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


Channel log from 31 August 2009   (all times are UTC)

00:10
<alkisg>
Good morning
00:34
<elias_a>
It is a good morning, isn't it? :)
00:35map7 is now known as map7|away
00:35
<alkisg>
Tommorow the vacations end, and today is raining.... so :-/
00:43
<elias_a>
Well - nice day for jogging... keep it positive! :D
01:13vagrantc has quit IRC
01:17alkisg has quit IRC
01:18
<ftherese1>
thank you nubae... I am starting over today, hopefully I'll have more luck!
01:19alkisg has joined #ltsp
01:27jkenney_ has joined #ltsp
01:27
<jkenney_>
hey
01:28
I need to create a default profile in ltsp
01:28
that is locked down
01:28
something i can place in skel
01:28
so when I add user
01:28
its locked
01:28
can anyone help me
01:28artista-frustrad has quit IRC
01:28
<jkenney_>
i really would like to get this finished and have been banging my head against the wall
01:32
<Appiah>
what's the problem?
01:44
<jkenney_>
ko
01:44
ok
01:44
i need to create a locked down profile
01:45
and not sure how
01:48artista-frustrad has joined #ltsp
02:25Selveste1 has joined #ltsp
02:29Selveste1_ has joined #ltsp
02:41Selveste1 has quit IRC
02:49alkisg has quit IRC
03:05cyberorg has joined #ltsp
03:12nubae_ has joined #ltsp
03:14
<nubae_>
how can one calculate startup time of a thin client based on its image size? or can one do that at all?
03:17alkisg has joined #ltsp
03:18
<nubae_>
morning alkisg
03:18
<alkisg>
Hi nubae
03:18
<nubae_>
just asked a question before u came in, maybe u have an answer:
03:18
how can one calculate startup time of a thin client based on its image size? or can one do that at all?
03:19
<alkisg>
Nope. It doesn't depend on the image size at all
03:19
<nubae_>
its just someone is trying to convince me that if I had a room of 25 computers starting up at once on a 100mb network, it would take more than 5 mins for them all to start up
03:19
<alkisg>
It's like saying that if you /dev/sda1 is 100 Gb, your normal os startup time will be greater than one with 1 Gb os
03:19
<nubae_>
yeah, thats what I thought... theoretically u cant make such calculations
03:20
<alkisg>
One way to do some calculations would be to measure the actually transfered bytes per client
03:20
One here did that, I think he said that it needed ~20-30 Mb per client till the client boots
03:21
So on a 100mbps network, that's 2-3 seconds per client
03:21
So... 25 computers x 3 seconds => here's a lower limit to your boot time: 75 secs.
03:22
But if the server <=> switch link is gigabit, that is *not* the case
03:22
Then the lower theoretical limit is still 3 seconds or so.
03:22
<nubae_>
ok, now that makes sense
03:25
<alkisg>
Even if the network is 100mbps, it's really easy to add 2 more nics to the server and get the theoretical 75 seconds down to 25 seconds => not a bottleneck anymore.
03:25
That would also provide fault tolerance & more speed after the initial boot
03:29try2free has joined #ltsp
03:31Selveste1_ has quit IRC
03:38
<nubae_>
yep... makes sense
03:38
theoretically, if u are adding gigabit nics, bottleneck could never really be the network
03:40try2free has quit IRC
03:46
<alkisg>
For boot up times, yeah, but for later on even gigabit is slow
03:47
E.g. HD playback = 500mbps => x 25 clients => 12.5 Gbps :)
03:47
You'd need 12.5 gigabit cards on the server to serve them :P
04:10alkisg has quit IRC
04:10ftherese has joined #ltsp
04:11ftherese1 has quit IRC
04:22try2free has joined #ltsp
04:24intelliant has joined #ltsp
04:37try2free has left #ltsp
04:41cyberorg has quit IRC
04:45garymc has joined #ltsp
04:49Selveste1_ has joined #ltsp
04:59alkisg has joined #ltsp
05:01cyberorg has joined #ltsp
05:21Selveste1_ has quit IRC
05:23alkisg has quit IRC
05:23alkisg has joined #ltsp
05:25mraudsepp has joined #ltsp
05:31Selveste1_ has joined #ltsp
05:40autoditac has joined #ltsp
05:41
<autoditac>
hi. does anyone know of any problems with subpixel hinting on debian lenny ltsp clients?
05:41
mine looks horrible ...
05:43autoditac has quit IRC
05:43
<alkisg>
Fast one :)
05:45artista-frustrad has quit IRC
05:48artista-frustrad has joined #ltsp
05:49panthera has quit IRC
05:49panthera has joined #ltsp
06:02Selveste1_ has quit IRC
06:11alkisg has quit IRC
06:15alkisg has joined #ltsp
06:24pmatulis has joined #ltsp
06:27alkisg1 has joined #ltsp
06:31sene has joined #ltsp
06:42alkisg has quit IRC
06:45autoditac has joined #ltsp
06:51Blinny has joined #ltsp
06:52autoditac has left #ltsp
06:55autoditac has joined #ltsp
06:56
<autoditac>
alkisg1: other thin client problems :-(
06:56
any hints for smooth subpixel hining?
07:01bf has joined #ltsp
07:02ftherese has quit IRC
07:23alkisg1 is now known as alkisg
07:24
<alkisg>
autoditac: I've barely seen debian, I wouldn't know... Is that in all applications, or just openoffice?
07:50IBC_jkenney has quit IRC
07:50jkenney_ has quit IRC
07:51maginot has joined #ltsp
07:56nubae_ has quit IRC
08:03alkisg has quit IRC
08:05intelliant has quit IRC
08:15otavio has quit IRC
08:19lieven1 has joined #ltsp
08:24
<nubae>
the subpixel hinting should be the same on ubuntu and debian
08:24
does a normal client behave properly, ie... its just the thin client u notice being crappy?
08:26
<lieven1>
Hi, need some debugging tips on ltsp 5 on Jaunty... seems like XSERVER variable is not honoured / or . (some way after tweaking pxelinux using vesamenu, xserver on ltspclient wants to load fbhwdev module after nv driver and I can't seem to force using the proprietary nvidia driver).. can anyone hint me on this one?
08:32otavio has joined #ltsp
08:41
<nubae>
lieven1: u have X_CONFIGURE=True ?
08:42
<lieven1>
yes I do, relevant portion from lts.conf:
08:42
[default]
08:42
LDM_GUESTLOGIN = True
08:42
LOCALDEV=True
08:42
SOUND=True
08:42
TIMESERVER=ntp.d4t.lan
08:42
[DIMENSION521]
08:42
LDM_DIRECTX = True
08:42
CONFIGURE_X = True
08:42
X_CONF = /etc/X11/xorg.conf.e521
08:42
XSERVER = nvidia
08:42
[00:13:72:37:97:02]
08:42
LDM_USERNAME = lg3-1
08:42
LDM_PASSWORD = plg3-1
08:42
LIKE = DIMENSION521
08:42
SYSLOG_HOST = 192.168.100.100
08:42
ouch sorry for that....
08:43otavio has quit IRC
08:44
<nubae>
ah... u need to use mac address
08:45
for me, hostnames have never worked in lts.conf
08:46
<ogra>
because they are not supposed to :)
08:46
<lieven1>
I know, the DIMENSION521 section is a grouping section for similar workstations. I use it to refer using LIKE variable in actual host section...
08:47
<ogra>
right, for LIKE it should work
08:47
its likely that vesamenu needs a generic framebuffer loaded
08:47
<nubae>
shouldnt it be the other way round then though?
08:48
<ogra>
there are reasones why we dont include it in ubuntu :)
08:48
<nubae>
or usage is LIKE=hostname?
08:48
<ogra>
LIKE can take anything you define
08:49
i think its rather caused by vesamenu ... its a shitty implementation ...
08:50
(sorry for the strong words ... i'm not a fan of it :) )
08:50
<nubae>
yeah... u can try putting in the mac for dimension521 to confirm what ogra says I guess
08:51korcan has quit IRC
08:51
<lieven1>
from ref guide: The LIKE keyword allows you to define a general set of parameters under a unique identifier, and then assign individual workstations that set of parameters using the LIKE keyword.
08:51korcan has joined #ltsp
08:52
<lieven1>
so that should work... I also think it is related to vesamenu, .... cause since then, trouble has begun...
08:53
But still, I should find a way to find out why the nvidia driver doens(t want to load...
08:54
client boots but gets stuck in text mode
08:54
<ogra>
is it installed in your chroot ?
08:54
<lieven1>
it is installed, cause I can load the module when login into the clients text terminal
08:55
<ogra>
you could add it to /etc/modules in your chroot (and rebuild the image)
09:01otavio has joined #ltsp
09:01
<lieven1>
I think I could get it to work, if I can find out what jaunty's restricted driver manager does when installing the nvidia driver. I figured out it installs the package (nvidia-glx) - which I installed in the chroot (did need manual dkms build), creates an xorg.conf - which I created in the chroot and refed in lts.conf. But... I'm sure it does something else to.....
09:03
<ogra>
did you run ltsp-update-image ?
09:03
<lieven1>
yep
09:05
how (which script) is xserver startup initiated on the client? I could try playing with params once logged into text terminal.
09:05bf has left #ltsp
09:15
<_UsUrPeR_>
hey all. testing out Karmic, just updated and LTSP works again. I have a question about how localapps w/firefox is handled now.
09:16
I am getting an error that line 14 is broken: http://pastebin.com/m5b8e0a73
09:16
Previously, that was SOP to put that in the lts.conf in order to ensure that /etc/resolv.conf was writeable.
09:16
is that no longer needed?
09:17irule has joined #ltsp
09:18otavio has quit IRC
09:18otavio has joined #ltsp
09:23
<zamba>
when i log in on my thin client i get an error about HAL
09:24
ah, sabayon's the culprit
09:27Gadi has joined #ltsp
09:27_UsUrPeR_ has quit IRC
09:31_UsUrPeR_ has joined #ltsp
09:31alkisg has joined #ltsp
09:32cliebow has joined #ltsp
09:35
<lieven1>
Hi, I can't find the script in the ltsp-client chroot that starts X/ldm .... I must be overlooking something...
09:37
<ogra>
ltsp-client-core in /etc/init.d in the chroot
09:44
<rm-rf>
i'm running into something that i'm not sure if it is a feature, or a bug
09:44
if i login to ltsp, then log out, nobody else can login until the session is rebooted
09:44
is this normal behavior?
09:44irule has quit IRC
09:45jammcq has joined #ltsp
09:45
<jammcq>
good morning friends
09:47
<alkisg>
rm-rf: huh? you login to a terminal T1 as user U1, and when you logoff, a user U2 can't logon to a terminal T2?
09:47* ogra wonders what "the session is rebooted" might mean
09:48
<jammcq>
ogra: howdie
09:48
<ogra>
hey jammcq
09:48
<jammcq>
ogra: did you buy your plane ticket to Maine yet?
09:48
<ogra>
nope, still no word from management
09:48
<jammcq>
hmm, maybe it's time to shake that tree
09:48artista-frustrad has quit IRC
09:49
<ogra>
my prob is that my HW is restricted, i cant take it with me to the US
09:49
<jammcq>
oh really?
09:49
<ogra>
so i wouldnt be able to do image testing and fixes for RC
09:49
<jammcq>
wow
09:49
<ogra>
yeah, patent issues
09:49
that makes the decision difficult since i have to train someone else on al the stuff
09:49
<jammcq>
oh, that's right... you've crossed over
09:50
<ogra>
no idea what that patent is about though, but you cant import that specific board into the US atm
09:51artista-frustrad has joined #ltsp
09:51
<jammcq>
so it's from a non-us company?
09:51
<ogra>
(no idea if any customs person would know though)
09:51
the company is US based, bt the production is asia based
09:51
<jammcq>
heh
09:51
makes it even funnier
09:51
<ogra>
yeah
09:51
<cliebow>
GOOD GREIF
09:52
heh
09:52
i before e
09:52
<ogra>
american law *shrug*
09:53
<cliebow>
thars all we need..ogra at Guantanamo
09:53
<rm-rf>
alkisg: i pxe boot the client, login with my account, then log out. if another user tries to login, they get kicked back out to the login screen after the error "no response from server. restarting." appears. if i reboot the client, then the user can login
09:53
<ogra>
lol
09:53
cliebow, i'd love to see cuba, lots of ubuntu users there :)
09:53ajohnson has joined #ltsp
09:54
<alkisg>
rm-rf: (just so to understand this better): can they login from other clients? can the same user login twice from the same pc?
09:55
<rm-rf>
the same user can login twice from the same terminal
09:56
and my user can login after they've logged in, but if i log out and try logging in as the previous user, it fails
09:56
it's like my account messes up the machine somehow
09:58
<alkisg>
Ah, so other users can login/logout many times if you don't logon? Is the problem just for your user account?
09:59
<rm-rf>
i can login as many times as i want
09:59
but as soon as i login to a terminal, nobody else can login
09:59rjune_wrk has joined #ltsp
10:03
<alkisg>
rm-rf: after that happens, get a local terminal and try: ssh -l username server
10:03
...and see what ssh tells you there.
10:03
(change "username" with the user that experiences the problem; leave "server" as it is)
10:06
<rm-rf>
wtf
10:06
now it's working just fine
10:07
lemme reboot and try it again
10:07shawnp0wers has joined #ltsp
10:08
<rm-rf>
i don't know what happened, but something "broke loose" and now it's working
10:09staffencasa has joined #ltsp
10:11keygod has joined #ltsp
10:12bieb has joined #ltsp
10:29bobby_C has joined #ltsp
10:31Anon4210 has joined #ltsp
10:34autoditac has quit IRC
10:41otavio has quit IRC
10:44shawnp0wers has quit IRC
10:45cnc has joined #ltsp
10:48Anon4210 has quit IRC
10:54shawnp0wers has joined #ltsp
10:57mraudsepp has quit IRC
11:03cnc has quit IRC
11:04cnc has joined #ltsp
11:04
<cnc>
!Seen Gadi
11:04
<ltspbot>
cnc: Gadi was last seen in #ltsp 3 days, 0 hours, 1 minute, and 10 seconds ago: <Gadi> ah, movable type
11:07
<Gadi>
wow... I have been quiet lately...
11:24autoditac has joined #ltsp
11:25
<autoditac>
does anybody know of requirements for proper font anti-aliasing on the client side (ltsp5 - debian lenny)
11:25
mine looks horrible?
11:35autoditac has quit IRC
11:39Lns has joined #ltsp
11:39fotanus has joined #ltsp
11:42
<Lns>
Hello fellow LTSPers
11:44
<cliebow>
Lns!!!
11:44
<Lns>
hey cliebow =)
11:44
<cliebow>
8~)
11:45
<Lns>
How's it going?
11:46
<johnny>
gentlemen
11:46
<cliebow>
ok..my new job didnt come through..
11:47
cause they wouldny pay enough alone..and werent keen on letting me continue my other job
11:47* Lns greets johnny
11:47
<knipwim>
johnny: what do you think of this:
11:47
http://www.pastebin.org/13704
11:48
<Lns>
cliebow: :( you looking for work?
11:48
<johnny>
i think you're insane :) jk :)
11:48
<knipwim>
:)
11:48
<johnny>
knipwim, you can't use pastebin.* for patches
11:48
<cliebow>
Lns: well..trying to get out of this schlemmozle i am in now
11:48
<johnny>
it eats up the @@ signs
11:48
<knipwim>
ah
11:49
<johnny>
try pastie.org instead or dpaste
11:49
or anything else :)
11:49autoditac has joined #ltsp
11:49
<knipwim>
next time
11:49
<johnny>
i just use wgetpaste :)
11:49
<knipwim>
you know what i'm getting at
11:49
<johnny>
-h ? hmm.. that's a new one to me
11:50
knipwim, btw.. we try to make sure these files do not contain bashisms
11:50
<knipwim>
what's a bashism?
11:50keygod has quit IRC
11:50keygod_ has joined #ltsp
11:50
<johnny>
somethign that only works in the bash shell
11:50
actually.. more importantly.. they aren't specified in the POSIX spec
11:51
as other shells do similiar things.. it's just common for people to put things that only work in bash because they don't know better
11:51
as they have probably never used a shell other than bash by default
11:51
this is important for both quickstart and ltsp scripts
11:51vagrantc has joined #ltsp
11:51
<johnny>
since quickstart tries to stay agnostic internally anyways
11:52
<knipwim>
makes sense
11:52
<johnny>
i don't think i've personally used -h before
11:55
so, you want to remove the stage file symlink from inside sure.. that sounds fine
11:55
trivial, but fine
11:55
setup a launchpad account and fork ltsp-trunk
11:56
or host bzr yourself..whatever
11:56autoditac has quit IRC
11:56
<johnny>
it is custom in this project to prefix distro specific changes in the commit message, so for us Gentoo:
12:00cliebow has quit IRC
12:00cliebow has joined #ltsp
12:01
<knipwim>
is the -h the bashism you're getting at?
12:01alexqwesa has quit IRC
12:01
<johnny>
i don't know if it is, i was just mentioning it
12:01
as i wasn't directly familiar with that one
12:04
<knipwim>
it is mentioned in man test, as is the -L option
12:04Ryan52 has quit IRC
12:06keygod_ has quit IRC
12:06keygod has joined #ltsp
12:09
<johnny>
this is almost what i do ..
12:09
http://xkcd.com/627/
12:10
i wish that it had a step for "i don't have the program" and go directly to the google thing on the side
12:12shawnp0wers has quit IRC
12:13otavio_ has joined #ltsp
12:15keygod has quit IRC
12:15keygod has joined #ltsp
12:17alexqwesa has joined #ltsp
12:18otavio__ has joined #ltsp
12:20shawnp0wers has joined #ltsp
12:22Blinny has quit IRC
12:24
<knipwim>
johnny: should i pastebin/email or bugreport the patch?
12:24
<johnny>
i just told you what you should do
12:24
<knipwim>
as i don't have a commit account for ltsp on launchpad
12:25
<johnny>
use your own account and fork it
12:25
and then i'll merge your work
12:25spectra has joined #ltsp
12:25
<knipwim>
kk
12:26Blinny has joined #ltsp
12:28
<Blinny>
So I have several users who have trouble with flash audio. It's not client-specific, it's user-specific. I've rm -rf .macromedia but have yet to touch .mozilla -- any ideas?
12:28otavio___ has joined #ltsp
12:28keygod has quit IRC
12:28otavio__ has quit IRC
12:29
<Blinny>
I did a round of updates, updating the chroot, as well as going to the non-distro 64-bit libflashplayer so I'm sure it's connected to one of those. Sound used to work for these users. Normal .wavs and local sound work great.
12:31garymc has quit IRC
12:31Anon2943 has joined #ltsp
12:34otavio_ has quit IRC
12:35
<Lns>
Blinny: maybe some users tried installing flash before you had the global libflashplayer.so file installed? Try 'mv'ing their .mozilla (and .adobe?) or simply search for libflashplayer.so in their homedirs
12:36
<Blinny>
.adobe!
12:37
My users wouldn't know a libflashplayer.so from their own shoes.
12:38
Let me try removing .adobe
12:38
Thanks Lns - back in a flash.
12:38
heh heh
12:38
<lieven1>
hi, did anyone succeed in installing and running NVIDIA's proprietary drivers in the clients chroot (jaunty)?
12:38Blinny has quit IRC
12:40
<alkisg>
lieven1: I did that once, I don't remember if it was with jaunty alpha/beta or intrepid...
12:41Blinny has joined #ltsp
12:41
<Blinny>
Lns: No love removing both .adobe and .mozilla -- still no sound.
12:41
<Lns>
Blinny: is this on *all* flash sites?
12:41
<Blinny>
Is there a user-specific setting for pulse/alsa/whatever?
12:41
Lns: All I've tried, yes.
12:41
<lieven1>
alkisg: do you rember using ubuntu's dkms mechanism for that?
12:41
<Lns>
shouldn't be, if you're getting sound in other apps
12:41
<Blinny>
(the major ones.. cnn, msnbc, redsox)
12:42
<Lns>
Blinny: you get video though?
12:42
<Blinny>
Lns: Yup.
12:42
And like I said, normal system sounds play fine. Login sound, .wavs, .mp3 and .m3us
12:42
<Gadi>
Blinny: check the user's .asoundrc*
12:42
<Lns>
Blinny: have you done about:config and verified your browser is using the same libflashplayer.so version that you thought you installed?
12:42
Gadi: !
12:43
<Gadi>
hidey ho
12:43
<vmlintu>
lieven1: I've got the nvidia module working by first installing the kernel+dkms+nvidia-source to a different box that compiled the module and then copied the module to the image
12:43
<alkisg>
lieven1: yeah, both with the nvidia proprietary drivers & with the nouveau drivers dkms is needed
12:43
<Lns>
Gadi: a bit Cab Calloway today are we?
12:43
<alkisg>
I did it directly on the chroot, with no different pc or anything
12:43
<Gadi>
hehe
12:44
<Blinny>
Hrm. .asoundrc references the skel account's .asoundrc.asoundconf
12:45
<lieven1>
I did to, manual dkms build in the chroot to use the clients kernel. All seems fine, except, xorg keeps loading the nv driver instead of nvidia's...
12:45
<alkisg>
Put XSERVER=nvidia in lts.conf
12:45
<Blinny>
One sec, lemme give another try. Be right back.
12:45Blinny has quit IRC
12:46
<lieven1>
I did that... problem stays, still nv driver loaded
12:47
<johnny>
guess you should drop to the console and try to manually load the module
12:47Blinny has joined #ltsp
12:47
<johnny>
make sure it is loaded at all..
12:47
usually xorg prefers the nvidia over nv
12:47
<Blinny>
Lns Gadi: .asoundrc was it
12:48
<Gadi>
cool
12:48
<Lns>
Blinny: good =)
12:48
<Blinny>
They were sourcing the skel account's .asoundrc.asoundconf file. When I switched it to their own home dir's .asoundrc.asoundconf it started working.
12:48
<lieven1>
johnny: modprobe nvidia works, but then... any way to manually start x (ldm)
12:48
<johnny>
sure..
12:49
<Blinny>
I guess I also removed the skel account during the same testing period. Thanks guys, appreciate it!
12:49
<alkisg>
lieven1: ah right you also need a MODULE=xx line in lts.conf, let me find it...
12:49
<johnny>
i forget the correct options.. something like ldm vt7 :0
12:50
it's weird.. but i don't have ldm here to look at the manpage to remind myself
12:50otavio___ has quit IRC
12:51
<alkisg>
lieven1: if my notes are correct, it's: MODULE_01=nvidia X4_MODULE_01=glx
12:51
<johnny>
are you sure you have to specify glx ?
12:51
that would be weird
12:51
<alkisg>
nOPE
12:51
Nope
12:51
<johnny>
i dont' think you do
12:51
<lieven1>
I'm trying....
12:53otavio_ has joined #ltsp
12:56
<lieven1>
hmmmm... very strange - nvidia module is not loaded after rebooting client
12:57
<alkisg>
lieven1: so what's the generated xorg.conf like on the client?
12:58
<vagrantc>
nvidia is a kernel module?
12:58otavio_ has quit IRC
12:59* Gadi would have guessed XSERVER=nvidia
12:59
<lieven1>
is there a way to check generated xorg.conf (except reading the xorg logs)?
12:59
<Gadi>
cat /var/run/ltsp-xorg.conf
12:59
<lieven1>
log says it's using nv driver
12:59
<Gadi>
iirc
13:00shawnp0wers has quit IRC
13:00
<lieven1>
no /var/run/ltsp-xorg.conf here....
13:00
<alkisg>
lieven1: so what do you have in lts.conf now? xserver and module_01 ? and that (with jaunty) didn't generate a xorg.conf?
13:02
<lieven1>
indeed: CONFIGURE_X = True; XSERVER = nvidia; MODULE_01=nvidia; X4_MODULE_01=glx
13:02panthera has quit IRC
13:02
<Gadi>
lieven1: do you have [default] ?
13:02panthera has joined #ltsp
13:02
<lieven1>
yes, those vars are in host section though
13:02
<alkisg>
try getltscfg -a on the client to see if they get there
13:03
<lieven1>
getltscfg -a: nothing
13:03
<alkisg>
Urm, then you got a problem with lts.conf not being sent to the client
13:03
<Gadi>
there's your problem: cat /etc/lts.conf
13:03otavio_ has joined #ltsp
13:04cliebow has quit IRC
13:04
<lieven1>
indeed, sample file in /etc/lts.conf
13:04cliebow has joined #ltsp
13:05
<alkisg>
Where's your normal lts.conf on the server? In /var/lib/tftpboot/ltsp/i386/lts.conf ?
13:05
<lieven1>
yes
13:06
<alkisg>
Maybe there's a problem in your dhcpd.conf, and it doesn't send the "filename" dhcp option when the client vendor identifier is "Linux ipconfig"
13:06
Can you upload the relevant settings in /etc/ltsp/dhcpd.conf ?
13:07* alkisg thinks we should tell ipconfig to use "PXEClient" :-/
13:07cnc has quit IRC
13:08
<ogra>
alkisg, pfft
13:08* ogra points to bzr
13:08
<ogra>
added:
13:08
client/initramfs/hooks/udhcp
13:08
client/initramfs/scripts/init-premount/
13:08
client/initramfs/scripts/init-premount/udhcp
13:08
<alkisg>
Yeah!!! :D
13:08
<Gadi>
some folks do use Etherboot, ya know
13:08
<ogra>
send flowers to stgraber :)
13:08
<Gadi>
:P
13:09otavio_ has quit IRC
13:09
<ogra>
Gadi, udhcp can handle that i think
13:09otavio_ has joined #ltsp
13:09
<alkisg>
I sure will. But we'll need a little testing/cleaning to make it ready for karmic
13:09
<ogra>
ipconfig is dead :)
13:09
<lieven1>
host lg3-1 {hardware ethernet 00:13:72:37:97:02;fixed-address lg3-1;filename "pxelinux.0"; }
13:09
<alkisg>
Gadi: etherboot is no problem
13:09
<Gadi>
right - but I wouldn't set the VCI to PXEClient
13:09
:)
13:10
lieven1: how does that even work?
13:10
you should have: filename "/ltsp/i386/pxelinux.0";
13:10
<alkisg>
Gadi: now with your patch that makes filename default to /ltsp/i386/pxelinux.0, it's ok. But without that, locating lts.conf by using filename *and* not telling ipconfig to use "PXEClient" was a little misleading...
13:10
<lieven1>
why shouldnt it?
13:11
<alkisg>
lieven1: the lts.conf location is guessed by looking the filename dhcp option
13:11
It's searched for in the same dir as pxelinux.0
13:11
<lieven1>
no, i moved pxelinux.0 and pxelinux.cfg to /var/lib/tftpboot/
13:11
<Gadi>
alkisg: but if you etherboot, ipconfig should say Etherboot, following ur model
13:11
<alkisg>
So move lts.conf there
13:11
<Gadi>
personally, I think ipconfig should say "LTSP-i486"
13:11
<alkisg>
Gadi: nah, that's only for the lts.conf filename...
13:11
<Gadi>
er, LTSP-i386"
13:12
my point is, ipconfig has nothing to do with PXE or Etherboot
13:12
<alkisg>
Sure. I'm talking about making dhcpd.conf easier for people to understand
13:12
<Gadi>
and if you are going to mess with the VCI, it should be done in a more meaningful way
13:13
right - setting it to PXEClient would make it more confusing
13:13
especially to Etherboot folk
13:13
:)
13:13
<alkisg>
Urm... the default dhcpd.conf is working by... chance!!! if substring( option vendor-class-identifier, 0, 9 ) = "PXEClient" { filename "/ltsp/i386/pxelinux.0"; } else { filename "/ltsp/i386/nbi.img"; }
13:14
<ogra>
lieven1, and where exactly does your lts.conf live ?
13:14
<alkisg>
When a PXE client boots, it gets "filename "/ltsp/i386/pxelinux.0"; " in the beginning, and filename "/ltsp/i386/nbi.img"; } later on with ipconfig
13:14otavio_ has quit IRC
13:14
<alkisg>
Is that *not* confusing?
13:14otavio_ has joined #ltsp
13:15
<Lns>
alkisg: well its all required for booting from the network, really iirc :)
13:15
isn't it mostly to do with the limitations of bootloaders, initramfs, etc?
13:16
<alkisg>
Not really...
13:16
<lieven1>
it lives in /var/lib/tftboot/ltsp/i386/lts.conf , i tried to copy to /var/lib/tftpboot (next to my pxelinux.0) but to no avail
13:16
<Gadi>
what is confusing to have a packet coming from an etherboot system over the wire identifying itself as a PXEclient
13:17
<alkisg>
Gadi: this is later than the initial booting. And, the newest etherboot version (=gpxe) *do* send "PXEClient" as the vci
13:17
<Gadi>
right - then why should we be forcing the VCI in the initramfs?
13:17
:)
13:18
<lieven1>
Problem is, I'm using tftpboot server also for network installs so I have /var/lib/tftpboot/[pxelinux.[0|cfg/]|install/|ltsp/]
13:18CAN-o-SPAM has joined #ltsp
13:19
<Gadi>
so, lieven1, with lts.conf next to pxelinux.cfg, you still get default lts.conf on client?
13:19
ah, I know why
13:19
<lieven1>
yes, indeed
13:19otavio has joined #ltsp
13:19
<Gadi>
add a forward slash before the pxelinux.0 in filename
13:20
<alkisg>
lieven1: put break=init in pxelinux.cfg/default, and cat /tmp/net-eth0.cfg
13:20
<Gadi>
and restart the dhcp server
13:20
<alkisg>
Ah right
13:20* alkisg had reported that a while ago :)
13:25
<lieven1>
IT WORKS! Stupid mistake! Now nvidia driver is loaded like it should! Thanks a lot guys....
13:26
I'll never forget the fwd slash before filename in dhcpd.conf anymore!
13:27
(though it is strange it even did a netboot. The forward slash causing the tftp-server to find the lts.conf now... )
13:28otavio has quit IRC
13:29
<ogra>
the code that downloads lts.conf reads the filename directive and cuts off the actual file
13:29
so in your case the path was empty ... adding the forward slash made it "/lts.conf"
13:30
<lieven1>
that explains everything... thank you!
13:31
<alkisg>
Still, "get lts.conf" should have worked - this is a bug: https://bugs.launchpad.net/ubuntu/+source/ltsp/+bug/238010
13:32
if filename=ltsp3.pxe with no path, tftpath becomes ltsp3.pxe instead of / or ./
13:32
So it was looking at /pxelinux.0/lts.conf :)
13:34otavio__ has joined #ltsp
13:34
<ogra>
damned, yeah, my bad coding skills ... bah
13:34* ogra is to blame for that one
13:34
<ogra>
but looks like scottie is already on it if i read the bug
13:35
<lieven1>
is there a way to leave to file in ltsp/i386/lts.conf, that is somewhat cleaner then leaving it in the root of my tftp server
13:35otavio_ has quit IRC
13:35shawnp0wers has joined #ltsp
13:36
<ogra>
you could hack up the ltsp_nbd script
13:36
usr/share/initramfs-tools/scripts/ltsp_nbd in your chroot
13:37
<lieven1>
yes but I mean a clean way, like passing it as a dhcp option or something
13:37
<ogra>
nope
13:37
tftpath=$(echo ${filename%/*}|tr -d '"'); (echo "get ${tftpath}/lts.conf"|tftp ${ROOTSERVER} >/dev/null 2>&1)
13:37
thats the current code
13:38
<rjune_wrk>
ogra: focus not on blame, but rather on solutions.
13:38
<Gadi>
(echo "get ${tftpath:-/ltsp/i386}/lts.conf"|/bin/tftp ${NBD_ROOT_SERVER} >/dev/null 2>&1)
13:38
that's the real current code
13:38
:)
13:39
<ogra>
right, i c/p'ed from the bug
13:39
<Gadi>
you could have an: unset filename before that line
13:39
<ogra>
some kind person fixed it already :)
13:39
<Gadi>
if you want it to always go to /ltsp/i386
13:40
er, unset tftpath
13:40
which, i always thought needed an extra "p"
13:40
;)
13:40
<lieven1>
:) thanks...
13:41
<ogra>
Gadi, trivial file transfer path :)
13:41otavio__ has quit IRC
13:41
<Gadi>
in lieven1's case - very trivial
13:41
:)
13:42
<alkisg>
lieven1: you could also symlink ltsp/i386/pxelinux.0 to /pxelinux.0...
13:42
<lieven1>
that's what I did for now... thx
13:45artista-frustrad has quit IRC
13:47artista-frustrad has joined #ltsp
13:47otavio_ has joined #ltsp
13:58
<vagrantc>
stgraber: we you thinking of an ldm uploader sooner than later?
13:58
stgraber: well, new upstream tarball, that is
13:59
<ogra>
vagrantc, alpha freeze ahead
13:59
A5 is on thu
14:00
<vagrantc>
ah, so probably not
14:01shawnp0wers has quit IRC
14:01* vagrantc is debating weather to upload a new upstream version to debian, as slow buildd's may stall migration to testing on the current version
14:02shawnp0wers has joined #ltsp
14:05|Ryan52 has joined #ltsp
14:06
<ogra>
vagrantc, yeah, i saw the bug
14:06
err, mail ...
14:07* vagrantc wishes there were virtual bugs for each of the buildd's.
14:07
<ogra>
heh, ia64 ... who cares anyway :)
14:08
<vagrantc>
well, all architectures are "supported" again
14:08
<ogra>
evil
14:08
and you still have the december freeze ?
14:08
<vagrantc>
ogra: doubtful
14:08
<ogra>
oh, really? i though there was such a big agreement
14:08
<vagrantc>
i figure with the architecture restrictions... who am i to limit the creativity of obscure architecture users
14:09
ogra: agreements made without widespread community agreement tend to have a difficult time moving forward
14:09
<ogra>
heh
14:26
<Lns>
vagrantc: hmm, wonder why =p
14:38
Anyone using VMWare guests on Linux hosts (especially ltsp servers) that have tips to reduce disk i/o? It seems to be a *major* bottleneck for me, causing horrible hangs/delays during some operations (especially during the first ~10min startup of a win2k3 guest).
14:39
<jammcq>
Lns: yeah, get faster/more disks
14:39
<Lns>
jammcq: =p
14:39
nice
14:39
<jammcq>
we've got a big setup with a SAN and vmware with an LTSP server as one of the guests, and it's working VERY well
14:39
<Lns>
well i guess that's a possibility actually, i could throw another drive and convert to raid 5 (from 1).. it's on a SAS channel
14:40
<jammcq>
you have just 1 pair of drives now?
14:40
<Lns>
yea
14:40
poor disks are constantly hammered =)
14:40
<jammcq>
well... yeah, it's gonna be slow, and adding a 3rd disk prolly won't improve it much
14:40
<vagrantc>
Lns: be sure to use noatime
14:40
<jammcq>
I typically go with Raid 10, with 12 or more drives
14:40
<Lns>
but the only real noticeable thing is when vmware is going
14:41
vagrantc: ty, will do that too
14:41
<vagrantc>
especially with raid-1
14:42
<Lns>
vagrantc: stability gotchas? won't noatime increase possibility of corruption during a crash?
14:42
<jammcq>
hmm, I don't think so
14:43
<ogra>
nah
14:43
<jammcq>
if anything, it might make it better
14:43
<Lns>
I can definitely see that vmware would be constantly writing small bits to the same .vmdk file, causing a bottleneck possibly
14:43|Ryan52 is now known as Ryan52
14:43
<jammcq>
'noatime' tells the kernel not to update the 'time last accessed' when a file is read
14:43
and vmware is doing mostly reading
14:44
lots and lots and lots of reading
14:44
<Lns>
how about converting to ext4? Thoughts on stability in a production env?
14:44
<Gadi>
wow - raid 10 with 12 drives? how often do you have to rebuild the array?
14:44
<jammcq>
never used ext4, so no opinion
14:44
Gadi: umm.... never
14:44
<Gadi>
is it a raid 1 over two stripes?
14:44
<vagrantc>
Lns: if you use mutt with mbox files, you *might* have issues seeing new mail. if you use popularity-contest, it won't be able to collect as much useful data.
14:44* Lns uses popcon for all servers :)
14:45
<jammcq>
Gadi: it's either 6 drives striped, mirrored against the other 6 drives striped
14:45
<Lns>
but other than that it's just students hammering away at ff/ooo
14:45
<vagrantc>
Lns: a middle-ground compromise is relatime
14:45
<Lns>
or win ts :(
14:45
<jammcq>
or, it's 6 mirrored pairs all striped together
14:45
<Lns>
vagrantc: ty for that
14:45
<vagrantc>
Lns: but i'd just put your vmware disk images on a partition with noatime, for sure.
14:45
there's no sane reason to need that for disk images
14:45
<Gadi>
jammcq: roght - so if one of the six in the stripes fails, you need to rebuild because the stripe is dead, no?
14:45
<jammcq>
Lns: how's the RAM ?
14:46
Gadi: yeah, but pop a new drive in, and it rebuilds in place
14:46
<Gadi>
right
14:46
guess you have spares on hand
14:46
<jammcq>
yep
14:46
<vagrantc>
mdadm raid10 is actually crazier than that.
14:46
you can have a raid10 array of 5 disks, for example.
14:46
<jammcq>
md lets you do all sorts of crazy variants
14:46
<Gadi>
you must have great access times
14:46
<jammcq>
Gadi: yeah, crazy fast
14:47
but now we're using SAN's. that's just a big box full of drives. I don't even know how they're all arranged, but it's wicked fast
14:48
<Lns>
jammcq: 8gb
14:49
<jammcq>
Lns: how much ram is allocated to vmware?
14:49
<Lns>
jammcq: 3gb
14:49
<jammcq>
and specifically, how much to that guest
14:49
<Lns>
same :)
14:49
<jammcq>
cuz swapping inside the guest wouldn't be good
14:49
<Lns>
hmm, would it be sane to just 'noatime' the vmdk file itself?
14:50
<jammcq>
can you do individual files?
14:50
<Lns>
chattr +A filename
14:50
http://tldp.org/LDP/solrhe/Securing-Optimizing-Linux-RH-Edition-v1.3/chap6sec73.html
14:51
<jammcq>
hmm
14:51
<Lns>
of course noatime would be good for chroot / client boots simultaneously i think
14:51
err
14:51Blinny has quit IRC
14:51
<Lns>
yeah..for the image probably
14:52
anyway, i'm off to observe a lab full, bbl :) thanks for the help all
14:53
<jammcq>
sbalneav: ping
14:54
<alkisg>
Lns: causing horrible hangs/delays during some operations (especially during the first ~10min startup of a win2k3 guest ==> that's usually the windows update service
14:54
Installing a local WSUS server helps a lot
14:55cliebow has quit IRC
14:56
<ogra>
vagrantc, lol, i wasnt aware lamont still does debian ia64
14:59* vagrantc is struggling to write a proper debian/copyright for ltsp-docs
15:03garymc has joined #ltsp
15:04nubae has left #ltsp
15:04pmatulis has quit IRC
15:09garymc has quit IRC
15:23lieven1 has left #ltsp
15:25* vagrantc just changed some of the title information on ltsp-docs ...
15:25
<vagrantc>
hopefully nobody minds
15:26
this clause also kind of worries me: "If you modify the software in any way, identify your software as a variant of LTSP."
15:26bieb has left #ltsp
15:27
<vagrantc>
do i need to actually patch the xml file and packaging information if i make debian-specific patches that don't make sense upstream?
15:29maginot has quit IRC
15:30wimiko has joined #ltsp
15:31warren has joined #ltsp
15:31
<warren>
Somebody with Ubuntu or Debian can check something for me?
15:32
<jammcq>
hey warren, what do you need looked up?
15:34
<warren>
jammcq: /usr/lib/pulse-*/modules/module-detect.so
15:34
jammcq: does this exist?
15:35
<Gadi>
it does here
15:35* Gadi waves to warren
15:35
<warren>
Gadi: versioned directory too?
15:35
<_UsUrPeR_>
warren: I am seeing /usr/lib/pulse-0.9/module-detect.so
15:35
<jammcq>
umm
15:35
<_UsUrPeR_>
no modules dir
15:35
<Gadi>
/usr/lib/pulse-0.9/modules/module-detect.so
15:35
<_UsUrPeR_>
that's in 9.04
15:35
<warren>
Gadi: I'm still trying to figure out a solution to the module-detect vs. module-udev-detect problem
15:36
<_UsUrPeR_>
woop, nm. modules dir was glazed over
15:36
<vagrantc>
warren: i've got /usr/lib/pulse-0.9.15/module-detect.so on debian unstable.
15:36
<warren>
I guess looking for the file isn't a good solution.
15:36
/usr/lib64/pulse-0.9.16/modules/module-udev-detect.so
15:36
<ogra>
/usr/lib/pulse-0.9.16/modules/module-detect.so
15:36
<warren>
/usr/lib/pulse-0.9.16/modules/module-udev-detect.so
15:36
<ogra>
on karmic
15:36
<vagrantc>
warren: don't have module-udev-detect
15:36
warren: with 0.9.15
15:37
<warren>
module-udev-detect is pulesaudio-0.9.16+
15:37
module-detect is gone
15:37
<ogra>
/usr/lib/pulse-0.9.16/modules/module-udev-detect.so
15:37
i have both
15:37
<vagrantc>
warren: sure, so if it's present, use module-udev-detect, if it's not present, use module-detect ?
15:37
<ogra>
but not .16+
15:37
<warren>
OK, how should we make it pass the right option?
15:37
<Gadi>
vagrantc++
15:37
<warren>
vagrantc: detect present is not trivial becuase they are in different locations in different distributions
15:37
<ogra>
yeah
15:38
<vagrantc>
warren: not exceedingly different, from anything i've heard thus far...
15:38
<warren>
[warren@newcaprica ltsp-trunk]$ pulseaudio --version
15:38
<Gadi>
if $(find ...)
15:38
<warren>
pulseaudio 0.9.16-test6
15:38
how do we use that?
15:38skeemer has joined #ltsp
15:38
<ogra>
split off pulseaudio 0.9.
15:39
if the remaining stuff is bigger than 16 use it
15:40
<skeemer>
hello, got hopefully an easy question. i'm running ubuntu jaunty. if i add a printer on the server what do i have to do to make it available to the clients.
15:42
<warren>
ogra: perhaps if 0.9 isn't there, assume it is some future version
15:43
<Gadi>
testme=$(pulseaudio --version|sed -e 's/pulseaudio\s0\.9\.\([0-9]\+\).*$/\1/')
15:43
<warren>
how does dash handle arithemetic?
15:43
<Gadi>
if [ $testme -gt 15 ]; then ...
15:44
try that
15:44
<ogra>
dash handles arithmetic as posix defines it
15:45
<Gadi>
a little dash of ALU a little dash of registers...
15:45
:)
15:47
<alkisg>
skeemer: nothing, as the programs that the users run actually run on the server
15:48
<skeemer>
alkisg: that's what i thought, but they don't show up when they run openoffice
15:48* vagrantc would use a case statement
15:49
<ogra>
yeah
15:49
its faster
15:49
<skeemer>
alkisg: would it matter if i added via the cups interface and not in gnome?
15:50
<alkisg>
Ah, it wasn't autodetected? Maybe you added it only for the current user?
15:50* alkisg only used auto-detected printers so far... they work fine with openoffice and all.
15:50
<skeemer>
alkisg: i added it via http://localhost:631
15:50
it's a network printer
15:51nubae has joined #ltsp
15:52
<nubae>
hmmm... Quassel is a pretty decent irc client
15:52
xchat is probably more mature, but this isnt bad
15:53
<alkisg>
Urm what about empathy? :D
15:54
<warren>
vagrantc: what would that look like as a case statement?
15:54Q-FUNK has joined #ltsp
15:54
<ltsppbot>
"vagrantc" pasted "pulseaudio versioning with case" (7 lines) at http://pastebot.ltsp.org/492
15:54
<Q-FUNK>
re
15:54
<vagrantc>
warren: something like that ^^
15:55
<Q-FUNK>
anybody with GX2 hardware who wants to test our latest and greatest xf86-video-geode ? ;)
15:55
<vagrantc>
Q-FUNK: always the same with you :P
15:55
<Q-FUNK>
vagrantc: I have a one-track mind :)
15:56
<warren>
Q-FUNK: I don't have GX2 anymore
15:56
<Q-FUNK>
oh
15:56
<warren>
Q-FUNK: i'm sure gadi still supports customers using it
15:56
<Q-FUNK>
Gadi: hello dear ;)
15:56
<Gadi>
yes, dear
15:56
<Q-FUNK>
warren: yeah, he probably does
15:56
<Gadi>
:P
15:56
<Q-FUNK>
Gadi: any GX2 hardware on hand?
15:57
<Gadi>
plenty - just a lack of test environments
15:57
<Q-FUNK>
arf
15:57
<Gadi>
but, maybe I will stick one in my bag and test at home
15:57CAN-o-SPAM has quit IRC
15:58
<Gadi>
I tried to do a liveCD the other day, but I needed to use a really old GX2 guy with a bios for usb cdrom and then it was painfully slow
15:58
<Q-FUNK>
Gadi: GX2 support has been broken for a bit, but one X core dev just found a stupid typo in one of our commits. I pushed ubuntu packages to my PPA and need testers to confirm whether it finally fixes it.
15:58
<Gadi>
plus, ubuntu doesn't blacklist gxfb
15:58
<Q-FUNK>
doesn't it?
15:58
<Gadi>
which usually cases issues
15:58
I don't think so
15:58
<Q-FUNK>
at least, it does here
15:58
<Gadi>
I think it still only blacklists gx1fb
15:58
oh yeah?
15:58
oh ok
15:59
well, the USB CDROM approach was not the way to go ;)
15:59
I need to PXE boot it at home
15:59
<warren>
vagrantc: I'll commit something based on that case statement
16:01
<vagrantc>
case is often the easiest way to do pattern matching in shell
16:01
and flexible
16:01
<ogra>
and fast
16:02
<nubae>
Q-FUNK: hi!
16:03
<Q-FUNK>
nubae: hi!
16:10
<warren>
vagrantc: what is the best way to isolate the numbers in shell?
16:10
vagrantc: of pulseaudio --version
16:11
<vagrantc>
warren: could do a case in a for loop... but that's kind of ugly.
16:11
<warren>
[warren@newcaprica devel]$ pulseaudio --version |awk {'print $2'}
16:11
?
16:12
<vagrantc>
for v in $(pulseaudio --version) ; do case $v in ... pulseaudio ) ;; *) pulse_version=$v ;; esac ; done
16:12
yeah, awk is more readable.
16:12
<warren>
I'm using it
16:12
<vagrantc>
external binary ... but i'd prefer the readability
16:12
hopefully pulseaudio never gets fancy with the output of --version
16:13
<Q-FUNK>
I'm getting more and more itred of pulseaudio. it keeps on trying to reinvent the wheel and it makes a worse attempt at it, at every release.
16:16ftherese has joined #ltsp
16:17
<ftherese>
I am using ltsp-localapps with xubuntu jaunty but for some reason the menu does not launch the localapps
16:18
I can launch a localapps xterm
16:18
and I can launch firefox locally
16:19
as a localapps that is
16:19
but I have to do it command line style
16:19
<Gadi>
do you have LOCAL_APPS_MENU=True
16:19
<ftherese>
the menus do not update, despite my having followed the instructions to modify the lts.conf file
16:20
it has clearly modified my user apps list
16:20
for example, i enabled openoffice.org-writer in the LOCAL_APPS_MENU_ITEMS
16:21
along with calc an gimp
16:21
<vagrantc>
is that the name of the .desktop file?
16:21
<ftherese>
yup
16:21
<vagrantc>
and is openoffice.org-writer installed in the chroot?
16:21
<ftherese>
yes
16:21
they all show up in the menu
16:21
but when I run them it runs the servers copy... strange
16:21
<warren>
I'm tagging ltsp-trunk later today.
16:22
<ftherese>
only the ones I selected through the lts.conf file MENU_ITEMS show up in the applications menu
16:23
but when I run them, the server's copy runs
16:23pmatulis has joined #ltsp
16:23
<ftherese>
could this be a problem with xubuntu/xfce4 menu working differently?
16:23
<vagrantc>
ftherese: you've rebooted the thin client after changing the lts.conf options?
16:23
<ftherese>
yes...
16:24
<vagrantc>
ftherese: easiest way to test that is to try gnome...
16:24
<ftherese>
but I have to install that then... what a pain
16:24
and it will slow everything down
16:24
I am already working with slow equipment
16:24alumno01 has joined #ltsp
16:25
<alumno01>
Hi, when I print with open office writer the upper margin configuration doesnt take effect and the document prints with upper margin 0
16:25
<vagrantc>
ftherese: well, no sense beating your head against a wall when something is testable.
16:25
<ftherese>
is there anything else I can try before installing gnome?
16:25
<vagrantc>
ftherese: of course it's going to be more work to get an initial setup.
16:25
<ftherese>
if not... I will just do that
16:26
<vagrantc>
ftherese: it's the simplest way to test if it's just broken on your server, or broken in xubuntu
16:26
<ftherese>
okie dokie
16:27
do I need to install gnome on both the server and /opt/ltsp/i386 ?
16:28
probably not
16:28
never mind
16:28
just on the server...
16:29
168MB... 1hour away
16:29
<alumno01>
Hi, when I print with open office writer the upper margin configuration doesnt take effect and the document prints with upper margin 0
16:30
<ftherese>
alumno01: does your issue relate to ltsp?
16:30
<alumno01>
in some users it prints ok
16:30
but in others the config just doesnt take effect
16:30
all users have the same permissions
16:31
<ftherese>
alumno01: sorry... that is beyond me...
16:31
<Lns>
alkisg: thank you for that!! I want to say that I *thought* it might be windows update, but wasn't totally sure..i'm going to do the noatime, as well as disabling windows update service in the VMs. :)
16:32
<vagrantc>
stgraber, nubae: pretty impressively simple LTSP_FATCLIENT support! :)
16:32
<alkisg>
Lns: another "common suspect" is the antivirus scanning & update
16:32
<Lns>
alkisg: yeah..stupid windows! I thought i was getting away from it by doing ltsp ;)
16:33
<alkisg>
If you have a windows server around, though, do install a wsus server, it'll take much of the heat away
16:33
<nubae>
its all strgraber's work... think he just based it on the packages I install/dont install and guess printer/sound/local device support
16:35
<Lns>
alkisg: yeah - that's another consultant's job... /me wonders if debuntu has a wsus daemon ;)
16:35
<ftherese>
another thing that would be nice for local apps is the ability to install wine stuff
16:35* Lns just got done building an 8.04lts local repository server on lenny
16:36jammcq has quit IRC
16:36
<alkisg>
Lns: I think I've seen an open source repository for proprietary windows apps, though... :D
16:36
<Lns>
heh
16:36
<alkisg>
apt-get install msword :P
16:36
<ftherese>
if I do a chroot to /opt/ltsp/i386, how do I export the proper DISPLAY for wine, and which user should it install as?
16:37
wine is generally not supposed to install as root
16:37
but there is no information about adding users within the chroot
16:38
<nubae>
alkisg: wow... what a wonderful contradiction.. an open source repo for closed sourced binaries... love it
16:39
or a closed source repo for open source projects would also be equally funny
16:40* Lns is thinking an opensource citrix server would be cool instead of having to muck around with wine
16:40
<alkisg>
If I ever hear that MS replaced the MSI technology with something like the repositories used in Linux, I may boot windows once more to check it out. :)
16:41Egyptian[Home] has quit IRC
16:41
<ftherese>
so how would you install a wine app under the ltsp client (to run as a localapp)?
16:41
<Lns>
alkisg: i doubt that'll ever happen.. just think of the 3rd pty devs that will completely ignore it for eons
16:41
no turning back now ;)
16:41PMantis has joined #ltsp
16:42
<PMantis>
sbalneav: Have a minute?
16:42
<Lns>
the community is already built with opensource..microsoft would be fighting an uphill battle for years before anything even remotely comparable would surface
16:42
<ftherese>
*wonders if anyone can hear him
16:42
<alkisg>
Lns, it's not that impossible... each vendor would have his own server, something like the PPAs in launchpad. And installing/updating software would be as easy as installing the vendor's ppa...
16:43
<Lns>
alkisg: lol..yeah, imagine the legal headaches of that for a second
16:43
<alkisg>
Why? Vendors would be distributing their own apps
16:43
<Lns>
ftherese: we hear you, but most likely nobody knows the answer :)
16:43
<alkisg>
They won't be microsoft hosted or anything
16:44
ftherese: wine isn't designed for multiple users
16:44
<Lns>
alkisg: but it's tied to the microsoft name
16:44
<alkisg>
If you install it for a user, I bet you can run it as a localapp
16:44Egyptian[Home] has joined #ltsp
16:44
<alkisg>
Because all the .wine directory would be visible locally with the sshfs
16:44
<Lns>
you'd be using an MS interface to install third party apps..i dunno, i guess it could be possible but i highly doubt they're going to even try to be as cool as us ;)
16:45
<alkisg>
Lns, everyone is already using an MS interface... the MSI installer :)
16:45
but anyway we're in #ltsp, let's leave it there.
16:45
<Lns>
alkisg: agreed.. dont need to waste cycles ;)
16:46
I already do that enough w/ms
16:47
<ftherese>
well.. how about this question then
16:47
when I do a chroot into /opt/ltsp
16:47
how am I supposed to export my DISPLAY correctly?
16:48
<alkisg>
ftherese: to run wine apps locally, you *shouldn't* install them on the chroot.
16:49
<ftherese>
well... that is what I do when I install the other apps...
16:49
<alkisg>
I've never actually done it, but wine apps are installed on a per-user basis, so you'd need to install them on the server
16:49
<ftherese>
I just figured I would follow the same procedure?
16:49
<alkisg>
Then, sshfs would transfer them on the client and they'd run locally if you made an appropriate .desktop file
16:49
<ftherese>
ok
16:50
<alkisg>
No, wine apps generally aren't installed in the same way as other linux apps
16:50
<ftherese>
ok
16:50
so If I install it on the server
16:50
what do I have to do to the desktop file?
16:51epsas is now known as SeaweedStardom
16:51
<alkisg>
Well, first you'd have to have installed wine in the chroot. Not the windows-app; just wine.
16:51
<ftherese>
that is done
16:52
<alkisg>
Then if the desktop file ran `ltsp-localapps wine windows-app` I guess it would run locally.
16:53
Again, I've never tried this, so I'm not sure that would work. That's what I would try, though...
16:53
Try it manually first: ltsp-localapps wine /home/user/.wine/path/to/app.
16:54
<ftherese>
the problem is, the file to run is an installation file
16:54
<alkisg>
Can you be more specific?
16:54
<PMantis>
setup.exe ?
16:54
<vagrantc>
ftherese: install wine on the server, install the application *not* as a localapps, make a .desktop icon to run it as a localapp, and you should be done.
16:55
<ftherese>
for example... but this one is update.exe
16:55
<alkisg>
" install the application *not* as a localapps" ==> means that you'd just logon as the user and run update.exe. No chroot involved in the windows-app installation.
16:55
<vagrantc>
ftherese: or run "ltsp-localapps x-terminal-emulator" and run "wine path/to/your.exe
16:56
ftherese: main thing is just start trying *something*.
16:56
<ftherese>
ok
16:57
I will run wine update.exe on the server
16:57* vagrantc doesn't really know how wine works
16:57* alkisg is a little worried about sshfs not following symlinks. I think wine uses them...
16:57* Lns likes Pinos
16:58
<ftherese>
and then the problem is that the installed program depends upon the environment installed in the server
16:58
<alkisg>
...can you be more specific? :)
16:58
<vagrantc>
ftherese: well, you're the one trying to use wine.
16:58
<ftherese>
when you run wine
16:59
wine sets up a windows environment in /home/user/.wine
16:59
and then
17:00
your programs depend upon this environment
17:00
<alkisg>
Yes, that's a per user dir. That's why you'll have to install the app for each user seperately
17:00
It doesn't depend on the server environment. It depends on the user ~/.wine dir.
17:00
<Lns>
alkisg: you can't symlink ~/.wine ?
17:00* Lns remembers this conversation a couple weeks ago
17:01
<ftherese>
so you would have to move the .wine directory to the chroot
17:01
<alkisg>
ftherese: All the /home/user dir is visible to localapps, so /home/user/.wine is also visible.
17:01
That's why I was referring to sshfs. As vagrantc said, start trying :)
17:01
Lns: sshfs with the default options doesn't follow symlinks that point outside of the home dir
17:02
<Lns>
ah
17:02
<alkisg>
So if there's a link from ~/.wine/.../gecko to /usr/share/wine/gecko, it won't use it (unless it's also installed localy?)
17:02fotanus has left #ltsp
17:03
<vagrantc>
so it's probably better to run the install program as a localapp ...
17:03
easiest way is from a terminal started as a localapp
17:03
<alkisg>
No, I don't think that matters
17:03
<vagrantc>
then you'll have the client environment
17:03* alkisg looks for symlinks
17:03
<vagrantc>
gah. how did i get dragged into a conversation about supporting wine?
17:04
<Lns>
ah, irc ;)
17:04
<alumno01>
fixed, i imported a template with the new margin for each session
17:04alumno01 has quit IRC
17:04
<alkisg>
OK no symlinks pointing out of ~/.wine :)
17:04* Lns thinks irc has given me a ~20sec. attention span
17:06Sarten-X has quit IRC
17:06
<ftherese>
I am installing on the server
17:07* alkisg calls it a day. G'night all...
17:07
<ftherese>
good night
17:07
thanks
17:07alkisg has quit IRC
17:11
<ftherese>
trying a few things... i'll be back
17:12ftherese has quit IRC
17:14PMantis has quit IRC
17:14Q-FUNK has left #ltsp
17:18sene has quit IRC
17:18ftherese has joined #ltsp
17:18brftherese has joined #ltsp
17:19
<brftherese>
Ok... tested the APPS_MENU on gnome
17:19
it works... so there is a bug with xubuntu-dekstop/xfce
17:21
any ideas why the lts.conf wouldn't export the MENU_ITEMS correctly to xfce4?
17:25artista-frustrad has quit IRC
17:29
<vagrantc>
brftherese: it's more like that xfce isn't respecting the xdg configurations properly.
17:29
likely
17:30
as the LTSP code that generates these things does no checks whatsoever for which desktop you're using. it simply follows the xdg specifications
17:34Sarten-X has joined #ltsp
17:38Sarten-X has quit IRC
17:42wimiko has quit IRC
17:43artista-frustrad has joined #ltsp
17:55shawnp0wers has joined #ltsp
17:57
<brftherese>
well... I have checked through the xdg stuff in xfce4, but I don't know what to look for... so I will try to compare it with gnome stuff
17:57
<Lns>
heh, going back to my disk i/o bottleneck thing, saw this while going through the lists :) http://blog.gavinmc.com/?p=86
17:58Gadi has quit IRC
17:58Gadi has joined #ltsp
17:58bobby_C has quit IRC
18:00FGXR6 has joined #ltsp
18:00F-GT has quit IRC
18:01Gadi has left #ltsp
18:01shawnp0wers has quit IRC
18:01shawnp0wers has joined #ltsp
18:01nubae has quit IRC
18:03pmatulis has quit IRC
18:06pmatulis has joined #ltsp
18:07pmatulis has quit IRC
18:22FGXR6 has quit IRC
18:26shawnp0wers has quit IRC
18:27skeemer has quit IRC
18:32FGXR6 has joined #ltsp
18:44staffencasa has quit IRC
18:50* vagrantc finally got around to testing the udhcpc hooks
18:59vagrantc has quit IRC
19:06
<Lns>
FYI: Added https://help.ubuntu.com/community/UbuntuLTSP/FilesystemOptimize for hints on optimizing the filesystem for LTSP servers...additions/edits welcome, as always :)
19:07
Doesn't have a whole heckofalot of info, just what I picked up today
19:07
but i'm sure it's going to help out a lot for people tweaking for performance
19:25shawnp0wers has joined #ltsp
19:25Lns has quit IRC
19:53cliebow has joined #ltsp
19:59cliebow has quit IRC
20:23shawnp0wers has quit IRC
20:23shawnp0wers has joined #ltsp
20:30shawnp0wers has quit IRC
20:48jammcq has joined #ltsp
20:48
<jammcq>
hello friends
21:02Topslakr has quit IRC
21:26
<sbalneav>
Evening all
21:30warren has quit IRC
21:30ninkendo has quit IRC
21:30
<jammcq>
sbalneav: Scotty !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
21:33
<sbalneav>
Hey jammcq!
21:36warren has joined #ltsp
21:36ninkendo has joined #ltsp
21:42artista-frustrad has quit IRC
21:43artista-frustrad has joined #ltsp
21:48Topslakr has joined #ltsp
21:57vagrantc has joined #ltsp
22:09
<chupacabra>
jammcq? waddup?
22:10
jammcq: waddup?
22:15
<jammcq>
chupacabra: HEY !!!!
22:15
wow, how have you been?
23:17
<chupacabra>
good, looking for you.
23:18
jammcq: did I miss you?
23:25
email me
23:28yanu has quit IRC
23:30yanu has joined #ltsp
23:41alkisg has joined #ltsp
23:52alkisg has left #ltsp