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


Channel log from 20 May 2010   (all times are UTC)

00:05waldo323 has quit IRC
00:05alkisg has joined #ltsp
00:12highvoltage has quit IRC
00:24Faithful has quit IRC
00:27highvoltage has joined #ltsp
00:37alkisg has quit IRC
00:48highvoltage has joined #ltsp
00:52highvoltage has quit IRC
00:59highvoltage has joined #ltsp
01:00highvoltage has quit IRC
01:04highvoltage has joined #ltsp
01:14highvoltage has joined #ltsp
01:15highvoltage has quit IRC
01:16highvoltage has joined #ltsp
01:27alkisg has joined #ltsp
01:49F-GT has quit IRC
01:50dobber has joined #ltsp
01:50Guerdal82_ has joined #ltsp
01:55highvoltage has quit IRC
02:01highvoltage has joined #ltsp
02:02ogra_cmpc has quit IRC
02:04ogra_cmpc has joined #ltsp
02:06F-GT has joined #ltsp
02:23highvoltage has quit IRC
02:24highvoltage has joined #ltsp
02:27
<dobber>
anyone developing ltsp here ?
02:28
i've noticed that if you logout incorectly or hard restart the client, alot of the running processes become zombies
02:28
i would like to write a skript that cleans those zombies
02:39
does ltsp prefere specific language for those scripts? bash/perl/python/c ?
02:44Selveste1 has joined #ltsp
02:44highvoltage has quit IRC
02:45highvoltage has joined #ltsp
02:57
<alkisg>
dobber: a shell script would be fine.
02:57
I think sbalneav has written a "when xorg dies run a script" C wrapper
02:58
If you could contact him and make this all somehow integrated as an option to ltsp, it'd be very nice
02:59
<dobber>
i would like that
02:59
<alkisg>
dobber: http://bazaar.launchpad.net/~sbalneav/ltsp/xexit/changes
03:00
<dobber>
i'm new to contributing
03:00
<alkisg>
I think the functionality should already be there, so mostly testing + bug fixing would be needed.
03:01
<dobber>
ok let me check that wrapper first and try it
03:01
<alkisg>
If you come back in 8+ hours, sbalneav will likely be here to walk you through the required steps to put this upstream
03:01
<dobber>
are most of the devs in US time ?
03:01
<alkisg>
Yup
03:01
!seen sbalneav
03:01
<ltspbot`>
alkisg: sbalneav was last seen in #ltsp 16 hours, 56 minutes, and 10 seconds ago: <sbalneav> yup
03:04
<dobber>
alkisg i'll try to come back later tonight
03:04
in the mean time, i'll try to finish my script
03:04
<alkisg>
dobber: I think, using sbalneav's wrapper, the script would be really small,
03:05
<dobber>
digging in...
03:05
<alkisg>
i.e. drop a sh script in the /etc/Xexit.d directory,
03:05
which would source ltsp_config to get the lts.conf variables,
03:05
and if $LTSP_END_PROCESSES_AT_LOGOUT is True, it would just killall $USER..
03:05
or something along those steps
03:05
bbl.
03:05
<dobber>
thanks
03:28Lumiere has quit IRC
03:31Lumiere has joined #ltsp
03:49Lumiere has quit IRC
03:49Lumiere has joined #ltsp
04:04bobby_C has joined #ltsp
04:38bix0r has joined #ltsp
05:19Eghie has quit IRC
05:21gnunux has joined #ltsp
05:21
<gnunux>
hi
05:32Eghie has joined #ltsp
05:38t0mmii_ has joined #ltsp
05:39t0mmii has quit IRC
05:39t0mmii_ is now known as t0mmii
05:39t0mmii has joined #ltsp
05:40bobby_C has quit IRC
05:44bobby_C has joined #ltsp
05:55japerry has quit IRC
05:56japerry has joined #ltsp
06:08alkisg has quit IRC
06:20toscalix has joined #ltsp
06:22pmatulis has joined #ltsp
06:23artista_frustrad has quit IRC
06:36artista_frustrad has joined #ltsp
06:48artista_frustrad has quit IRC
06:52scottmaccal has joined #ltsp
06:52slidesinger has joined #ltsp
07:01artista_frustrad has joined #ltsp
07:02frgds has joined #ltsp
07:11knipwim has quit IRC
07:11stgraber has quit IRC
07:11zamba has quit IRC
07:14knipwim has joined #ltsp
07:14stgraber has joined #ltsp
07:14zamba has joined #ltsp
07:19slidesinger has joined #ltsp
07:20reynolds has quit IRC
07:32scottmaccal has quit IRC
08:12komunista has joined #ltsp
08:26Lumiere has quit IRC
08:26Lumiere has joined #ltsp
08:35mgariepy has joined #ltsp
08:54toscalix has quit IRC
08:55bobby_C has quit IRC
09:00toscalix has joined #ltsp
09:05frgds has quit IRC
09:12ogra has quit IRC
09:13ogra has joined #ltsp
09:15otavio has quit IRC
09:15otavio has joined #ltsp
09:16LedHed has joined #ltsp
09:32frgds has joined #ltsp
09:43frgds has quit IRC
09:45otavio has quit IRC
09:45otavio has joined #ltsp
09:46otavio has joined #ltsp
09:46otavio has joined #ltsp
10:05
<johnny>
dobber, what about gnome-watchdog?
10:09
<dobber>
i haven't used it
10:11
<johnny>
might solve the problem you mentioned
10:11
<dobber>
what does it do ?
10:11
<johnny>
kills leftover processes at some point
10:12
<dobber>
it runs on user login and stays as daemon ?
10:12
one daemon per user ?
10:15
xexit is sort of ok-ish
10:15
it works if the client logs out
10:15
<sbalneav>
_UsUrPeR_: Pong
10:15
<dobber>
but if the client is hard shutdown, it does not work
10:18
<sbalneav>
dobber: Yes it will, it will just take a while.
10:18
You'll have to wait for the TCP_KEEPALIVE timeout to realize the display's disappeared.
10:19
Usually, on most systems, this defaults to 10 minutes.
10:19
Sorry, 20
10:19
<dobber>
i didn't wait 20 minutes in none of my tests :))
10:19
will wait now ;)
10:22
where do i set this ?
10:22
sysctl ?
10:23
./proc/sys/net/ipv4/tcp_keepalive_time
10:24
2 hours for me
10:25
<sbalneav>
The only other option would be to have a thread running in xexit which would keep "pinging" the server periodically, which, of course, would add slightly to network traffic/client load.
10:25komunista has quit IRC
10:25
<dobber>
it works
10:25bobby_C has joined #ltsp
10:25
<dobber>
i guess it's not 2 hours and i'm looking at the wrong proc file
10:26
<sbalneav>
I may decide to add that, and make it an option, i.e. command line switch to set "active" thin client disconnect behaviour
10:26komunista has joined #ltsp
10:27
<dobber>
also, it just killed xexit (probably parrent process)
10:27
it did not execute anything
10:28
<sbalneav>
well, it should execute something, as xexit properly daemonizes.
10:28
I can have a further look at it.
10:28
<dobber>
i could look at this too, tomorow
10:28
it's time for beer
10:29
also, proftpd sucks
10:31dobber has quit IRC
10:33bobby_C has quit IRC
10:34gnunux has quit IRC
10:36bobby_C has joined #ltsp
10:41japerry has quit IRC
10:43japerry has joined #ltsp
10:43artista_frustrad has quit IRC
10:46
<mgariepy>
hi guys
10:47F-GT has quit IRC
10:48
<mgariepy>
sbalneav, i have a problem with jetpipe on thin client
10:49
http://pastebin.com/HV7SDR9F
10:49
jetpipe crash with this python trace.
10:52
and when this happen, we need to launch jetpipe again
10:52
if it's any help, we talk to jetpipe through a windows print server.
10:55artista_frustrad has joined #ltsp
11:00bobby_C has quit IRC
11:04F-GT has joined #ltsp
11:07muttl3y has joined #ltsp
11:07staffencasa has joined #ltsp
11:08
<muttl3y>
where do i start for make xbox and LTSP client?!
11:08artista_frustrad has quit IRC
11:08
<muttl3y>
MS xbox 1 just have etherboot support...
11:20The_Code has quit IRC
11:21artista_frustrad has joined #ltsp
11:22
<muttl3y>
make DHCP request, load nbi.img ( for etherboot ) from LTSP serve and freeze :(
11:43t0mmii has quit IRC
11:48t0mmii has joined #ltsp
11:53alkisg has joined #ltsp
11:58
<_UsUrPeR_>
sbalneav
11:59
nm. I think I got it
12:01toscalix has quit IRC
12:10cliebow has joined #ltsp
12:20Egyptian[Home] has quit IRC
12:28Lns has joined #ltsp
12:29
<sbalneav>
_UsUrPeR_: Okie
12:29
mgariepy: lemme seeeeeee......
12:30muttl3y has quit IRC
12:30
<mgariepy>
sbalneav, thanks ;)
12:31
maybe line number have changed a bit, i had to modify the code to log the output a bit ;)
12:37
<sbalneav>
mgariepy: Well, it's just dying on a flush, really, we should just wrap it in a try/expect block, so...
12:37
try:
12:37
self.device.flush()
12:38
execpt:
12:38
pass
12:38
<mgariepy>
ok, i'll except:\n pass ;)
12:38
<sbalneav>
right
12:38
give that a go, see what happens.
12:38
<mgariepy>
i wasn't sure if it was important
12:38
<sbalneav>
prolly not.
12:39
<mgariepy>
ok i'll keep you posted
12:39
<sbalneav>
sure! Lemme know.
12:39
I'm just looking at adding a pinger thread to xexit to rapidly detect a turned-off thin client.
12:40Egyptian[Home] has joined #ltsp
13:00vagrantc has joined #ltsp
13:00
<alkisg>
sbalneav: what does one need to do to compile xexit?
13:00* alkisg tried ./autogen.sh and ./configure but that failed...
13:04
<alkisg>
Ah, ChangeLog was missing...
13:04highvoltage has quit IRC
13:04thunsucker has joined #ltsp
13:18bix0r has quit IRC
13:24thunsucker has quit IRC
13:26thunsucker has joined #ltsp
13:26highvoltage has joined #ltsp
13:33dobber_ has joined #ltsp
13:33bobby_C has joined #ltsp
13:36
<sbalneav>
alkisg: I'm just playing with a threaded xexit now
13:36
<alkisg>
sbalneav: when X crashes, do X programs receive a SIGTERM or something? Wouldn't it be simpler then to just install a signal handler and wait for X to call it?
13:36
(or a XSetIOErrorHandler or similar...)
13:37
<sbalneav>
Well, that's what xexit does..., set XsetIOErrorHandler
13:37
<alkisg>
Why is an X property needed then?
13:37
<sbalneav>
problem is, you have to wait for the TCP timeout to happen
13:37
<alkisg>
(and a thread?)
13:37
Ah
13:38
Hmmm well if we're looking for "when X crashed, run a script to kill all user processes" handler, waiting wouldn't matter much, would it?
13:38
<sbalneav>
so, if I have some kind of thread in the background, that checks to see if the display's still valid every once in a while, that will handle the "turned off thin client" problem
13:39
No, so I'm figuring if we check every 10 or 15 secods or so.
13:39
that should be sufficient
13:39
<alkisg>
How long is the normal tcp timeout? a few minutes?
13:43
<sbalneav>
Depending on the linux distro? Anywhere from 20 minutes to 2 hours
13:48
<dobber_>
i think it was /proc/sys/net/ipv4/tcp_keepalive_idle
13:49
in seconds
14:00rad4Christ has joined #ltsp
14:01
<rad4Christ>
Good afternoon guys and gals. Thank you so much for your work on this project. We're looking to completely replace our school districts computer labs with LTSP and seemless RDP for our windows apps
14:02
<Lns>
rad4Christ: awesome!
14:02
<rad4Christ>
I do need a small bit of help. We're in the "lock down" phase of testing, and I've been able to successfully create a workable profile for users that is satisfactory for lockdown.
14:02
However, the login window allows me to choose my session, and thereby breaks this security.
14:03
<Lns>
rad4Christ: afaik there's no option (yet) to disable the choosing of a session.
14:03
<rad4Christ>
How can I FORCE it to only use the default windowmanager I want, or more to the point, remove that from under preferences altogether.
14:03
Lns: BAH! Is there a way I can remove/reduce the options inside?
14:04
<Lns>
rad4Christ: I don't think it's really configurable outside of the ldm code
14:04
<alkisg>
LDM_SESSION?
14:05
<rad4Christ>
Well, here's whereI am. I'm completely new to LTSP, and as I'm trying to configure, I can't even find the correct ltsp.conf file
14:05
<sbalneav>
I've had it on my TODO list for a while to add a variable to not show the sessions menu, but I haven't gotten around to it yet.
14:05
<rad4Christ>
I've looked everywhere I've seen documented, and I don't see it.
14:05
<alkisg>
distro/version?
14:05
<sbalneav>
rad4Christ: by default, there's usually not one shipped.
14:05
<alkisg>
sbalneav: won't LDM_SESSION force a session?
14:06
<rad4Christ>
If I could simply remove it, or remove options UNDER it, or somehow force the default, I'd be happy.
14:06
<sbalneav>
alkisg: it'll force a session, but the menu's still there, I beleive.
14:06
<alkisg>
I don't think that would matter much, would it rad4Christ?
14:06
(i.e. if you forced a session while the menu was still there...)
14:06
<rad4Christ>
BTW, netbook-remix runs really well over LTSP.
14:06
alkisg: true, removal would be best.
14:06
<sbalneav>
rad4Christ: Like I say, at the moment, there's no option to make the menu disappear.
14:07
<alkisg>
Even if the menu is there, won't .dmrc be processed?
14:07
I.e. LDM_SESSION would be required even then...
14:07
*if the menu *isn't there
14:07
<rad4Christ>
sbalneav: Thanks. As a work around, any ideas why, if it is changed, I can no longer get the user to login with the settings I've set up in sabayon?
14:07
<sbalneav>
alkisg: yeah, you'd be right there.
14:08
rad4Christ: What do you mean?
14:09
<rad4Christ>
Using the Proflle editor (sabayon), I've set up a custom interface/menu structure/lockdown policies for a usergroup. If a user in that user group logs in, they get the layout I prepared...
14:09
However, if I change the session at login, it always logs into a basic GNOME session, with none of my customizations.
14:10
Then, after logging out of the user, setting the session to "Default" or ANY other session, I can no longer get to the custom profile setting
14:11
Knowing my limited knowledge of linux and users, I basically delete the user and all settings, and recreate it to get it to work functionaly.
14:11
As much as we want to use LTSP, this is kind of a big thing, as it completely undo's what we're trying to put together for the student's experience.
14:12
<sbalneav>
rad4Christ: Well, whatever session you're setting isn't calling the standard XSession script, and sabayon's only started from that.
14:12
as sabayon's called within the /etc/X11/Xsession.d script dir.
14:13bobby_C has quit IRC
14:14
<rad4Christ>
I gotcha.
14:15
sbalneav: Thanks for the help.
14:16
I guess until we can disable the menu, I'm not sure what we can do. My coworkers are much more knowledgeable of linux than I, I'll pass this onto them to figure out. Thanks for your time!
14:16
<cliebow>
rad4Christ, and sbalneav takes donations of moosemeat 8~)
14:16
delivered to his camp
14:17
<rad4Christ>
Plenty of deer around here, even a small farm with buffalo, but no moose.
14:17
I'll get it special ordered!
14:18
<sbalneav>
rad4Christ: you could always make a quick patch to the greeter code for LDM, simply remove the session menu.
14:18
<rad4Christ>
Where's it located, I don't mind takign a swing at it.
14:19
sbalneav: I saw this: http://ubuntuforums.org/showthread.php?t=889951
14:20bobby_C has joined #ltsp
14:22
<sbalneav>
rad4Christ: That's an older thread, the code's changed since then, but it is the ldmgtkgreet.c file that needs tweaking.
14:23
<rad4Christ>
OK. Thanks
14:29
<Lns>
sbalneav: how tied to ubuntu specifically is sabayon? Would it be a major pita to get it to work in debian?
14:29
<johnny>
not at all
14:29
it's not at all tied to ubuntu
14:29
it's generic
14:29
<Lns>
I'd love to try it out on my lenny server
14:29
<johnny>
it's only tied to gnome
14:30
<Lns>
i gues i'll just downlaod the deb ;)
14:31
man... globalwarming.house.gov is freakin hammered right now
14:31
i guess everyone wants to watch the oil cam
14:33bobby_C has quit IRC
14:33
<sbalneav>
Lns: it does work in Debian
14:33
Also Fedora
14:34
in fact, it should work on any Linux that supports gnome.
14:34
<Lns>
wootzorz!
14:34
thx =)
14:35
<alkisg>
Lns, how's Debian working out for you?
14:35
Got any ltsp labs working with it?
14:35
<Lns>
loving it
14:35
no hardcore labs besides my own office
14:35
but i'm planning on converting them this summer
14:35
<alkisg>
Woah, nice... reasons/comparison etc?
14:35
<Lns>
i'm also running ubu 10.04 at home for my media client downstairs and that is working out really well too
14:36
debian doesn't change much. :) It just gets more stable.
14:36
i need that kind of consistency for my techs/teachers not to breathe fire down my neck.
14:36
<alkisg>
Coming from the "I'm sticking to LTS releases" guy, that means a lot...
14:36
<Lns>
well
14:37
debian stable is what lts SHOULD be imho..but it's not.
14:37
<johnny>
sad thing is.. ubuntu could be more stable by using newer packages in many cases
14:37
<Lns>
Ubuntu LTS is just another release, supported for longer. Debian stable is stable WHEN it's released.
14:37
<johnny>
since they don't seem to have the ability to backport all the fixes that are necessary
14:38
<sbalneav>
HAHAHAHAHAHAHAHA
14:38
<alkisg>
Well lts releases sync from testing, and that supposedly makes them more stable even before freeze, but yeah, some features aren't ready for prime time in ubuntu...
14:38
<sbalneav>
\o/
14:38
bbiab, got xexit pinger going!!! :)
14:38
<alkisg>
Go go sbalneav :)
14:38
<Lns>
hehe!
14:40
alkisg: I am by far, no ubuntu basher now that i chose debian over it... hell i still use it! Its just a bit too much for me to handle with multiuser servers.
14:40
especially ones that i get paid to support.
14:40
I probably won't go with replacing ubuntu at home like i was going to initially. Ubuntu is great.
14:41
<alkisg>
Lns, don't worry, we're all grown ups here, just choosing what best fits our needs for each particular installation. I just wanted to listen your experience, as we have somewhat similar background on ubuntu...
14:41
<Lns>
for sure
14:42
it's all linux underneath anyway.
14:42
<alkisg>
And you're using debian stable? Doesn't that feel too old?
14:42
<Lns>
not with backports for gnome, firefox(iceweasel), ltsp and openoffice :) brand spanking new in fact
14:42
<alkisg>
Ah, yup, backports sound nice
14:43
Are you using the default theme?
14:43
(just curious :D)
14:43
<Lns>
gtk/gnome theme?
14:43
<alkisg>
Yes
14:43
<Lns>
i think so. though i need to find out whether you can import the human/light themes into debian. anyone do that?
14:43
<alkisg>
I tried the testing live cd today, and I thought it was a little '90s...
14:44
<Lns>
i'm just using clearlooks
14:44pmatulis has quit IRC
14:45muttley has joined #ltsp
14:45
<muttley>
how can start for make xbox 1 ltsp client
14:47
<Lns>
gotta jet..ttyl guys
14:47
<muttley>
(xbox 1 have many linux distribution (http://www.xbox-linux.org/wiki/Download) and etherboot work)
14:49bobby_C has joined #ltsp
14:50cliebow has quit IRC
14:52
<muttley>
etherboot, after load nbi.img freeze ...
14:53
<sbalneav>
muttley: Not sure, I don't think anyone here's ever done any work with trying to turn an XBOX into a thin client.
14:53
<muttley>
ok...
14:54
however, i think xbox is good thin client: cheap, silent, low power and a nomrla x86 architetture
14:54
<alkisg>
muttley: nbi.img? why don't you try pxelinux.cfg?
14:54
*pxelinux.0?
14:55artista_frustrad has quit IRC
14:55
<alkisg>
!rom-o-matic
14:55
<ltspbot`>
alkisg: Error: "rom-o-matic" is not a valid command.
14:55
<alkisg>
!romomatic
14:55
<ltspbot`>
alkisg: Error: "romomatic" is not a valid command.
14:55
<alkisg>
Damn
14:55
!gpxe
14:55
<ltspbot`>
alkisg: "gpxe" :: the successor to etherboot (http://etherboot.org). To download bootable floppy, cd, usb, kernel etc images, visit http://rom-o-matic.net/
14:55
<muttley>
xbox have etherboot...
14:56
<alkisg>
? not sure what you mean...
14:56
Do you mean "boot from lan"?
14:56
<muttley>
a program called Xromwell...it's a linux bootloader
14:56
<alkisg>
Right, that's not etherboot
14:56
OK, just change your dhcpd.conf then, to provide pxelinux.0 instead of nbi.img
14:57
<muttley>
xromwell support boot from cd, fatx (xbox filesystem) and lan (etherboot)
14:57
<alkisg>
Aaaah got it. Which version?
14:57
<muttley>
i try...but xbox (xromwell) say "bad image"
14:57
<alkisg>
Hrm. OK, temporarily, can you try with a cd?
14:57
(gpxe from rom-o-matic)
14:58
<muttley>
if i send in dhcp.conf (dnsmasq.conf in my case) nbi.img load, but freeze
14:58
<alkisg>
There are some known problems with nbd.img... which distro/version are you using?
14:58
*nbi.img
14:58
<muttley>
ubuntu
14:58
8.4
14:58
<alkisg>
Hrm... it should be working there...
14:59
Do try with that cd: http://rom-o-matic.net/gpxe/gpxe-1.0.0/contrib/rom-o-matic/
14:59
<muttley>
but i thought there are problem in kernel
14:59
<alkisg>
What's the last message that you see?
14:59
<muttley>
on xbox?!
14:59
<alkisg>
Yes
15:00
<muttley>
dhcp find....
15:00
load /ltsp/i386/nbi.img ............................................
15:00
...........................
15:00
.....................
15:00
and many athers point :)
15:00
<alkisg>
Heh
15:01
OK, can you try with the gpxe cd?
15:02
<muttley>
mhhh...
15:02
gpxe make a bootable img for cd...bootable for my xbox?!?!
15:03
<alkisg>
I believe so. Isn't it a standard x86 pc?
15:03
gpxe is the newer version of etherboot...
15:04
<muttley>
is a celeron...but boot depend from bios isn't it?!
15:04
i don't know the xbox bios
15:04
sorry...my english is bad
15:05
if i insert a normal livecd (linux) in my xbox NOT work!
15:05
<alkisg>
No it doesn't depend on the bios when you boot from a cd
15:06
Well it wouldn't take more than a few minutes to try it... and if etherboot works, gpxe should also work.
15:07
<muttley>
i have search along on google...but no someone try your procedure
15:07
i don't think is possible
15:07
for start i must use xromwell
15:07artista_frustrad has joined #ltsp
15:07
<muttley>
and support only etherboot
15:07
<alkisg>
Didn't you say that xromwell supports booting from a cd?
15:08
<abeehc>
http://www.xbox-linux.org/wiki/Making_a_Linux_distribution_Xbox_compatible_HOWTO
15:08
<muttley>
boot for cd yes...but only for modded linux distribution!
15:09
i read...but i'd made modded ubuntu distribution with lstp client support...i can't :(
15:09
<alkisg>
Erm, and you think it'll be able to boot a modded linux distribution but not gpxe? weird..
15:09
Anyway, I don't have to propose anything else, so bbl.
15:09
<muttley>
too difficult for me
15:10
can you read this post: http://forums.gentoo.org/viewtopic-t-292982-highlight-diskless+xbox.html
15:11
this man make what i want do...but not for ltsp
15:15
use xbox for MythFrontend..."My xBox now boots Etherboot from Cromwell and loads Gentoo with an NFS root and a FatX Swap."
15:16
<sbalneav>
aaaargh. Why is getting pthreads support so %!&% hard in autotools?!
15:30rad4Christ has left #ltsp
15:31
<muttley>
http://www.xbox-linux.org/wiki/FAQ#Can_I_use_my_XYZ_Linux_CD_to_install_Linux.3F
15:32
http://www.xbox-linux.org/wiki/FAQ#The_Xbox_is_a_standard_PC.2C_isn.27t_it.3F
15:33
(this is beacause gPXE don't work on xbox...i think...)
15:34
i think it's needed recompile kernel sended by ltsp server to client for be compatible with xbox....
15:45
<sbalneav>
well, no, it's probably a recompiled gpxe you'd need.
15:47
If anyone's interested, the current version of xexit on the site now detects a "dead" display (i.e. unplugged). I created a "30killall" script for Xexit.d that just does a "pkill -9 -u $USER", and within 15 seconds of unplugging the thin client, all of it's processes are gone.
15:47
later on tonight I'll "productionize" the code a bit.
15:49komunista has quit IRC
15:49
<muttley>
ok sorry for the insistence ... and thanks for the fantastic LTSProject!
15:50
<sbalneav>
muttley: No problem, I think it's just a case that none of us here have ever done anything with an xbox
15:51
'sides, the entire point of a thin client is that it's reliable. The whole "Red Ring of Death" thing doesn't sort of fit into that :)
15:51mgariepy has quit IRC
15:54
<muttley>
ok..you have right :) but RROD is the x360...xbox1 is reliable :)
15:54bur has joined #ltsp
15:58Mava has quit IRC
16:14Eghie has quit IRC
16:16thunsucker has quit IRC
16:16bur has quit IRC
16:26bobby_C has quit IRC
16:32Mava has joined #ltsp
16:44muttley has quit IRC
16:46vagrantc has quit IRC
16:51Egyptian[Home] has quit IRC
16:54pmatulis has joined #ltsp
16:55johnny has left #ltsp
17:01
<alkisg>
man tcp
17:01
tcp_keepalive_time (integer; default: 7200; since Linux 2.2)
17:01
The number of seconds a connection needs to be idle before TCP begins sending out keep-alive probes. Keep-alives are only sent when the
17:01
SO_KEEPALIVE socket option is enabled. The default value is 7200 seconds (2 hours). An idle connection is terminated after approximately
17:01
an additional 11 minutes (9 probes an interval of 75 seconds apart) when keep-alive is enabled.
17:02
After exactly that amount of time (2 hours + 11 minutes), my thin client "ghost processes" were terminated automatically
17:02
sbalneav, dobber_: so, if the processes terminate after 2 hours, then there's no need for a "gnome-watchdog"?
17:04artista_frustrad has quit IRC
17:05leio_ has joined #ltsp
17:08Guerdal82_ has quit IRC
17:08leio has quit IRC
17:08
<ogra>
alkisg, right ... the prob is that you pointlessly occupy resources for 2h
17:09
and lowering the default might have unexpected sideeffects for other server processes that rely on the 7200 secs
17:20johnny has joined #ltsp
17:32
<alkisg>
ogra, no, I wasn't thinking to lower it. But it's not a big deal if the processes are only there for 2 hours. The problem is with the processes that don't have tcp sockets or are not childs of the ssh process, which stay around even after those 2 hours
17:33
So, one simple process which would kill the others when that one died, would be enough I think....
17:33artista_frustrad has joined #ltsp
17:38
<alkisg>
Hmmm and maybe killing only the processes with the same XDG_SESSION_COOKIE would be better.
18:00bbd has quit IRC
18:07alkisg has quit IRC
18:12otavio has quit IRC
18:23bbd has joined #ltsp
18:25staffencasa has quit IRC
18:26F-GT has quit IRC
18:30dobber_ has quit IRC
18:30artista_frustrad has quit IRC
18:35vagrantc has joined #ltsp
18:42F-GT has joined #ltsp
18:44artista_frustrad has joined #ltsp
19:01Guerdal82_ has joined #ltsp
19:02artista_frustrad has quit IRC
19:13BenChapman has joined #ltsp
19:16artista_frustrad has joined #ltsp
19:18vagrantc has quit IRC
19:22leio_ is now known as leio
19:31artista_frustrad has quit IRC
19:44artista_frustrad has joined #ltsp
19:45lucascoala has joined #ltsp
20:01artista_frustrad has quit IRC
20:14artista_frustrad has joined #ltsp
20:39artista_frustrad has quit IRC
20:53artista_frustrad has joined #ltsp
21:02artista_frustrad has quit IRC
21:02pmatulis has quit IRC
21:04slidesinger has joined #ltsp
21:15artista_frustrad has joined #ltsp
21:59johnny has left #ltsp
22:03BenChapman has quit IRC
22:06artista_frustrad has quit IRC
22:16ogra_cmpc has quit IRC
22:18BenChapman has joined #ltsp
22:19artista_frustrad has joined #ltsp
22:30ogra_cmpc has joined #ltsp
22:35artista_frustrad has quit IRC
22:48artista_frustrad has joined #ltsp
23:03artista_frustrad has quit IRC
23:16artista_frustrad has joined #ltsp
23:21Guerdal82_ has quit IRC
23:21artista_frustrad has quit IRC
23:34artista_frustrad has joined #ltsp
23:42artista_frustrad has quit IRC
23:55artista_frustrad has joined #ltsp