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


Channel log from 12 March 2010   (all times are UTC)

00:07ogra_cmpc has quit IRC
00:08ogra_cmpc has joined #ltsp
00:12alkisg has joined #ltsp
00:17daya has joined #ltsp
00:18
<alkisg>
Devices that are mounted from a user U1 on a fat client persist logoffs, and cannot be unmounted by another user U2.
00:18
I can think of 2 ways to solve that - either unmount all user-mounted devices at logoff, or enable (with policykit) any user to unmount devices mounted by other users
00:19
Any thoughts? I'm thinking of going with unmounting everything at logoff (I just hope nothing else gets unmounted by mistake...)
00:21
Maybe `mount | grep uhelper=udisks` gives the proper list...
00:22
(or I could just look at /media/*...)
00:27artista_frustrad has joined #ltsp
00:32artista_frustrad has quit IRC
00:35daya has quit IRC
00:37daya has joined #ltsp
00:42daya has quit IRC
00:56daya has joined #ltsp
00:58daya_ has joined #ltsp
01:09ogra_cmpc has quit IRC
01:11ogra_cmpc has joined #ltsp
01:37laron has quit IRC
02:02gnunux has joined #ltsp
02:03
<gnunux>
hi
02:15alkisg has quit IRC
02:29alkisg has joined #ltsp
03:20ltsplogbot has quit IRC
03:22ltsplogbot has joined #ltsp
03:34ogra_cmpc has quit IRC
03:36ogra_cmpc has joined #ltsp
03:40try2free has joined #ltsp
03:51try2free has left #ltsp
03:58nubae_ has joined #ltsp
03:59nubae has quit IRC
04:27mikkel has joined #ltsp
04:33nubae has joined #ltsp
04:33nubae_ has quit IRC
04:46tthorb_ has joined #ltsp
04:49tthorb has quit IRC
04:52hersonls has joined #ltsp
04:53Selveste1_ has joined #ltsp
04:58Selveste1_ has quit IRC
04:59Selveste1_ has joined #ltsp
05:01johnny has quit IRC
05:19daya has quit IRC
05:26plasticdoc has joined #ltsp
05:27_Techie_ has quit IRC
05:33
<plasticdoc>
I am using Ubuntu 9.10 and LTSP 5.2 from Stéphane Graber's PPA, is here anybody whiling to help me with the new --fat-client switch?
05:35
<stgraber>
plasticdoc: hey there, what's the issue with --fat-client ?
05:35
<plasticdoc>
Hello, Neither Stéphane Graber's sugested 'LTSP handbook' <https://code.launchpad.net/~ltsp-docwriters/ltsp/ltsp-docs-trunk>, nor Ubuntu's 9.10 'man ltsp-build-client', provide any useful info on how to setup (and configure) the new --fat-client
05:36
<stgraber>
btw, I'm Stéphane Graber :)
05:36
<plasticdoc>
Oh, ok.
05:36
Is it ok to use something like sudo ltsp-build-client --arch i386 --dist karmic --fat-client --fat-client ubuntu-desktop
05:37
<stgraber>
yes but you'll need to add the PPA inside the chroot at the end
05:37
to avoid doing that you can do: --copy-sourceslist
05:37
and I know we have another option to bypass the gpg key check
05:37
hang on a sec
05:38
--accept-unsigned-packages that'd be
05:38
so ltsp-build-client --arch i386 --dist karmic --fat-client --fat-client ubuntu-desktop --accept-unsigned-packages --copy-sourceslist
05:39
that should create a new karmic i386 chroot with fat-client support and with an up to date LTSP (that actually supports fat-client)
05:42
<plasticdoc>
So let me see if I got it (English is not my mother tongue),
05:42
I already had: sudo add-apt-repository ppa:stgraber/ppa
05:43
If i do sudo rm -fr /opt/ltsp/i386
05:43
and then:
05:43
sudo ltsp-build-client --arch i386 --dist karmic --fat-client --fat-client ubuntu-desktop --accept-unsigned-packages --copy-sourceslist
05:43
<stgraber>
yep, that should work
05:44
<plasticdoc>
It seems so simply that it gets scary!
05:45
<alkisg>
plasticdoc: https://help.ubuntu.com/community/UbuntuLTSP/FatClients
05:45
<stgraber>
well, to be honnest I never tried fat client in Karmic so it very well may fail :)
05:45
so far most tests have been done with Lucid in the chroot
05:45
<alkisg>
2 people said that they followed that in karmic and it worked..
05:45
<stgraber>
though maybe Alkis did some tests with Karmic
05:45
alkisg: oh, great
05:45
<alkisg>
Not myself, but 2 other persons here
05:46
<stgraber>
btw, anything that I should wait for before tagging ltsp 5.2.1 and ldm 2.1.1 ?
05:46
my current plan is to test + tag them on Sunday and get them uploaded to Lucid + backport immediately
05:46
<alkisg>
stgraber: I think it's in a good shape now for tagging.
05:46
Not bulletproof, i.e. I'd like some fixes, but I don't think I'll find them soon... :-/
05:46
<stgraber>
as we are already frozen for Beta1. I'll poke Steve to get an exception and get them on the first candidate image on Monday or Tuesday
05:46
<alkisg>
stgraber: ah, only one thing, the vt switching bug
05:47
It's not properly resolved yet, but it's working for the default case so it's not a very big deal
05:47
<stgraber>
alkisg: what vt switching bug ? didn't Marc solved it ?
05:48
alkisg: ah, right, the: Do not openvt when it's X thing ?
05:48
<alkisg>
No not that one
05:48
I wanted that for fat clients, but I found another way around it
05:48
There are 2 problems: e.g. if I specify SCREEN_02=shell and SCREEN_07=ldm, then there's a chance that X won't start until I switch to vt7
05:49
So it'd be nice if we had exactly one chvt, for the first X screen
05:49
<stgraber>
yeah, X does that ... it needs you to be on the vt to start ...
05:49
well, I'd fix that another way for now. Switch to the highest SCREEN unless we have mentioned another one in the config
05:49
<alkisg>
We had a lot of chvt's before, marc removed all of them, but I think we should keep 1
05:50
That would work fine for me
05:50
<stgraber>
so we run the code that spawns all screens, then check what the current VT is. If it's already the one needed for SCREEN_max we don't switch, if not, we switch to it
05:51
<alkisg>
Right
05:51
<stgraber>
I can commit a change upstream that does that though I can't test it here. Can you test it ?
05:51
<alkisg>
Yup
05:51
Ping me when you do
05:53
<plasticdoc>
stgraber: In LTSP 5.2 do we still need to install gnome-watchdog, or have a server cron job to keep rebooting the server to cope with all those zombie logged but non-active users ?
05:54
<stgraber>
there shouldn't be that much remaining processes, though I mostly work with ltsp-cluster and our cluster agent does the cleaning up
05:54
so just to be on the safe side, keep gnome-watchdog
05:54
<plasticdoc>
I'll do
05:55
<stgraber>
alkisg: http://pastebin.com/D4hSxATp
05:56
<alkisg>
testing...
05:56Egyptian[Home] has quit IRC
06:00Selveste1_ has quit IRC
06:01
<alkisg>
stgraber: wouldn't that always switch to vt12? I think we should use "num" instead of screen...
06:02
erm, or maybe add a new variable in the loop
06:02
highest_vt=screen <== inside the "if"
06:10* alkisg tries with http://paste.ubuntu.com/393977/
06:11
<stgraber>
alkisg: argh, indeed
06:11
alkisg: we should set some variable in the if and use that instead
06:12
<alkisg>
stgraber: see the link just above
06:12
<stgraber>
lunch time, will be back just after
06:12
<alkisg>
ok
06:13
I think there are some times where X will "steal the focus", but I wouldn't mind for that...
06:13
I.e. if Gadi puts SCREEN_07=ldm and SCREEN_08=shell, and we chvt to vt8, sometimes X will switch back to vt7
06:13
(that should teach him to put X in the highest screen :P)
06:14artista_frustrad has joined #ltsp
06:15
<alkisg>
Uh, my bad, chvt $highest_vt, not chvt $screen...
06:18pmatulis has joined #ltsp
06:20artista_frustrad has quit IRC
06:42MaRX-Mode has quit IRC
06:45MaRX-Mode has joined #ltsp
06:46
<alkisg>
Hmm I think plymouth does an additional chvt and voids all of our efforts :)
06:46
...putting our `chvt` lower...
06:47knipwim_ has quit IRC
06:48knipwim has joined #ltsp
07:02
<alkisg>
Nope, it still doesn't cut it
07:03
<stgraber>
alkisg: http://pastebin.com/WyhV4Hzz ?
07:03
<alkisg>
I put a debug statement next to the one and only chvt, e.g. "fgconsole=7, switching to vt11", so I'm sure the calculations are correct, but it doesn't work
07:03
<stgraber>
ah, plymouth ...
07:03
<alkisg>
Something else is using chvt after us
07:03
Yup
07:04
<stgraber>
I guess we'll need to wait for the fixed plymouth as we have vt bugs even in regular ubuntu
07:04
once it's there, then we can focus on making that part of the code work as it should
07:04
<alkisg>
Nice. Should I submit this patch, as it makes the problem a little more rare?
07:04
<ogra>
stgraber, there should be plymouth test packages in keybuks PPA
07:05
<alkisg>
http://paste.ubuntu.com/394000/
07:05
(without the debug `echo`)
07:06
<stgraber>
alkisg: isn't that the same as my last pastebin ?
07:06
<alkisg>
stgraber: I haven't tried your method...
07:07
chvt on the same console is as fast as fgconsole, though
07:07
So I don't think there's any need to complicate this...
07:07
<stgraber>
alkisg: it'll kill plymouth and flicker if you don't do the check
07:08
<alkisg>
OK, any reason for the "case" there?
07:08
Is that to put an extra zero in front?
07:08
<stgraber>
yep
07:08
<alkisg>
So, that isn't needed, as we can compare integers...
07:08
<stgraber>
otherwise I'll get 1 != 01 and it'll fail. If we always add it, we'll get 011 != 11 and it'll fail again
07:09
hmm, right
07:09
<alkisg>
[ 007 -eq 7 ] && echo yup
07:09
<ogra>
case is always faster than test :)
07:09
<alkisg>
ogra, are you sure about that?!
07:09
<ogra>
yes
07:09
<alkisg>
OK, I'll time it...
07:10
<ogra>
in ubuntu we have an unwritten policy to use case in all scripts because of that
07:10
instead of if/then/blah with test
07:10
beyond that it makes the code better readable
07:11* dba nods; for both speed and readability
07:11
<alkisg>
Are we talking about the same thing?!!!!
07:11
How are 4 lines in a case more readable than a simple comparison?
07:12
(btw, we're compare a case and an if with a single if...)
07:12
not a case with an if...
07:13
So in this particular problem, we can just _remove_ the case, not substitute it with an if...
07:13
<ogra>
get rid of the if in the case :)
07:13
<alkisg>
I'll try the timing though, as you got me curious...
07:13weenus has joined #ltsp
07:13Selveste1_ has joined #ltsp
07:13
<jammcq>
good morning friends
07:14
<stgraber>
ogra: currently we have a case statement to append an extra 0 to a variable if it isn't made of two digits, thne we compare it as a string using if. What alkisg suggests is to use the variables as integers and do the check in the if, getting rid of the case in the process.
07:14
<ogra>
yo ho jammcq
07:15
stgraber, yeah, that might be smaller ... not sure about faster
07:15weenus has quit IRC
07:20jammcq has quit IRC
07:22
<alkisg>
(1) Just a loop: 0m3.434s, (2) If: 0m6.668s, (3) case + if: 0m8.122s
07:23
So case + if is about twice as slow as a plain if...
07:29Selveste1__ has joined #ltsp
07:29Selveste1_ has quit IRC
07:33wima1 has joined #ltsp
07:33mgariepy has joined #ltsp
07:34
<mgariepy>
morning all
07:35
<alkisg>
Hi Marc
07:37weenus_ has joined #ltsp
07:40Selveste1__ has quit IRC
07:41
<mgariepy>
how you doing alkisg ?
07:41* alkisg is struggling with the VTs problem - but it must be due to plymouth after all...
07:42
<mgariepy>
if you add like chvt 7 to the ldm script does it do the trick ?
07:42
<alkisg>
I tried on ltsp-client-core, but it doesn't cut it
07:43
I think plymouth is also using chvt after us...
07:43slidesinger has joined #ltsp
07:44
<mgariepy>
maybe, on my laptop, X starts on :0 then when i switch to :7 it kills X and respawn it. it's kinda weird too.
07:44slidesinger has quit IRC
07:44timborn has quit IRC
07:45Selveste1__ has joined #ltsp
07:46wima1 has quit IRC
07:47scottmaccal has joined #ltsp
07:47GodFather has joined #ltsp
07:48
<alkisg>
Starting the screen scripts in the reverse order seems to help: for screen in 12 11 10 09 08 07 06 05 04 03 02 01; do
07:49
(assuming X in on the highest)
07:49
Anyway, I'll test more later... bbl.
07:53alkisg has quit IRC
07:53wima1 has joined #ltsp
07:54pths has quit IRC
07:58Selveste1___ has joined #ltsp
07:59Selveste1__ has quit IRC
08:00Gadi has joined #ltsp
08:05alexqwesa has quit IRC
08:07alexqwesa has joined #ltsp
08:17Egyptian[Home] has joined #ltsp
08:21jammcq has joined #ltsp
08:21
<jammcq>
good morning friends
08:25
<wima1>
can anyone point met to the latest and greatest way to get a scanner running attached to a thinclient? Is it sufficient to just install saned in the chroot?
08:26
<jammcq>
it's pretty much the same as setting up a scanner on full linux machine
08:28
<wima1>
and is there some kind of broadcasting going on so that applications on the server can find the scanner?
08:29
<jammcq>
no, you have to point your scanning software to the IP address of the thin client that has saned running
08:29alexqwesa has quit IRC
08:29Kicer86 has joined #ltsp
08:29
<wima1>
ok, so i better make sure that it has a fixed ip
08:30
i'll try that. thanks!
08:31alexqwesa has joined #ltsp
08:36
<wima1>
if i update an image that is in use, will it mess up the connected clients?
08:37
<jammcq>
umm, I don't think so, but i'm not sure
08:37
<wima1>
i'm working on a live system :)
08:39plasticdoc has left #ltsp
08:39plasticdoc has joined #ltsp
08:41
<wima1>
and another question :) : if i rebuild the imge, does it also rebuild the kernel or do i have to do it separately?
08:43
<Gadi>
wima1: SCANNER=True in lts.conf
08:43Selveste1___ has quit IRC
08:43
<Gadi>
(if you are using a relatively recent version of LTSP)
08:44
<wima1>
gadi: what does this do?
08:44
<Gadi>
enables scanner support
08:44
for a locally attached scanner
08:45
5.1.85 or higher
08:45
(LTSP version that is)
08:46
<wima1>
where are these variables used? in the startup scripts?
08:48
<Gadi>
http://bazaar.launchpad.net/~ltsp-upstream/ltsp/ltsp-trunk/revision/1363
08:49
<wima1>
ah, ok
08:49
thanks
08:49Egyptian[Home] has quit IRC
08:50
<wima1>
so it just makes sure saned is running
08:50
<Gadi>
hmm... doesn't seem like he added SANE_NET_HOSTS to the env var
08:50* Gadi thinks he should have done that, too
08:50
<wima1>
what does that do?
08:51
<Gadi>
wima1: if you set SANE_NET_HOSTS=${LTSP_CLIENT} it will tell the sane client on the server to look for a scanner on your thin client
08:51
($LTSP_CLIENT is set to the thin client IP)
08:52
this way, you don't need a fixed IP or anything
08:52
you can add an Xsession.d/ script on the server to set this env variable for now
08:54
<wima1>
ok thanks
09:03alexqwesa_ has joined #ltsp
09:03alexqwesa has quit IRC
09:09wima1 has quit IRC
09:11plasticdoc has left #ltsp
09:13wima has joined #ltsp
09:23roba-au has joined #ltsp
09:25
<roba-au>
Greetings: can anyone point me at doco on how to handcraft an ltsp installation into a non-supported distro (centOS 4 based)
09:26roba-au has quit IRC
09:38Faithful has quit IRC
09:40alkisg has joined #ltsp
09:53Faithful has joined #ltsp
10:06gnunux has quit IRC
10:08johnny has joined #ltsp
10:13wima has quit IRC
10:26
<alkisg>
stgraber, mgariepy: putting a `chvt ${TTY}` right before the last line of ldm (=xinit...) seems to solve the problem for me, at least for ldm screens.
10:28
We could put that at least as a temporary workaround, until any plymouth problems are resolved and we can have another look at it
10:29
stgraber: so if you want, do commit that line, and tag ltsp/ldm... bbl.
10:30alkisg has quit IRC
10:30
<stgraber>
I'd still include a test to make sure we aren't already on that VT.
10:36
<johnny>
stgraber, when will you update your ppa for ltsp?
10:36
after things are tagged?
10:37
<stgraber>
yeah, it's currently containing a snapshot of bzr that's pretty up to date
10:37
I'll update with the final packages once tagged
10:38
<johnny>
stgraber, alkisg has been doing awesome stuff i wanted to use :)
10:38
nfs home.. and proper fixes for policykit
10:39
<stgraber>
ah, I can try to make a new snapshot today depending on how much time I have left :)
10:40
<johnny>
no rush
10:40
i made a hack to the pk files to tide me over
10:40
plus.. he might add something else cool in the meantime
10:49thunsucker has joined #ltsp
10:51litlebuda has joined #ltsp
10:52staffencasa has joined #ltsp
10:56timborn has joined #ltsp
11:01
<Gadi>
stgraber: if plymouth is the problem, why not simply call: plymouth --quit before we run through the screen scripts?
11:23vagrantc has joined #ltsp
11:29thunsucker has quit IRC
11:31
<johnny>
NEAT dudes..
11:31
http://www.hadess.net/2010/03/speaker-testing.html
11:43
<_UsUrPeR_>
johnny: cool!
11:43* _UsUrPeR_ is still having problems getting HDMI to work properly in ubuntu 9.10
11:44
<_UsUrPeR_>
err... s/HDMI/SP-DIF
12:11litlebuda has quit IRC
12:12GodFather has quit IRC
12:12GodFather has joined #ltsp
12:14ogra_cmpc has quit IRC
12:20krbaum has joined #ltsp
12:26ogra_cmpc has joined #ltsp
12:35
<krbaum>
Hi, I'm having troubles getting USB drives to work on my Ubuntu fat client. I think it may be related to the fact that ck-list-sessions shows active = FALSE. When the USB drive is inserted, I get the message "Unable to mount 4.0 GB Filesystem Not Authorized". Are USB drives supposed to be working out-of-the-box on fat clients?
12:45
<Gadi>
krbaum: alkisg has been busy fixing that exact problem (amongst others) - best to try and get him when he gets back
12:45
fat client stuff is really new, and alkisg has been working through the finer points
12:45* vagrantc waves to Gadi
12:46* Gadi waves back
12:47
<krbaum>
Ok, I'll ask again when he's back. Thanks.
12:59litlebuda has joined #ltsp
13:02plasticdoc has joined #ltsp
13:28the_fronny has joined #ltsp
13:29laron has joined #ltsp
13:30
<laron>
how can i add firefox extensions to ltsp clients?
13:30
or what is the easiest
13:30
way
13:31
<the_fronny>
We have several ltsp-4.x systems as well as several ltsp-5 systems. The latter are all based on Ubuntu 9.04 or 9.10. They work fine except that you cannot Xnest the client environment to your own machine. This was an invaluable diagnostic tool for us and now it's broken. Anybody know of a fix or where the breakage lives; ltsp or Ubuntu? thanks
13:31alkisg has joined #ltsp
13:32
<jammcq>
I think Xnest is broken in Xorg
13:32
i've heard lots of comlaints about it being horribly broken
13:32
s/comlaints/complaints/
13:34
<the_fronny>
jammcq: Ouch! That's the wrong answer :) Is there and app that serves that purpose? These systems are remote and admin-ing was a lot easier with that functionality.
13:36
<Gadi>
laron: I believe you can download the extension (.xpi file) and run: firefox -install-global-extension <filename>
13:37
laron: if you run firefox locally, do this in the chroot, as well and reroll the image
13:37
<jammcq>
the_fronny: you could use vnc or NX
13:38
<Gadi>
the_fronny: in LTSP 4.2, where you use XDMCP, I can see how Xnest would help, but in LTSP5, we use ssh, so it is less useful for diagnostics
13:43
<krbaum>
Can I give users a choice on what type of client (fat or thin) to boot? I would like to create two boot options in the pxelinux configuration, one for fat and one for thin client. I know that I can use LTSP_FATCLIENT=True|False in lts.conf, but that would not make it user-selectable. Has anyone done this?
13:45
<Gadi>
why would a user choose one or the other?
13:46
its usually a question of hardware resources
13:46
<krbaum>
Depending on the hardware he/she happens to plug in to the network. Or simply by personal preference.
13:46
<Gadi>
if it is hardware, that is an lts.conf thing
13:47
hardware doesnt change
13:47
(ie specify by mac addy)
13:47
<the_fronny>
Gadi: I don't know if it's "less useful" but being able to see what the (remote) user would see is powerful indeed. I don't see the relevence of the protocol used to the usefulness of the tool unless the protocol is the cause of the breakage and that doesn't appear to be the case. Everything I've read so far points to an X problem. I just wondered if that was Xorg proper or Ubuntu's "version" of X (if there actually is one.) Thanks.
13:48
<Gadi>
the_fronny: I did not mean to make a statement of criticism at all - it is just that in an XDMCP environment, Xnest makes a -query to the XDM to get a login screen
13:48
the_fronny: in LTSP5, we do nothing of the sort
13:49
we run a local display manager on the client (LDM)
13:49
collect username/password
13:49
and execute:
13:49
ssh -X ... user@server /etc/X11/Xsession
13:49
<krbaum>
Yes, I know I can do that, but why should an administrator have to do this for all clients that are plugged into the network. I think that it would be better if the users gets to choose.
13:50the_fronny has quit IRC
13:50
<Gadi>
hmm.... maybe it was something I said.
13:51
I really didn't mean anything negative
13:51the_fronny has joined #ltsp
13:51plasticdoc has left #ltsp
13:52
<Gadi>
krbaum: you may want to modify the "menu" scren script
13:53
(which is in LTSP 5.2 - if you have that)
13:53
<alkisg>
Gadi: after several tries, putting `chvt ${TTY}` right before `xinit` in the ldm screen script was the only one that worked for me. Any objections if I commit it, at least until all the plymouth problems in lucid are resolved?
13:53
<Gadi>
otherwise, you can find it on launchpad - it is basically a whiptail menu, but you can build off of it (probably a better way than the pxelinux menu)
13:54
alkisg: are we sure that the issues are qith plymouth?
13:54
<alkisg>
Gadi: nope
13:54
But I haven't find anything else that actually works...
13:54
<Gadi>
alkisg: if so, can we simply call: plymouth --quit before the for loop launching screen scripts?
13:55
<alkisg>
Won't that break the "smooth booting" of lucid?
13:55
:D
13:55
<Gadi>
I don't think so - its right before X launches
13:55
plymouth should fade and then X should start
13:55
<alkisg>
OK, let me test if that works...
13:56
Is that the exact line?
13:56
plymouth --quit ?
13:56
<Gadi>
I think so
13:56
according to the source code (man page is horrible)
13:56
<the_fronny>
Gadi: You know, I just tried that command and it blew out the X on my machine! LOL! Well, that's progress, so it's good.
13:57
<Gadi>
hehe - well, at least you're laughing - I thought you were going to beat me up
13:57
:)
13:57
the_fronny: are you on the same subnet as the thin clients?
13:58
(if so, the best thing to recreate the client env is to run a virtual machine that simply PXE boots)
13:58
then, aside from drivers, you actually have the client environment
13:59
<krbaum>
Gadi: I don't quite follow. What "menu" screen script are you referring to? Yes I have LTSP 5.2.
13:59
<Gadi>
krbaum: we through in a menu screen script (/opt/ltsp/i386/usr/share/ltsp/screen.d/menu)
14:00
krbaum: this allows you to set a whiptail menu of screen scripts for the user to choose from
14:00
say, "Windows Server", "Linux Server1", "Linux Server2", etc
14:00
all lts.conf-configurable
14:00
but, currently just lets you choose amongst screen scripts.
14:01
oh, wait
14:01
but fror fat client, you need boot-time settings changed, as well
14:01
sorry
14:01
ur right,
14:01
you are better off doing it as a pxelinux menu
14:02
you will need to write a script that reads /proc/cmdline early on and sets LTSP_FATCLIENT according to a kernel command line argument
14:02mischko has joined #ltsp
14:04
<krbaum>
Yes, I was looking in that direction, but I was hoping for a simpler solution. Thank you!
14:08
Gadi: Would the best way for me to fix it be to check /proc/cmdline for a "fat|thin" parameter in /usr/share/ltsp/ltsp_config and if found override the LTSP_FATCLIENT variable? Would that work?
14:08
<alkisg>
Hmm maybe it's possible to just pass LTSP_FATCLIENT=True in the kernel command line?
14:08
<krbaum>
Tried that, but no.
14:08
<alkisg>
We could make a simple patch for that, though
14:08
<krbaum>
Well, Actually I tried LTSP_FATCLIENT=False
14:09
<Gadi>
krbaum: you should probably have an initscript that runs before ltsp-client-setup
14:09
krbaum: in this script, you can do something like:
14:09
<alkisg>
(how about using the ltsp_config.d folder?)
14:09
<Gadi>
if grep fat /proc/cmdline >/dev/null 2>&1; then
14:10
echo LTSP_FATCLIENT=True > /var/cache/ltsp_config.d/userfat.conf
14:10
fi
14:10
alkisg: that's the right dir, right?
14:10* alkisg isn't sure
14:11
<Gadi>
yeah, I have to check, as well
14:11
/usr/share/ltsp/ltsp_config.d
14:11
doh
14:11
<alkisg>
Yeah, I think he could just drop the script there
14:11
No need to generate other files..
14:12
<Gadi>
oh thats right
14:12
we execute scrpts there
14:12
nm
14:12
krbaum: just add a script in the chroot's /usr/share/ltsp/ltsp_config.d/
14:12
that runs:
14:12
if grep fat /proc/cmdline >/dev/null 2>&1; then
14:12
export LTSP_FATCLIENT=True
14:12
fi
14:13
or you can make it a case statement
14:13
<alkisg>
Maybe the opposite of that
14:13
<Gadi>
for fat and thin
14:13
and if nothing do nothing
14:13
<alkisg>
As LTSP_FATCLIENT is true for fat chroots...
14:13
<Gadi>
just use lts.conf
14:13
I would force it either way if there's a cmdline arg
14:13laron has quit IRC
14:14
<krbaum>
I'll try this. Thank you.
14:14
<Gadi>
and do nothing otherwise
14:14
<alkisg>
if ! fgrep -q fat /proc/cmdline; then export LTSP_FATCLIENT=False; fi
14:14
k
14:15
<Gadi>
case "$(cat /proc/cndline)" in
14:15
...
14:15
:)
14:16the_fronny has left #ltsp
14:16
<highvoltage>
heh, I like how there's always shell 1-upping in #ltsp :)
14:18
<krbaum>
Shouldn't it be the other way around: if grep fat.... then export LTSP_FATCLIENT=False ? Since fat is the default on a fat client?
14:19
Forget that last comment
14:26
<alkisg>
Gadi: plymouth --quit in ltsp-client-core didn't make any difference... :-/
14:26
<Gadi>
did plymouth go away?
14:28
<krbaum>
alkisg: Gadi told me you are working on getting USB devices to work on fat clients. How is that going? I thought it might have to do with ck-list-session displaying active = FALSE. Have you got an idea on what's wrong?
14:29
<alkisg>
Gadi, yes, and I think it went away sooner than usual...
14:29
krbaum: I've fixed that, you can get the sources from ltsp/ldm_trunk
14:29
It works ok for one X screen, but it might have problems for multiple x screens.
14:30
<krbaum>
Wow! Great.
14:30
<alkisg>
The workaround I've found to make the CK session active works with system-wide env vars, so.. :-/
14:30
Anyway it should be fine for most use cases
14:31
krbaum: it now works as usual, i.e. non-admins can mount removeable devices, and admins can mount internal disks as well
14:31
<krbaum>
Ok, great.
14:32
<alkisg>
All user mounts are unmounted at the end of each session.
14:34
Gadi, anyway I'm gonna commit a `chvt ${TTY}` in the ldm screen script, because stgraber was looking to tag ltsp, and if we find a better fix we can revert it...
14:37
<kusznir>
So what are other people doing for authentication with LTSP in cases where they need to integrate windows and several linux "servers"?
14:38
<alkisg>
kusznir: since you're testing fat clients, you might also want to check the new NFS_HOME lts.conf variable. E.g. evolution doesn't work with sshfs, but it does work with nfs exported home.
14:39
(that's the only program i've seen so far that doesn't works with sshfs, though)
14:39
(there may be others...)
14:39
<kusznir>
At this point, I'm not really messing with fat clients...I've got a mix of "normal workstations", windows terminal servers, and LTS servers that need a common authentication and fileserving backend.
14:40
I've been trying for nearly 3 weeks now to make something work....
14:40
My single standalone lts server is getting overloaded; I need to get another server online real soon now...And it needs to share files and authentication with the windows server (that's also been too long in coming).
14:41
<nubae>
so after some time running mobil internet activated devides.... is there any particular modul that coumes to min=
14:41GodFather has quit IRC
14:41
<nubae>
mind?
14:43
<Gadi>
kusznir: I have been using likewise-opne (version 4) as per here: https://help.ubuntu.com/community/UbuntuLTSP/ActiveDirectoryIntegration
14:43
but am in the process of trying likewise-open5
14:43
as it populates getent appropriately
14:44Kicer86 has quit IRC
14:47
<kusznir>
Gadi: is likewise just a configuration frontend to samba+winbind, or a completely separate implementation? (I've never heard of it before)
14:47
<Gadi>
seaparate
14:47
*separate
14:54GodFather has joined #ltsp
14:56
<kusznir>
Gadi: found the likewise main site...do you know how it handles UID/GID assignment/resolution? Specifically, will I get the same UID/GID on different computers joined to the same domain?
14:58pmatulis has quit IRC
14:58
<Gadi>
I believe it is based on the AD domain SID
14:58scottmaccal has quit IRC
14:58
<Gadi>
so, it should be consistent across a single domain
15:00kusznir_ has joined #ltsp
15:03
<krbaum>
Your suggestion on how to select thin|fat upon boot worked really well. Thank you!
15:06
<kusznir_>
krbaum: Hmm..I missed the discussion, but this sounds intertesting..do you have a summary version?
15:07
<alkisg>
(10:38:25 μμ) alkisg: kusznir: since you're testing fat clients, you might also want to check the new NFS_HOME lts.conf variable. E.g. evolution doesn't work with sshfs, but it does work with nfs exported home.
15:07
^^^ I mixed up the names
15:07
^^^ I wanted to say that to krbaum... :D
15:08
<kusznir_>
ok, so its just that...I thought there was some way to select which mode on boot
15:09
<krbaum>
Sure: In the chroot, add a script in /usr/share/ltsp/ltsp_config.d/ (create the directory if it doesn't exist) containing "if grep thin /proc/cmdline >/dev/null 2>&1; then export LTSP_FATCLIENT=False ; fi". Then you can add the kernel parameter "thin" to boot a thin client, the default is to boot a fat client.
15:10
This is nice way to override LTSP_FATCLIENT.
15:13hersonls has quit IRC
15:13
<krbaum>
Now time for a beer. Bye!
15:15otavio has quit IRC
15:18efra has joined #ltsp
15:19
<vagrantc>
krbaum: i just recent set up a separate fat client chroot, and added it to the boot menu.
15:19
krbaum: the pxelinux menu
15:20
a little bit of an unusual use-case (basically, needed to run lshw and a few sudo commands).
15:24shawnp0wers has joined #ltsp
15:41Gadi has left #ltsp
15:46alkisg has quit IRC
15:47lfaraone has left #ltsp
15:53shawnp0wers has quit IRC
16:11
<nubae>
hmmm wanna hear something weird... I changed my thin client's lts from fglrx to radeonhd in xSERVER= and it still ltries for fglrx
16:11
I've restarted inetd, dhcp
16:11
strage behaviour... only thing I can think is tht the chroot somehow has some fglrx stuff hardwired to the kernel
16:12jammcq has quit IRC
16:12
<nubae>
and I neeed to remove that, add the radeonhd modules and rebuild image....
16:12
am I right?
16:12
lts above means lts.conf
16:16
<johnny>
you probably do
16:18mgariepy has quit IRC
16:32vagrantc has quit IRC
16:46
<kusznir>
Gadi: I just installed likewise-open5, set up the AD stuff, but it didn't appear to actually allow anyone else to log in...is there anything else to it? There doesn't seem to be much in docs.
16:51
<johnny>
hmm.. i wonder what i should try for my one monitor that isn't getting edid correctly..
17:03mischko has quit IRC
17:04artista_frustrad has joined #ltsp
17:04nubae has quit IRC
17:04nubae has joined #ltsp
17:12johnny has left #ltsp
17:22johnny has joined #ltsp
17:23laron has joined #ltsp
17:39efra has quit IRC
17:54pmatulis has joined #ltsp
17:59
<johnny>
how do you set the computer to use gmt time?
17:59
terminal*
18:14laron has quit IRC
18:15
<johnny>
ah.. got it..
18:17mikkel has quit IRC
18:23rjune has joined #ltsp
18:28timborn has quit IRC
18:49mordocai has joined #ltsp
18:53GodFather has quit IRC
19:03johnny has left #ltsp
19:10
<AndyGraybeal>
_UsUrPeR_: offtopic , the box came in today, it's hooked up! works great. thank you.
19:11
i think the monitor runs hotter than it :)
19:31staffencasa has quit IRC
19:33litlebuda has quit IRC
19:37johnny has joined #ltsp
20:29hookdr has joined #ltsp
20:31
<hookdr>
anyone have a client setup to use svideo display
20:36mordocai has quit IRC
20:43_Techie_ has joined #ltsp
20:43
<_Techie_>
will an ubuntu LTSP server work properly if i tell X not to start?
20:49
<hookdr>
Has for me
20:50
<_Techie_>
cool
20:51
as far as installing applications, does this have to be done at the server or can it be done by a client workstation?
20:57
<hookdr>
either its the same box
20:59
<_Techie_>
cool
21:01
<hookdr>
_techie_ you have not run one atall just looking into?
21:01
<_Techie_>
first time running one
21:02
so i understand the basics but im still unsure on some parts
21:02Lns has joined #ltsp
21:02
<hookdr>
you use linux much b4?
21:03
<_Techie_>
yeah
21:03Lns has quit IRC
21:03
<_Techie_>
im very familiar with the CLI
21:03
<hookdr>
ah
21:03
k
21:04
<_Techie_>
mind if i throw a few questions at you?
21:04
<hookdr>
shoot
21:04
<_Techie_>
i understand how X forwarding works
21:04
<hookdr>
k
21:05
<_Techie_>
but im unclear as to whether things such as gnome are run on the workstation, or whether the server is doing all the grunt work
21:06
<hookdr>
server
21:06
XDMCP‎
21:06
<_Techie_>
so watching movies over LTSP is not a great idea
21:06
<hookdr>
i have in my LAN
21:07
<_Techie_>
so if i was to run say... mplayer
21:07
it would run on the server and the display would be forwarded to the LTSP client
21:09
<hookdr>
are you fimmiruar with X11
21:09
<_Techie_>
yes
21:09
not 100%, but pretty familiar
21:09
<pmatulis>
LTSP4 used XDMCP, LTSP5 uses X-forwarding
21:10jammcq has joined #ltsp
21:10
<hookdr>
ah k clears me up
21:10
<jammcq>
evening gents
21:11
<hookdr>
server dose all the work
21:15
what application where you putting it work as? a clinet media station?
21:16
<_Techie_>
was gonna start out with general web browsing etc
21:16
and was aiming ot set it up for xbmc
21:20
<hookdr>
ic
21:21
what kinda display are you useing on client?
21:21
<_Techie_>
VGA
21:22
when i progress to xbmc it will suddenly get very complicated as its for media extender boxes
21:22
<hookdr>
ah k should be NB trying to get mine to go svideo
21:22
never playerd with it some more with myth i have
21:23
gen for videos to enetertainem system had gread go with pms
21:23
off a ps3 client
21:24
<_Techie_>
im looking at selling media boxes
21:24
<hookdr>
media boxes?
21:25
<_Techie_>
think like DVD players but only with the functionality of mythTV
21:26
and completely open source
21:26
<hookdr>
ah homebrews or STB
21:26
<_Techie_>
STB???
21:26
<hookdr>
set top box
21:26
<_Techie_>
aiming at STB
21:27
<hookdr>
ah i here yeah there some beuty cases but im kinda cheap 2
21:27
<_Techie_>
im not thinking of beauty, im thinking of functionality
21:28
also they are most likely going to be micro ATX systems
21:29
<hookdr>
yeah gotta old p4 1.8 1GB ram on LTSP working NP
21:30
just want to get it to push out through svideo
21:31try2free has joined #ltsp
21:33
<_Techie_>
ati or nvidia
21:33
<hookdr>
ati
21:34
<_Techie_>
i know youve probably seen this sort of info before but ill link it anyway, https://help.ubuntu.com/community/RadeonDriver#Configuring%20your%20graphics%20card%20%22Device%22%20section
21:35
i think theres a section for svideo
21:36
<hookdr>
yeah i had working fine with os installed on box but HDD died and went LTSP sould of kept X11 conf to put in LTSP config
21:36
<_Techie_>
i havent exactly had the need to force any of my ATI cards to use S-Video as i have HDMI
21:37
<hookdr>
ah yeah lol much easier to config but im 2 cheap to but new display -- the frustation might buy me one yet LOL
21:38
<_Techie_>
haha, not my area
21:38
im running 2x 17" CRTs
21:39
<hookdr>
ah my cmputer cornner/ i have 2 19s but want the 27 in bedroom to pull it
21:39
27 TV
21:39
<johnny>
mythbuntu is an official plugin
21:39
no need for xmbc is there?
21:39
<_Techie_>
oh great, my LTSP install of ubuntu isnt booting
21:39
<johnny>
it's an option to ltsp-build-client
21:40
<_Techie_>
i chose to use XBMC over myth due to the ease of use and the eyecandy
21:40
makes for a much nicer system to sell
21:41
<hookdr>
ah yeah
21:41
<johnny>
the point is that ease of use is easy ..
21:41
since it installs automatically via ltsp-build-client :)
21:43
<hookdr>
ah not able to use livestation through it thou?
21:44
<_Techie_>
when i meant ease of use, i wasnt meaning ease of implementation via LTSP, ive got all the time in the world to get that right
21:44
i meant ease of use as a media player
21:51try2free has left #ltsp
21:58alkisg has joined #ltsp
22:10
<hookdr>
Any one successfully boot a intel imac from pxe
22:10
<_Techie_>
i dont have access to an iMAC
22:20
hookdr, how did you stop X from starting, my GDM persists on starting even after i run "sudo update-rc.d -f gdm remove"
22:26
<hookdr>
SCREEN_02 = shell
22:27
<_Techie_>
umm
22:29
i dont quite think tahts of help
22:37
i think i have found the solution
22:38
<hookdr>
add in lts.conf
22:38hookdr has quit IRC
22:51rjune has quit IRC
22:53
<pmatulis>
_Techie_: configure GRUB
22:56
<_Techie_>
thanks pmatulis already found that
23:04
<pmatulis>
good night everybody
23:05
<_Techie_>
night
23:05pmatulis has quit IRC
23:42alkisg has quit IRC
23:46__Techie__ has joined #ltsp
23:49_Techie_ has quit IRC
23:50_Techie_ has joined #ltsp
23:53__Techie__ has quit IRC