IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 24 December 2010 (all times are UTC)
| 00:14 | MorningSon has quit IRC | |
| 00:41 | alkisg has quit IRC | |
| 00:51 | daya has quit IRC | |
| 00:51 | daya has joined #ltsp | |
| 01:41 | Mava has quit IRC | |
| 01:48 | gnunux has joined #ltsp | |
| 01:49 | <gnunux> hi
| |
| 02:21 | komunista has joined #ltsp | |
| 02:31 | alkisg has joined #ltsp | |
| 02:44 | alkisg has quit IRC | |
| 02:45 | alkisg has joined #ltsp | |
| 02:46 | alkisg has quit IRC | |
| 03:22 | komunista has quit IRC | |
| 03:37 | alkisg has joined #ltsp | |
| 03:59 | <alkisg> What's the purpose of the 000-init-whitelist plugin? It deletes some rc.d links and so some services don't start on thin/fat clients: http://ltsp.pastebin.com/4tyEJhSw
| |
| 03:59 | Gremble has joined #ltsp | |
| 04:09 | <alkisg> That's probably the reason why some people here complain about cups, ondemand, sendsigs, rc.local, binfmt-support, openbsd-inetd, winbind etc not working with localapps and fat clients...
| |
| 04:09 | If that list is indeed needed, it should be revisited
| |
| 04:11 | I think Debian doesn't use that list at all, it's commented out: #RCS_WHITELIST=$ ...
| |
| 04:12 | So maybe Ubuntu should sync with Debian on this, i.e. in Ubuntu/000-basic-configuration
| |
| 04:20 | Gremble has quit IRC | |
| 04:25 | <alkisg> I reported it at https://bugs.launchpad.net/ubuntu/+source/ltsp/+bug/694066
| |
| 04:36 | I'll try commenting them out, building a fat chroot, and if it works OK I'll commit it...
| |
| 04:38 | jimjimovich has joined #ltsp | |
| 04:38 | <jimjimovich> good morning everyone
| |
| 04:39 | what's the best way to put run a script each time a user logs into ltsp?
| |
| 04:39 | the script needs to be run as that user
| |
| 04:40 | <alkisg> Put it in the server /etc/xdg/autostart
| |
| 04:41 | (as a .desktop file, see examples there)
| |
| 04:41 | <jimjimovich> thanks, that's exactly what i was looking for!
| |
| 04:45 | alexqwesa has quit IRC | |
| 04:46 | alexqwesa has joined #ltsp | |
| 04:50 | baiki has joined #ltsp | |
| 04:55 | baiki has quit IRC | |
| 05:11 | komunista has joined #ltsp | |
| 05:14 | daya has quit IRC | |
| 05:14 | daya has joined #ltsp | |
| 05:21 | Gremble has joined #ltsp | |
| 05:25 | gnunux has quit IRC | |
| 05:29 | baiki has joined #ltsp | |
| 05:32 | <baiki> @alkisg hi. i updated this bugreport with the graca output from the thin client
| |
| 05:32 | 01:00.0 VGA compatible controller [0300]: VIA Technologies, Inc. CN700/P4M800 Pro/P4M800 CE/VN800 [S3 UniChrome Pro] [1106:3344] (rev 01)
| |
| 05:33 | <alkisg> baiki: nice, it seems that *only* unichrome cards are affected
| |
| 05:34 | <baiki> one more question: since I can't see any different on the thin clients screen resolution i assume i am modifying the wrong lts.conf. Is this the right path: /var/lib/tftpboot/ltsp/i386/lts.conf
| |
| 05:36 | i try to set this value "X_MODE_0 = 1024x768" in /var/lib/tftpboot/ltsp/i386/lts.conf and afterwards an image update but.... nope :-(
| |
| 05:36 | <alkisg> Use XRANDR_MODE_0=1024x768
| |
| 05:36 | Don't forget to put [Default] on top of it
| |
| 05:37 | <baiki> ok, will try right now. thanks a lot.
| |
| 05:39 | <alkisg> baiki: you don't need to update the image when you modify lts.conf
| |
| 05:40 | <baiki> ok
| |
| 05:44 | baiki has quit IRC | |
| 05:52 | baiki has joined #ltsp | |
| 05:53 | <baiki> @alkisg nope. did not work out for me. i used this: http://pastebin.com/p8NRL3Kr
| |
| 06:00 | !compiz
| |
| 06:00 | <ltspbot> baiki: "compiz" :: the default window manager in gnome is gnome-wm, which automatically chooses compiz if it thinks that the card supports it. Compiz is causing login problems to some clients (LP #673072). To disable it, see !disable_compiz. To restore it, see !restore_compiz
| |
| 06:00 | <baiki> !XRANDR_MODE_0
| |
| 06:01 | <ltspbot> baiki: Error: "XRANDR_MODE_0" is not a valid command.
| |
| 06:01 | <alkisg> !localxterm
| |
| 06:01 | <ltspbot> alkisg: "localxterm" :: while sitting on a thin client, open a gnome terminal. In that, run: ltsp-localapps xterm. An xterm will open. That xterm runs locally, so any commands you enter there are executed directly on the client.
| |
| 06:01 | <baiki> !disable_compiz
| |
| 06:01 | <ltspbot> baiki: "disable_compiz" :: To disable compiz, try: sudo gconftool-2 --direct --config-source xml:readwrite:/etc/gconf/gconf.xml.mandatory --type string --set /desktop/gnome/session/required_components/windowmanager metacity
| |
| 06:01 | <alkisg> baiki: open a localxterm (on the client) and run: getltscfg -a
| |
| 06:01 | <baiki> ok
| |
| 06:06 | baiki has quit IRC | |
| 06:06 | Trixboxer has joined #ltsp | |
| 06:08 | baiki has joined #ltsp | |
| 06:09 | <baiki> @alkisg here is the output of getltscfg -a: XRANDR_MODE_0="1024x768" (newline) export XRANDR_MODE_0
| |
| 06:10 | <alkisg> baiki: ok, now run this: xrandr
| |
| 06:10 | and pastebin the output
| |
| 06:10 | <baiki> ok
| |
| 06:10 | ok
| |
| 06:15 | baiki has quit IRC | |
| 06:20 | <alkisg> !xauthority
| |
| 06:20 | <ltspbot> alkisg: "xauthority" :: to access the client X session from a local root shell, try: eval $(tr '\0' '\n' < /proc/$(pidof -s ldm gdm-simple-greeter gnome-session | cut -d' ' -f1)/environ | egrep '^DISPLAY=|^XAUTHORITY=') && export DISPLAY XAUTHORITY
| |
| 06:20 | baiki has joined #ltsp | |
| 06:21 | <baiki> @alkisg can't open pastebin
| |
| 06:22 | Screen 0: minimum 640 x 480, current 800 x 600, maximum 800 x 600 default connected 800x600+0+0 0mm x 0mm 800x600 60.0* 56.0 640x480 60.0
| |
| 06:22 | <alkisg> baiki: ok the problem is that your monitor timings aren't correctly detected
| |
| 06:22 | Is that on the openchrome thin client?
| |
| 06:22 | <baiki> possible. it's a CRT (and old) 17"
| |
| 06:23 | <alkisg> What's the graphics card driver? openchrome?
| |
| 06:23 | <baiki> yes, on the hpt3350. if i do this on the TC itself i get: can't open display if i enter xrandr
| |
| 06:23 | <alkisg> Wait wait what?
| |
| 06:23 | if i do this on the TC itself i get: can't open display if i enter xrandr ==> ?
| |
| 06:24 | Just open a gnome-terminal on the thin client (=the TC) and run xrandr...
| |
| 06:24 | That shouldn't say "can't open display"...
| |
| 06:24 | <baiki> yes, the 800x600 output is from the gnome-terminal
| |
| 06:25 | <alkisg> ok
| |
| 06:25 | Try putting those in lts.conf:
| |
| 06:25 | X_HORZSYNC=30-88
| |
| 06:25 | X_VERTREFRESH=30-88
| |
| 06:26 | (also keep the XRANDR_MODE_0)
| |
| 06:26 | and reboot the client
| |
| 06:31 | baiki has quit IRC | |
| 06:36 | jimjimovich has left #ltsp | |
| 06:41 | Gremble has quit IRC | |
| 06:43 | baiki has joined #ltsp | |
| 06:44 | <baiki> @alkisg yes man!! the h/v sync and refresh did the job. thanks a lot. got i question for today left: LDM_ALLOW_USER does not work. can login with all users
| |
| 06:45 | <alkisg> Try LDM_USER_ALLOW
| |
| 06:51 | baiki has quit IRC | |
| 07:06 | alkisg has quit IRC | |
| 07:09 | jimjimovich has joined #ltsp | |
| 07:10 | <jimjimovich> okay, I now have a startup script running when a user logs in, how can I run a script when they log out?
| |
| 07:16 | baiki has joined #ltsp | |
| 07:19 | komunista has quit IRC | |
| 07:20 | baiki has quit IRC | |
| 07:22 | Mobe_ has joined #ltsp | |
| 07:24 | Mobe has quit IRC | |
| 07:27 | litlebuda has joined #ltsp | |
| 07:41 | alkisg has joined #ltsp | |
| 07:44 | <jimjimovich> alkisg: thanks for your tip on running a script at login. how would i run one at logout?
| |
| 07:46 | <alkisg> !xexit
| |
| 07:46 | <ltspbot> alkisg: Error: "xexit" is not a valid command.
| |
| 07:47 | <jimjimovich> !exit
| |
| 07:47 | <ltspbot> jimjimovich: Error: "exit" is not a valid command.
| |
| 07:53 | <jimjimovich> i see there's an xexit thing in the ltsp bazaar code on launchpad
| |
| 07:57 | <alkisg> There's no standard way to run scripts at logout, because e.g. xorg might crash so those scripts wouldn't be called
| |
| 07:58 | One ltsp dev, sbalneav, made the xexit package which tries to execute stuff at logout
| |
| 07:58 | Why do you want to run scripts at logout?
| |
| 07:58 | <jimjimovich> i need to be able to clean up some apps that don't shut down correctly, etc
| |
| 07:59 | <alkisg> For example? Can't you clean them on logon?
| |
| 07:59 | <jimjimovich> I guess I could, they're just sitting around taking up memory while the user is not logged in
| |
| 08:00 | <alkisg> Which ones? I'm asking because there's a difference if the app crashed or xorg crashed or even if the user unplugged the power cord of the client...
| |
| 08:01 | <jimjimovich> yeah, i know. the one that's bothering me right now is my own app. i probably just don't know how to launch it correctly so that it shuts down when the user logs out. it's a ruby app
| |
| 08:01 | but I've had problems with Evolution not shutting down right too and the process hanging around after logotu
| |
| 08:02 | <alkisg> evolution-data-server does that yeah
| |
| 08:02 | <jimjimovich> also ubuntu-one
| |
| 08:02 | <alkisg> Normally programs are launched inside the x-session, so they're closed when it finishes
| |
| 08:02 | If you made your program run within the x session, it would be automatically killed
| |
| 08:03 | <jimjimovich> i'm launching a background ruby process when the user logs in, how could i make it run within their x session?
| |
| 08:03 | <alkisg> So anyway yeah xexit will kill those apps for you
| |
| 08:03 | x apps, gtk apps etc belong to the session, and e.g. gtk and qt automatically declare themselves as session clients
| |
| 08:03 | I don't know about console apps...
| |
| 08:04 | <jimjimovich> the ruby process runs as that user, so i don't quite understand why it keeps running when the user logs out (i'm not that much of a guru)
| |
| 08:04 | <alkisg> Noone tells it to close
| |
| 08:04 | The session only keeps track of processes that declare that belong to it
| |
| 08:05 | OK anyway the xexit package will close those apps for you
| |
| 08:05 | <jimjimovich> okay, thanks. i'll also look into making my app behave correctly so that i don't need xexit :)
| |
| 08:06 | <alkisg> If you find a solution for console apps for this, please notify me :)
| |
| 08:06 | <jimjimovich> will do!
| |
| 08:07 | Mobe_ has quit IRC | |
| 08:13 | litlebuda has quit IRC | |
| 08:17 | jimjimovich has quit IRC | |
| 08:20 | BWMerlin has joined #ltsp | |
| 08:26 | BWMerlin has quit IRC | |
| 08:35 | artista_frustrad has quit IRC | |
| 08:37 | jd42 has quit IRC | |
| 08:39 | jd42 has joined #ltsp | |
| 08:42 | jd42 has quit IRC | |
| 08:48 | jd42 has joined #ltsp | |
| 09:13 | Mobe has joined #ltsp | |
| 10:17 | m4xx has quit IRC | |
| 10:22 | [censored] has joined #ltsp | |
| 10:26 | MorningSon has joined #ltsp | |
| 10:42 | Gremble has joined #ltsp | |
| 10:43 | Gremble has quit IRC | |
| 11:02 | [censored] is now known as m4xx | |
| 11:38 | Kicer86 has joined #ltsp | |
| 11:39 | vagrantc has joined #ltsp | |
| 11:48 | <alkisg> vagrantc, what does commit #1757 mean? (comment out RC*_WHITELIST defaults, as they are not respected properly in squeeze.)
| |
| 11:48 | There are a lot of useful but disabled services in thin/fat clients so I'm thinking it might be better if the RC*_WHITELISTs were disabled in Ubuntu too...
| |
| 11:50 | Did you notice any problems after commenting them out?
| |
| 12:05 | bbl, will see the irclogs
| |
| 12:09 | alkisg has quit IRC | |
| 12:20 | <vagrantc> alkisg: ah, irc logs ... well ... debian's insserv (which manages /etc/rc?.d/ symlinks) ignores the whitelisting, and the implementation of whitelisting in ltsp was pretty ugly and hackish anyways, so i just commented it out.
| |
| 12:21 | alkisg: didn't notice any problems, but they weren't being respected at all... might be totally different with Ubuntu's upstart or whatever manages init.
| |
| 12:27 | otavio has joined #ltsp | |
| 12:27 | otavio has joined #ltsp | |
| 12:29 | otavio has quit IRC | |
| 12:29 | otavio has joined #ltsp | |
| 12:29 | otavio has joined #ltsp | |
| 12:41 | alkisg has joined #ltsp | |
| 12:53 | alkisg has quit IRC | |
| 13:13 | alkisg has joined #ltsp | |
| 13:19 | <alkisg> The whitelisting works in Ubuntu, i.e. ltsp-build-client disables all non-whitelisted services. And that's the problem, because some services are needed...
| |
| 13:19 | If some services really need to be disabled (didn't see any so far), then blacklisting might be better than whitelisting, especially with localapps and fat clients
| |
| 13:20 | And if it could be done from an initramfs script or an initscript, it'd be even better, as now purging and re-installing a service re-enables it
| |
| 13:20 | But I didn't find any service that needs blacklisting so far...
| |
| 13:32 | <vagrantc> the whitelisting was introduced in the frenzy to speed up boot times
| |
| 13:32 | but i always thought it was too heavy-handed
| |
| 13:33 | alkisg: but how the debian scripts work, i'm not sure there's a way to even blacklist something
| |
| 13:33 | <alkisg> vagrantc: update-rc.d doesn't work?
| |
| 13:33 | <vagrantc> short of removing or modifying the scripts in /etc/init.d/
| |
| 13:33 | alkisg: not really
| |
| 13:34 | might work until the next package update
| |
| 13:34 | <alkisg> (removing the script from an initramfs script was what I thought some time ago before seeing the ltsp whitelists...)
| |
| 13:34 | <vagrantc> but then insserv reorders all the rc?.d scripts
| |
| 13:34 | <alkisg> *the scripts in /etc/init.d
| |
| 13:34 | Or even replacing them with symlinks to a dummy script
| |
| 13:34 | So that the rc.d/* symlinks are not broken...
| |
| 13:35 | <vagrantc> to do that for read-only NFS would eat nearly half a meg of ram
| |
| 13:35 | <alkisg> Why? Just for a dozen symlinks?
| |
| 13:36 | <vagrantc> i guess could bind-mount the blacklisted ones with dummy scripts
| |
| 13:36 | that would be the most efficient approach
| |
| 13:37 | <alkisg> But I didn't find yet any service that needs to be disabled, so I'm guessing much less than a dozen bind-mounted symlinks will be needed...
| |
| 13:41 | <vagrantc> can't bind-mount symlinks
| |
| 13:41 | but you can bind-mount a dummy file
| |
| 13:41 | <alkisg> Ah. OK, dummy files then
| |
| 13:41 | List of services that were disabled in my fat chroot:
| |
| 13:41 | apparmor binfmt-support bluetooth brltty cups dns-clean fancontrol kerneloops lm-sensors nbd-client networking ondemand openbsd-inetd pcmciautils pppd-dns pulseaudio rc.local rsync saned sendsigs speech-dispatcher umountfs umountnfs.sh umountroot unattended-upgrades urandom winbind wpa-ifupdown x11-common
| |
| 13:42 | Which of those need blacklisting? pulseaudio, unattented-upgrades...
| |
| 13:42 | <vagrantc> i've heard network-manager and gdm do evil
| |
| 13:43 | <alkisg> network-manager can be used to control vpns etc
| |
| 13:43 | If we write an "eth0 manual" in /etc/network/interfaces, then it doesn't touch that interface
| |
| 13:44 | gdm doesn't do evil if ldm writes the default-display-manager in debconf and in /etc
| |
| 13:44 | <vagrantc> which will get overriden on upgrades
| |
| 13:45 | of gdm, at least
| |
| 13:45 | <alkisg> Why?
| |
| 13:45 | It doesn't do that for me
| |
| 13:45 | <vagrantc> ldm doesn't implement pax displayicus managerius
| |
| 13:45 | <alkisg> Let me check the code I put, I think in the fat client plugin...
| |
| 13:45 | <vagrantc> http://bugs.debian.org/267198
| |
| 13:46 | <alkisg> echo '/usr/sbin/ldm' > "$ROOT/etc/X11/default-display-manager"
| |
| 13:46 | echo 'ldm shared/default-x-display-manager select ldm' | chroot "$ROOT" debconf-set-selections
| |
| 13:46 | Those should be part of the ldm code
| |
| 13:47 | When gdm was upgraded for me, ldm was kept as the default display manager
| |
| 13:47 | <vagrantc> it's been years since i looked into it, so maybe it respects local changes now
| |
| 13:48 | <alkisg> And that would also solve any other *dm problems, if kdm or some else dm is installed (assuming they all read that shared/default-x-display-manager debconf setting)
| |
| 13:48 | <vagrantc> they're supposed to
| |
| 13:48 | and ldm should as well
| |
| 13:49 | <alkisg> And reconfiguring gdm showed ldm and gdm in the debconf dialog, so afaik it worked ok...
| |
| 13:50 | So what I'm saying is, those services like network-manager or gdm that can "properly" be worked around, don't need to be disabled...
| |
| 13:50 | <vagrantc> sure
| |
| 13:50 | <alkisg> (btw in ubuntu those are upstart scripts so the current whitelisting doesn't remove them)
| |
| 13:51 | <vagrantc> if pax displayicus managerius is any easier to implement than it used to be, we should probably do it in ldm's postinst
| |
| 14:14 | dlezcano has joined #ltsp | |
| 14:16 | Trixboxer has quit IRC | |
| 14:17 | alexqwesa has quit IRC | |
| 14:18 | alexqwesa has joined #ltsp | |
| 14:19 | dlezcano has quit IRC | |
| 14:31 | dlezcano has joined #ltsp | |
| 15:23 | Kicer86 has quit IRC | |
| 15:32 | alkisg has quit IRC | |
| 15:34 | vagrantc has quit IRC | |
| 15:36 | baiki has joined #ltsp | |
| 15:44 | Mobe has quit IRC | |
| 16:02 | alkisg has joined #ltsp | |
| 17:20 | alkisg has quit IRC | |
| 19:07 | ltspbot` has joined #ltsp | |
| 19:09 | ltspbot has quit IRC | |
| 19:24 | otavio has quit IRC | |
| 21:36 | yeedl has joined #ltsp | |
| 21:36 | <yeedl> hey ya'll
| |
| 22:23 | yeedl has quit IRC | |
| 22:28 | cyberorg has quit IRC | |
| 22:35 | cyberorg has joined #ltsp | |