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


Channel log from 8 March 2010   (all times are UTC)

00:01ogra_cmpc has joined #ltsp
00:21try2free has joined #ltsp
00:24try2free has left #ltsp
00:52
<alkisg>
I've submitted the "Cannot umount /proc after using update-binfmts in a chroot" problem as a bug: https://bugs.launchpad.net/ubuntu/+source/binfmt-support/+bug/534211
00:56Eghie has quit IRC
01:17cyberorg has quit IRC
01:18cyberorg has joined #ltsp
01:33pths has quit IRC
01:43pts has joined #ltsp
01:45pts is now known as pths
01:51pths has quit IRC
01:53pts has joined #ltsp
01:57pts has quit IRC
01:58pts has joined #ltsp
02:04pts is now known as pths
02:04gnunux has joined #ltsp
02:11
<gnunux>
hi
02:12
<alkisg>
vagrantc: hi, would diverting /sbin/modprobe be acceptable? (like we do with start-stop-daemon)?
02:13
<vagrantc>
alkisg: as long as we can reliably pass arguments on...
02:14
in general, diversions should be avoided, if at all possible ...
02:16
<alkisg>
It causes problem, e.g. that one: https://bugs.launchpad.net/ubuntu/+source/binfmt-support/+bug/534211
02:17
I might be able to divert update-binfmts instead, but I think modprobe is the root of the problem
02:17
So it would be more generic to disallow module loading on the chroot
02:18
<vagrantc>
right... which seems like a good thing.
02:18
i.e. loading modules from the chroot seems like a bad idea.
02:19
though most cases would just break, no?
02:19
<alkisg>
I'm not sure
02:19* alkisg looks if he can conditionally blacklist things...
02:21
<alkisg>
Hmmm we could use the "install" modprob.conf command
02:21
I.e. 'if LTSP_HANDLE_DAEMONS=False then /sbin/modprobe ... else don't do anything'
02:24try2free has joined #ltsp
02:27
<alkisg>
Bah, nope, that doesn't solve the umount problem :(
02:28vagrantc has quit IRC
02:44bobby_C has joined #ltsp
02:52ogra has joined #ltsp
02:57mikkel has joined #ltsp
03:20try2free has left #ltsp
03:47ogra_cmpc has quit IRC
03:59otavio has quit IRC
04:04otavio has joined #ltsp
04:44alkisg has quit IRC
04:54ogra_ has joined #ltsp
04:55ogra has quit IRC
04:55ogra_ is now known as ogra
04:55ogra has joined #ltsp
05:20Q-FUNK has joined #ltsp
05:20
<Q-FUNK>
re
05:22daya has quit IRC
05:48sene has joined #ltsp
05:50sene has quit IRC
05:57bobby_C has quit IRC
06:13alkisg has joined #ltsp
06:16pmatulis has joined #ltsp
06:37scottmaccal has joined #ltsp
06:38scottmaccal has quit IRC
06:41hersonls has joined #ltsp
06:44rjune has quit IRC
06:51scottmaccal has joined #ltsp
06:54scottmaccal has quit IRC
06:56scottmaccal has joined #ltsp
06:59scottmaccal has quit IRC
07:01scottmaccal has joined #ltsp
07:02ogra_cmpc has joined #ltsp
07:04scottmaccal has quit IRC
07:06scottmaccal has joined #ltsp
07:09scottmaccal has quit IRC
07:49scottmaccal has joined #ltsp
07:54MaRX-Mode has quit IRC
07:55evilx has quit IRC
07:56scottmaccal has quit IRC
08:01scottmaccal has joined #ltsp
08:03shawnp0wers has joined #ltsp
08:13etyack has joined #ltsp
08:14otavio has quit IRC
08:15Gadi has joined #ltsp
08:18otavio has joined #ltsp
08:18otavio has joined #ltsp
08:33bobby_C has joined #ltsp
08:35mikkel has quit IRC
08:42evilx has joined #ltsp
08:43Q-FUNK has quit IRC
08:50litlebuda has joined #ltsp
08:53Pulga has joined #ltsp
09:02mikkel has joined #ltsp
09:03
<Pulga>
I installed ltsp 5 in ubuntu 9.10, to the login okay, but the letters appear upside down on the screen in gnome, any idea??
09:03
<Gadi>
!compiz
09:03
<ltspbot`>
Gadi: "compiz" :: if compiz is giving you problems, one way to disable it for all users is: sudo gconftool-2 --direct --config-source xml:readwrite:/etc/gconf/gconf.xml.mandatory --type string --set /desktop/gnome/session/required_components/windowmanager metacity
09:03
<Gadi>
Pulga: ^^
09:04
<Pulga>
hi
09:04
this is the problem
09:04
<Gadi>
Pulga: you must disable compiz
09:04
<Pulga>
very crazy
09:04
<Gadi>
yes
09:04
<Pulga>
how to..
09:04
<Gadi>
!compiz
09:04
<ltspbot`>
Gadi: "compiz" :: if compiz is giving you problems, one way to disable it for all users is: sudo gconftool-2 --direct --config-source xml:readwrite:/etc/gconf/gconf.xml.mandatory --type string --set /desktop/gnome/session/required_components/windowmanager metacity
09:05
<Gadi>
Pulga: ^^^^
09:06CAN-o-SPAM has joined #ltsp
09:06
<alkisg>
Or you may remove the proprieratyr nvidia drivers from the server :)
09:06
<Pulga>
I have a nvidia Inside
09:06
!!!
09:06
<ltspbot`>
Pulga: Error: "!!" is not a valid command.
09:07
<alkisg>
Yes, nvidia on the server, intel on the clients
09:07
<Pulga>
running : sudo gconftool-2 --direct --config-source xml:readwrite:/etc/gconf/gconf.xml.mandatory --type string --set /desktop/gnome/session/required_components/windowmanager metacity solves the proble?
09:07
<alkisg>
Yup
09:07
But that disables compiz
09:07
So either of those will do
09:07
(disable compiz or remove the nvidia drivers)
09:07
<Pulga>
what is compiz for??
09:08
<alkisg>
Fancy window moving
09:08
<Pulga>
ahh
09:08
I will remove nvidia
09:08
i think its better
09:08
<Gadi>
sure, until the next thin client with a different driver that won't let you login due to compiz
09:09
<alkisg>
Heh :D
09:09
<Pulga>
thanks, I will try..... i doens't work i'm back
09:09
<alkisg>
We need an anti-compiz moto
09:09
<Gadi>
compiz would be fine if it just did proper checks
09:09
<alkisg>
We do need to start filing bugs, though...
09:10
I don't think anyone has...
09:10
<Gadi>
I sent in a patch
09:10
<alkisg>
And? any responces?
09:10
<Gadi>
that disables compiz for remote connections unless someone specifies SKIP_CHECKS="yes"
09:10
never got merged
09:10
:(
09:11
<alkisg>
Yeah, I guess they need per-driver-fixes
09:11
...which obviously we don't have the time for
09:11
<Gadi>
well, it would be nice if you could *disable* compiz by env var at least
09:11
then, we could set the env var in LDM by default
09:21Q-FUNK has joined #ltsp
09:22mgariepy has joined #ltsp
09:25
<moldy>
hi
09:25
on my terminal server, when a bunch of people is logged in, dbus runs crazy (100%cpu) and makes the system unuseable
09:25
is that a known issue? does anyone know anything about this?
09:28
<AndyGraybeal>
moldy from colorado?
09:30
<alkisg>
mgariepy: hi, would you have any clues about https://bugs.launchpad.net/ubuntu/+source/ltsp/+bug/533964 ?
09:32
<mgariepy>
hi
09:34Q-FUNK has quit IRC
09:34
<mgariepy>
i don't use screen_02=shell since i already have a shell on vt1 by default.
09:34
<alkisg>
It happens with vt1 as well
09:35
I.e. with no lts.conf at all, I sometimes get ldm over vt1
09:35GodFather has joined #ltsp
09:35
<alkisg>
So I lose the shell there...
09:35
<mgariepy>
last week i fixed the openvt to switch to add the -s
09:35
<alkisg>
And it improved things, but I think there's still a problem...
09:36
<mgariepy>
with my tests last week this was working great, i might have an hour or so to test if i get the same issue here.
09:36
but i don't think it's gonna be today
09:36
<alkisg>
It seems to be some race condition, I'm only having the problem half of the times (or less)
09:37
<mgariepy>
last week i noticed that when i was logged on VT1, when killing Xorg, x was starting on the current vt.
09:37
<alkisg>
Right
09:37
<mgariepy>
it might have something to do with that.
09:37
<alkisg>
And if start-stop-daemon 07 is called before start-stop-daemon 01, then the problem occurs - or something like that...
09:37etyack has quit IRC
09:38
<alkisg>
So I'm wondering if we should "prefer" switching to the vt with ldm?
09:38
(or rdesktop or...) :-/
09:38
<mgariepy>
i think that would duplicate the code no ?
09:39
i'll do some testing this week, i'll let you know about my results
09:39
<alkisg>
I mean, to only call chvt for that one vt, changing this code
09:39
Thank you :)
09:41
<moldy>
AndyGraybeal: umm, no, sorry :)
09:41
<AndyGraybeal>
k :)
09:45GodFather has quit IRC
09:45MaRX-Mode has joined #ltsp
09:54The_Code has joined #ltsp
09:59jammcq has joined #ltsp
09:59
<jammcq>
good morning friends
10:02timborn has joined #ltsp
10:06
<CAN-o-SPAM>
hi jammcq
10:13thunsucker has joined #ltsp
10:13staffencasa has joined #ltsp
10:14Selveste1_ has quit IRC
10:15
<mgariepy>
hey alkisg, for your shell/ openvt problem the problem arrive when you kill X or on logout ?
10:15
<alkisg>
mgariepy: on boot
10:16
<mgariepy>
ok
10:16
<alkisg>
*but I've also seen it on kill x _and_ on logout
10:20
<mgariepy>
alkisg, ok, i'm working ono this right now. i tought that the -s was working, but it doesn;t seam to be the case ;)
10:21
<alkisg>
mgariepy: I'm more than willing to test whatever you may find :)
10:25gnunux has quit IRC
10:25
<mgariepy>
i think i know what's happening, the change on screen_session is working the way it should.
10:25
this script isn't reexecuted on logout
10:26
<alkisg>
But then why it isn't working correctly sometimes on the first boot?
10:26
(without even logging in?)
10:30
<mgariepy>
if somehow the vt change, between the screen_session and the execution of the screen.d script, it's just get messed up.
10:31
so technicaly it's possible that both script are executed in parallele and the vt get confused.
10:33shogunx has quit IRC
10:34etyack has joined #ltsp
10:54shogunx has joined #ltsp
10:55thunsucker has quit IRC
10:56thunsucker has joined #ltsp
11:03pmatulis has quit IRC
11:05pmatulis has joined #ltsp
11:11gentgeen__ has quit IRC
11:12khris4 has joined #ltsp
11:13
<khris4>
hello
11:13khris4 has quit IRC
11:14
<alkisg>
Hi :)
11:21gentgeen__ has joined #ltsp
11:28etyack has quit IRC
11:36thunsucker has quit IRC
11:37thunsucker has joined #ltsp
11:45
<alkisg>
I want to rename 090-clean-package-cache to 089-clean-package-cache. In plugins/Ubuntu, that is a symlink to the debian dir. I need to delete it and create a new symlink, right?
11:58MorpheusDe has joined #ltsp
12:00
<AndyGraybeal>
i don't know how to wrap my head around this, but i'm wondering -- is it possible to run a virtual machine kvm/qemu/libvirt as an ltsp server?
12:01
<MorpheusDe>
Hi Folks - New to LTSP and using irc for the first time - hope you'll be a little bit patient with a noob....
12:01
Can someone point out to me whether it is possible or not to connect to a thin client's desktop via RDP from another box?
12:03
<alkisg>
AndyGraybeal: sure. Personally I've used vbox for that, but I heard of others using kvm etc...
12:03
MorpheusDe: is that gnome?
12:03
<AndyGraybeal>
thank you alkisg
12:03
<MorpheusDe>
Working on Ubuntu (9.10) I have two clients booted up right now but when I choose the remote desktop preferences of course the IP of the server is shown - so how can I actually connect to one specific client?
12:04
Jepp - it's Gnome
12:04
<alkisg>
MorpheusDe: you need to change the ip where vino-server is listening on
12:04
Otherwise all users would use 5900 - which isn't really possible
12:04
To do that, you need to run gconf-editor locally
12:05
And navigate to /desktop/gnome/remote_access/alternative_port
12:05
<MorpheusDe>
Locally means on each thin client?
12:05
<alkisg>
Yes
12:05
On each *user*
12:05
There are other solutions, though, e.g. iTalc
12:05
Or x11vnc, or ...
12:06
<MorpheusDe>
Being new to the whole topic I would like to stick to "standard" tools include by default first...
12:06
<alkisg>
italc is the standard tool for ubuntu/ltsp
12:06
So that would be your best bet
12:06
<MorpheusDe>
The tools you mention must be installed in the chroot, right?
12:06
<alkisg>
No
12:06
So you can just disable *ALL* vino-servers (the one in the remote desktop preferences),
12:07thunsucker_ has joined #ltsp
12:07
<MorpheusDe>
So on the server itself, right?
12:07
<alkisg>
and just install the italc-master package
12:07
Yes
12:07thunsucker_ has quit IRC
12:07thunsucker_ has joined #ltsp
12:08
<MorpheusDe>
Please give me a sec - Will try adjusting the gconf-editor
12:08thunsucker has quit IRC
12:08
<alkisg>
OK. Again, this is the *non standard* way to do it.
12:11ogra_cmpc has quit IRC
12:12litlebuda has quit IRC
12:13cliebow has joined #ltsp
12:15
<Gadi>
alkisg, mgariepy: why do we do all this gymnastics with the vt switching? Why not simply switch vt to the lowest number of SCREEN_XX for all screens?
12:16
that way, you get a consistent default screen and you don't need to magically calculate things
12:16
<MorpheusDe>
alksig: Ok changed the port using gconf-editor but I still can't access the specific client and because you've again pointed out that it is *not* the standard way do I get you right that your reccomended tool would be italc
12:16
<alkisg>
Gadi, I think that if we switch to vt1, and we (supposedly) have ldm on vt7, then ldm will start on vt1 instead
12:16
<Gadi>
alkisg: I am not suggesting switch to vt1
12:17
<alkisg>
MorpheusDe: yes, I'd recomment italc over vino-server for ltsp.
12:17
<Gadi>
I am suggesting that if SCREEN_07 is the only one defined, switch to vt7
12:17
if SCREEN_08 is also defined, switch to vt7
12:17
<alkisg>
Gadi, and what if we have multiple screens defined?
12:17
E.g. SCREEN_02=shell SCREEN_07=ldm
12:17
<Gadi>
if 09 -> still switch to vt7
12:17
<alkisg>
So the lowest, except for vt1?
12:18
(I think vt1 is usable on fedora...)
12:18
<Gadi>
the lowest of anything
12:18
we should not make exceptions
12:18
<alkisg>
Gadi, in that case, I *think* that X will start on vt1, independed of where you want it to start
12:18
(i.e. it uses the active console)
12:19
<Gadi>
in what case?
12:20
<alkisg>
In the default case, for example, where vt1=logs and vt7=ldm
12:20
<mgariepy>
alkisg, i had the same problem, x as started on vt 2 instead of vt 7..
12:20
<Gadi>
what Im saying is in the default case, SCREEN_07=ldm
12:20
so, it should switch to vt7
12:21
(if it even needs to switch - since openvt is already called with -c 7)
12:21
<alkisg>
Gadi, if you run openvt -c 7 ldm, and quickly switch to vt2, you get ldm on vt2
12:21
(I think that) X uses the active vt
12:21
<Gadi>
then, dont switch vt
12:21
:)
12:22
<alkisg>
So we need to "guess" on which screen X is, _or_ we could do it all serially, not with parallel start-stop-daemons..
12:22
Gadi, I don't - but start-stop-daemon calls the screen scripts in parallel
12:23
<Gadi>
we chvt in our code atm
12:23
what happens if we don't?
12:23johnny has joined #ltsp
12:23
<alkisg>
I believe ldm will start on vt1 then.
12:23
(haven't tried, though)
12:23
<Gadi>
I think ur wrong
12:23
:)
12:24ogra_cmpc has joined #ltsp
12:24
<Gadi>
afaik, we introduced the chvt to make the transition prettier
12:24
<alkisg>
I hope so :)
12:24
(that I'm wrong)
12:24
<Gadi>
in any event, I think we *should* have a default screen when multiple ones are defined
12:24
and not leave it as a race
12:25
<alkisg>
Choosing the highest would make more sense to me than the lowest, as vt1 containst text usually
12:25
<Gadi>
which means, if we have 02, 07, 08, 09, .... by default it always goes to 02 OR an lts.conf-specified preferred screen
12:26
vt1 is only a choice if someone does SCREEN_01
12:26
<alkisg>
vt1 is active, it's opened by the initscripts (not ours)
12:26
<Gadi>
right, but I am saying limit the choices to those SCREEN_XX defined
12:28
<alkisg>
I'd still use the highest one for X...
12:28
...but anyway, I don't think my knowledge on VTs is good enough to allow me to express any opinions :)
12:28
<Gadi>
lol
12:28
chicken
12:28
<mgariepy>
hehe lol
12:28
<alkisg>
...so... I can test whatever you guys choose :)
12:29
<mgariepy>
to spawn screen.d scripts to to the correct VT, you suggest we do what exactly Gadi ?
12:29
<alkisg>
Well, I could tell you all about consoles on Windows, though :P
12:30
<Gadi>
mgariepy: so, first we need to know whether in the absence of any additional "chvt" calls, openvt -c <vt_num> will attach the resulting screen script to <vt_num>
12:30
<alkisg>
btw, (07:45:58 μμ) alkisg: I want to rename 090-clean-package-cache to 089-clean-package-cache. In plugins/Ubuntu, that is a symlink to the debian dir. I need to delete it and create a new symlink, right?
12:31
<Gadi>
mgariepy: right now, with all the added vt siwtching code, that much is unclear
12:32etyack has joined #ltsp
12:32
<Gadi>
mgariepy: if it does, then we are causing a problem by running openvt and quickly switching vts away from where openvt puts us
12:33
mgariepy: if we *need* our chvts to make openvt attach to the right place, then the chvt needs to happen in the screen session itself, meaning in a screen-session.d/ scipt
12:33
but, I don't believe the latter to be the case
12:34
mgariepy: if at the end of the day, we are looking for "prettiness of transition", we should *only* do chvt in screen_session.d/ AND have all chvts go to the same vt
12:34
so, whichever screen script is initialized, they all init and switch
12:34
tho, maybe the switch should happen post-X init, meanining in xinitrc.d/
12:35Pulga has quit IRC
12:36
<mgariepy>
-w wait for command to complete. If -w and -s are used together then openvt will switch back to the controlling terminal when the command completes.
12:38
i'll remove all openvt option but the -c and do some testing, i'll post my result on lp:533964
12:39
<alkisg>
I'd like to commit that (http://alkisg.pastebin.com/DxrEsbSJ) as plugins/common/090-umount-dirs. Does it look cross-distro "enough"? johnny? cyberorg?
12:39
<Gadi>
ah, yeah, thats kinda what we want
12:39
try adding -s
12:39
we don't use that at the moment
12:39
<mgariepy>
-s is already there
12:40
;) i added it last week
12:40
r1697
12:40
<Gadi>
alkisg: what version were you using?
12:40
<alkisg>
trunk
12:41
<mgariepy>
but it doesn't behave correctly. when the terminal is started, if i do chvt 1, pkill X, then X start on vt 1
12:41
<alkisg>
(i.e. with mgariepy's fix - it's a little better, but still suffers from race conditions)
12:42* Gadi doesn't think we should have all this code with "if current tty is different than screen_num, switch, etc"
12:42
<mgariepy>
if i kill screen_session 07, start it over again with start-stop-deamon the vt is always right ;)
12:42
<Gadi>
it does not make sense to me
12:43
<alkisg>
I think vagrantc has remove some of that for debian
12:44
<Gadi>
I don't believe X initializes on a vt until you switch to it when it is run from openvt
12:45
<alkisg>
Not using chvt at all would be great, if it works.
12:45litlebuda has joined #ltsp
12:45hersonls has quit IRC
12:46hersonls has joined #ltsp
12:47
<mgariepy>
give me 10 or so min i'll tell you if it works
12:47
<Gadi>
actually, the cleanest code might be to only add the "-s" flag to openvt for a particular screen_num
12:48
(that screen_num being the default)
12:49Lns has joined #ltsp
12:50
<mgariepy>
i thinks that the problem is that screen_session calculate "CURTTY_NUM, TTY_NUM, etc.. " outside of the main loop. and it's not true when it gets into the loop.
12:51
<MorpheusDe>
alksig: Thank you for recommendig italc looks like a pretty powerful software - Now I only got to figure out how it works - I can start it on the lts server itself but that was not what I wanted. Don't get me wrong I don't want you guys help me figure out how to deal with that piece of software - just to clarify that I can do with it what I want - The main point for me is to control thin clients from another physically separated workstation
12:51
<mgariepy>
maybe we would just need to refresh those.
12:51
<Gadi>
mgariepy: I think "-s" obviates the need to calculate any vts
12:52
unless we need to chvt
12:52
<CAN-o-SPAM>
MorpheusDe: do you need anymore functionality aside from the simple means of controlling another thin client? i.e messaging, screen locking, monitoring but not controlling?
12:52
<MorpheusDe>
It's just meant for remote support....
12:53
sorry - that was not very clear - remote control is the main feature I want
12:53
this is the reason I asked for remote desktop in the first place
12:54
<alkisg>
MorpheusDe: you wrote "alksig" so I didn't know you were talking to me - use <tab> :)
12:54
MorpheusDe: you can use italc to remote support your clients even from a windows workstation in a different subnet
12:55
MorpheusDe: do you want to use it from an ltsp client? or from another ubuntu workstation? or from another os?
12:56
<johnny>
i upgraded to karmic last night..
12:56
<MorpheusDe>
alright - thank you guys - you got more important stuff to discuss here - so the only thing left todo is to figure out how to use it on the remote workstation now - right now it complains about the service that's not running - but the service is of course running on the ltsp machine.... I gues rtfm is now the thing todo :-)
12:56
<johnny>
i have some questions for fat clients for you now alkisg ...
12:56
but i think i'm gonna wait til tomorrow to ask em
12:57
<alkisg>
MorpheusDe: some quick tips:
12:57
<MorpheusDe>
from another ubuntu workstation
12:57
<alkisg>
MorpheusDe: (1) copy the italc keys from /etc/italc, from the ltsp server to the other workstation,
12:57
(2) add the "controller" user to the admin group
12:57
(3) logoff/logon for the changes to take effect
12:58
(1) is a little hard, you need to get the permissions + the owner right.
12:58
johnny: shoot :)
12:59
<MorpheusDe>
alright - thank you - give me some minutes - will try it...
12:59
<mgariepy>
Gadi, alkisg updating the image now, we'll be fixed soon if it works great (i think) ;)
12:59
<alkisg>
Great
13:04
<johnny>
there are now 3 options..
13:05
--fatclient, --high-fat, --fatclient-desktop
13:05
<mgariepy>
Gadi, alkisg my ttys got mixed again.
13:05
<johnny>
until i upgraded to stgraber ppa in karmic.. there were only the first 2
13:05
<alkisg>
johnny: There's no --high-fat option - that's nubae's plugin
13:05
Do remove that one..
13:06
--fatclient == autodetection, --fatclient-desktop==specify a specific desktop (e.g. ubuntu-desktop, edubuntu-desktop-gnome etc)
13:06
<johnny>
ah..
13:06
so.. if i have 512mb ram.. what should i use :)
13:06
<alkisg>
I'm guessing... lubuntu? :D
13:06
I'd go for standard ubuntu ;)
13:06
<johnny>
ok
13:07
so.. how hard is it to have 2 computers fat client and 1 not?
13:07
<alkisg>
How much ram on the "not" one?
13:08
<johnny>
256 or 128.. i can't recall
13:08
<alkisg>
With 256, it's pretty easy, you just use LTSP_FATCLIENT=false in lts.conf.
13:08
<johnny>
ah
13:09
<alkisg>
With 128 you either need to disable some services, or use a different chroot
13:09
<johnny>
ok..
13:09
<alkisg>
(a thin one)
13:09
<johnny>
thanks alkisg .. that's all i needed to know
13:09
<alkisg>
johnny: I don't really like the ltsp-build-client --extra-help description of --fatclient, could you come up with something better?
13:10
Now they're both the same... I didn't want to change them myself, being non-native speaker, but I think someone should
13:10
<mgariepy>
Gadi, http://pastebin.com/vdrNShuG
13:10
<MorpheusDe>
alkisg: Controller user is my primary user that is a member of the admin group on the workstation anyway - Do the keys have to be owned by that user?
13:10
<johnny>
so.. what about --fat-client-desktop vs --fat-client ?
13:10
<mgariepy>
i think i'm missing something do i ?
13:10
<alkisg>
MorpheusDe: ls -lha -R /etc/italc/keys on the server ==> see the permissions there, and duplicate them
13:11
(09:07:02 μμ) alkisg: --fatclient == autodetection, --fatclient-desktop==specify a specific desktop (e.g. ubuntu-desktop, edubuntu-desktop-gnome etc)
13:11
johnny: so, --fatclient would use xubuntu-desktop, if your server already had xubuntu
13:11
<mgariepy>
Gadi, with this on my first boot, vt 2 and vt7 (shell and ldm) ldm got spawned on vt2 instead of vt7.
13:12
<alkisg>
Right, I think X uses the active vt
13:12
So we need to do at least one switching
13:14
<johnny>
alkisg, ok..
13:14
ah.. now i get it
13:14
ok.. fatclient it is ..
13:14
thanks alkisg ..
13:14
<alkisg>
np
13:14
<mgariepy>
screen-x-common, line 41
13:15
damn
13:15
fgconsole to calculate the vt to use.. the problem isn't X is this script.
13:17
<MorpheusDe>
alkisg: Alright permissions fixed - but do I not have to tell the italc app to look on the remote server ip for the running service - can't find an option for that...
13:17
<alkisg>
mgariepy: http://bazaar.launchpad.net/~vagrantc/ltsp/ltsp-debian-packaging/annotate/head%3A/patches/revert-fgconsole.dpatch
13:17
MorpheusDe: is that on the local network (same subnet)?
13:18
<MorpheusDe>
alkisg: Yepp
13:18
<alkisg>
MorpheusDe: restart everything, to make sure everything's running properly. It should autodetect all the clients on the local network.
13:18
So you don't need to specify any IPs at all.
13:18
<MorpheusDe>
alkisg: wks 192.168.250.130 --> ltsp machine: 192.18.250.223
13:19
alkisg: Opps - Sorry - wks 192.168.250.130 --> ltsp machine: 192.168.250.223
13:19
<alkisg>
MorpheusDe: it only works for logged on users, though, so you need to login for the clients to get autodetected
13:21
<MorpheusDe>
alkisg: Clients are up and when I start italc on one of the thin clients itself it works - When i start the italc software on the separate workstation it keeps complaining about the fact that there's no service running on the local (the workstation) machine....
13:22
<alkisg>
MorpheusDe: and you've rebooted that workstation?
13:24etyack has quit IRC
13:24
<MorpheusDe>
alkisg: Sorry - no - A reboot of the whole machine's necessary? It's the machine I'm writing from right now... I don't want to keep bugging you with so much non development related stuff.... I will check again that I have everything set up correctly and will try again 2mrow - Had a long day so far - Thank's again for all your help
13:25
<alkisg>
MorpheusDe: ok, open a terminal, and run: ica-launcher &
13:25
then, try to run italc again.
13:28
<MorpheusDe>
alkisg: thx - interface's up...
13:28
<mgariepy>
alkisg, do you know why this is not in ltsp-trunk ?
13:29
<alkisg>
MorpheusDe: do you seee the other clients?
13:29
mgariepy: no - I think vagrantc didn't like the ubuntu changes, and he reverted them, but I've no idea who did the ubuntu changes, didn't look...
13:29
<MorpheusDe>
alkisg: right now I only see the workstation itself - just wanted to try to use the support button and give it a try...
13:30
<alkisg>
MorpheusDe: do you have any other users logged on?
13:30* Gadi has no idea why we use fgconsole at all if we specify SCREEN_XX
13:30
<Gadi>
we should always just use XX
13:30
where XX=tty
13:31
<MorpheusDe>
alkisg: Yepp I do - two thin clients are up
13:31
<alkisg>
MorpheusDe: what do you get with: avahi-browse -trp _italc._tcp
13:32scottmaccal has quit IRC
13:33
<MorpheusDe>
alkisg: Among others: =;eth0;IPv4;italc\032root;_italc._tcp;local;lts.local;192.168.250.223;5902;"lts" "fbffecaafcdc7cd39357cf34d61f9f37" "fc6ca0d44c2d800d38cf6d452c7c1c19" "70a2424e2f427abf08d9df9acd8f6db0"
13:33
<alkisg>
MorpheusDe: that one only?
13:33
You should be seeing 3 of them
13:34
If you're seeing more, put all of them to pastebin
13:34
!pastebot
13:34
<ltspbot`>
alkisg: "pastebot" :: The LTSP pastebot is at http://ltsp.pastebin.com. Please paste all text longer than a line or two to the pastebot, as it helps to reduce traffic in the channel. Don't forget to paste the URL of the text here.
13:34
<MorpheusDe>
alkisg: Nope - the other ones are all related to other networks
13:34
<alkisg>
MorpheusDe: this command only looks for italc clients
13:35
There are italc clients on other networks near you?
13:35
MorpheusDe: on the thin clients that are logged on, did you log them on *after* italc was installed?
13:35
<MorpheusDe>
alkisg: No - there all just all the IPs of the workstation listed: Two physical nics and two nics from a local VM ware server
13:36
<alkisg>
They need to logoff/logon for the service to get loaded.
13:36
<MorpheusDe>
alkisg: yepp - clients were logged on before the local (workstation) italc was installed - log out - log in again on the clients?
13:37
<alkisg>
The server italc installation is the important part here, as the clients get that installation
13:37
Do try to log them off/on in any case...
13:38
<MorpheusDe>
alkisg: Awesome!
13:39
You guys rock!
13:39* alkisg guesses that it works.. :)
13:39
<Gadi>
mgariepy: can you try calling this in screen_session: ${openvt} -s -f -w -c ${TTY_NUM} -- TTY_NUM=${TTY_NUM} ${SCRIPT_DIR}/${SCREEN_SCRIPT} ${SCREEN_ARGS}&
13:39
<MorpheusDe>
Thank's a lot - Way more than I asked for - That's a great piece of software!
13:39
<Gadi>
mgariepy: and then have screen-x-session use TTY_NUM
13:40
mgariepy: instead of fgconsole
13:41
<mgariepy>
ok i'll test this as soon as i have my thin client up and running again. i had a crash ;)
13:43ogra_cmpc has quit IRC
13:44etyack has joined #ltsp
13:45ogra_cmpc has joined #ltsp
13:47MorpheusDe has quit IRC
13:56X-Raimo has joined #ltsp
14:00
<X-Raimo>
hi all. I have server with sound card, it works as LTSP Server. Currently, when client plays music it plays from server's speakers. And my ltsp-clients with their sound cards need to play music via their own sound cards.
14:09
<AndyGraybeal>
my staff wants to use their jumpydrives (stickdrive, thumbdrive, etc) but it's not letting them 'safely remove' them. is there a configuration i can change to make it so they can safely unmount them?
14:12
<alkisg>
X-Raimo: distro/version?
14:12
AndyGraybeal: you can safely remove them after 2 seconds - it flushes every 2 secs
14:12
<AndyGraybeal>
aah thank you!
14:14
<X-Raimo>
alkisg: debian lenny and ltsp Version: 5.1.98-1
14:14
<alkisg>
X-Raimo: gnome?
14:15
<johnny>
probably kde?
14:15
nubae had the same problem
14:15
<X-Raimo>
alkisg: kde3
14:16
<alkisg>
k, no idea there. It should be using pulse and redirecting it to the clients, but i've no idea how that should work on kde
14:16
<johnny>
do we have any ltsp devs who use kde at all?
14:16thunsucker_ has quit IRC
14:18
<alkisg>
Some day kde will be using pulse by default and will make things easier :D
14:18
<X-Raimo>
alkisg: I prefer solution without pulse if possible. If it's not possible, then how and where (at server or at chrooted /opt/ltsp/) to setup pulseaudio?
14:20
<alkisg>
X-Raimo: try again later on, others, more experienced on debian/kde might be here. I guess you'll need pulseaudio on the server, but I don't know how you're supposed to set it up.
14:21
<X-Raimo>
alkisg: can nas or esd can be alternative to pulse?
14:21
<mgariepy>
Gadi, i won't have time to fix this today. i got an emergency to attend to, i'll be able to do more testing later this week.
14:21
<alkisg>
I think esd worked at some time, but I also think it's been years since someone tested it.
14:21
mgariepy: thanks though :)
14:22
<mgariepy>
maybe tonight but i don't have ltsp setup at home.
14:22
i'll check what i can do.
14:24Kicer86 has joined #ltsp
14:25
<johnny>
X-Raimo, pulse is the future..
14:25
so. don't bother fighting it
14:28
<Lns>
johnny is borg...resistance to pa is futile
14:28
=p
14:32pmatulis has quit IRC
14:35
<X-Raimo>
alkisg: how to understand whatever LTSP server uses pulse to play audio or not?
14:38
<Lns>
X-Raimo: SOUND_DAEMON variable in lts.conf should tell you if it's there, not sure what default is in debian
14:39avena has joined #ltsp
14:42kusznir has quit IRC
14:44etyack has quit IRC
14:44cliebow has quit IRC
14:50
<Gadi>
X-Raimo: you don't need to muck with the chroot - it should have everything you need. If you are using KDE, you should set the server sessions to use pulse
14:50
(or esd - as pulse is backwards compatible with esd)
14:51
alsa will most likely work, as well
14:51
if it is properly set to use a pulse output
14:54
<X-Raimo>
Gadi: ok. Am I right: I need to make pulse as my default sound daemon at server. And in my KDE3 settings
14:55
<Gadi>
default sound output or system (the *daemon* actually runs on the thin client)
14:55kusznir has joined #ltsp
14:55
<Gadi>
but, in general, yes
14:56
<X-Raimo>
Gadi: I will try as soon as possible (in next 15-20 minutes)
14:56
<Gadi>
when you log in through LDM, an environment variable is set for the session that tells it which daemon to connect to
14:56
(set to the IP of the thin client, of course)
14:57
if sound is coming out the server, it is most likely because your session talks to artsd (default for kde) which talks to the server erroneously
14:57
you may also want to check your settings for ~/.asoundrc
14:59
(which configures alsa -> this file can be used to redirect alsa to pulse)
14:59
<X-Raimo>
Gadi: can you give a sample of ~/.asoundrc please?
15:00
<Gadi>
actually, it looks like on debian, you may be better off with /etc/asound.conf as per: http://www.mail-archive.com/ltsp-discuss@lists.sourceforge.net/msg35007.html
15:02mordocai has joined #ltsp
15:04hersonls has quit IRC
15:11litlebuda has quit IRC
15:20shawnp0wers has quit IRC
15:21Kicer86 has quit IRC
15:25mordocai has quit IRC
15:29Gadi has left #ltsp
15:30mordocai has joined #ltsp
15:31alkisg has quit IRC
15:35dlandry has joined #ltsp
15:36
<dlandry>
Anyone available for a localapps question?
15:37
I have localapps set up on karmic. I am able to run xterm, evince, abiword. mtr as localapps but not firefox.
15:37
When I do ltsp-localapps firefox I get nothing.
15:38
<johnny>
did you install firefox in the chroot?
15:38
and then rebuild your image?
15:38
hmm..
15:38
dlandry, did you also try typing firefox in the local xterm?
15:39
that's what i did to test mine..
15:39
<dlandry>
yes. firefox is in the chroot and I tried it from the local xterm.
15:39
...and I rebuilt the image
15:39
<johnny>
did it work from local xterm?
15:40
it takes like 10 minutes for it to come up for me..
15:40
<dlandry>
No, but all the other apps I mentioned did
15:40
<johnny>
no console output either?
15:40
sorry.. not 10 minutes.. more like 3
15:41
<dlandry>
No - just empty space
15:41
<johnny>
ah.. no idea
15:42avena_ has joined #ltsp
16:06mgariepy has quit IRC
16:18shawnp0wers has joined #ltsp
16:24mgariepy has joined #ltsp
16:27avena_ has quit IRC
16:28avena_ has joined #ltsp
16:29Barbosa_ has joined #ltsp
16:32Barbosa has quit IRC
16:32highvoltage has quit IRC
16:34highvoltage has joined #ltsp
16:41
<AndyGraybeal>
my video is really glitchy after i changed a setting in lts.conf
16:41
:(
16:41
<Appiah>
change it back!
16:41
<AndyGraybeal>
which i backed up my image :(
16:42
can i use # symbols to remark?
16:42
in lts.conf?
16:42
<X-Raimo>
AndyGraybeal: yes
16:44
<AndyGraybeal>
thanks X-Raimo
16:44
<X-Raimo>
AndyGraybeal: you're welcome
16:45Selveste1 has joined #ltsp
16:45
<X-Raimo>
I have problems with pulseaudio http://paste.org.ru/?dyd9jg Both at server and clients as well
16:47
<AndyGraybeal>
do you guys backup images before you do an ltsp-update-image ?
16:47
cause i think that's wha i should have done :(
16:48Barbosa_ has quit IRC
16:49bobby_C has quit IRC
16:51X-Raimo has left #ltsp
17:13Selveste1 has quit IRC
17:17jammcq has quit IRC
17:17Selveste1 has joined #ltsp
17:18shawnp0wers has quit IRC
17:19avena_ has quit IRC
17:22
<johnny>
AndyGraybeal, i backup my chroot.. not the image
17:36CAN-o-SPAM has quit IRC
17:47shawnp0wers has joined #ltsp
17:53
<AndyGraybeal>
johnny: okay nice
17:57mikkel has quit IRC
17:57shawnp0wers has quit IRC
17:59lucascoala has joined #ltsp
17:59litlebuda has joined #ltsp
18:03shawnp0wers has joined #ltsp
18:11japerry has joined #ltsp
18:14
<AndyGraybeal>
does anyone knwo if LDM_AUTOLOGIN=true works in ubuntu 8.10?
18:16
<abeehc>
i know it works in 9.04
18:16
<AndyGraybeal>
thank you abeehc
18:17
<abeehc>
you want LDM_USERNAME and LDM_PASSWORD defined as well of course
18:19
<AndyGraybeal>
yea, how do you define your machiens with MAC hardware addresses or IP addresses?
18:19
and is the password in cleartext?
18:21
<map7>
I'm getting the error: 'mount: according to mtab, aufs is already mounted on /' when trying to boot my thin client can anyone help me?
18:21
I'm using Ubuntu 9.10 64bit on the server and have created a 32bit client image
18:25
<AndyGraybeal>
abeehc: when i try to autologin in 8.10, my video gets all glitchy on the client.
18:35
<abeehc>
i do it with mac addresses, yes plaintext as far as the lts.conf file goes
18:36
but normal login works alright?
18:36
make sure you got the right password set and the like
18:36
i would run into something similar when my likewise auth was busted and wouldn't accept the password
18:36
the autologin happens so quick you can't tell what's going on from the client as i recall
18:36steve__ has joined #ltsp
18:37
<steve__>
can anyone help me with ltsp-cluster?
18:40
<AndyGraybeal>
abeehc: yea, normal login works fine, plus it actually looks like it logs in.. but the screen goes to hell.. like completely glitched out... like the display adapter got fried.
18:41staffencasa has quit IRC
18:42
<abeehc>
i'm not sure about that.. you shoudl enable ssh and look around the clients' /var/log
18:42
<AndyGraybeal>
aah okay
18:43
abeehc: i'm sorry, but i don't know how to do that -- install openssh server locally onto that machine?
18:43
can i do a ctrl-alt-f2 or something to get to a console?
18:44
<abeehc>
you can
18:44
but i imagine X might be looping and it'll mess around with you
18:44
maybe not
18:44
<AndyGraybeal>
okay, i don't know eitherway.
18:44
tell me about this ssh you were talking about please.
18:45
steve__: i asked that question before and i think i was referred to sbalneav but i could be wrong.
18:45
<mgariepy>
AndyGraybeal, does your autologin accounts are parts of the groups audio/video/fuse ?
18:45
<abeehc>
AndyGraybeal: have you unlocked the client root account?
18:46
<AndyGraybeal>
mgariepy: hmm audio/video... but yes fuse ... i didn't realize there was an audio/video group
18:46
abeehc: i don't think i've unlocked the client root account, no -- i'm sorry this is new territory for me.
18:46
<abeehc>
no worries
18:47
https://help.ubuntu.com/community/UbuntuLTSP/UnlockChrootRootAcct
18:47
pretty much gotta do that before ssh i think
18:47
<AndyGraybeal>
k thank yuo.
18:47
<mgariepy>
AndyGraybeal, does the autologin account belong to the same groups then your regular account ?
18:48
<abeehc>
when I asked if normal login worked I assumed you'd be using the same account credentials as you have configured for autologin
18:49
<AndyGraybeal>
autologin is for existing users, not for it's own account ... i'm sorry -- i have them in fuse.
18:50
<steve__>
thanks andy, my question concerns load balancing rdp clients... if anyone knows anything that would be great
18:50
<mgariepy>
anyone here having problem with vt + ltps5.2 ?
18:50
<AndyGraybeal>
mgariepy: your question is a little confusign to me - my account has admin and other groups. these accounts have 'workers' and 'fuse'
18:50
<mgariepy>
i would like a tester for my patch.
18:51
AndyGraybeal, what are the other groups ?
18:51
<AndyGraybeal>
why are you asking? you think i should try to autologin as myself to see what happens?
18:52
<mgariepy>
AndyGraybeal, if you use the same account like abeehc like said, for autologin and for regular login do you have the same issue ?
18:52
<AndyGraybeal>
aah no no.. it is the same login
18:53
so say - 'mariposa' is the login i would nromally use -- and 'mariposa' is the one i'm usign to autologin - it's the same -- i don't know why we got off on this tangent, i'm sorry -- it's because i don't fully understand what your saying.
18:53
thye can login just fine when i type the passwords.
18:54
<abeehc>
ok right on
18:55
and in lts.conf you have the mac address or ip defined, and then the three autologin lines?
18:55
<AndyGraybeal>
i swear the passwords are okay -- and it looks like they DO login... automatically - but i can't really be sure because the screen has gone mostly to static... like i've busted my video card.
18:56
abeehc: i've tried both ip and mac addresses, plus #LDM_AUTOLOGIN=true
18:56
<abeehc>
ok, try to get a console via ctrl+alt+f1
18:56
<AndyGraybeal>
er i didn't mean to hit enter
18:56
<abeehc>
once the video is messing up
18:56
<AndyGraybeal>
pls LDM_AUTOLOGIN=true in either [default] or under the address (just for kicks)
18:56
okay thank you abeehc
18:57
<abeehc>
i don't think you ever want that under default
18:58
sorry I'm pretty sure you need to first unlock the root account for ctrl+alt+f1 console to be usefull
18:59
<mgariepy>
or do a SCREEN_02=shell ?
18:59
<abeehc>
ah yeah that is likely even easier
18:59
i've never tried that though
19:00
<mgariepy>
sbalneav, are you around ?
19:00
<AndyGraybeal>
abeehc: thank you guys, this is great information.
19:02
<mgariepy>
AndyGraybeal, you are welcome ;)
19:08
<AndyGraybeal>
okay this is off-topic, how do you get a group of people to work on the same documents; for instance when someone creates a document it doesn't assign write permissions for the group by default, just group read.
19:09
<johnny>
umask is one way..
19:09
editing the umask..
19:09
but depending on the document..
19:09
<AndyGraybeal>
okay, i've never heard of that
19:09
<johnny>
it's easier to just use a webservice
19:09
like etherpad
19:09
<AndyGraybeal>
we are using openoffice
19:09
<johnny>
perhaps you might want to just use google docs
19:09
or something similiar
19:09
<AndyGraybeal>
we also use scribus and other things
19:10
<johnny>
well.. if you plan on tracking changes.. you might want to store them in a version control system
19:10
<AndyGraybeal>
yea, we're not there yet - but thank you for that bit of informaiton because it is important.
19:12dlandry has quit IRC
19:16litlebuda has quit IRC
19:18slidesinger has joined #ltsp
19:22Faithful has quit IRC
19:24Lns has quit IRC
19:26F-GT has joined #ltsp
19:31
<mgariepy>
if someone can review this to make sure i don't break anything :) http://bazaar.launchpad.net/~ltsp-upstream/ltsp/ltsp-trunk/revision/1705
19:38mordocai has quit IRC
19:41nubae has joined #ltsp
20:06lucascoala has quit IRC
20:21shawnp0wers has quit IRC
21:25steve__ has quit IRC
21:28daya has joined #ltsp
21:31F-GT has quit IRC
21:48F-GT has joined #ltsp
22:04mgariepy has quit IRC
22:04Faithful has joined #ltsp
22:13F-GT has quit IRC
22:26map7 has quit IRC
22:30F-GT has joined #ltsp
22:48alkisg has joined #ltsp
22:54ogra_cmpc has quit IRC
22:56ogra_cmpc has joined #ltsp