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


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

00:06subsume has quit IRC
00:06subsume has joined #ltsp
00:42
<subsume>
Got passed my problem!
00:50
Getting file not found errors. Can someone help debug?
00:52
<daya>
subsume, whats your problem
00:52
<subsume>
I'm getting TFTP File Not Found errors
00:53ogra_ has quit IRC
00:53highvolt1ge has quit IRC
00:53mikesh has quit IRC
00:53mnemoc has quit IRC
00:53twinprism has quit IRC
00:53Shingoshi has quit IRC
00:53iMacGyver has quit IRC
00:53ufo_ has quit IRC
00:53
<subsume>
daya: got booted. say something?
00:53
<daya>
subsume, check the status, if it is started by inetd or standaloine, /etc/default/tftpd-hpa
00:53
subsume, in which system r u
00:53
<subsume>
ubuntu
00:54
RUN_DAEMON = no
00:54
OPTIONS="-l -s /var/lib/tftpboot"
00:54
its started ny inetd
00:54
<daya>
switch that to yes, if it works and restart the service from /etc/init.d
00:55
subsume, is ur inetd service is started
00:55
<subsume>
daya: it was.
00:55
daya: now tftp is hanging
00:55
tftp open timeout now.
00:55ufo_ has joined #ltsp
00:55iMacGyver has joined #ltsp
00:56mnemoc has joined #ltsp
00:56
<daya>
subsume, reboot your server and check it again
00:56mikesh has joined #ltsp
00:56
<subsume>
shutdown -r NOW?
00:56highvolt1ge has joined #ltsp
00:57subsume has quit IRC
00:58subsume has joined #ltsp
00:58subsume has left #ltsp
00:58subsume has joined #ltsp
00:58
<subsume>
daya: TFTP still hangs
00:59
<daya>
subsume, it must work, is inetd is running
01:00
<subsume>
Yes. I just did tftpd-hpa restart
01:00
still hanging
01:00
daya: ^
01:00
wireshark is seeing requests from client to server
01:03subsume has left #ltsp
01:03subsume has joined #ltsp
01:03
<subsume>
daya: network problem. did you say something?
01:03
<daya>
subsume, network problem ?
01:03
<subsume>
I got booted off IRC for a sec.
01:04
did you get my last response?
01:04
<daya>
which response
01:04
<subsume>
Yes. I just did tftpd-hpa restart
01:04
still hanging
01:04
wireshark is seeing requests from client to server
01:04
<daya>
subsume, ye ya,
01:05
<subsume>
daya: no clue?
01:06
<daya>
subsume, one try, change the setting in /etc/default/tftpd-hpa to yes, and restart the service
01:06
<subsume>
daya: didn't I just do that?
01:07
it is = "yes"
01:07
<daya>
subsume, ok, then try to run tftpd-hpa thur xinetd, install it, apt-get install xinetd, but I think it shouldn't required.
01:07
just check it, I have run tftpd-hpa thru xinetd in lenny
01:07
<subsume>
daya: how to run it through xinetd?
01:09Shingoshi has joined #ltsp
01:11ogra_ has joined #ltsp
01:11twinprism has joined #ltsp
01:11K_O-Gnom has joined #ltsp
01:12Shingoshi has quit IRC
01:12
<daya>
don't forget to change the parameter to no in default/tftpd-hpa file after u install xinetd
01:12
Any one have idea about the X error in lenny for client.
01:12
<subsume>
daya: https://help.ubuntu.com/community/Installation/Netboot
01:13
<daya>
subsume, no just install it,
01:13Shingoshi has joined #ltsp
01:13
<subsume>
daya: I did install it......
01:13
daya but it doesn't come with a conf
01:14
<daya>
subsume, In my case I haven't configured it,
01:15
<subsume>
daya: BINGO
01:18
daya: works great.... why am I getting initramfs? any idea? I installed edubuntu into the client image
01:19
<daya>
subsume, initialise RAM disk,
01:20
<subsume>
daya: ((not sure what you mean))
01:20
<daya>
subsume, https://wiki.ubuntu.com/Initramfs
01:21
<subsume>
daya: why does this start instead of my netboot client image?
01:22
<daya>
subsume, netboot client image
01:22
<subsume>
daya: why do i get initramfs instead of my /opt/ltsp/i386?
01:23
<daya>
subsume, I think they run squsahfs and nbd not requiring nfs
01:23
ltsp w/o nfs
01:24
<subsume>
daya: is initramfs happening because of a problem with nfs?
01:30
<daya>
subsume, don't know completely about it,
01:38
<subsume>
http://dpaste.com/39773/ daya
01:40
<daya>
subsume, not seen such error, isn't your client system booted well
01:40
<subsume>
Nope. Its booting the contents of /var/lib/tftpboot/ltsp/i386/ instead of /opt/ltsp/i386
01:42
<daya>
subsume, yes
01:42
<subsume>
that should not be.
01:43
<daya>
go thru the content in lts.conf and dhcpd.conf
01:44
<johnny>
there's a way to make it use nfs
01:44
i forgot which file it is tho
01:44
but it is slower
01:48
<daya>
johnny, have u used lenny for ltsp
01:48
<johnny>
no
01:48
ubuntu
01:48
and hopefully soon gentoo
01:49
<daya>
ok
01:50tux_440volt has joined #ltsp
02:09tux_440volt is now known as Subhodip
02:13rcy has joined #ltsp
02:14deavid has joined #ltsp
02:17subsume has quit IRC
02:19daya has quit IRC
02:24subsume has joined #ltsp
02:38Pascal_1 has joined #ltsp
02:44
<Pascal_1>
Bonjour !
02:52subsume has quit IRC
02:53subsume has joined #ltsp
02:58rcy has quit IRC
02:59daya has joined #ltsp
03:02subhodip_ has joined #ltsp
03:03Subhodip has quit IRC
03:10mikkel has joined #ltsp
03:12subhodip_ is now known as tux_440volt
03:30subsume has quit IRC
03:36sep has quit IRC
03:36sep has joined #ltsp
03:51daya has quit IRC
04:18makghosh has joined #ltsp
04:32tux_440volt is now known as Subhodip
04:37Subhodip is now known as tux_440volt
04:42makghosh has quit IRC
04:47makghosh has joined #ltsp
04:53makghosh has joined #ltsp
05:01yopla has joined #ltsp
05:18
<yopla>
morning all
05:19
I wonder if ther's a way to set X server color depth on some ltsp client.
05:19
<johnny>
yes
05:19
see the file lts-parameters on your system
05:19
lts-parameters.txt perhaps
05:21
<yopla>
Thank you...
05:42
<Pascal_1>
clear
05:42
oups
05:47
ogra_??
05:48
what is the difference between connecting on the ltsp server from ssh and connecting to the server from a thin client ?
06:09tux_440volt has quit IRC
06:14
<yopla>
thin client uses by default ldm. It displays greeter and session manager. then it launches the desktop linked to your session through ssh. With ssh, you just have console then eventually launch manually some X apps or desktop.
06:27ogra_cmpc has quit IRC
06:37ogra_cmpc has joined #ltsp
06:58elisboa has quit IRC
06:58elisboa has joined #ltsp
07:01mhterres has joined #ltsp
07:10indradg has quit IRC
07:11indradg has joined #ltsp
07:15indradg has quit IRC
07:15indradg has joined #ltsp
07:21daya has joined #ltsp
07:25
<daya>
any have tried ltsp in lenny,
07:27otavio has joined #ltsp
07:28makghosh has quit IRC
07:29
<daya>
ogra_, any idea about ltsp in lenny, I am not able to get X in client,
07:33
<ogra_cmpc>
no idea, no
07:33
it works in hardy
07:33
<johnny>
so what's the word on aufs ?
07:33
<daya>
ogra_, is there and clue why so, is ldm is buggy in lenny
07:33
<johnny>
daya, you're talking to the wrong person
07:34
<ogra_cmpc>
daya, wait for vagrant to drop by
07:34
<daya>
johnny, :P
07:34
<ogra_cmpc>
he's the one maintaing the debian port
07:35
<daya>
ogra_cmpc, Ok, But I am not meeeting him for the past few days,
07:35Guaraldo has joined #ltsp
07:35
<ogra_cmpc>
he was here last nigh and the night before
07:35
and he'll surely be here tonight
07:35
<daya>
ogra_cmpc, whats make ubuntu and debian a more gaps
07:35jammcq has quit IRC
07:36
<ogra_cmpc>
"night" from a european TZ view :)
07:36
daya, they are different distros
07:36* johnny wonders whether aufs or unionfs will make it to mainlien
07:36* laga hopes for aufs
07:36
<johnny>
unionfs is in mm.. but who knows..
07:37
aufs should get in mm if they want to be in mainline
07:37
<daya>
ogra_cmpc, yes I do realize that, But ubuntu is based on debian isn't it.
07:37
<johnny>
otherwise i doubt it will happen.
07:37
<ogra_cmpc>
ltsp5 is implemented to use the distros system tools and packaging system ... while the packaging system is still not to much divereted between deb and ubuntu (even though we have many differences there as well) the system setup is completely diverted
07:38
daya, and ltsp5 was developed in ubuntu with different focus than vagrant puts it into in debian
07:39
s/puts it into/puts at it/
07:39
<johnny>
for example?
07:39
<ogra_cmpc>
ubuntu = all out of the box ... debian = as configurable as possible
07:39
<daya>
amm
07:39
<ogra_cmpc>
which contradicts in some areas and encanches each other in others
07:40
we're getting closer over time though ... i.e. ubuntu now uses debians ltspfs packages right away
07:41
in the near future the same will happen with ldm
07:41
but ltsp itself will always differ
07:42
<daya>
ogra_cmpc, I am in confusion does sdm perform same as ldm or its used for different purpose
07:42
<ogra_cmpc>
sdm is dead
07:43
<daya>
ok
07:43
<ogra_cmpc>
it predates ldm and was recently abandoned in favor of ldm
07:43
it had better autologin functions and was a tad smalled than ldm
07:44
but with the recent update to ldms autologin that wasnt needed anymore
07:44
<cyberorg>
ogra_cmpc, what is ltsp5's local app project called?
07:44subsume has joined #ltsp
07:45
<ogra_cmpc>
cyberorg, no idea, i wasnt affiliated with all that
07:45
<cyberorg>
what happened of this? http://ltspthinclient.blogspot.com/
07:45
<ogra_cmpc>
there is a spec in launchpad that has all info and code examples afaik
07:46
but nobody touched any code after iot was specded and scott had a proof of concept ready
07:46
i will look into it after may
07:47
if scpott doesnt do anything before :)
07:48
<cyberorg>
xrexecd which is in ltsp-trunk is the launcher?
07:48
<johnny>
i wish i could get to that :)
07:48
stupid genkernel :(
07:52daya has quit IRC
07:56
<yopla>
hello !
07:56Guaraldo has quit IRC
07:56
<yopla>
I'm back for some testing
07:57Guaraldo has joined #ltsp
07:57
<yopla>
can we disable nbd and use nfs instead?
07:57
ltsp-server gutsy bin package
07:57
<laga>
yopla: yes.. it's described somewhere in the wiki
07:58joebaker has joined #ltsp
08:01
<mikesh>
hello, i have ltsp5 at debian lenny, is there some easy way how to fix ldm autologin bug?
08:04subsume has quit IRC
08:06deavid has quit IRC
08:07* ogra_cmpc wasnt aware there was a bug
08:07
<lilian>
mikesh: what kind of bug are you talking about? we use the etch backports from http://pkg-ltsp.alioth.debian.org/debian/, and autologin is working fine.
08:08
<mikesh>
it just didnt work and there is some about it .. http://lists.alioth.debian.org/pipermail/pkg-ltsp-devel/2008-January/001208.html
08:09slidesinger has joined #ltsp
08:09
<mikesh>
i put this lines to my lts.conf
08:09
LDM_USERNAME=user
08:09
LDM_PASSWORD=pass
08:10
and ldm didnt open loginscreen and wm doesnt start
08:10
<lilian>
mikesh: you need LDM_AUTOLOGIN = true too
08:10
<mikesh>
oh
08:10
ok, my false
08:10
sorry
08:11
<lilian>
mikesh: no problem.
08:11
<ogra_cmpc>
a fix like that was pllied (a different one i thonk though)
08:11
*applied
08:11soneyka has quit IRC
08:15soneyka has joined #ltsp
08:17
<mikesh>
lilian: its same, i see just cross (X) :(
08:18
<yopla>
What LDM_GUEST does mean? In dlm-scott branch?
08:18
<lilian>
mikesh: how old is your ldm? autologin was fixed about 4 weeks ago..
08:19
<mikesh>
lilian: my ltsp is 5.0.40~bzr20071229-1
08:21
<yopla>
ho, is there any source.list for gutsy? I still have 5.0.29 one
08:21
<lilian>
mikesh: this is too old for sure. it was fixed in late Feb/early Mar
08:22
<laga>
you won't be able to run hardy's LTSP on gutsy without some (minor) changes.. eg remove the 'no lzma' argument from the mksquashfs call in ltsp-update-image
08:22
and probably some other things
08:23
<yopla>
hardy is the latest distrib?
08:23
<johnny>
hardy isn't out
08:23
it will be out sometime next month
08:23
you should prolly wait
08:24
<yopla>
So I have to use ldm-scott branch...
08:24
and manually copy binaries int chrot
08:25
<johnny>
i wouldn't do that..
08:25
you're gonna break it
08:25
<mikesh>
lilian: ok, thanks
08:26lilian is now known as daduke
08:26
<yopla>
yes my fisrt try was unsuccessfull
08:26
<johnny>
just wait til hardy comes out
08:31
<yopla>
l
08:32
haven't found how to disable nbd
08:33Pascal_2 has joined #ltsp
08:34cliebow_ has joined #ltsp
08:36
<johnny>
it's in there
08:36ogra_ has quit IRC
08:36ogra_cmpc has quit IRC
08:36
<johnny>
but you shouldn't do it
08:37
unelss you are editing the chroot alot
08:37ogra_ has joined #ltsp
08:37ogra_cmpc has joined #ltsp
08:39deavid has joined #ltsp
08:41abadger1999 has quit IRC
08:41
<yopla>
johnny : yes that's what I want for testing
08:42elisboa has quit IRC
08:42
<yopla>
it's in there where?
08:43elisboa has joined #ltsp
08:44
<johnny>
check some file in your $chroot/etc/defaults perhaps?
08:45slidesinger has quit IRC
08:47
<yopla>
humm I think it's initrd related. Perhaps an option in pxe.conf?
08:50slidesinger has joined #ltsp
08:52Pascal_1 has quit IRC
08:54tux_440volt has joined #ltsp
08:59vagrantc has joined #ltsp
09:00
<johnny>
no
09:00prpplague has joined #ltsp
09:01prpplague has left #ltsp
09:04Pascal_2 has quit IRC
09:05mccann has joined #ltsp
09:06GiantPickle has quit IRC
09:06GiantPickle has joined #ltsp
09:08cdealer has joined #ltsp
09:09
<cdealer>
hi all. Some clients are having a mouse issue, when they click one time the click go 2 times, like a dual click.... only with clients, and are not all
09:15abadger1999 has joined #ltsp
09:15Gadi has joined #ltsp
09:21
<mikesh>
daduke: please, can you help me? i have still same problem, but now i have ldm 2:0.1~bzr20080308-1~40.etch.0
09:21Pascal_1 has joined #ltsp
09:21
<Pascal_1>
hello
09:21
gadi ?
09:23
you remember my problem with pam_mount ? i made test on a debian etch , local connection works fine, ssh connection works fine but now when i try to connect from a thin client (ssh also ) the same problem appear : smbumount doesnt works
09:23
have you got an idea ?
09:23
i thought that the connection from a thin client was the same than an ssh connection
09:23
<vagrantc>
cyberorg: glad to hear LTSP5 is going well for you! :)
09:24
<cyberorg>
vagrantc, thanks, now i understand how all pieces come together :)
09:24
<vagrantc>
:)
09:26
<Gadi>
Pascal_1: what was the distro that you used the first time when it was working?
09:27
<Pascal_1>
it never works :-(
09:27
<Gadi>
ah
09:27
hmm...
09:27
<ogra_cmpc>
thats not a helpful bugreport
09:27
<Pascal_1>
it works last week with your help but i didnt try with thin client
09:27
<ogra_cmpc>
:)
09:27
<Gadi>
does etch use the C-based ldm?
09:27
<vagrantc>
no
09:27
<ogra_cmpc>
the backport afaik
09:27
<Pascal_1>
hmm i dont know
09:28
<vagrantc>
unless you use the backport
09:28
<Pascal_1>
what is the way to know that ?
09:28
<Gadi>
file /opt/ltsp/i386/usr/sbin/ldm
09:28
<vagrantc>
Pascal_1: dpkg -l 'ltsp*' | egrep ^ii ; dpkg --root=/opt/ltsp/i386 -l 'ltsp*' ldm | egrep ^ii
09:28
gadi's works well too :)
09:28
<Gadi>
:)
09:29
<ogra_cmpc>
way to short command ... :P
09:29
<Gadi>
call me old fashioned
09:29
:)
09:29
<Pascal_1>
ii ltsp-server 0.99debian11 Basic LTSP server environment
09:29
ii ldm 0.99debian11 LTSP display manager
09:29
ii ltsp-client 0.99debian11 LTSP client environment
09:29
i thought ldm was like ssh
09:29
<ogra_cmpc>
0.99 sounds like the original pkg
09:29
<vagrantc>
ltspbot: learn version as to get version info on debian/ubuntu (please use the pastebot): dpkg -l 'ltsp*' | egrep ^ii ; dpkg --root=/opt/ltsp/i386 -l 'ltsp*' ldm | egrep ^ii
09:29
<ltspbot>
vagrantc: The operation succeeded.
09:30
<vagrantc>
Pascal_1: ldm is basically a GUI frontend to ssh
09:30
<Gadi>
Pascal_1: what do you see in the logs?
09:30
<Pascal_1>
then why it doesnt works as ssh ?
09:30
<Gadi>
is there a proper logout of ssh?
09:30
<ogra_cmpc>
ldm uses: ssh -X user@server /etc/X11/Xsession by default
09:30
make sure to use the -X in your manual tests
09:30
<Pascal_1>
Gadi, my problem is when i use ldm on thin client there is nothing in auth.log at the disconnect
09:31
<vagrantc>
hmmm ....
09:31
<Gadi>
yeah - sounds like it does not disconnect properly
09:31
<vagrantc>
we do do a "kill $PPID" ...
09:31
<Pascal_1>
it's the unique case where there is nothing
09:31
<vagrantc>
not sure if that actually logs out properly.
09:31
<mikesh>
ogra_cmpc: hello, please can you help me with ldm autologin?
09:32
<Gadi>
vagrantc: is 0.99 python?
09:32
<ogra_cmpc>
vagrantc, erm
09:32
i patched that to be ${PPID}
09:32
else it has no effect
09:32* ogra_cmpc thought he did that in the branch
09:33
<ogra_cmpc>
mikesh, i dont use autologin, whats the issue ?
09:33
<warren>
vagrantc, request, could you please preface the commit comment with "Debian: " if the checkin is clearly debian specific?
09:33
<ogra_cmpc>
it should work in the latest versions
09:33
<Pascal_1>
Gadi, any idea ?
09:33
<warren>
vagrantc, I try to use "Fedora: " if the checkin is clearly fedora specific, and don't if it could possibly relevant to other people
09:33
<mikesh>
ogra_cmpc: i have setup ltsp5 with ldm, it works, but it doesnt work with autologin
09:33
<Gadi>
Pascal_1: can you pastebot /opt/ltsp/i386/usr/sbin/ldm ?
09:33* Gadi forgets the code
09:33
<ogra_cmpc>
warren, kudos for so much self-disciplien
09:34* ogra_cmpc aready noticed the prefix
09:34
<warren>
vagrantc, hmm... "call chroot without full path"
09:34
- /usr/sbin/chroot $ROOT passwd root
09:34
+ chroot $ROOT passwd root
09:35
<Pascal_1>
Gadi, ldm is a file what i pastebot ?
09:35
<Gadi>
yes
09:35
it is a file
09:35
<ogra_cmpc>
paste that file
09:35
<Gadi>
in your case a python script
09:35
(hopefully)
09:35
:)
09:35
<warren>
vagrantc, I just had a bug filed where somebody did "sudo ltsp-build-client" and it failed because I wasn't calling chroot with full path
09:35
<cdealer>
what may be causing my ltsp server loadaverage to go higher ... now its on 4.3 if I dont reboot the machine, after 3~4 days the load goes to 17 or more... O.o
09:35
<warren>
where is chroot located on your distros?
09:36
Not /usr/sbin/chroot?
09:36
cyberorg, ping
09:36
<ogra_cmpc>
/usr/sbin is in my path
09:36
<Gadi>
how about: sudo $(which chroot) /opt/ltsp/i386 ltsp-build-client
09:36
:)
09:36
<warren>
/usr/sbin isn't in the PATH of non-root users in Fedora
09:36
<cyberorg>
warren, pong
09:36
/usr/bin/chroot
09:36
<warren>
cyberorg, is chroot at /usr/sbin/chroot on suse?
09:36
<Pascal_1>
Gadi, i send you the file ?
09:36
<ogra_cmpc>
Gadi, rather in a common function and feed $CHROOTEXEC or so with it
09:36
<warren>
/usr/bin? crap.
09:37
<cyberorg>
why?
09:37
<ogra_cmpc>
warren, gentoo is also in /usr/bin it seems
09:37
<Gadi>
Pascal_1: why not paste?
09:37
I like your pastes
09:37
:)
09:38
<warren>
OK, how about this: an early common/ plugin detects the location of chroot and sets it to CHROOTEXEC
09:38
agreed?
09:38
<ogra_cmpc>
no
09:38
not a plugin
09:38
<warren>
why?
09:38
<ogra_cmpc>
lets put it in ltsp_functions or so
09:38
<warren>
ltsp-build-client itself?
09:38
<ogra_cmpc>
or even that
09:38vagrantc has quit IRC
09:39* warren checks
09:39
<ogra_cmpc>
seems we scared vagrant :)
09:39
<Pascal_1>
iwhat is the url of pastebot
09:39
<Gadi>
!pastebot
09:39
<ltspbot>
Gadi: "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.
09:40
<warren>
ogra_, i'm not entirely happy about setting it in /usr/share/ltsp/plugins/functions
09:40
<ogra_cmpc>
hum ... where did ltsp_functions go ?
09:40
<ltsppbot>
"pascal" pasted "#!/usr/bin/python import os im" (214 lines) at http://pastebot.ltsp.org/473
09:41
<warren>
ogra_cmpc, because nothing else gets executed in there when it is merely sourced
09:41
<ogra_cmpc>
no, i was talking about a function script we used to source from ltsp-build-client
09:41
seems that vanished
09:41vagrantc has joined #ltsp
09:41
<Gadi>
Pascal_1: try changing the line that reads: 'kill -1 $PPID']
09:41
to: 'kill -1 ${PPID}']
09:42
<ogra_cmpc>
you can also make it -9 :)
09:42
<warren>
ogra_cmpc, what's wrong with an early plugin?
09:42
<ogra_cmpc>
if you feel evil enough
09:42
<vagrantc>
warren: yeah, sorry about that. makes good sense.
09:42
<warren>
vagrantc, seems that we do need absolute path for chroot, but different distros have different paths
09:42
vagrantc, thus we're talking about where to detect the location of chroot and set $CHROOTEXEC
09:42
<vagrantc>
build time.
09:42
<warren>
vagrantc, ltsp-build-client itself, or an early plugin
09:42
<ogra_cmpc>
vagrantc, we're a comon plugin to determine the full path of the chroot command or a function in ltsp-build-client
09:43
<vagrantc>
well, i guess with ltsp-build-client, i don't care if it's run-time or build time
09:43
<ogra_cmpc>
s/we're/we're discussing/
09:43* vagrantc is doing laundry at a laundromat
09:43
<warren>
I'm against putting it in /usr/share/ltsp/plugins/functions because nothing else in there is executed or set by merely sourcing it.
09:43
<Pascal_1>
gadi i try
09:43
<vagrantc>
so i'm only like 35% here
09:43
<ogra_cmpc>
vagrantc, we used to source something like ltsp_functions in ltsp-build-client, wher did that go ?
09:43
<vagrantc>
ogra_cmpc: it's just plugin-specific functions
09:43
<warren>
ogra_cmpc, that must have been long before I began hacking on it
09:44
<ogra_cmpc>
vagrantc, there was a differnt thing we had in ltsp-build-client directly
09:44
<Pascal_1>
ogra_cmpc, it's better to make -9 than -1 ?
09:44
<warren>
I think ltsp-build-client or an early plugin is the best place to put this.
09:44
Just pick one and I'll write it
09:44* ogra_cmpc votes for ltsp-build-client
09:44* warren votes for ltsp-build-client
09:44
<vagrantc>
i like plugin
09:45
<warren>
gah!!
09:45
=)
09:45
<vagrantc>
because then it's possible for the distro to override it
09:45
<ogra_cmpc>
actually i would suggest to have something like _functions back that finds all executable pathst for us
09:45
<warren>
what other paths do we need?
09:45
<ogra_cmpc>
plugins can still override that
09:45
<Pascal_1>
Gadi, kill -9 or kill -1 ?
09:45
<vagrantc>
FOO and BAR
09:45
you never know what crazy things other distros are going to do.
09:45
<ogra_cmpc>
warren, netcat
09:46
<Gadi>
Pascal_1: hmm... try both
09:46
<Pascal_1>
ok
09:46
<ogra_cmpc>
to give a recent example
09:46
<Gadi>
(not at the same time)
09:46
<warren>
ogra_cmpc, netcat isn't used during ltsp-build-client
09:46
<Pascal_1>
Gadi, ;-)
09:46
<warren>
ogra_cmpc, and we don't share any code that uses netcat
09:46
<ogra_cmpc>
ldm does
09:46
<Gadi>
Pascal_1: /me tries to recall what he did - bec I was definitely using the python ldm when I was playing with this....
09:46
<vagrantc>
the screen script does
09:46
<warren>
OK, I'm with vagrantc on plugin.
09:46
<ogra_cmpc>
right
09:46
<warren>
vagrantc, which screen script?
09:46
<vagrantc>
warren: you don't use the ldm screen script?
09:47
<ogra_cmpc>
well, i give up then, being overruled here
09:47
<vagrantc>
brb
09:47
<ogra_cmpc>
make it a plugin
09:47
<Pascal_1>
Gadi, i try and i tell you
09:47
<warren>
oh!
09:47
<ogra_cmpc>
warren, i'm pretty sure we'll get more like these in the future
09:48
<Pascal_1>
Gadi, do i have to reboot the ltsp server ?
09:48
<Gadi>
Pascal_1: no
09:48
just the client
09:48
<Pascal_1>
ok
09:48
<warren>
ogra_cmpc, it isn't relevant for the client to spend any cycles looking for the location of chroot.
09:48
<Gadi>
but make sure you don't have stray ssh processes on the server
09:48
from previous logins
09:48
and make sure you dont have stray mounts
09:50
<Pascal_1>
stray ?
09:50
remember i'm french ;-)
09:50
<Gadi>
left over
09:50
remaining
09:50
residual
09:50
present but unused and unwaanted
09:51
*unwanted
09:51
<vagrantc>
still running
09:51
<Gadi>
un amok
09:51
*run
09:51
indefatigable
09:51* vagrantc wonders how long thesaurus-fest will continue
09:51
<Gadi>
:)
09:52
<vagrantc>
ltsp thesaurus-a-thon
09:52
<Gadi>
its about as long as the toilet paper roll
09:52
;)
09:52
<Pascal_1>
;-)
09:52* warren silent.wav
09:53
<Gadi>
:)
09:54ogra_cmpc_ has joined #ltsp
09:55
<ogra_cmpc_>
grmbl
09:55ogra_cmpc has quit IRC
09:57
<Pascal_1>
Gadi, kill -1 ${PPID}'] doesnt works i try with kill -9 ${PPID}']
09:58
kill -9 ${PPID}'] doesnt works any more
09:58
and always no log in auth.log
09:59
<Gadi>
hmm
09:59
<Pascal_1>
it'is the only case that there is no log
10:01
with ssh , with your help it works fine and the logs are here
10:02
on the thin client (ldm) i've got only the connection logs
10:03
<Gadi>
and on the client, it returns to login screen after logout?
10:04
<Pascal_1>
yes
10:04
i change the theme
10:04
it's not a problem ?
10:05
<vagrantc>
hmmm....
10:05
<cyberorg>
warren, here is our process diagram http://cyberorg.kicks-ass.org/~cyberorg/kiwi-ltsp-diagram.png
10:05
i don't understand any of the plugin thing :(
10:06
here is the script that does this https://forgesvn1.novell.com/viewsvn/kiwi-ltsp/trunk/kiwi-ltsp/ltsp/suse-11.0/kiwi-ltsp-setup?revision=117&view=markup
10:06
<Gadi>
Pascal_1: shouldnt be
10:06
Pascal_1: I have a mtg - bbiab
10:06
<Pascal_1>
a what ?
10:07
<Gadi>
meeting
10:07
<vagrantc>
cyberorg: since you have an existing infrastructure to install ltsp, you just drop a plugin in to call that.
10:07
<cyberorg>
vagrantc, that is what i sent on the ML
10:07
<vagrantc>
cyberorg: if there's any commandline options you could pass to that command, then add commandline options in the plugin
10:07
cyberorg: what are you not understanding with the plugin system?
10:07
<cyberorg>
vagrantc, the diagram shows what each commandline options do
10:08
<vagrantc>
cyberorg: i can't check the diagram right now
10:08
<Pascal_1>
then Gadi any idea ?
10:08
:-(
10:09
<cyberorg>
vagrantc, remember you and me came up with the plugin that called kiwi-ltsp-setup -s? i sent that on the list
10:09
<vagrantc>
cyberorg: yes, i know.
10:09
cyberorg: we should give you commit access to the bzr repository so you can commit it yourself.
10:09
<warren>
so he can commit a single plugin of 1 line?
10:10
<cyberorg>
that would be suse plugin that does everything in correct order, the script can be called with different switches to do any process in arbitrarily
10:10
<vagrantc>
warren: and fix other bugs as needed
10:10
<warren>
Are we reasonably certain that he knows what not to change that might break other people?
10:10
He would also need to know how not to merge and push destructively.
10:11* ogra_cmpc_ would suggest to work on a branch for a while
10:11
<vagrantc>
yes, unfortunately, bzr doesn't take care of that...
10:11
<cyberorg>
lol, warren i have full access to compiz on fd.o and compiz-fusion git without breaking anything :)
10:11
<ogra_cmpc_>
we can regulary merge
10:11
<vagrantc>
maybe ogra_cmpc_'s suggestion is the way to go
10:11
<warren>
oh?
10:11
<vagrantc>
just publish a branch and request to merge
10:12
i'd prefer bzr merges to applying patches
10:12
<warren>
you want *me* to do that too?
10:12
<cyberorg>
vagrantc, i don't mind sending patches to mailing list though, if i do not have to wait long for them to be commited
10:12
<ogra_cmpc_>
warren, i think we should have a probation time for everyone actually ...
10:12
<vagrantc>
warren: in general, that's probably a good working model... although i trust you to handle your merges yourself
10:12
<ogra_cmpc_>
commit to a branch for a while and come back after two months or so
10:13
<vagrantc>
cyberorg: i would rather see a bzr branch i can merge than patches, but if you are absolutely opposed to publishing a bzr branch, then i guess we'll have to work with each other.
10:14
i've probably gotten in the habit of committing directly too often
10:14
<cyberorg>
vagrantc, i already have svn tree of whatever i do, i would not like to keep one more place to maintain if it can be avoided
10:14
<vagrantc>
cyberorg: well, if you want changes to be accepted upstream... you might want to work with the tools upstream has chosen...
10:14
<ogra_cmpc_>
cyberorg, well, if you get commit acces you will also be the responsible person for adding other suse devs to the ACLs
10:15
<cyberorg>
and without much coding knowledge there wont be serious amount of code coming from me
10:15
<vagrantc>
cyberorg: there's also bzr-svn ... where you can sync between bzr and svn ...
10:16
<cyberorg>
ogra_cmpc_, my account name is cyberorg on launchpad
10:16
<warren>
ogra_cmpc_, I am NOT waiting 2 months
10:17
<cyberorg>
how do we go about doing it the right way :)
10:18
<ogra_cmpc_>
warren, i'm not talking about you, but i want to make sure cyberorg unserstands all the code, if there is some suse dude he added to the acl who changes things in common areas (for example rewrites ltsp-build-client) cyberorg needs to be able to judge if thats suitable or not
10:18
warren, i know you know the code since two years
10:18
<warren>
ogra_cmpc_, well, not 2 years
10:18
<ogra_cmpc_>
well, since detroit at least
10:19
in any case we need some QA for people we add to the ACL
10:19
probation time would be one option
10:20
i.e. knowing and understanding the plugin system would be a prerequisite imho simply bevcause you may need to work on common plugins that affect all of us
10:22abadger1999 has quit IRC
10:22
<cyberorg>
for now i just need a plugin added, one line change in lsb_release command and Xsession path check for suse in ldm.c
10:23
<warren>
the Xsession thing might become more dicey
10:23abadger1999 has joined #ltsp
10:23
<warren>
I need to change our ldm to launch *any* session via Xsession
10:23
<vagrantc>
default Xsession should be handled at build time
10:24
debian's already supports that, although we don't actually use it :(
10:24
<ogra_cmpc_>
warren, fedora doesnt do that ?
10:24
<vagrantc>
we just call the binary directly ....
10:24
<ogra_cmpc_>
vagrantc, all old releases did rely on alternatives and used it like that
10:25
vagrantc, the binary is a script that walks the users home and the alterantives
10:25
<cyberorg>
suse's /etc/X11/xdm/Xsession supports any Xsession, i have kde4 clients booting here :)
10:25
<ogra_cmpc_>
cyberorg, right thats what ubuntu and debian have as well
10:25
<Pascal_1>
Gadi, after your meeting could you help me ?
10:25
<ogra_cmpc_>
warren, i saw fedora plans to switch to upstart ?
10:26
<vagrantc>
ogra_cmpc_: the current code still calls the binnary directly, and not as an argument to Xsession
10:26
<cyberorg>
i get nice dropdown to select session after i fixed nc to netcat
10:26
<ogra_cmpc_>
vagrantc, indeed, it calls whatever alternative is set
10:26
<vagrantc>
ogra_cmpc_: which means that if you select your session, rather than using the default, you don't get all the stuff Xsession sets up
10:26
<ogra_cmpc_>
right
10:26
thats true
10:26
<vagrantc>
ogra_cmpc_: well, LDM calls whatever ldminfod tells it to call :)
10:27
could ldminfod tell it /etc/X11/Xsession /path/to/selected/foosession
10:27
<cyberorg>
there we fixed one more potential bug :)
10:27
<warren>
ogra_cmpc_, you asked about this a week or two again
10:27
<vagrantc>
that way it could be implemented server-side ... so you could log into a fedora server and get the right options
10:28
<ogra_cmpc_>
vagrantc, you could dump a teporary ~/.xsession into place (ugly, i know)
10:28
<warren>
vagrantc, yeah but ldm also displays the entire name too, which isn't too nice.
10:28
<ogra_cmpc_>
it overrides the system default
10:28
<vagrantc>
ogra_cmpc_: no. :P
10:28
<ogra_cmpc_>
warren, nope, i asked aboyut alternatives last time
10:28
<warren>
ogra_cmpc_, we were using upstart for weeks now
10:28
<ogra_cmpc_>
i got aware of upstart pon the weekend only ,,,
10:28
funny
10:28
<vagrantc>
warren: well, long-term it requires more changes. short term we could do a basename on it or something
10:29
<ogra_cmpc_>
huh
10:30
whats the prob here ? we have a user readable name in the .desktop files
10:30
<warren>
common/000-check-paths ?
10:30joebaker has left #ltsp
10:30
<warren>
any better name?
10:30
<ogra_cmpc_>
(if you had listened to me and took xdg desktop file utils it would be one more option to just get the real name :P )
10:31
<vagrantc>
actually, we could just set it in 000-basic-configuration
10:31
<ogra_cmpc_>
but you can indeed grep it out there as well
10:31
<vagrantc>
chroot and whatever else...
10:31
<warren>
vagrantc, doesn't exist in common
10:31
<ogra_cmpc_>
does it need to ?
10:31
<vagrantc>
i guess it's no big deal to set it in common.
10:31
000-check-paths is probably good
10:32
<ogra_cmpc_>
sounds sane
10:32staffencasa has joined #ltsp
10:32
<vagrantc>
would it be evil to write a function called "chroot" that handles the detection and sets a variable ?
10:33
<ogra_cmpc_>
sounds evil
10:33
<warren>
yes
10:33
<vagrantc>
ok
10:33
<Pascal_1>
ogra_cmpc_, any idea why i disconnect from a thin client there is no log in auth.log ?
10:34
<vagrantc>
gotta run. see y'all later.
10:34
<ogra_cmpc_>
Pascal_1, no
10:34vagrantc has quit IRC
10:35
<warren>
ogra_cmpc_, Ubuntu has chroot in /usr/sbin/chroot ?
10:35
<Pascal_1>
is there a way to log all it happen during connection and disconnection ?
10:35
<ogra_cmpc_>
warren, same as debian, yes
10:36
<warren>
ok, then I'll check for /usr/sbin/chroot and /usr/bin/chroot
10:36
<ogra_cmpc_>
firwst for $(which chroot) :)
10:37
only if its not in $PATH you actually need to search
10:37
and /usr/bin should be in every systems user PATH
10:38
so if $(which chroot) doesnt return andthing we check if its in /usr/sbin ... if its missing there we exit
10:38
*anything
10:40
<Pascal_1>
yopla, you told me something about the difference between ssh connection and ldm connection, could you told me again ?
10:42slidesinger has quit IRC
10:42
<warren>
ogra_cmpc_, checking for paths would be faster
10:42
ogra_cmpc_, there's no fork involved
10:43
<ogra_cmpc_>
but the code gets bigger, i dont think speed is an issue during -build-client
10:43
<warren>
which fails in my case
10:43
=)
10:44
<ogra_cmpc_>
??
10:44
<warren>
ogra_cmpc_, chroot is not in PATH
10:44
which wont work here
10:44
<ogra_cmpc_>
thats what its supposed to :)
10:44
right
10:46
if [ ! -n "$(which chroot)" ]; then if [ -x /usr/sbin/chroot ]; then CHROOTEXEC=/usr/sbin/chroot .... else echo "no chroot found"
10:46
<warren>
it's more code than that
10:47
less code the other way
10:47
<ogra_cmpc_>
then do it your way :)
10:49
<warren>
heh
10:49
-dd if=/dev/zero of=$SWAP bs=1024k count=$SIZE 2> /dev/null
10:49
+dd if=/dev/zero of=$SWAP bs=1024k count=0 seek=$SIZE 2> /dev/null
10:49
vagrant's change to nbdswapd
10:50
<ogra_cmpc_>
heh
10:50
<warren>
that's actually a clever idea
10:51indradg has quit IRC
10:51
<warren>
I just pushed CHROOTEXEC to ltsp-trunk
10:51
you'll want to update Debian and Ubuntu directories later
10:51
<ogra_cmpc_>
i dont really care atm :)
10:51* ogra_cmpc_ is in pre release stress
10:52
<warren>
nod
10:52slidesinger has joined #ltsp
10:52
<warren>
ogra_cmpc_, you work on other parts of Ubuntu as well?
10:52
ogra_cmpc_, or own?
10:52
<ogra_cmpc_>
and the classmate image builder needs my full attention atm
10:52
i just moved to the platform team after the boston summit
10:53
<warren>
is classmate still relevant given that competitors are better?
10:53
<ogra_cmpc_>
so theoretically i am responsible for the layer between kernel and desktop
10:53
there are millions of pre orders for classmates
10:54
i guess at least the existing contracts will be worked through
10:54
but i cant talk on behalf of intel ... no idea what they will do
10:55
but all that stuff i do here will work on other subnotebooks with some adaption
10:56
(the hardcoded module lists etc need to be changed and the predefined partition sizes would need to be adjusted, but then it would make a fast tiny OS for teh eeePC as well)
10:57gentgeen__ has joined #ltsp
10:59
<ogra_cmpc_>
warren, did i show you that ? http://code.google.com/p/compcache/
10:59
its pretty cool, i plan to experiment wiht it a bit in the next release
10:59mcfloppy is now known as floppybaer
10:59
<ogra_cmpc_>
we could drop the minimal requs for ltsp a lot if it works as advertised
11:03alekibango has quit IRC
11:05vagrantc has joined #ltsp
11:09indradg has joined #ltsp
11:12
<Pascal_1>
vagrantc, any idea for my problem ?
11:13alekibango has joined #ltsp
11:16lns has joined #ltsp
11:16alekibango has quit IRC
11:17Pascal_1 has quit IRC
11:19spectra has joined #ltsp
11:22
<warren>
ogra_cmpc_, userspace or kernel space compression?
11:22alekibango has joined #ltsp
11:23
<ogra_cmpc_>
kernel
11:23
as i understand the explanation at the top at least
11:24
though i wonder why the redirectiion through a swapfile is needed, you could as well just add the compression to the standard ram handling
11:24
(with a config option indeed)
11:26
<warren>
will that be submitted to upstream kernel?
11:26
At the very least it would have an upstream review that might improve the code
11:29
<vagrantc>
compcache ?
11:31
<cdealer>
how can I know how much disk i/o my server is doing ?
11:38
<ogra_cmpc_>
i doubt it will go upstream
11:38twinprism has quit IRC
11:38twinprism has joined #ltsp
11:41alekibango has quit IRC
11:48Faithful has quit IRC
11:48sbalneav has joined #ltsp
11:48
<sbalneav>
Morning all
11:50
<ogra_cmpc_>
!s
11:50
<ltspbot>
ogra_cmpc_: "s" is Scotty!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
11:50* vagrantc waves to sbalneav
11:50alekibango has joined #ltsp
11:53
<vagrantc>
sbalneav: you figured out the xauth stuff? :)
11:53
<sbalneav>
Yes I did
11:53
:)
11:53
Got it going late last night.
11:53
See my post?
11:53
<vagrantc>
yes, reading it now.
11:53
just couldn't contain my excitement.
11:53
<sbalneav>
Hey ogra_cmpc!
11:54
If we could get going what's in that post, it would HUGELY clean up a whole lot of things.
11:54
make ogra happy for the CK stuff
11:55
eliminate the entire LTSPROCKS ssh chat ugliness.
11:55
<ogra_cmpc_>
sbalneav, how much is it ? i'm unlikely to get changes with more then ten lines past the release manager now
11:55
<vagrantc>
LTSPROCKS makes me smile, though.
11:55
<sbalneav>
it's just a short script added to the rc.d
11:56
<ogra_cmpc_>
ah, cool
11:56
<sbalneav>
hold on, I'll paste the gist of is.
11:56* ogra_cmpc_ has no mailer handy, to busy with cmpc atm
11:57
<vagrantc>
UFDA. new options to sshd
11:58
<ltsppbot>
"sbalneav" pasted "S01-setup-xauth" (15 lines) at http://pastebot.ltsp.org/474
11:58
<sbalneav>
UFDA?
11:58
<vagrantc>
if it requires modifications to sshd ... i really can't envision it being included in debian this release cycle.
11:59
<sbalneav>
Right, I think this is more of a long term goal.
11:59
<vagrantc>
which will hopefully close by the end of the year.
11:59
ldm3
11:59
<ogra_cmpc_>
should that be foople_authority ?
11:59
<vagrantc>
:)
11:59
<ogra_cmpc_>
:)
11:59
<sbalneav>
Certainly, we'd want to convince the ssh authors that this is necessary.
12:00
<vagrantc>
in the short term, would it be feasible to still handle xauth properly, but break expiry?
12:00
<sbalneav>
Oh, well, this will both not break expiry, and handle xauth.
12:00
We just have to keep doing the password chat for now.
12:00
<vagrantc>
right, but it's not feasible in the near future
12:01
i.e. using ssh-askpass
12:01
<sbalneav>
Right. If we switch to ssh-askpass, we can't do expiry.
12:02
which was the whole reason I did a chat instead of ssh-askpass in the first place.
12:02
<vagrantc>
but with ssh-askpass, it's easier to properly handle xauth ?
12:03
<sbalneav>
No, no different. The two aren't related. Sorry, I was talking about 2 different things in the email: xauth will be fixed regardless of wether we use a chat, or ssh-askpass.
12:04
That email was sort of a response to Warren about the whole chat/lang thing being hacky.
12:04
<ogra_cmpc_>
right, it hooks into the second ssh call anyway
12:04
<sbalneav>
this would be the way to eliminate it.
12:04
be back in 10 minutes or so: workping....
12:05
<vagrantc>
sbalneav: ah, great!
12:08alekibango has quit IRC
12:11
<warren>
sbalneav, chat/lang?
12:15
<sbalneav>
Back
12:16
<ogra_cmpc_>
warren, the ssh chat
12:16
<sbalneav>
warren: remember on friday we talked about the ssh chatting, and detecting strings to chat on, not being language specific, etc.
12:17
Basically, I put some thought as to what would be needed to eliminate chatting with ssh altogether.
12:18
Hm, isn't ltspbot around?
12:18
<vagrantc>
ltsppbot: hi
12:18
ltspbot:
12:19
!status
12:19
<ltspbot>
vagrantc: I am connected to freenode as ltspbot.
12:19
<sbalneav>
Ah, there he is
12:23floppybaer has quit IRC
12:32
<warren>
Are you folks with NFS root using no_root_squash in /etc/exportfs?
12:32
I mean /etc/exports
12:37* Gadi wave to sbalneav! Welcome back!
12:37
<Gadi>
saw ur post
12:39Phisbut has joined #ltsp
12:40alekibango has joined #ltsp
12:40alekibango has joined #ltsp
12:43
<vagrantc>
warren: yes, as far as i know, no_root_squash is absolutely necessary.
12:44
<warren>
ah, nobody told me until now
12:44
<vagrantc>
warren: otherwise the UID's all come up as nobody
12:44
warren: well, it was in the common plugins, although ogra may have moved it to Debian since Ubuntu no longer uses NFS
12:46
<ogra_cmpc_>
i didnt move anything
12:46Guaraldo has quit IRC
12:46daduke_ has joined #ltsp
12:46
<ogra_cmpc_>
i want nfs configured even i dont use it
12:46
so users can switch easily
12:47
<cliebow_>
!seen sbalneav
12:47
<ltspbot>
cliebow_: sbalneav was last seen in #ltsp 27 minutes and 23 seconds ago: <sbalneav> Ah, there he is
12:47
<vagrantc>
ogra_cmpc_: you talked about moving it
12:47Guaraldo has joined #ltsp
12:47
<ogra_cmpc_>
hmm
12:48
<Phisbut>
Hello people. I'm new to LTSP and I have a little problem with my setup. I just finished installing Ubuntu Dapper (Server) and configuring LTSP.
12:48
<vagrantc>
ogra_cmpc_: and the commit logs confirm that you did it :P
12:48
<ogra_cmpc_>
meh
12:48
silly me
12:48
<Phisbut>
I can boot a workstation with PXE and get to the login screen. However, I won't let me log it graphically (no error message).
12:49
When I go to a console (Ctrl-Alt-F1), I can login, but when I try to write anything to my home directory, I get the following error : Cannot touch `test`: read-only filesystem
12:49
<ogra_cmpc_>
Phisbut, which ltsp release, distro ?
12:49
<Phisbut>
Ubuntu 6.06 Server, comes with LTSP-5
12:49
On the server, my /etc/exports file specifies "read-write" : /opt/ltsp *(rw,no_root_squash,sync)
12:50
<ogra_cmpc_>
ubuntu server ?
12:50
but you installed the ubuntu-desktop package rught ?
12:50
*right
12:50
<Phisbut>
uhh... not that I know of... the server edition boots to a command line
12:50
<ogra_cmpc_>
(ltsp doesnt work without a desktop installer on the server machine)
12:51
!ubuntu
12:51
<ltspbot>
ogra_cmpc_: "ubuntu" is a Debian based GNU/Linux distrubution that includes a Ubuntu specific version of LTSP. It can be found at http://www.ubuntulinux.org. to install ltsp on ubuntu: https://help.ubuntu.com/community/UbuntuLTSP/LTSPQuickInstall
12:51
<ogra_cmpc_>
see the latter URL
12:51
<Phisbut>
Then what's the point of having a "server edition" if you need the desktop stuff anyway?
12:53
<vagrantc>
Phisbut: LTSP is essentially a "Desktop Server" ... it serves up desktops.
12:53
<ogra_cmpc_>
well, ltsp requires a desktop to log in to
12:53
<vagrantc>
Phisbut: unless you want plain console logins.
12:53
<ogra_cmpc_>
indeed :)
12:53
<vagrantc>
then you don't need the desktop packages on the server.
12:53jammcq has joined #ltsp
12:53
<jammcq>
hey kids
12:54
<ogra_cmpc_>
yo jammcq
12:54* vagrantc waves to jammcq
12:54
<Phisbut>
I thought the client desktop packages were built from the "ltsp-build-client" command
12:54
<jammcq>
what's new?
12:54
<warren>
hey
12:54
<ogra_cmpc_>
Phisbut, the thin client OS is
12:54yopla has quit IRC
12:54
<ogra_cmpc_>
Phisbut, that only boots an X server and ldm
12:54
(the ltsp login manager)
12:54
<Phisbut>
ok, I got that
12:55
<ogra_cmpc_>
if you log in ldm does essentiallys: ssh -X user@server /etc/X11/Xsession
12:55
<Phisbut>
however, that doesn't explain why, even in a console login, I can't write to my own $HOME
12:55
<ogra_cmpc_>
how do you manage to log in
12:55
?
12:55
<Phisbut>
ctrl-alt-f1
12:55
<ogra_cmpc_>
there are no userws inside the client system and root is locked on purpose :)
12:56
<Phisbut>
ok, so the $HOME directory of a user logged on a workstation would be the $HOME directory of that same user on the server?
12:57
<vagrantc>
if logged in through LDM, yes.
12:57
<ogra_cmpc_>
since you shouldnt log in on the console ... if you want console logins thorugh telnet you can define that in lts.conf so it gets started by the appropriate screen.d script
12:58
(not sure which other console methods are there, i have actually never tried them)
12:58* vagrantc uses shell all the time
12:58
<ogra_cmpc_>
but i dont understand how you could log in at all
12:58
<vagrantc>
but that just gives a root shell without any authentication
12:58
<ogra_cmpc_>
right
12:59
<vagrantc>
unless dapper pre-dates the locking of root password?
12:59
<ogra_cmpc_>
nope
12:59
<vagrantc>
but i thought i got that in much earlier
12:59
<ogra_cmpc_>
that was in since the first day
12:59
its ubuntu policy
12:59
<Phisbut>
well... I don't see why I shouldn't be able to log in... ltsp-build-client builds a whole system, and then the boot through PXE roots to that system, which is able of logging in
13:00
<ogra_cmpc_>
Phisbut, well, if you can log in wiht an account there is clearly something very broken
13:00
since thats disabled in all areas on a thin client system (ion purpose as i said)
13:01
<Phisbut>
I can login with a *new* account created from the server once chrooted to /opt/ltsp/i386... I *can't* login with an account that is on the server
13:01
<ogra_cmpc_>
no users and disabled root account so i dont understand how you can log in
13:01
right
13:01
so you created an account
13:01
<Phisbut>
I guess I shouldn't have done the "create a new account" part then...
13:02
<ogra_cmpc_>
no, but you should have mentioned it :)
13:02
<Phisbut>
lol... ok, now I know, so here goes : By the way, I created a new account while chrooted ;-)
13:02* Phisbut uncreates the account...
13:02
<ogra_cmpc_>
anyway, if you want a desktop login, there needs to be a desktop on the LTSP server, thats how ltsp works
13:03
any specific reason why you use dapper ?
13:03
<Phisbut>
ok, that wasn't clear to me, I thought the whole file hierarchy created under /opt/ltsp/i386 included that desktop stuff
13:03
<ogra_cmpc_>
no, its only for getting HW support and the login manager up
13:04
everything else happens on the server
13:04
else you would need quite beefy clients :)
13:10tux_440volt is now known as Subhodip
13:13
<Phisbut>
in retrospect, did I miss something or is there really no documentation on how to get a *graphical* LTSP install from Ubuntu Server Edition? (all the doc I've seen start from a desktop install, which obviously already has gnome)
13:17Egyptian[Home] has quit IRC
13:19
<ogra_cmpc_>
Phisbut, i wasnt even aware that it is on the dapper server CD ... its an installer item in the latest laternate CDs for hardy now
13:19
*alternate
13:19abadger1999 has quit IRC
13:19
<ogra_cmpc_>
and not included in -server nowadays
13:20
<Phisbut>
It's actually a "selling point" of the Server Edition... http://www.ubuntu.com/products/WhatIsUbuntu/serveredition
13:20
Quote "Ubuntu Server edition includes thin client support using LTSP (Linux Terminal Server Project). LTSP-5, the latest release, offers a simple installation and easy maintenance."
13:20
<ogra_cmpc_>
silly marketing
13:20
i'll talk to the guys to rephrase that
13:21
<Phisbut>
which is why I picked the server edition for the *drum roll* ... server part of the terminal-server setup ;-)
13:21
<ogra_cmpc_>
heh
13:21abadger1999 has joined #ltsp
13:21
<ogra_cmpc_>
you are really better off with the desktop CD
13:21
<Phisbut>
duly noted
13:21
<ogra_cmpc_>
sorry for that the text is clearly confusing
13:22
<Phisbut>
not your fault
13:22
(unless of course you wrote those lines yourself, in that case, you can rot in hell for all I care :-P )
13:22
<ogra_cmpc_>
well, i'm in charge of ltsp in ubuntu :)
13:22
and they use to ask me before making statements .... (normally)
13:22Subhodip has quit IRC
13:23
<Phisbut>
don't we all love the marketing dept... I feel your pain
13:23
<ogra_cmpc_>
heh
13:23
so whats the reason for taking dapper ?
13:23
<Phisbut>
LTS
13:23
<ogra_cmpc_>
do you plan to buy support ?
13:24
all later ltsp versions are far beyond dapper ... if you are not bound by company policy to support contract to use old software i'd really pick something new
13:24
<Phisbut>
not yet. actually, we're an ISV, and we're deploying test environments to mirror what our clients (current and potential) might have, to make sure our software works well in such setups
13:25
<ogra_cmpc_>
ah, k
13:25
then LTS is inded the better choice
13:25
<Phisbut>
since some of our potential clients are in some gov branches, we figured they'd be using LTS
13:25
<ogra_cmpc_>
next LTS is due in 4 weeks btw
13:25Egyptian[Home] has joined #ltsp
13:25
<Phisbut>
tell that to my boss ;-)
13:25
<ogra_cmpc_>
betarelease will be this week
13:26
<Phisbut>
I'll probably have to test both (6.06 and 8.04) anyway, for at least as long as 6.06 is supported by Canonical
13:26
<cliebow_>
Phisbut, it should be worth the wrath of khan to look to the hardy release for lts
13:27
<ogra_cmpc_>
cliebow_, well, he cant choose his customers ... makes sense to know both :)
13:27Egyptian[Home] has quit IRC
13:27
<cliebow_>
ah..well
13:28
<Phisbut>
yeah... what I'd give to have the right to tell any customer "just upgrade your distro and quit bugging us"...
13:28
<cliebow_>
heh..wouldnt thsat be nice..
13:29Pascal_1 has joined #ltsp
13:29Gadi has left #ltsp
13:29
<Phisbut>
Dapper ain't that bad... I still have to support Fedora Core 4 :-(
13:30
<ogra_cmpc_>
ltsp will only be supported for another year on dapper btw
13:30
<Phisbut>
cool
13:31
<ogra_cmpc_>
since its bound to the desktop and desktop support for dapper runs out mid 2009
13:31
<Phisbut>
but why ding just this one feature while supporting the rest?
13:31
well... it doesn't look like it's bound to the desktop anymore, thanks to marketing
13:31
<ogra_cmpc_>
its not about features :) its about packages
13:32
packages that belong to server (dhcpd and friends) will be supported for 5 years indeed
13:32
but desktop is limited to 3 ... ltsp is tied to desktop functionallity, so dies with the rest of the desktop packages wrt support
13:33
<Phisbut>
and when I get Hardy, I still should get the Desktop for LTSP instead of the server, right?
13:34
<ogra_cmpc_>
on hardy you should take the alternate CD and select "thin client server" from the modes menu on the cd bootscreen
13:34
<Phisbut>
ok, thanks
13:35
<ogra_cmpc_>
if your server has two NICs it will work out of the box after first reboot
13:35
<Phisbut>
great, you've been very helpful, thanks
13:35
<ogra_cmpc_>
youre welcome :)
13:35
<Phisbut>
and off I go downloading the alternate Dapper CD...
13:37
oh, LTSP isn't listed on the Alternate feature for Dapper, is it only there for Hardy?
13:39
<ogra_cmpc_>
right
13:39
on dapper you need to follow the wiki
13:40
the only flavour that had ltsp fully included in dapper was edubuntu
13:40
<Phisbut>
... and server ;-)
13:40
<ogra_cmpc_>
the functionallity you find in hardys installer is based on that
13:44Pascal_1 has quit IRC
13:47
<lns>
ogra_, if ppl are using Dapper will they be able to upgrade directly to hardy?
13:47
just out of curiosity
13:48
<ogra_cmpc_>
lns, with the general upgrade rule of "rebuild your client environment", yes
13:48
<moquist>
ogra_cmpc_: hey-a
13:48
<ogra_cmpc_>
hey moquist
13:48
<lns>
coolio
13:49* ogra_cmpc_ just tries out a classmate install in urdu language ...
13:49
<moquist>
I'm experimenting with diskless thick clients. I've done 'apt-get ubuntu-desktop' in the chroot (this is feisty, actually) and I'm wondering if anybody has any OTOHH advice for me about problems I'll have before I have a workable kiosk-like system.
13:49
For example, right now avahi is upset.
13:50
<ogra_cmpc_>
moquist, Nubae in #edubuntu is experimenting with that since some time
13:50
and on the edubuntu ML there is jim kronenbush
13:50
<moquist>
And maybe this was dumb, but I bind mounted /proc from the server under the chroot during the ubuntu-desktop installation.
13:50
ogra_cmpc_: I'll hop over there and ask; thx.
13:52
<vagrantc>
moquist: bind-mounting proc shouldn't cause any problems. we do it during ltsp-build-client.
13:52mccann has quit IRC
13:53
<moquist>
vagrantc: Ah, cool. I was a bit worried the chroot might think it's got the server proc, then, but I guess that's not an issue.
13:53
<ogra_cmpc_>
it has
13:54
which is why it will create a matching xorg.conf for the servers card for example
13:54
<moquist>
Right, but apparently that's OK.
13:54
<ogra_cmpc_>
if you boot it on a client it will indeed have the kernels proc :)
13:56ogra_cmpc_ is now known as ogra_cmpc
13:56BadMagic has quit IRC
13:59BadMagic has joined #ltsp
14:04
<moquist>
ogra_cmpc: so far so good; I took avahi and NetworkManager out of startup and now GDM is unhappy b/c it can't write to /var/lib/gdm, but at least I can log in to TTY1.
14:04
<vagrantc>
ogra_cmpc: what do you mean it will have the kernel's proc ?
14:04
ogra_cmpc: NFS doesn't export sub-mounts
14:04
<ogra_cmpc>
vagrantc, the /proc of the running kernel
14:04
<vagrantc>
ogra_cmpc: and it's excluded from the squashfs images
14:05
ah, yes.
14:05
<ogra_cmpc>
:)
14:05
OMG i wasnt aware urdu is a RTL language
14:05
<vagrantc>
my brain somehow saw server instead of kernel
14:06* ogra_cmpc tries to find his way around on a mirrored desktop
14:06* moquist has been mispronouncing urdu all this time...and to think it's "Uh-Drew"
14:07
<ogra_cmpc>
man, thats wrecking my brain
14:08
working from behind the mirror
14:08milesd has joined #ltsp
14:12
<vagrantc>
ogra_cmpc: i'm thinking of making ltsp-client strictly depend on ldm ...
14:12
if they want some other setup without ldm and it's dependencies, they can install ltsp-client-core
14:13
<ogra_cmpc>
well, i never dropped that dep :)
14:13
oh, i had an "or" in there though
14:13
for sdm
14:13
<vagrantc>
yes
14:13
i dropped sdm-terminal in debian, but i'm thinking of dropping x-display-manager as well.
14:13
<ogra_cmpc>
well, no more sdm you said
14:14
if gdm has a sane ssh mode we can re-add it
14:14
until then i agree
14:14
<vagrantc>
alright.
14:14* moquist logs into an Ubuntu Desktop on a diskless thick client
14:14
<vagrantc>
i'll drop it in my next upload.
14:15* ogra_cmpc applauds moquist
14:15
<moquist>
Hmm. Gnome is unhappy.
14:15
So not complete success yet...
14:27K_O-Gnom has quit IRC
14:31DonSilver has joined #ltsp
14:40daduke_ has quit IRC
14:53elisboa has quit IRC
15:10cdealer has quit IRC
15:13abadger1999 has quit IRC
15:15cliebow_ has quit IRC
15:15mhterres has quit IRC
15:15mhterres has joined #ltsp
15:16abadger1999 has joined #ltsp
15:16
<ogra_cmpc>
vagrantc, not sure you plan xen support, but here is something that helps: https://bugs.launchpad.net/ubuntu/+source/ltsp/+bug/134865
15:19
<vagrantc>
ogra_cmpc: i've done many installs inside xen instances with no problem.
15:19
<ogra_cmpc>
hmm
15:19
<vagrantc>
oh...
15:19
--early-packages "ltsp-client libc6-xen" ...
15:19
<ogra_cmpc>
heh
15:19
<vagrantc>
although the poster forgot to add ltsp-client
15:19
<ogra_cmpc>
thats a default ? or local ?
15:20
<vagrantc>
if you specify --early-packages it overrides the default
15:20
which is "ltsp-client"
15:21
<ogra_cmpc>
the plugin doesnt do that
15:21
<vagrantc>
no, but a comment did
15:21
<ogra_cmpc>
he just adss it separately in after-install
15:21
<vagrantc>
i think a plugin is a way to handle it, though.
15:21
though i'd probably add it so that it gets installed via debootstrap and is available as early as possible.
15:21
<ogra_cmpc>
at least a way that doesnt get in myv way :)
15:22
<vagrantc>
ogra_cmpc: oh, libc6-xen is in universe?
15:22
<ogra_cmpc>
if i add it as separate new plugin no RM will complain, if i add it to existing code they will look closer
15:23
yeah
15:23
<vagrantc>
i thought xen was supported out of the box ...
15:23
anyways ...
15:23
<ogra_cmpc>
with jepo i think
15:23
dont askl me, i'm no virtualization expert
15:23
<vagrantc>
heh
15:23
<ogra_cmpc>
*jeos
15:23
weird keyboard
15:24
<vagrantc>
well, basically you need to install libc6-xen in the chroot ... should be possible to handle that all in a single additional plugin
15:24
<ogra_cmpc>
i'm happy if i get my virtualbox running fine
15:24
<vagrantc>
unless debootstrap doesn't handle INCLUDES ...
15:25
yup... code should have include support
15:27
i'll test if using includes works on debian
15:28
and then we can just add it there instead of in after-install ... otherwise do as the novaki (the poster) did
15:28
and i'd probably do a simpler test for libc6-xen :)
15:29
<ogra_cmpc>
yeah
15:30
<vagrantc>
can actually just test the return status of dpkg
15:30
dpkg apparently doesn't have a quiet option :(
15:30
<ogra_cmpc>
well, a grep ^ii should be added :)
15:31
<vagrantc>
if dpkg -l libc6-xen > /dev/null ; then ...
15:31
should be sufficient
15:31
<ogra_cmpc>
even with config files left ?
15:31
<vagrantc>
i can try it
15:32
but my guess is that will still have a non-0 return
15:33
<ogra_cmpc>
yeah
15:33
doesnt help, you need grep ^ii
15:33
<vagrantc>
why?
15:33
<ogra_cmpc>
and grep has a -q :)
15:34
<vagrantc>
it returns 0 ... ugh.
15:34
ok.
15:34
grep it is
15:34
actually, i need to work on this security update for ldm in lenny
15:34
<ogra_cmpc>
right
15:35* vagrantc grumbles about slow buildd's
15:36* ogra_cmpc sadly doesnt like sbalneav's proposal on th ML
15:36* jammcq too
15:37
<ogra_cmpc>
it means we need server side stuff and wiont be able to just connect to any sshd we want
15:37bobby_C has joined #ltsp
15:37
<jammcq>
what we need is good ole network auth
15:38
<vagrantc>
just fire off an ssh in an xterm :)
15:38
maybe even a custom xterm, so it's not quite as ugly.
15:38
<jammcq>
that's ugly too
15:39
<ogra_cmpc>
vagrantc, we have that already :)
15:39
<vagrantc>
yeah, i guess in a different way
15:45Phisbut has quit IRC
15:50
<vagrantc>
even if upstream openssh accepts those changes, my guess is you won't be able to make them default in most distros
15:50
so it will require manual configuration ...
15:50
<ogra_cmpc>
thats what i meant
15:51
<vagrantc>
although, that's to get password expiry support ... maybe it's ok to not have that *part* of it working out of the box
15:55
password expiry isn't enabled by default... or is it?
15:55Guaraldo has left #ltsp
15:56
<ogra_cmpc>
it is
15:56
<vagrantc>
oh, ok.
15:57* ogra_cmpc wonders hwy tzdata"s debconf selection offers him Sytem5 and Etc at the timezone selection
15:58
<ogra_cmpc>
err SystemV
15:59
vagrantc, do you have that too if you do dpkg-reconfigure tzdata on debian ?
15:59
<vagrantc>
ogra_cmpc: yup
15:59
<ogra_cmpc>
weird
15:59
<vagrantc>
2008a-1
15:59
on sid
16:00
<ogra_cmpc>
2007k-1ubuntu4
16:00BadMagic has quit IRC
16:01
<vagrantc>
on etch, it doesn't present me with anything
16:01mhterres has quit IRC
16:01
<vagrantc>
i think it wasn't debconfized till later or something
16:01
<ogra_cmpc>
i think it was in another package
16:02
tzadmin ?
16:02
or so
16:02
<vagrantc>
it says to use tzconfig
16:02
<ogra_cmpc>
or the command was called like that
16:02
yeah
16:02
<vagrantc>
somewhere there was tzsetup
16:02
seems hard to keep track of all the advances in timezone configuration these days
16:04
<ogra_cmpc>
yeah, except that i have to offer confguration in my installer :)
16:04BadMagic has joined #ltsp
16:06
<johnny>
hi
16:10rcy has joined #ltsp
16:13mccann has joined #ltsp
16:14alekibango has quit IRC
16:17jammcq has quit IRC
16:30rcy has quit IRC
16:37mccann_ has joined #ltsp
16:37mccann_ has quit IRC
16:40slidesinger has quit IRC
16:48bobby_C has quit IRC
16:49alekibango has joined #ltsp
16:50
<johnny>
dberkholz, i've been messing with genkernel4 for awhile, can't get the thing to compile gcc
16:50
big long nasty error
17:01mccann has quit IRC
17:17
<johnny>
laga, it seems that neither aufs nor unionfs will make it into the kernel anytime soon
17:17
that we won't see anything like that in kernel untion the union mount stuff happens..
17:17
which deals with the issue at the vfs layer i think, instead of trying to work around it/on top of it
17:18
s/untion/until/
17:18
lol
17:20mikkel has quit IRC
17:25tessier__ has joined #ltsp
17:25
<tessier__>
Hello all!
17:26
I am trying to network boot with mboot.c32 but after it tftp's mboot.c32 the server just hangs. Any ideas for how to troubleshoot? Anyone know if mboot.c32 works with Nvidia chipset?
17:26
I've done this before and have copied the old config as a baseline. I don't understand why this wouldn't work.
17:29chupa has quit IRC
17:31
<tessier__>
ah...
17:33
hm...I'm afraid the pxe in these mobo's might be hosed. :( It is supposed to ask for the config file pxelinux.cfg/<mac> but instead looks for pxelinux.cfg/0A010168
17:34* tessier__ wonders if pxelinux.0 is different between 32 bit and 64 bit machines
17:41
<tessier__>
ah!
17:41
pxelinux.0 is different for 32 and 64 bit
17:41
Now it works! :)
17:41
Funny how it mostly worked when running the 32 bit version. It got as far as getting mboot.c32 anyway
17:45
<vagrantc>
tessier__: what are you network-booting mboot.c32 for? xen ?
17:45chupa has joined #ltsp
17:45
<tessier__>
vagrantc: Yep!
17:45
Ah....and now it grabs the correct config file named by mac as well
17:45
<vagrantc>
i tried that a little over a year ago...
17:46
<tessier__>
vagrantc: I've been doing it for a couple of years now. Works great. I use xen with aoe. http://xenaoe.org
17:46
I am building virtualization cluster with AoE backed SAN for companies.
17:46
<vagrantc>
ata-over-ethernet?
17:46
<tessier__>
Yes
17:47
<vagrantc>
i'm guessing you're not using LTSP for this ... ?
17:47
<tessier__>
I usually run LTSP on the cluster. But the booting of the cluster itself is not directly related to LTSP
17:48
<vagrantc>
so the cluster is your LTSP application serveR?
17:51
<tessier__>
Exactly
17:53
<vagrantc>
nice.
18:17J45p3r has joined #ltsp
18:18Faithful has joined #ltsp
18:21Shingoshi has quit IRC
18:41elisboa has joined #ltsp
18:42DonSilver has quit IRC
18:46cyberorg has quit IRC
18:47steph_ has joined #ltsp
18:48subsume has joined #ltsp
18:56staffencasa has quit IRC
18:57K_O-Gnom has joined #ltsp
19:03jammcq has joined #ltsp
19:04
<jammcq>
hey kidz
19:04
<johnny>
ogra_cmpc, ogra_ sup with grawert.org? :)
19:14steph1 has joined #ltsp
19:15steph1 has left #ltsp
19:17milesd has quit IRC
19:17subsume has quit IRC
19:17abadger1999 has quit IRC
19:21subsume has joined #ltsp
19:28K_O-Gnom has quit IRC
19:38subsume has quit IRC
19:45mccann has joined #ltsp
19:53
<ogra_cmpc>
johnny, i didnt find the time to care for finishing the move yet ....
19:56J45p3r has left #ltsp
20:20steph1 has joined #ltsp
20:22steph1 has left #ltsp
20:42otavio has quit IRC
20:44subsume has joined #ltsp
20:49steph_ has quit IRC
20:50rjune__ has quit IRC
20:56Shingoshi has joined #ltsp
21:03mischko has joined #ltsp
21:04mischko has quit IRC
21:28mccann has quit IRC
21:32tux_440volt has joined #ltsp
21:38abadger1999 has joined #ltsp
21:45MacIver has quit IRC
22:05MacIver has joined #ltsp
22:07subsume has left #ltsp
22:13vagrantc has quit IRC
22:43spectra has quit IRC
23:20elisboa has quit IRC
23:29basanta has joined #ltsp
23:29daya has joined #ltsp
23:35basanta has quit IRC
23:36subir has joined #ltsp
23:36subhodip__ has joined #ltsp
23:36tux_440volt has quit IRC
23:36subhodip__ has quit IRC
23:42tux_440volt has joined #ltsp
23:45indradg has quit IRC
23:47Faithful has quit IRC
23:50latarsky has joined #ltsp