have to do that manually
00:43
<lns>
oh ok
00:43
i always sudo ltsp-build-client
00:43
<vagrantc>
at some point, i should just tell white lies
00:44
<lns>
oooooook..? =p
00:44
don't follow
00:44
of course that's easy for me to do ;)
00:44
<vagrantc>
technically, if you run ltsp-build-client --early-packages ltsp-client-core, you could get an install without pulseaudio
00:45
but the default is to install all the bells and whistles
00:45
<lns>
ahh
00:45
what about on the server side?
00:45
i still remember having to do stuff manually in gutsy (ubuntu not edubuntu/xubuntu)
00:46
<vagrantc>
not having used ubuntu i can't say for sure, but it shouldn't require anything special
00:46
<lns>
ala https://wiki.ubuntu.com/PulseAudio
00:47
just that i would think - my clients' audio isn't detected until this is in place (at least afaik - been a few weeks since i installed my last svr)
00:48
<vagrantc>
lns: if you look at the dependencies of ltsp-server-standalone and ltsp-client, you should find some of those packages in there.
00:49
<lns>
vagrantc, right - but not just ltsp-server i'm assuming (which I installed last)
00:49
<vagrantc>
lns: correct.
00:49
<lns>
ok
00:49
=) I want to make my own wiki and collect all of the things i'm doing so i don't keep having to search around ;)
00:50
feels like the hardest part is keeping all the pieces together
01:50Joris_ has joined #ltsp
01:52Joris has quit IRC
01:57infoshopkeeper has joined #ltsp
01:58ogra has quit IRC
01:59ogra has joined #ltsp
01:59Q-FUNK has joined #ltsp
02:09wdg has joined #ltsp
02:27vagrantc has quit IRC
02:32open_lad has quit IRC
02:33open_lad has joined #ltsp
02:35open_lad has joined #ltsp
02:55infoshopkeeper is now known as johnny_
02:56
<johnny_>
beep beep ltsp peeps
02:56* johnny_ wonders who's still awake
02:56cyberorg has quit IRC
03:43putty_thing has joined #ltsp
03:43putty_thing is now known as jim``
03:50Traveler6 has joined #ltsp
03:54
<Traveler6>
hello
03:55
where can i find explanations on "how" ltsp5 works? i mean how dhcp + tftp + nfs + ssh come together. ideally with diagrams ;)
04:03mikkel has joined #ltsp
04:07rcy has quit IRC
04:10soneyka has quit IRC
04:17
<johnny_>
uhmm... maybe the edubuntu handbook?
04:18
otherwise.. it's pretty simple to explain.. as long as you know what those services really are
04:18
client gets IP via dhcp and also gets pointed to a tftp server where it downloads the kernel
04:18rcy has joined #ltsp
04:18
<johnny_>
then it mounts the rootfs via nfs or nbd
04:18
and starts X via ssh tunnelling
04:18
that's the short answer anyways
04:23cyberorg has joined #ltsp
04:28vernier_ has quit IRC
04:30
<Traveler6>
thanks johnny
04:31
i understand that, i wanted more detail. i will check out the edubuntu handbook
04:34
<johnny_>
what more detail?
04:34
the handbook isn't meant for more detail
04:35
it's meant for users
04:35
i don't know about any "official" diagram for ltsp5
04:35
any diagram would only show the basics
04:35
anyways
04:40
<Traveler6>
especially about the SSH bits
04:40
and LDM
04:40
how it really works
04:40otavio has joined #ltsp
04:41
<johnny_>
that involves reading about ssh x forwarding
04:41
there are no ltsp specifics except for ltspfs
04:41
and there are no nice docs
04:42
it's all nice
04:42
err
04:42
it's all new
04:43
<Traveler6>
ok
04:43
thanks
04:43
btw
04:43
how do i add users? chroot /opt/ltsp/i386 first?
04:43
<johnny_>
no
04:43
they run off the server
04:43
just add users to it
04:43
<Traveler6>
ok
04:43
how about adding software?
04:44
<johnny_>
server too
04:44
until the localapps stuff is easier
04:44
<Traveler6>
ok
04:44
<johnny_>
until now you'll have to hack it up
04:44
err then*
04:44
but i don't have any localapps yet
04:44
it's a spec on launchpad for the next ltsp steps
04:45
the edubuntu handbook should mention that kinda stuff tho
04:45
<Traveler6>
i saw it
04:45
thanks
04:45
i'll just install it on the server
04:45
btw
04:45
right now i have ltsp5 + riplinux on the same network segment
04:45
clients boot one or the other depending on mac address
04:45
it's dhcp3-server that does the filtering
04:46
do you know by chance if i can let the clients choose what to boot with a menu or something like that?
04:46
<johnny_>
uhmm.. i have heard of such things with pxelinux
04:46
but i have no idea how to do it :)
04:46
<Traveler6>
ok :)
04:46
<johnny_>
but that should get you started
04:47
and if you come back in thsi room a few hours from now
04:47
somebody might even be able to point you even farther
04:47
<Traveler6>
great
04:47
<johnny_>
when all the folks wake up
04:47
but search "pxelinux menu" see if that gets you anywhere
04:49
<Traveler6>
actually the day b4 yesterday i did a pxelinux menu
04:49
but... it didn't really work
04:50
i'll try again
04:56
not sure i can do it since these distros have different root-path
04:56
and ltsp has no "filename" option in dhcp3 ...
04:56
etc
04:56
<johnny_>
hmm.. usually you can force the filename
04:56
and tftp can just use the rootpath and request anything under it
04:56
of course.. i just use dnsmasq
04:57
not dhcp
04:57* johnny_ heads home
04:57
<johnny_>
bbin20
04:57johnny_ has quit IRC
04:57Traveler6 has quit IRC
05:09wdg has quit IRC
05:36open_lad has quit IRC
05:40subir has quit IRC
05:41exodos has joined #ltsp
05:48misui has joined #ltsp
05:50plamengr has joined #ltsp
05:56plamengr has quit IRC
06:02Egyptian[Home] has quit IRC
06:05Egyptian[Home] has joined #ltsp
06:06Guaraldo has joined #ltsp
06:22johnny has quit IRC
06:34johnny has joined #ltsp
07:39cdealer__ has joined #ltsp
07:40
<cdealer__>
need some help ... when booting my client its not passing the tftp screen, the dhcp server is giving a ip to the client and gateway, but stops on tftp
07:40
im using ltsp5 on ubuntu 7.10
07:45
anyone ???
08:09prpplague has quit IRC
08:09jammcq has joined #ltsp
08:13cliebow_ has joined #ltsp
08:17Q-FUNK has quit IRC
08:19
<exodos>
cdealer__: is your tftp server running on the same machine as tftp?
08:19
s/as tftp/as dhcp/
08:20Q-FUNK has joined #ltsp
08:45
<cdealer__>
exodos: yeah... the tftp is running on port 69 like supposed to be...
08:46
using tcpdump there is no data going from the tftp to the client and vice-versa
08:46
is like the client isnt seeing the tftp ...
08:50
<jammcq>
cdealer__: what distro are you using?
08:53
<cliebow_>
08:53
Kinda broad at the shoulder and narrow at the hip
08:53
And everybody knew you didn't give no lip to Big Jim
08:53
<jammcq>
:)
08:56Gadi has joined #ltsp
09:03
<cliebow_>
Through the dust and the smoke of this man made hell
09:03
Walked a giant of a man that the coders knew well
09:09cimas has joined #ltsp
09:11
<cdealer__>
jammcq: ubuntu 7.10 and ltsp5
09:21MeBadMagic is now known as BadMagic
09:22
<cimas>
hi guys
09:22
I got a problem with LTSP5
09:23
I installed ltsp 5 on my ubuntu gutsy
09:24
the istallation was smooth, but when I boot the client it keeps me to the login screen (LDM)
09:24
and when i insert user and pass it keeps me back to the login screen after a while
09:25
my /var/log/auth.log says
09:26
Jan 17 16:04:12 adminserver sshd[4618]: Received signal 15; terminating.
09:26
Jan 17 16:04:12 adminserver sshd[6596]: Server listening on :: port 22.
09:26
Jan 17 16:04:12 adminserver sshd[6596]: error: Bind to port 22 on 0.0.0.0 failed: Address already in use.
09:26
Jan 17 16:04:32 adminserver sshd[7031]: Accepted password for provaltsp from 192.168.0.250 port 51054 ssh2
09:26
Jan 17 16:04:32 adminserver sshd[7039]: pam_unix(ssh:session): session opened for user provaltsp by (uid=0)
09:26
pass is accepted but cannot get into X
09:26
any hint?
09:28
what is this error about port 22?
09:34
<cliebow_>
cimas:generally return to login screen means you should sudo ltsp-update-sshkeys followed by sudo ltsp-update-image
09:35
<cimas>
ok ... let me check
09:40
cliebow: no luck ... again back to login screen
09:42
<cliebow_>
ogra mentioned looking in .xsession-errors in the users homedir..
09:42
<cimas>
ok ... let me see
09:43spectra has joined #ltsp
09:44
<cimas>
here is my .xsession-errors
09:44
(process:5596): Gtk-WARNING **: This process is currently running setuid or setgid.
09:44
This is not a supported use of GTK+. You must create a helper
09:44
program instead. For further details, see:
09:44
http://www.gtk.org/setuid.html
09:44
Refusing to initialize GTK+.
09:44
(process:5600): Gtk-WARNING **: This process is currently running setuid or setgid.
09:44
This is not a supported use of GTK+. You must create a helper
09:45
program instead. For further details, see:
09:45
http://www.gtk.org/setuid.html
09:45
Refusing to initialize GTK+.
09:45
/etc/gdm/Xsession: Beginning session setup...
09:45
SESSION_MANAGER=local/adminserver:/tmp/.ICE-unix/5593
09:45
Avviso del window manager: Lettura del file della sessione /home/cimas/.metacity/sessions/default0.ms fallita: Apertura del file "/home/cimas/.metacity/sessions/default0.ms" fallita: Nessun file o directory
09:45
/home/cimas/.gtkrc-2.0:2: Impossibile trovare il file da includere: "(null)/Clearlooks/gtk-2.0/gtkrc"
09:45
/home/cimas/.gtkrc-2.0:2: Impossibile trovare il file da includere: "(null)/Clearlooks/gtk-2.0/gtkrc"
09:45
/home/cimas/.gtkrc-2.0:2: Impossibile trovare il file da includere: "(null)/Clearlooks/gtk-2.0/gtkrc"
09:45
/home/cimas/.gtkrc-2.0:2: Impossibile trovare il file da includere: "(null)/Clearlooks/gtk-2.0/gtkrc"
09:45
Tracker version 0.6.3 Copyright (c) 2005-2007 by Jamie McCracken (jamiemcc@gnome.org)
09:45
This program is free software and comes without any warranty.
09:45
It is licensed under version 2 or later of the General Public License which can be viewed at http://www.gnu.org/licenses/gpl.txt
09:45
Initialising tracker...
09:45
Could not set idle IO priority...attempting best effort 7 priority
09:45
/home/cimas/.gtkrc-2.0:2: Impossibile trovare il file da includere: "(null)/Clearlooks/gtk-2.0/gtkrc"
09:45
/home/cimas/.gtkrc-2.0:2: Impossibile trovare il file da includere: "(null)/Clearlooks/gtk-2.0/gtkrc"
09:45
** Message: Server desktop remoto non avviato
09:45
/home/cimas/.gtkrc-2.0:2: Impossibile trovare il file da includere: "(null)/Clearlooks/gtk-2.0/gtkrc"
09:45
/home/cimas/.gtkrc-2.0:2: Impossibile trovare il file da includere: "(null)/Clearlooks/gtk-2.0/gtk
09:45
<cliebow_>
cimas:whoa!
09:45
<lns>
pastebot
09:45
<jammcq>
DUDE !!!!!! please don't paste big stuff like that to the channel
09:45
<lns>
jeez
09:46
<jammcq>
!pastebot
09:46
<ltspbot>
jammcq: "pastebot" is The LTSP pastebot is at http://pastebot.ltsp.org. Please paste all text longer than a line or two to the pastebot, as it helps to reduce traffic in the channel. A link to the content will be pasted in the channel.
09:46
<jammcq>
use the pastebot
09:46
<cliebow_>
you should use pastebot.ltsp.org
09:46
<cimas>
sorry
09:48
<ltsppbot>
"cimas" pasted "content of my .xsession-errors" (103 lines) at http://pastebot.ltsp.org/420
09:48
<cimas>
done
09:56
<cliebow_>
cimas:i cant make any sense of it..
09:57
<cimas>
me neither
09:58
:(
09:59
opppss
09:59
i made a mistake
10:00
I posted the content of the .xsession-error in the admin user home
10:00
not the one in the client user
10:00
<cliebow_>
pastebot..what you got 8~)
10:01K_O-Gnom has joined #ltsp
10:01
<cimas>
gonna past the right one
10:02
<ltsppbot>
"cimas" pasted "content of user .xsession-errors" (6 lines) at http://pastebot.ltsp.org/421
10:03
<cimas>
ccliebow: can u make sense of that one?
10:03
it says xrdb: Connection refused
10:03
xrdb: Can't open display 'localhost:11.0'
10:03
<cliebow_>
sort of..
10:03
is this completely stock ltsp5 gutsy
10:04
?
10:05yanu has quit IRC
10:05
<cimas>
yes
10:06
<cliebow_>
and x is certainly installed on server..
10:06daya has joined #ltsp
10:06
<cimas>
yes
10:06daya has quit IRC
10:07daya has joined #ltsp
10:12
<cimas>
geeee!!!!
10:12
that what's happened...
10:12
I installed xfce
10:13
and then login my client in xfce.... and know what? ... It works
10:13
!!???
10:13
<ltspbot>
cimas: Error: "!???" is not a valid command.
10:13
<cimas>
whoops
10:13
does that make any sense to ya?
10:15
<cliebow_>
heh..not really..
10:15
<cimas>
and now gnome works as well!!!
10:15
<cliebow_>
just a kickstart
10:17
<cimas>
is it possible that installing xfce added some sort of config file that was missing before?
10:23
anyway... many thanks for the help guys
10:27cimas has quit IRC
10:30sepski has joined #ltsp
10:37staffencasa has joined #ltsp
10:48misui has quit IRC
10:58vagrantc has joined #ltsp
11:00yanu has joined #ltsp
11:06yanu_ has joined #ltsp
11:20Briareos1 has joined #ltsp
11:21
<Briareos1>
hi! can anyone give me a hint how to set up the DHCP of a WINDOWS 2000 server to forward the clients the to the ltsp-server correctly?
11:21yanu has quit IRC
11:21
<vagrantc>
!dhcp
11:21
<ltspbot>
vagrantc: Error: "dhcp" is not a valid command.
11:22
<vagrantc>
http://wiki.ltsp.org/twiki/bin/view/Ltsp/DHCP
11:22
you'll need to find the windows equivalents for next-server and root-path
11:23
and filename
11:23
<Briareos1>
that's what i am trying to find here *gg*
11:23
<vagrantc>
i think the example for dnsmasq lists the numeric equivalents. ...
11:24
ltspbot: learn dhcp as http://wiki.ltsp.org/twiki/bin/view/Ltsp/DHCP
11:24
<ltspbot>
vagrantc: The operation succeeded.
11:24
<Briareos1>
cheers - i'll have a look
11:24
<vagrantc>
Briareos1: that URL listed mentions some windows dhcp stuff
11:29exodos has quit IRC
11:38
<Briareos1>
grr all these dialogs are so confusing
11:38yanu_ has quit IRC
11:43
<Briareos1>
i guess i'll just replace the win2k DHCP by a newly setup linux-one
11:44
<vagrantc>
we would better be able to support you :)
11:51
hrm.
11:51
the fedora libexecdir changes don't work for debian
11:51
some of the things belong in /usr/lib and some in /usr/share
11:51
warren: is toshio kuratomi on irc?
11:52
<warren>
vagrantc, he is abadger1999
11:52
<abadger1999>
vagrantc: Sure. What needs to be split?
11:52
rc.d?
11:53
<vagrantc>
abadger1999: well, the ldm-script and rc.d stuff belongs in /usr/share on debian
11:53
<abadger1999>
ldm-script belongs in libexecdir.
11:53
<vagrantc>
abadger1999: not on debian
11:53
<abadger1999>
Guideline?
11:53
<vagrantc>
FHS
11:53* vagrantc rummages for a url
11:53
<abadger1999>
Entry in FHS?
11:54
I was reading that /usr/share is for data.
11:54
ldm-script is not really data.
11:54
I can see things in rc.d being data, though.
11:54* vagrantc reads
11:54
<abadger1999>
As they're being sourced rather than invoked.
11:54
<vagrantc>
rc.d are executed script
11:54
ah...
11:55yanu has joined #ltsp
11:55
<vagrantc>
http://www.debian.org/doc/packaging-manuals/fhs/fhs-2.3.html#USRSHAREARCHITECTUREINDEPENDENTDATA
11:56
"It is recommended that application-specific, architecture-independent directories be placed here.It is recommended that application-specific, architecture-independent directories be placed here."
11:56
ldm-script is definitely architecture-independent
11:56
<warren>
vagrantc, ugh. ogra agreed to libexecdir.
11:57
<vagrantc>
let me check a few more things
11:57
<abadger1999>
It is architecture independent but I disagree that it fits the definition of "data file".
11:57
<vagrantc>
yeah, /usr/share does seem to be focused on "data files"
11:58
<warren>
vagrantc, I was happy keeping it in /usr/share until I realized that it isn't data.
12:03
<vagrantc>
hm. maybe it's not as much of a problem as i thought
12:03
changing the rc.d directory kind of sucks, but that's not a big deal for debian since it was never present in a stable release.
12:04
<abadger1999>
Cool. And like I say, I'm not attached to where rc.d goes. ogra thought it should go in the same hierarchy as ldm-script so I moved it too.
12:06
<warren>
abadger1999, did you already commit the gcc warnings patch?
12:06
<abadger1999>
warren: No. I felt it was better for you to look it over first.
12:06
Two eyes are better than one.
12:06
<lns>
FYI Everyone: If you want to help fix the OpenOffice pixmap LTSP crash issue, register and comment on the following issue in the OOo issue:
12:07
http://www.openoffice.org/issues/show_bug.cgi?id=85321
12:07
<abadger1999>
warren: I approved the review request whether you commit that patch or not, though.
12:07
<warren>
abadger1999, it looks safe
12:07
abadger1999, go ahead and commit it
12:07
<abadger1999>
k
12:07
<warren>
abadger1999, given that we haven't acutally *executed* any of this we'll know for sure later =)
12:08
lns, hmm, firefox doens't have the same problem?
12:09Shingoshi has quit IRC
12:10
<lns>
warren, yes it does
12:10
but it's being actively worked on in FF3
12:10Shingoshi has joined #ltsp
12:10
<lns>
We need to get some momentum going with OOo too
12:11
and any other apps that we find that don't pay attention to pixmap caching (at least that's what I think the real issue is)
12:11
<abadger1999>
warren: ldm changes pushed.
12:14
<lns>
Apparently "voting" for the issue in OOo will help it get more attention as well, so if anyone having this issue could do that as well, it will probably give the devs more to work with
12:21
<vagrantc>
hrm. also moved /usr/lib/ltsp/screen.d
12:21
and that was actually part of a stable debian release
12:22
<PerfDave>
Would a workaround for the OOo issue be to use fullscreen VNC rather than XDMCP for the clients?
12:22
<vagrantc>
to /usr/share/ltsp/screen.d ... (although i suggested moving it to /usr/share years ago)
12:23yanu has quit IRC
12:24
<vagrantc>
hrm. putting this stuff in /usr/lib seems wrong also ... it's no more a library than it is data
12:24sonjag has joined #ltsp
12:36sgonzalez has joined #ltsp
12:45
<warren>
lns, I voted for it
12:46
lns, it seems that only one other person voted for it, and they only voted half of their possible votes
12:46RiXtEr has joined #ltsp
12:46
<warren>
lns, (each person may vote 2 times)
12:46
<Briareos1>
basic question (as i just got worried when i read about the OOo crash ins wrote about): would you run ltsp in a company? I've so far only read about usage in schools ...
12:46yanu has joined #ltsp
12:47
<warren>
Briareos1, many companies and government agencies have used it for years
12:48
<PerfDave>
Briareos1: I've been trying to talk my company into using LTSP to reduce desktop noise and cost.
12:49
<Briareos1>
ok I've now read the bugreport and it seems to be easily avoidable if i just size the client a little higher from the RAM-perspective for the beginning
12:50
perfdave: well i don't need to talk mine in, but if i change i have full responsability :)
12:50
<Q-FUNK>
PerfDave: simeple trick: talk to the person in charge of budget. once they see just the cost saving in electricity and computer maintenance, they'll beg you to find thin clients.
12:50
<Briareos1>
ggg
12:51
<PerfDave>
Q-FUNK: There is virtually no cost saving for us there.
12:51
<Briareos1>
One of my main reasons to change to thin is that with the current set up there is always some misconfiguration, some programs not up-to-date, client running out of diskspace on systempartition and blabla
12:51
if i calculate only my worktime against that it completely makes sense
12:52
<PerfDave>
Q-FUNK: The work involved in moving over to an LTSP system actually outweighs the benefits currently. We've got so many racks of servers that our desktop cost and power usage is only a tiny tiny fraction anyway.
12:52
<Briareos1>
warren: could you provide me some interesting links?
12:54daya has quit IRC
12:55
<warren>
Briareos1, these days though, most of the benefit of the LTSP model is the central management part.
12:55
Briareos1, people might be better off using central network managed workstations instead
12:55nicoAMG has joined #ltsp
12:55sonjag has quit IRC
12:56mcfloppy_ has quit IRC
12:57
<lns>
warren, ok - just voted my 2 votes
12:58
<warren>
lns, I'm guessing a few hundred people would have to do it in order to get any attention?
12:58
<Briareos1>
warren: what do you mean by "central network managed workstations"?
12:59
<warren>
Briareos1, install and configure software only once, all workstations use it without individual configuration
12:59
Briareos1, home directories are central as well
12:59
Briareos1, if a workstation dies, unplug it, plug in another (blank one, maybe no hard drive) and it just works without any configuration.
12:59
<Briareos1>
warren well isn't that what ltsp is about?
12:59
<vagrantc>
warren, abadger1999 : thanks for all the cleanup work on ldm
13:00
<Briareos1>
warren or did i just misunderstand your "though" in the sentence?
13:00
<warren>
vagrantc, abadger1999: more cleanup of the autotools stuff in those components are needed, for example there is currently no clean way to get rid of everything generated by autogen.sh
13:00
vagrantc, also it doens't help that Debian/Ubuntu doesn't actually use it. =)
13:00
Briareos1, LTSP is mostly thin clients, not diskless workstation
13:00
<vagrantc>
warren: well, we can move towards it
13:01
<warren>
Briareos1, but stuff going on now is enabling the latter.
13:02cdealer__ has left #ltsp
13:02
<Briareos1>
warren ah i get it - you mean putting the processing needs also mainly to the cpu/gpu of the client
13:03
<warren>
Briareos1, yes, with client hardware getting much cheaper
13:03
Briareos1, most costs of maintaining many clients are in the labor of maintaining individual machines anyway.
13:03mcfloppy_ has joined #ltsp
13:05GiantPickle has joined #ltsp
13:08yanu has quit IRC
13:10
<abadger1999>
vagrantc: No problem. Regarding screen.d... It's in /usr/share/ltsp in our ltsp-client package. I just looked and see that there's no Makefile so we're deciding on layout in the .spec file (our package build) rather than in an upstream script :-(
13:10
<vagrantc>
abadger1999: sure.
13:10
<abadger1999>
vagrantc, warren: Should we change ltsp-client to create screen.d in /usr/lib instead?
13:11
<vagrantc>
abadger1999: hard to say.
13:12
abadger1999: it currently supports /usr/lib/ltsp/screen.d and /etc/ltsp/screen.d (with preference for /etc) ... we could just add another layer ...
13:12
abadger1999: that would maintain backwards compatibility while moving forward at the same time ...
13:13
abadger1999: and a one-line patch to screen_session :)
13:13
<sutula>
vagrantc: Have you received complaints about the latest etch-backports ltsp-server (5.0.40~bzr20071229-1~40.etch.0) and users not being able to log in?
13:14
(this can wait if you're busy)
13:14
<warren>
vagrantc, we don't want to maintain a patch against the script that only applies to us.
13:14
<vagrantc>
sutula: i haven't
13:14
warren: well, i mean to put the patch upstream
13:15
the only cost is an additional directory checked for each call to screen_session, which should be minimal
13:15
<sutula>
vagrantc: I updated my otherwise working server this AM and users couldn't log in. The users were in group fuse, but not plugdev. When I added the users to plugdev as well, they can now log in.
13:15
<warren>
abadger1999, technically it isn't wrong in /usr/share because they are not executed?
13:15
(are they?)
13:16
<vagrantc>
sutula: that's weird
13:16
warren: they are executed
13:16
<sutula>
vagrantc: Of course, thumb drives don't work any more either, but I'll look further into that...just wanted to know if this was a known problem
13:17
<vagrantc>
otavio: do you have opinions regarding all this /usr/lib and /usr/share stuff ?
13:17
<sutula>
vagrantc: Anyway, when you get the chance, I'll be around if you want to try some experiments or something.
13:17
<vagrantc>
otavio: we've beentalking about moving some things around, and fedora also uses /usr/libexec (i think)
13:17
sutula: i'm at a loss as to what to even try ...
13:18
sutula: what was your previous version?
13:18
<sutula>
vagrantc: Hmmm ...I think the last backport before that
13:18
<warren>
vagrantc, we suggested libexecdir only because it resolves to the right place on both fedora and debian.
13:18
<sutula>
Didn't you change how local devices were being mounted?
13:18
<warren>
vagrantc, /usr/libexec and /usr/lib
13:18
<abadger1999>
yeah we do. Luckily configure makes it easy to redefine libexecdir to /usr/libexec or /usr/lib depending on where you're building
13:18
<warren>
vagrantc, well, and it is the correct
13:19
<abadger1999>
ltsp-client doesn't use autoconf/configure yet, though.
13:19
<warren>
doh
13:19
<abadger1999>
No Makefile so no configure either.
13:20
<warren>
abadger1999, while there is a desire to do this correctly, I also don't want to be so concerned about correctness on the first draft that this slows us down by weeks.
13:20
<vagrantc>
sutula: yes, there's a comment on known issues for the new backport on the wiki page
13:20
<abadger1999>
warren: Sure. We just have to remember to put the screen files in the right place when/if the location is changed in ltsp-client.
13:20
<Briareos1>
ok thanks for the talk - I am off for now. bye
13:21Briareos1 has quit IRC
13:23
<sutula>
vagrantc: My guess would be that when the user logs in, something in the automount stuff is dying and the user gets logged out. I don't know enough about how the local devices and mounting work to begin to track it down, but without being in plugdev, users seem to not be able to log in.
13:24
vagrantc: I did do the symlink as on the wiki page. By the way, the ln -s files are backwards on the wiki page.
13:24
<vagrantc>
sutula: well fix it then :P
13:24* sutula can do that
13:25
<vagrantc>
sutula: so the issue is that when the users are in group plugdev, it causes them to log out?
13:25
sutula: or when they aren't in plugdev ?
13:26
<sutula>
vagrantc: When they aren't in plugdev (Wiki fixed)
13:26
<vagrantc>
sutula: i'll test on my sid install ..
13:27
why that would matter is perplexing to me ...
13:27
<sutula>
vagrantc: My thin clients have several built-in disk devices, so my guess is that it has something to do with those devices getting mounted at login
13:27* johnny tries to think of of a clever codename for the first sabayon release of 2.22
13:30bobby_C has joined #ltsp
13:31yanu has joined #ltsp
13:32
<vagrantc>
sutula: non-removable media are now disabled by default
13:33* sutula wonders how the code tells what's non-removable and what isn't, and whether that's where plugdev comes in
13:33
<johnny>
will it ever be possible for non removable remedia to be invisible to the thin clients if mounted on the server?
13:33
<vagrantc>
sutula: we use udev attributes
13:33
<johnny>
but for the client's own cdroms to still show up
13:33
<vagrantc>
sutula: but that's on the client-side, not server-side ...
13:34
<sutula>
vagrantc: Hmmm...client side shouldn't matter
13:35
<vagrantc>
sutula: works for me with or without plugdev
13:35
sutula: on sid....
13:35
<sutula>
vagrantc: Any (simple) way you can test the etch-backport version?
13:35
<vagrantc>
sutula: will test on etch
13:35
sutula: yes :)
13:36* johnny tired of having to register on infinite bug trackers
13:36
<vagrantc>
sutula: i've got a virtualbox etch server for exactly this purpose :)
13:37
<Gadi>
johnny: using gnome?
13:37
there's a gconf setting to disable visible volumes
13:37
if thats what you want to do
13:38
<johnny>
Gadi, well.. i've been triaging bugs for sabayon.. and being annoyed
13:38
plus i wanted to vote on that OOo bug
13:38
<Gadi>
ah
13:38
<johnny>
got bugs on various distros
13:38
<Gadi>
yeah - the worst part about OSS bugs is you know they can be addressed if you ask enough times
13:39
:)
13:41sonjag has joined #ltsp
13:41sonjagonzalez has joined #ltsp
13:43yanu has quit IRC
13:46sgonzalez has quit IRC
13:48
<vagrantc>
sutula: which window manager/desktop are you using
13:49
<K_O-Gnom>
can i redirect with a ltsp to a windows image? (if it workes did somebody have a howto for me?)
13:50
<sutula>
vagrantc: gnome
13:51
vagrantc: I added a fresh, new user, and it doesn't appear to have trouble logging in
13:51* sutula wonders what, in a user's gnome history, is causing this :(
13:52
<vagrantc>
sutula: i had no problems with or without plugdev, fuse or any combination, using gnome or dwm on etch.
13:53
<sutula>
vagrantc: I'll experiment...backup a user, then start deleting hidden files
13:53
vagrantc: Thanks for trying
13:53yanu has joined #ltsp
13:55sgonzalez has joined #ltsp
14:02sonjag has quit IRC
14:02
<vagrantc>
sutula: regarding the white background issue ... it almost seems definite that when it displays the logo is when the screen goes white
14:03K_O-Gnom has quit IRC
14:03
<sutula>
vagrantc: Hmmm
14:04
<vagrantc>
sutula: i wonder if we switched the order of when they get displayed somehow ... but then it would probably hide the logo
14:04
<sutula>
vagrantc: The switch would be a good experiment, though
14:04
<vagrantc>
sutula: it's all in C, so it's a little over my head
14:05
<sutula>
vagrantc: Done by the ldm c code?
14:05
<vagrantc>
sutula: i believe so
14:05
sutula: well, the gtkgreet code ...
14:05
<sutula>
vagrantc: Once I get to the bottom of my login issue, I might have time to look at that
14:06
<vagrantc>
sutula: that would be great. it's kind of ugly as it stands.
14:07sonjag has joined #ltsp
14:08
<sutula>
vagrantc: OK...so my plugdev problem went away...I think when I rebooted the server
14:09
It shouldn't have been necessary after the update, but seems to be
14:09
<vagrantc>
sutula: strange
14:09* sutula agrees, but isn't willing to roll back to find that one :)
14:09
<vagrantc>
yeah, i guess i haven't actually done backport upgrades
14:11sonjagonzalez has quit IRC
14:14vagrantc has quit IRC
14:17sgonzalez has quit IRC
14:39sgonzalez has joined #ltsp
14:45Gadi has left #ltsp
14:47sonjag has quit IRC
14:58oh207 has joined #ltsp
14:58sgonzalez has quit IRC
15:12cliebow_ has quit IRC
15:19muh2000_ has joined #ltsp
15:24yanu has quit IRC
15:28mikkel has quit IRC
15:33muh2000 has quit IRC
15:42sepski has quit IRC
15:44slidesinger has joined #ltsp
15:45Guaraldo has left #ltsp
15:47RiXtEr has quit IRC
15:49Q-FUNK has quit IRC
15:52Q-FUNK has joined #ltsp
16:09linux2 has quit IRC
16:21jammcq has quit IRC
16:22jaffo has joined #ltsp
16:24nicoAMG has quit IRC
16:25
<jaffo>
i've just installed ubuntu 6.10 server and am following the intructions, when i run ltspcfg i get the following... Checking Runlevelgrep: /etc/inittab: No such file or directory
16:25
what have i missed
16:26
when trying to set the runlevel it bombs out of the ltspcfg utility
16:29mcfloppy_ has quit IRC
16:32
<jaffo>
anyone here to help me
16:32
<Lumiere>
jaffo: you should probably try the latest version of edubuntu
16:32
also... ltsp is more related to desktop not server
16:32
<jaffo>
i did and the installer is broken on the cd, two differnet times i burned it and tried to install
16:33
<Lumiere>
use md5sum
16:33
<jaffo>
i need a server and 5 clients
16:33
<Lumiere>
to check your cd
16:33
I do not see anyone on who usually gives support
16:33
<jaffo>
ok
16:34jaffo has quit IRC
16:38vagrantc has joined #ltsp
16:41chul has joined #ltsp
16:43yanu has joined #ltsp
17:01Q-FUNK has quit IRC
17:01bobby_C has quit IRC
17:07
<sutula>
vagrantc: So I'm getting back to trying to figure out why local devices aren't mounting on the latest Etch-backports. Following debugging stuff in https://wiki.ubuntu.com/DebugLocalDev
17:08
vagrantc: Did something change in Step 3? I don't see /var/run/ltspfs_fstab at all on the client
17:08otavio has quit IRC
17:13
<vagrantc>
sutula: what kind of devices do you have?
17:13
sutula: /var/run/ltspfs_fstab is still what's used
17:13
sutula: the only thing that changed client-side, that i know of, is the udev rules
17:14
that i can think of
17:14
<sutula>
vagrantc: So I see Step 2...it's a thumbdrive
17:14
sda: assuming drive cache: write through
17:14
sda: sda1
17:14
sd 0:0:0:0: Attached scsi removable disk sda
17:15
Any pointers at where I should look?
17:15
<vagrantc>
sutula: cat /sys/block/sda/removable on the thin client
17:15
it should say 1
17:15
presuming it is actually removable.
17:16
sutula: otherwise, follow the instructions in /opt/ltsp/i386/usr/share/doc/ltspfsd/NEWS.Debian (or is it README ...)
17:16
and re-enable non-removable devices.
17:17
<sutula>
It does say "1"
17:17
<vagrantc>
sutula: is there a floppy or cdrom drive on the system?
17:18
sutula: i haven't actually had a chance to test usb sticks
17:18
<sutula>
vagrantc: Not a real one, but in the past, a floppy showed on the desktop
17:18
<vagrantc>
heh
17:18* sutula suspects there's floppy hardware support on the MB
17:18
<sutula>
vagrantc: I can be your "test" :)
17:19Egyptian[Home] has quit IRC
17:20
<vagrantc>
sutula: well, having no idea what's wrong ... i think i'll have to test it myself
17:20
<sutula>
vagrantc: OK...if there's something I can do to help, or if you can point me to the code that ought to be running, I'd be happy to work on it
17:20
<vagrantc>
sutula: i don't know where to start really ... it works for cdroms and floppy drives in virtualbox ... and i can't test usb sticks in virtualbox
17:21Egyptian[Home] has joined #ltsp
17:21
<vagrantc>
sutula: on the thin client, the udev rules in /etc/udev/rules.d/*ltspfsd* upon device insertion/removable should run /lib/udev/*fstab_entry as appropriate
17:22
<sutula>
vagrantc: OK, I'll put some debugging stuff in there
17:23
<vagrantc>
s,removable,removal,
17:23
floppy rules are handled speciall, and cdroms are handled separately ...
17:23
<sutula>
k
17:38J45p3r has joined #ltsp
17:46mcfloppy_ has joined #ltsp
17:50vernier__ has joined #ltsp
17:50vernier__ has quit IRC
17:51vernier__ has joined #ltsp
17:56chul has left #ltsp
17:58
<sutula>
vagrantc: On my system, I had to remove "ATTRS{removable}==1, " to get the rule to fire. Is it possible that the ATTRS isn't working (correctly) on Etch?
18:00
vagrantc: Since I'm mucking about, I could try some more things if you want, but have to leave in 15 minutes.
18:08
<vagrantc>
sutula: it is definitely possible that it's broken on etch, yes.
18:11slidesinger has quit IRC
18:11
<sutula>
vagrantc: Anything I can do to try to debug it, or will you take a shot (and submit any bugs) later?
18:26
<vagrantc>
sutula: i might work on it saturday or sunday. by all means, if you can figure anything out, please do so :)
18:26
<sutula>
vagrantc: k
18:26
<vagrantc>
sutula: though bugs should really only be submitted if it's also a bug in the sid packages
18:27
<sutula>
vagrantc: Right...hopefully, it's like the util-linux thing, and there might be a workaround
18:34
<muh2000_>
hmm strange ltsp client starts to do some weird things, color is switching apart. wasnt like yesterday. resolution is not recognized yet ^^
18:38
<vagrantc>
sutula: i definitely tested the ATTRS{removable}==1 thing with a usb stick and sid ...
18:38
granted, it was a single usb stick
18:41ari_stress has joined #ltsp
18:41
<ari_stress>
morning all :D
18:48mischko has joined #ltsp
18:49staffencasa has quit IRC
18:53psycodad has quit IRC
19:09vagrantc has quit IRC
19:34ari_stress has quit IRC
19:42jammcq has joined #ltsp
20:05spectra has quit IRC
20:06mischko has quit IRC
20:13jaffo has joined #ltsp
20:23cliebow has joined #ltsp
20:24anything4cash has joined #ltsp
20:28
<anything4cash>
anyone here help me with connecting to a newly installed edubuntu server
20:28
from a windows machine how to i connect
20:28
ive received an ip address from the new edubuntu server
20:29
and im connecting to the net fine, andcan ssh to the edubuntu server fine
20:35jaffo has quit IRC
20:36
<anything4cash>
do i misunderstand.can i not connect to the ltsp/edubuntu from a windows machine. do i have to boot from a network card or bootable cd into the terminal server
20:38
<cliebow>
you "could " use cygwin to get a gui on the server
20:38
ir use nx to get a gui on the server..but th epoint is...ltsp is about "no" os on a client
20:40
<anything4cash>
ok i get it, is there a way to boot a usbkey that will connect me to it or a bootable cd orsomwthing like that
20:40
<cliebow>
you prob can use pxe to boot directly to the server..assuming your bios supports it..
20:41
or use a floppy..or an etherboot enabled network card..
20:41
or i think there is an etherboot cd..
20:42
but chance are an f12 on a modern bios will let you boot across the network'
20:42
<anything4cash>
i'm using a wireless cardso i don't know howto do that. the builtin intel card supports pxe boot in the bios but im not near the edubuntu server
20:42
<cliebow>
wireless is tough..double tough
20:42
<anything4cash>
do you know how to boot pxe from grub or vista's bootloader or know where to point me
20:42
<cliebow>
catch 22..
20:43
errr.pxe should just work
20:43
<anything4cash>
hmmm wireless problem again. i'll have to keep googleing to see if i can pxeboot from my wireless card
20:43
<cliebow>
no slimesucking vista bootloader
20:44
<anything4cash>
or bcd or something like that... waht about vmware does it do that in their bios
20:44
if you know
20:44
<cliebow>
sorry i don't
20:45
<anything4cash>
ok tahnks for your help i'll keep looking
20:45
<cliebow>
sure..you can ask the smart guys here...
20:56cliebow has quit IRC
21:01anything4cash has quit IRC
21:15J45p3r has quit IRC
21:47open_lad has joined #ltsp
22:21vagrantc has joined #ltsp
22:30Joris has joined #ltsp
22:46Joris_ has quit IRC
22:48subir has joined #ltsp
23:07aboo0ood has joined #ltsp
23:19hursand has joined #ltsp
23:20
<aboo0ood>
hi everybody, i wanna ask about the communication between the X server and the LTSP server, is it TCP/UDP/else ?
23:21
<vagrantc>
aboo0ood: it depends a lot on weather you're using ssh X11 forwarding or XDMCP or which point at which you're talking about
23:22
<aboo0ood>
vagrantc, aha, ssh is TCP, XDMCP is UDP ?
23:22
<vagrantc>
XDMCP is UDP ... ssh uses TCP ... X11 ... i'm not sure which regular old X11 uses.
23:23
i *think* it's UDP
23:23
<aboo0ood>
vagrantc, wut do u mean by X11 alone ?
23:23
<vagrantc>
maybe either
23:23open_lad has quit IRC
23:23
<aboo0ood>
vagrantc, shouldn't it be using either SSH forwarding or XDMCP ?
23:24
<vagrantc>
aboo0ood: well, LDM by default uses ssh's X11 forwarding. but it can be configured to use ssh just for login and plain X11 from there on out.
23:24
<aboo0ood>
vagrantc, what is plain X11 ?
23:24
<vagrantc>
X windows
23:24
<aboo0ood>
vagrantc, u mean the normal communical between X server and X client ?
23:24
<vagrantc>
correct
23:24
<aboo0ood>
communication *
23:25
<vagrantc>
ports 6000-6007 (at least) seem to suggest x11 traffic- udp or tcp.
23:25
<aboo0ood>
vagrantc, so in the case of the default LTSP5 , forwarding is done over SSH = TCP
23:26
<vagrantc>
aboo0ood: it *might* be that simple.
23:26
aboo0ood: can't say for sure.
23:26
aboo0ood: why do you ask?
23:26
<aboo0ood>
vagrantc, i am doing a presentation about LTSP and i want to be sure about this point
23:27
<vagrantc>
it gets tricky be ssh X11 forwarding set's up a proxy server... and i don't necessarily know what's going between the proxy server and the X server and the X clients
23:30
<aboo0ood>
vagrantc, hmmm, ok .. wut about XDMCP ?
23:30
<vagrantc>
aboo0ood: XDMCP is udp
23:30
aboo0ood: when but that's just the login... from there on out it's X11
23:31
<aboo0ood>
aha
23:31
vagrantc, that's why it's insecure
23:32
<vagrantc>
aboo0ood: it's insecure because it's a known protocol and therefore only slightly obfuscated plain text.
23:32
<aboo0ood>
vagrantc, another question , when i say i got a session on the LTSP server, it means that my X server got privaliges to use applications on the LTSP server ??
23:33
<vagrantc>
aboo0ood: well, more like your applications got privledge to run on the X server.
23:34
aboo0ood: but what does it matter?
23:34
er, correction
23:34
aboo0ood: well, more like your applications got privledge to display on the X server.
23:35
<aboo0ood>
vagrantc, cz i read that a session is opened on the X server, so i was confused
23:35
but this way it means that the session is opened on the thin-client's X server, right ?
23:38
<vagrantc>
aboo0ood: define session different ways and you'll get different answers
23:39* vagrantc can't really see the point of getting into that much detail
23:39
<aboo0ood>
vagrantc, thanx :)
23:45hursand has left #ltsp
23:46
<sutula>
vagrantc: Time for more info on the etch backport udev problem?
23:47* sutula whacks hisself for asking to ask :)
23:47
<sutula>
vagrantc: Though the Etch man page for udev says that "ATTRS" searches upwards for a match, it doesn't seem to be doing so.
23:48
vagrantc: The rule matches for sda since there's a "removable" entry in /sys/block/sda
23:49
vagrantc: But there's no "removable" file in /sys/block/sda/sda1, so the rule is not firing.
23:49
vagrantc: By adding another rule, I can get it to fire:
23:49
ACTION=="add", SUBSYSTEM=="block", ATTRS{removable}=="1", RUN+="add_fstab_entry%k"
23:49
ACTION=="add", SUBSYSTEM=="block", ATTRS{../removable}=="1", RUN+="add_fstab_entry %k"
23:50
<vagrantc>
sutula: well, i can easily make that a new backports upload, then :)
23:50
<sutula>
Of course, this is a horrid hack.
23:50
...but maybe good enough for backports
23:50
<vagrantc>
yeah, but what are backports if not hacks? :)
23:51
i can't fathom it really breaking anything.
23:51
<sutula>
I looked in udev changelog but didn't find anything mentioning that sort of thing, but am no udev expert
23:51
<vagrantc>
i'm pretty sure i tested it with sid and a usb stick ...
23:51
<sutula>
vagrantc: OK, I'll update the wiki, and you update backports whenever you feel like it
23:52
<vagrantc>
i wouldn't be surprised if that didn't work
23:52
on etch, that is
23:53
<sutula>
vagrantc: So, last question for tonight...Sometimes it still isn't showing on the desktop (gnome). I can see it mounted on the server with "mount". Where do I look next?
23:53
Maybe my daemons are all horked with all the messing around I did getting to this point, and I just need a reboot :)
23:53
<vagrantc>
hm.
23:53
might be worth it, yes.
23:53
<sutula>
vagrantc: K, I'll reboot when the rest of the kids go to bed :)
23:54
vagrantc: Thanks for all your help
23:54
<vagrantc>
sutula: of course :)
23:56
looks like i actually have to make a new branch for the backport
23:56
the current ldm and ltspfs backports are just etch re-builds
23:59
<sutula>
vagrantc: Darn