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


Channel log from 20 August 2008   (all times are UTC)

00:03shogunx has quit IRC
00:28shogunx has joined #ltsp
00:48jonnor has joined #ltsp
01:05Arch4ngel has quit IRC
01:05soneyka has quit IRC
01:20mikkel has joined #ltsp
01:20
<warren>
Too bad LTSP doesn't have a registered trademark
01:21
cannot defend uses of LTSP that are confusingly similar but different
01:26
<jcastro>
warren: just sent you a mail wrt. the kiwi thing
01:32cyberorg has joined #ltsp
01:47
<lejo>
euhm we don't have a IP_SERVER in our lts.conf (on opensuse)
01:48
the lts.conf is nothing different then what is upstream
01:58
<cyberorg>
lejo, it is just SERVER
01:58
<lejo>
indeed
02:06tjikkun_work has joined #ltsp
02:23Ryan52 has quit IRC
02:23Ryan52 has joined #ltsp
02:45vagrantc has quit IRC
02:54btil_ has joined #ltsp
02:55btil_ is now known as btil
02:59Ryan52 has quit IRC
03:00Ryan52 has joined #ltsp
03:01btil is now known as btil_
03:02btil_ is now known as btil
03:05btil has left #ltsp
03:07btil has joined #ltsp
03:27Q-FUNK has joined #ltsp
03:51Q-FUNK has quit IRC
05:22alekibango has joined #ltsp
05:47ogra has quit IRC
05:49Jenna has joined #ltsp
05:50Jenna has left #ltsp
05:54J45p3r___ has joined #ltsp
05:55etyack has joined #ltsp
06:01ogra has joined #ltsp
06:10Q-FUNK has joined #ltsp
06:32Pascal_1 has joined #ltsp
06:43coolio has joined #ltsp
06:45
<coolio>
how can i solve a server boot issue; "kinit: No resume image, doing normal boot...." and it just hangs
06:45Pascal_1 has left #ltsp
06:48gate_keeper_ has joined #ltsp
06:49bobby_C has joined #ltsp
06:52
<coolio>
i am using hardy : it also times out on another terminal login
06:57
<ogra>
thats a proper message
06:57
no error
06:57
it just tells you that there was no suspend before booting ... so it doesnt have a resume image
06:57
your prob lies elsewhere
07:02lejo has quit IRC
07:02J45p3r___ has left #ltsp
07:03mhterres has joined #ltsp
07:03
<ogra>
coolio, does the server properly boot to a login gui ?
07:07
<coolio>
nope no gui
07:07
hi ogra
07:07
i just tried fixing X server, that did not work either
07:08
i know have root login, any suggestions
07:08
<ogra>
so if it boots, do you get to a non graphical login prompt ?
07:08
(on the server)
07:08
<coolio>
yep
07:09
root@lnix:/#
07:09
<ogra>
and you had a gui login before ?
07:09
<coolio>
yes before we booted the machine
07:09etyack has quit IRC
07:09
<coolio>
been up for 7days
07:09
<ogra>
??
07:09
ah
07:09
sounds like you are out of diskspace or some such
07:10
<coolio>
did not want to read usb so we rebooted and then we got this message
07:10
<ogra>
which message ?
07:12
<coolio>
got this bug report from launchpad but we run hardy
07:12
https://bugs.launchpad.net/ubuntu/+bug/103148
07:12
<ogra>
did you check your diskspace ?
07:12
thats no error as i said above
07:13
you have it on every boot
07:13
df -h should tell you how much space you have left
07:13
the kinit message is unrelated
07:15
<coolio>
nope 160gb drive clean ubuntu+edubuntu install nothing else
07:16
i just rebooted and it hangs on " *Starting GNOME Display Manager
07:17
<ogra>
if you log in as a normal user, can you touch files in /tmp ?
07:18
and are your disks mounted rw ?
07:18
might be a HD failure ... ext3 will automatically be remounted readonly in such a case
07:18
<coolio>
it times out when I try to logon in terminal 2
07:19
<ogra>
(thats why using root is just silly, on normal usre accounts you would immediately get the proper error message on login)
07:19
<coolio>
so do you think it my hdd
07:19
<ogra>
might be
07:20
do you have /home on a separate partition ?
07:20
do you have any weird ldap setup that could block ?
07:20etyack has joined #ltsp
07:20
<coolio>
nope straight forward single partition
07:21
<ogra>
is there any error message if it times out on normal user logins ?
07:22
(we are only talking about the server here atm, right ? you dont have a thin client involved in all that atm ?)
07:23
<coolio>
nope
07:23
not at all
07:23
they wont boot as the server has not started up
07:24
think we need to reinstall
07:24
<ogra>
well, if there is a HW issue you should rather find it
07:25
the behavior is realy very similar to a full disk or a readonly filesystem, but if you are sure neither is the case and the usual logfiles like dmesg, messages and syslog dont tell you anything about errors ....
07:25
<coolio>
weird it just came back after a reboot
07:26jimjimovich has joined #ltsp
07:26
<coolio>
I have no clue what the heck just happend
07:27
<jimjimovich>
Hi Guys. Made lots of progress on our setup today, just one more problem. Need your help! I can't seem to make monitor resolutions for specific terminals work in my lts.conf file
07:30
something like X_MODE_0 = "1280x1024"; does not work. I did get it to work once, but then after logging into the terminal it reverted to 1024x768
07:30
<ogra>
did you specify the right X_HORZSYNC and X_VERTREFERSH values ?
07:30
<jimjimovich>
ogra: I haven't been able to get that to work at all :(
07:30
<ogra>
very likely your monitor is nor reporting the proper DDC values
07:30
so if you set the right ones i wil start to work
07:30
look in your monitor manual
07:30
<jimjimovich>
what would X_HORZSYNC and X_VERTREFERSH be for an LCD panel?
07:31
<ogra>
look in your monitor manual
07:31
*X_VERTREFRESH sorry, typo
07:31Q-FUNK has quit IRC
07:32
<jimjimovich>
what are all the values that can be set for a monitor? I only saw X_HORZSYNC and X_VERTSYNC before
07:32
<ogra>
where did you see X_VERTSYNC ?
07:32
that was definately never used anywhere
07:33
<jimjimovich>
ok, i'm dumb. sorry
07:33
i didn't see it
07:33
<ogra>
the values need to be n your monitor handbook
07:33
<coolio>
hi ogra tx
07:33
<jimjimovich>
i was messing with those values earlier on some older monitors, and was never able to get it to work. perhaps it will work with this lcd. will give it a try.
07:33
<ogra>
alternatively you can plug the monitor into a system where you know you get he desired resolution and run xvidtune ... it shows the values
07:35GodFather has joined #ltsp
07:41etyack_ has joined #ltsp
07:43leio_ is now known as leio
07:46alekibango has quit IRC
07:47lejo has joined #ltsp
07:52
<coolio>
hw do I check if dhcp is running
07:54Subhodip has joined #ltsp
07:55GodFather has quit IRC
07:57
<jimjimovich>
ogra: got the values from xvidtune, but still no luck. it seems that it makes no difference on any of my monitors
07:57etyack has quit IRC
07:59_UsUrPeR_ has joined #ltsp
08:03
<coolio>
weird problem now i cant even sudo to run ltsp-update-image
08:03
just hang with authorisation
08:03GodFather has joined #ltsp
08:04
<ogra>
any hints in the logs?
08:05makghosh has joined #ltsp
08:07
<jimjimovich>
ogra: logs for monitors?
08:07
<ogra>
jimjimovich, no, i was talking to coolio
08:08
<jimjimovich>
ogra: is there somewhere i can paste my lts.conf for you to look at? maybe i'm making some simple mistake ...
08:09
<ogra>
!pastebot
08:09
<ltspbot>
ogra: "pastebot" is The LTSP pastebot is at http://pastebot.ltsp.org. Please paste all text longer than a line or two to the pastebot, as it helps to reduce traffic in the channel. A link to the content will be pasted in the channel.
08:09
<ogra>
^^^^
08:10
<ltsppbot>
"jimjimovich" pasted "[Default] X_MODE_0 = "1024x768" (7 lines) at http://pastebot.ltsp.org/45
08:12
<ogra>
did you try without the default section ? and are you sure the HORZSYNC, VERTREFRESH values are actually correct ?
08:13
<jimjimovich>
yes, both from the manual and from xvidtune on a working computer
08:13
yes, tried without the default
08:13
<ogra>
and the MAC is correct as well ?
08:13
<jimjimovich>
i tried it with several CRT monitors earlier today, and also had no luck
08:13
yes
08:13
<ogra>
what distro is that ?
08:14
<jimjimovich>
ubuntu 8.10.1
08:14
8.04.1
08:14
<ogra>
8.10 ??
08:14
ah :)
08:14
<jimjimovich>
sorry
08:14
<ogra>
and your lts.conf is located properly in the tftp directory ?
08:14
<jimjimovich>
it's in /var/lib/tftpboot/ltsp/i386/lts.conf
08:14
<ogra>
ok
08:15
what graphics HW do your clients have ?
08:15
<jimjimovich>
well, that's the problem, they're all kind of random
08:15
but at this point, i'm just looking to find a newer motherboard (with integrated garphics) and LCD combo that works well together
08:16jammcq has quit IRC
08:16
<jimjimovich>
so that we can buy a bunch of the same ones
08:16
<ogra>
err ...
08:16
drop the smicolons from the lines in lts.conf
08:16
*semicolons
08:16jammcq has joined #ltsp
08:16
<jammcq>
hello friends
08:16
<ogra>
that definately wont work
08:16
hey jammcq
08:16
<jammcq>
hey
08:17
<ogra>
also add SCREEN_02=shell and SCREEN_07=ldm to your default section
08:17
<_UsUrPeR_>
morning all
08:17
<ogra>
that way you can swtich to tty2 on the booted client and check the generated xorg.conf that the values you have set actually end up in ther
08:18
<jimjimovich>
ogra: dang semicolons! that's what i get for working on lts.conf and dhcp conf at the same time
08:18
<ogra>
heh
08:18
<jimjimovich>
ok, great. will try that out now
08:22coolio has left #ltsp
08:26_UsUrPeR__ has joined #ltsp
08:39
<jimjimovich>
ogra: you're a genius! thanks!
08:39
<ogra>
;)
08:40
<jimjimovich>
now, if my hardware wasn't so old and junky. hehe
08:40
got some 12 year old computers here running at terminals. kinda nuts
08:40
<ogra>
well, butthey should do the job :)
08:41_UsUrPeR_ has quit IRC
08:41etyack__ has joined #ltsp
08:41
<jimjimovich>
they do for the most part. it's just a matter of finding the right monitor/video card combos
08:43
which leads me to another issue. lets say you have different monitors running at different resolutions. is it possible to make the user screen resolution stick to the machine. right now, if i log into a machine at 800x600, the next machine i go to also tries to use 800x600
08:43etyack_ has quit IRC
08:43
<ogra>
you mean when you select the screen in the gnome tool ?
08:43
s&screen/size/
08:43
<jimjimovich>
yes
08:44
<ogra>
thats a per user setting
08:44
stored in gconf
08:44
you could try to lock down the gconf setting for it ... dig a bit in gconf-editor
08:44
so users are not able to set it
08:44
<jimjimovich>
ok, but i still have the problem of having different resoltuions on different terminals
08:45
<ogra>
then it will always use the X setup
08:46
<jimjimovich>
ok, i'll have to mess with that tomorrow. thanks for the pointer. it's been a long day of LTSP!
08:49
oh, one last question. is there anything that can be done about the long login times? seems like it takes about 30-60 seconds to check the password
08:52etyack__ has left #ltsp
08:53elisboa has quit IRC
08:53elisboa has joined #ltsp
08:55_UsUrPeR__ has quit IRC
08:59mikkel has quit IRC
09:00_UsUrPeR__ has joined #ltsp
09:04cliebow has joined #ltsp
09:05nicoAMG has joined #ltsp
09:07
<jimjimovich>
anyone used any Intel® GMA 3000 integrated video? i know the 950 works well, so hopefully this would too
09:09hp has joined #ltsp
09:10sbalneav has joined #ltsp
09:10
<sbalneav>
Morning all
09:10
<hp>
hello ogra, thanks a lot, I tried ubuntu ltsp, and it's great
09:12
do have one more question though, i tried configuring a locally attached Epson LX300II to a thin client
09:13
on the lts.conf, I added the line under the mac address, PRINTER_0_DEVICE=/dev/lp0
09:13
don't notice anything happening at all
09:14
<ogra>
you need to configure th eprinter in the user session
09:14
PRINTER_0_DEVICE=/dev/lp0
09:14
<hp>
checked on the logs but am not really sure what to look for, there doesn't seem to be anything pointing to an error
09:14
<ogra>
will just run the jetdirect forwarder
09:15
is that a parallel printer ?
09:15
<hp>
yes, it is
09:15
<ogra>
ok
09:15
<hp>
it is a parallel printer
09:15
<ogra>
well, in the user session select to add a new printer in the print tool
09:15
<hp>
I did, and then I chose local printer
09:15
<ogra>
make that a jetdirect printer and point it to the IP of the client
09:16
no, thats wrong
09:16
local printer would be something attached to the server
09:16
<hp>
ok, will follow your advise
09:16
<ogra>
(your user session runs on the server)
09:16_UsUrPeR__ has quit IRC
09:16GodFather has quit IRC
09:17
<ogra>
it needs to be a "appsocket/jetdirect" printer
09:17_UsUrPeR__ has joined #ltsp
09:17
<hp>
I tried that
09:17
but go on ogra, I must've missed something earlier or did something wrong
09:17
<ogra>
and point to port 9100 of the client ... you find the ip of every client at the bottom right in the login window if nobody is logged in
09:17
<hp>
yup
09:18
<ogra>
i sadly have only usb printers
09:18
so i cant confirm parallel ones work, but usb works exactly like that here
09:18
(with /dev/usblp0 indeed)
09:18Q-FUNK has joined #ltsp
09:19
<hp>
I'm sure usb printers work, however, the company I work for here in the Philippines use mostly parallel dot matrix printers :)
09:19
<ogra>
well
09:19
add SCREEN_02=shell and SCREEN_07=ldm to your default section in the lts.conf
09:20
<hp>
one more thing, I cannot do the configuration via the thin client itself?
09:20
<ogra>
that will give you a shell on tty2 on the runing thin client
09:20
then check that /dev/lp0 actually exists
09:20
<hp>
will do :)
09:21
<ogra>
if thats the case check the processlist if the jetpipe program is running (thats the program doing the forwarding from port 9100 to /dev/lp<something>)
09:21
<hp>
I will
09:22
<ogra>
you can indeed edit the lts.conf file from a clent if you are logged in ... since you are actually logged in to the server (so you edit /var/lib/tftpboot/ltsp/i386/lts.conf on the server in fact)
09:24
<hp>
I meant I tried to configure the printer via a user session on a thin client, using appsocket..... option, I was wondering if I should've done it directly on the terminal server itself. But then, it's the same thing right?
09:30
<ogra>
right
09:31
and you should actually do it in a rinning client session, since you will sit next to the printer then :)
09:32
<hp>
I'm just happy I was on the right track earlier, I will re-do it tomorrow. Thanks so much. :)
09:34GodFather has joined #ltsp
09:35Q-FUNK has quit IRC
09:40bindo has joined #ltsp
09:41
<cliebow>
Hi Ron!!
09:44
<bindo>
Hi there, I need some help. I'm trying to migrate my whole company to and LTSP setup based on kubuntu. All up and running except twinkle
09:44
unfortunately we are totally dependent on VOIP
09:44
<sbalneav>
ogra: Hey, how you feeling today?
09:44
<jammcq>
scotty !!!!!!!!!!!!!!!!!!
09:44
<sbalneav>
Hey jammcq
09:45
<bindo>
I think the best thing would be to have the voip softphone run directly on the client but Localapps documentation looks quite old
09:45
<sbalneav>
How's it going? I replaced my cablemodem yesterday, so that should improve drevil.
09:45
<bindo>
any help for LOCALAPPS on LTSP 5
09:45
<sbalneav>
bindo: localapps is being actively developed right now, but isn't ready for prime time just yet.
09:46
<bindo>
I am willing to mess a little bit or do you think it really isn't reasonable ?
09:47
<sbalneav>
I've done some work on it, but I think wtogami's probably messed with it more at this point than I have. I'd say "not yet", but you might want to see if you can grab the latest source and bodge something together.
09:48
There's 1) no docs. and 2) it's not finished yet.
09:48
<bindo>
ok maybe a little too risky for a production environment :)
09:49
<sbalneav>
I'd say so yeah :)
09:58
Off to a meeting
09:58jimjimovich has left #ltsp
10:00bindo has left #ltsp
10:07hp has quit IRC
10:12tjikkun_work has quit IRC
10:21_UsUrPeR__ is now known as _UsUrPeR_
10:23btil has quit IRC
10:26staffencasa has joined #ltsp
10:31
<ogra>
sbalneav, fine, thanks
10:36etyack_ has joined #ltsp
10:40
<dberkholz>
johnny: wolf was looking for catalyst/genkernel feature suggestions on the gentoo-catalyst list a few weeks ago. if you joined it, you could suggest some more
10:41
johnny: http://archives.gentoo.org/gentoo-catalyst/msg_0669a9c02fccd976d61d8306766e52ec.xml
10:42yanu_ is now known as yanu
10:50mccann has joined #ltsp
11:19btil has joined #ltsp
11:21vagrantc has joined #ltsp
11:23etyack_ has left #ltsp
11:46Q-FUNK has joined #ltsp
11:54Subhodip has quit IRC
12:01bobby_C has quit IRC
12:10spectra has joined #ltsp
12:25gate_keeper__ has joined #ltsp
12:26Faithful has quit IRC
12:32_UsUrPeR_ has quit IRC
12:36mccann has quit IRC
12:38gate_keeper_ has quit IRC
12:38barbosa has joined #ltsp
12:39gate_keeper__ has quit IRC
12:46_UsUrPeR_ has joined #ltsp
13:06Q-FUNK has quit IRC
13:10Q-FUNK has joined #ltsp
13:10spective has joined #ltsp
13:10
<spective>
hrm
13:12
<Q-FUNK>
hm
13:12
<spective>
It's almost as if my client build is mocking me.
13:12
I'm trying to get local apps working again...
13:12
There's some issue with ssh...
13:12
I can't directly ssh into a client, since it can't connect to port 22 ...
13:13
Which leads me to believe that the ssh daemon is not running...
13:13
However, it is installed...
13:13
But I can't check if its actually running...
13:13
because I can't get a shell...
13:13
I have SCREEN_02=shell in /etc/lts.conf
13:14
But when I ctrl-alt-F2, it's blank.
13:14
On tty1 there's a login prompt. Which won't let me log on.
13:15
So I don't know what to do atm.
13:15
<cliebow>
spective:have you enabled root in the chroot?
13:15
<ogra>
/etc/lts.conf ??
13:16
<spective>
cliebow, erm, I believe so... It has a password...
13:16
ogra: is there somewhere else?
13:17
<ogra>
depends how old your ltsp is :) we use lts.conf from the tftp dir since a year or so
13:17
<spective>
This was an upgraded ltsp install from about 1.5 years ago. I assumed it still used /etc/lts.conf
13:17GodFather has quit IRC
13:17fie_wr0k has quit IRC
13:18GodFather has joined #ltsp
13:18fie_wr0k has joined #ltsp
13:18
<cliebow>
spective:you are using nbd to mount an image? or nfs?
13:18
<spective>
to be precise /opt/ltsp/i386/etc/lts.conf
13:18
The client "environment" is nfs mounted from /opt/ltsp/i386
13:19
I think...
13:19
<cliebow>
ubuntu??must be quite old then..
13:19
<spective>
It was em... feisty I think.
13:19
<cliebow>
you can look in /etc/inetd.conf or summat to see if nbd is enabld..
13:19
<spective>
I've just ripped it up to hardy, and well, it's been a bit spotty.
13:19
<vagrantc>
if it was feisty, that was before the nbdroot stuff
13:20
<spective>
meh.
13:20
<vagrantc>
you upgraded the chroot itself, or re-created it?
13:20
<spective>
Recreated the chroot (heh, after trying to upgrade it.)
13:21
<ogra>
right, then use /var/lib/tftpboot/ltsp/i386/lts.conf
13:21
<vagrantc>
then it should be using nbd
13:21
and lts.conf in tftp
13:23
<spective>
Just looked in /etc/inetd.conf...
13:23
9572 stream tcp nowait nobody /usr/sbin/tcpd /usr/sbin/nbdswapd
13:23
2000 stream tcp nowait nobody /usr/sbin/tcpd /usr/sbin/nbdrootd /opt/ltsp/images/i386.img
13:23
<ogra>
yeah, thats fine
13:24
<spective>
Is there some hardlinking going on between the directories /opt/ltsp and /var/lib/tftpboot/ltsp?
13:24
<ogra>
just make sure to run ltsp-update-image if you install stuff in the chroot and use the lts.conf from the tftp dir
13:24
no
13:24
<spective>
ltsp-update-image copies /opt/ltsp stuff into /var/lib/tftpboot/ltsp??
13:25
<ogra>
no
13:25
thats ltsp-update-kernels
13:25
(which was already in feisty and before)
13:25
ltsp-update-image just rolls a new image fro the chroot
13:25
<spective>
ok. Let me get this straight. heh. If I update /opt/ltsp/i386 ... I need to run ltsp-update-image? And the lts.conf that I use needs to be stored in /var/lib/tftpboot/ltsp/i386?
13:26
<cliebow>
yes
13:26
<spective>
ooohkay. brb.
13:27
<cliebow>
spective..if you use wireshark you can see lts.conf being requested and d/l ed via tftp
13:27
<spective>
I imagine so. I'll take your word for it.
13:28
Does a copy of lts.conf need to remain in /opt/ltsp/i386/etc ?
13:28
<cliebow>
nope..
13:28* cliebow cliebow duscks
13:29* vagrantc ponders if it wouldn't be simpler for ltsp-update-kernels to copy /opt/ltsp/i386/etc/lts.conf along with the kernels, so that people can just maintain it in the historical place...
13:29
<cliebow>
err ducs
13:29
err ducks
13:29
<spective>
quack.
13:29
<ogra>
vagrantc, ??
13:29
vagrantc, that wouldnt work
13:29
<spective>
warum?
13:29
<vagrantc>
ogra: it wouldn't work any worse than it works for the kernels...
13:30
<ogra>
the kernels are not inside the image
13:30
<vagrantc>
ogra: would have to run ltsp-update-kernels after editing it, of course.
13:30* vagrantc never liked moving lts.conf's location
13:30
<ogra>
the prob here is that you get two out of sync files if you roll an image
13:30* ogra would love to make /etc/lts.conf die die die :)
13:31
<spective>
Ok. So lts.conf is now in /var/lib/tftpboot/ltsp/i386 ... and I've run ltsp-update-image ... now lets see what happens.
13:31
<vagrantc>
it's impossible to kill it. reasonably simple to make it go away by default, but impossible to say that there will never be a situation in which you don't need to override the defaults.
13:32
well, easily said :)
13:32
but reality comes harder than words
13:32
<spective>
vagrantc, is english your first language?
13:32
<ogra>
well, but it doesnt necessarily need to reside in $CHROOT/etc/
13:33* vagrantc wonders why the ltsp-developer list archives haven't archived any posts since 2008-08-16
13:33
<vagrantc>
spective: indeed
13:34
ogra: i guess so, but i don't see any real reason to need to change it ...
13:34
<spective>
*sigh*... it's now completely non-functional.
13:34
<ogra>
might be your lts.conf ... tired without any ?
13:34
<spective>
just a moment...
13:35
just disabled lts.conf....
13:35
rebooting.
13:37fie_wr0k has quit IRC
13:37GodFather has quit IRC
13:37
<spective>
well, that seems to be the problem.
13:37
meh.
13:38
simplifying lts.conf to see if it'll boot...
13:38
<ogra>
!pastebot
13:38
<ltspbot>
ogra: "pastebot" is The LTSP pastebot is at http://pastebot.ltsp.org. Please paste all text longer than a line or two to the pastebot, as it helps to reduce traffic in the channel. A link to the content will be pasted in the channel.
13:38
<ogra>
paste it :)
13:39indradg has joined #ltsp
13:39GodFather has joined #ltsp
13:39fie_wr0k has joined #ltsp
13:40
<warren>
vagrantc: I dislike the idea of moving ldminfod
13:40
<spective>
ahh. success.
13:40
<vagrantc>
warren: why?
13:40
<spective>
at least so far. I now have a shell.
13:40
<warren>
vagrantc: for one, it doesn't hurt ltsp-server. Another reason is I hate subpackages that have only a single file.
13:41
<vagrantc>
warren: well, i hate having to manually copy ldminfod over to a server that has no reason to install ltsp-server other than to have ldminfod installed on it.
13:42
warren: and then manually installing the dependencies to get a workingm ldm server.
13:42
<spective>
Amusing... my dansguardian installation blocks access to the LTSP documentation.
13:42
<vagrantc>
it must realize how out of date much of it is
13:43
<warren>
vagrantc: hmm
13:43
vagrantc: ok fine.
13:44
<vagrantc>
warren: and if we implement some things as plugins, ldm-server might include two, or ever three files :)
13:45
warren: any thoughts on the ldm handling of logouts ?
13:45
<warren>
vagrantc: the conditional behavior, giving things through ldminfod... that part sounds overengineered
13:45
vagrantc: oh, and my laptop was running openssh-5.1 as the server and I DID have the logout problem
13:45
<vagrantc>
warren: well, i can't think of any easier way to handle it...
13:46
warren: hmmm... well, on my test install, i used openssh 5.1, and it worked fine, downgraded *only* openssh-server to 4.7, and it stopped working ... upgraded back to 5.1, and it worked ...
13:47
<warren>
vagrantc: are you sure no options changed along with that upgrade?
13:47
<spective>
hrm. openssh-server complains that port 22 is already in use.
13:47
<vagrantc>
warren: i think there may have been additional issues with the original "ssh -O exit" patch.
13:47
warren: none that didn't break again when downgrading.
13:47
<warren>
vagrantc: all I know is 5.1 server here, I was having the same problem as you with logout
13:48
vagrantc: and the conditional behavior thing is a lot of extra complication...
13:48
<vagrantc>
warren: hopefully it's just an issue with the original patch
13:48
warren: agreed on the conditional behavior... but the alternative is to drop support for any servver running an older sshd.
13:49
well, not necessarily drop support entirely ... they just won't be able to log out :)
13:50
warren: well, thanks for your thoughts ... we'll eventually come up with something workable, i'm sure.
13:51* ogra doesnt care where ldminfod is as long as it doesnt become a hard requirement
13:51
<vagrantc>
ogra: good :)
13:51
<warren>
vagrantc: at what point do the ldm logout scripts run?
13:51
<spective>
So bizarre. sshd is unable to bind to port 22, because the port is already in use. netstat -nlp shows nothing on port 22.
13:51
<vagrantc>
warren: after the "ssh -Y Xsession ..."
13:51
<warren>
spective: try lsof -i
13:51
<vagrantc>
warren: after it stops, that is
13:51* ogra would run them while the tunnel is still there
13:52
<cliebow>
spective:see tat ssh tunnel reference by ogra??
13:52
<ogra>
cliebow, thats the cinet side
13:52
<cliebow>
heh
13:52
<vagrantc>
ogra: well, the way we currently do that (with kill -1 $PPID) there's no way to run them while the tunnel is still there.
13:52
<ogra>
he wants tu run a server on the TC
13:53
vagrantc, right
13:53
that needs reworking before
13:53
<cliebow>
k
13:53
<ogra>
though we have two ssh's
13:53* cliebow cliebow hides under rock again..
13:53
<spective>
lsof -i shows nothing.
13:53
<warren>
vagrantc: i'm looking at the code
13:54
<ogra>
and as we run the scripts on session startup it should be possible to do the ame on shutdown
13:54
<vagrantc>
right. but there's no consistant way with older openssh-server that i can find to log out without the kill ...
13:54
<spective>
cliebow, I see the reference... don't know what it implies.
13:54
<ogra>
we can put the kill into a K script ;)
13:54
into K99 :)
13:54
<vagrantc>
the K scripts don't ever run unless the kill is in the same ssh call as Xsession
13:55
<ogra>
you just need to store $PPID somewhere
13:55
<vagrantc>
ogra: but it hangs, so it never gets run.
13:55
<spective>
How can port 22 be in use if it doesn't show up in netstat or lsof??
13:56
<cliebow>
spective:i am so far behind now.....cant help any moe..
13:56
<vagrantc>
ogra: the "ssh ... Xsession ... ; kill -1 $PPID" ... if you remove "kill -1 $PPID" ... it just sits there indefinitely once the user has logged out.
13:56
<ogra>
well, we can kill the sessin clinet side ... then have a kill-1 4PPID as last script, no ?
13:56
*$PPID
13:57
<vagrantc>
ogra: so storing the PPID bit doesn't really gain us anything... unless we somehow background the ssh process...
13:57
<spective>
cliebow, hehe, s'ok.
13:57* ogra tries to type one handed during dinner
13:57
<vagrantc>
ogra: when do you run the kill ?
13:58
ogra: currently, it relies on the "ssh ... Xsession ..." command to complete, but it doesn't ever return without the "kill -1 $PPID" part.
13:58
<ogra>
i kill the session fron the TC side ? and run the kill as last K script (safety net)
13:58
hmm
13:58
<vagrantc>
ogra: when do you decide to run the kill ?
13:58
<ogra>
yeah, right
13:58
<vagrantc>
the K scripts don't run until the "ssh ... Xsession ..." finishes.
13:59
and that's the part that hangs.
13:59
<ogra>
hmm
13:59
<vagrantc>
it's a really obnoxious bug.
14:00
if there was some way we could pass the ssh escape chars ... ~& would basically handle what we want:
14:00
~& Background ssh at logout when waiting for forwarded connection / X11 sessions to terminate.
14:00
<spective>
At this juncture, I'm going to change sshd's config to run on port 222.
14:00
<vagrantc>
i think it's that "waiting for X11 sessions to terminate." bit
14:01
<warren>
vagrantc: ldm-server would contain what files?
14:01
beyond just ldminfod
14:01
<vagrantc>
warren: ldminfod plugins
14:01
not yet developed, but my proposal was to use ldminfod plugins to detect teh openssh-server version
14:01
since it's probably distro-specific
14:05
<spective>
What was the rationale behind switching to nbd instead of nfs? other than nfs's historic locking weirdnesses and complexity?
14:05
<ogra>
cant we get that during connect anyway ?
14:05
without any ldminfo involvement ?
14:05* spective hopes that his local apps will work when this is done...
14:05
<vagrantc>
ogra: i'm open to ideas.
14:05
<ogra>
spective, speed and bandwith usage
14:05
<spective>
ogra, 'k.
14:05
<vagrantc>
ogra: but i haven't heard any yet.
14:06
well, any that don't have issues.
14:06
<ogra>
vagrantc, i *think* there was a way to telnet to port 22 ... similar to the EHLO mailserver command ... i'd have to ask cjwatson
14:06
<vagrantc>
the simplest i've heard, is to basically do: ssh $server ssh -V
14:06
<spective>
... how much bandwidth does the filesystem use after the client's connected to the X server??
14:06
<ogra>
(wh is off for the day ... which should be since some hours as well ... my doctor will shout at me )
14:06
<vagrantc>
and *assume* that the ssh client and ssh-server are the same.
14:07
ogra: take care!
14:07
<ogra>
well, i'm not really eager :P
14:08
(not to take care ... but to close the laptop lid)
14:08
<vagrantc>
heh
14:08
<spective>
?! I can't bind to port 222 on 0.0.0.0, the address is in use?
14:08
I'm lost. There's _nothing_ using port 222...
14:08
<vagrantc>
spective: that must be me. i open backdoors wherever i go on 222.
14:09
<spective>
vagrantc, heh. oh good.
14:09
So what in the world is going on?
14:09
Is my network card read-only now?
14:09
(joke.)
14:09
<ogra>
oh, didnt we tell you ? :P
14:10
thats anew feature
14:10
<spective>
yay.
14:10
seriously. I have no idea how this possible.
14:10
<ogra>
nah, but it likely didnt generate any ssh keys during install
14:10* spective thought he had the gist of linux. *sigh*
14:11
<ogra>
so sshd wont start
14:11
<spective>
I've generated the keys.
14:11
<ogra>
ah
14:11
<sbalneav>
Didn't we figure on doing the ssh -O shutdown command to close after the session ended?
14:11
<ogra>
sbalneav, tat broke with some ancient windowmanagers
14:11
<sbalneav>
spective: Are you running the command as root or a user?
14:11
<spective>
And it starts... and binds to the ipv6 0::0 address on port 22.
14:11
<ogra>
so vagrantc wasnt happy with it
14:11
<spective>
sbalneav, root!
14:12
<sbalneav>
0.0.0.0?
14:12
or 127.0.0.1?
14:12
<spective>
0.0.0.0
14:12
no point in listening on 127.
14:12
well, not much anyways.
14:13
<sbalneav>
0.0.0.0 isn't a valid ip address
14:13
<spective>
sbalneav, 0.0.0.0 means bind to all interfaces.
14:13
So it's listening on eth0, eth1, lo, etc.
14:13
<sbalneav>
errr, not as far as I know.
14:14
You're running a daemon?
14:14
<spective>
yes.
14:15
Do a netstat -nlp and look under local address.
14:15
netstat -lp shows it as *:22
14:15
or whatever port.
14:16
<vagrantc>
sbalneav: not just broke on oddball window managers(icewm, xfce), but also with versions of openssh-server older than 5.1
14:16
sbalneav: i just posted a message to ltsp-developer about it
14:16
<warren>
vagrantc: openssh-server-5.1 here didn't prevent the problem
14:16
it might be something else
14:16
<vagrantc>
and apparently it still is an issue with warren for something else.
14:16
<warren>
vagrantc: did ssh version on the client side matter?
14:16
<vagrantc>
warren: nope.
14:17
warren: i'm wondering if we're just experiencing a *different* problem that has similar symptoms ...
14:17
<ogra>
selinux ?
14:17
<warren>
no selinux on the clients
14:17
doesn't work with nfs root
14:17
<vagrantc>
warren: and again, i think you tried with an older patch.
14:17
<ogra>
but on the server
14:17
<warren>
ogra: why are you so quick to blame selinux for everything?
14:18
vagrantc: got a new patch? i'll test whatever now
14:18
<ogra>
i dont like it ? its an easy candidate to blame ? no idea :)
14:18
<vagrantc>
warren: i've been working with ldm 2.0.6 ... but shouldn't be hard to forward-port it.
14:19
<warren>
let me test it without conditional stuff
14:19
<vagrantc>
yeah.
14:19
i haven't actually worked on the conditionals yet anyways.
14:22cliebow has quit IRC
14:25spective_ has joined #ltsp
14:25
<ltsppbot>
"vagrantc" pasted "new untested ssh -O exit" (30 lines) at http://pastebot.ltsp.org/46
14:25
<vagrantc>
warren: ^^
14:25
<spective_>
Ugh. client crashed.
14:25
So did I miss any deep wisdom on why I can't bind to IPv4 ports w/ sshd?
14:28
<vagrantc>
only shallow wisdom that's not likely to be helpful
14:28
<spective_>
Is there a direction I should look? /proc kernel settings? kernel configs?
14:29
<ogra>
did you try to bind to the actual IP ?
14:29
just for laughs ?
14:29
<spective_>
ogra, no. I can try that.
14:29
brb.
14:31fie_wrk has joined #ltsp
14:31spective has quit IRC
14:34GodFather has quit IRC
14:34
<spective_>
just for laughs, it works.
14:34* ogra thought so
14:34GodFather has joined #ltsp
14:34
<spective_>
WHY???
14:34
excuse my shouting.
14:35
<warren>
are you using suse?
14:35
<ogra>
i dont think we set up massivley much more routing on clients than needed to connect to the server
14:35
<spective_>
ubuntu
14:35fie_wr0k has quit IRC
14:35
<spective_>
ogra: elaborate?
14:35
<ogra>
which is fine with the planned localapps implementation
14:35
<warren>
OK good, then I'm not annoyed.
14:35
<ogra>
warren, lets wait what jcastro achieves :) he is great at such things
14:36
<spective_>
however... I think I've borked by ssh key setup now.
14:36
<ogra>
just ignore the annoance for now
14:36
spective_, there is no defaulr route or gateway ... and we dont actually need it in our localdev implementation since we dont open extra ports for it
14:36makghosh has quit IRC
14:36
<warren>
why ssh server on a client?
14:37
<ogra>
its the old way of doing localapps
14:37
<spective_>
warren: local apps setup
14:37
<ogra>
massively painful
14:37
but works as well indeed
14:37
just has a lot maintenance overhead
14:38
<warren>
why can't you use the modern localapps?
14:38
<ogra>
they are not in hardy
14:38
<spective_>
yup.
14:38
<ogra>
(and no, i dont plan to backport)
14:38
<spective_>
And implementing them by hand looks daunting...
14:40spective_ is now known as spective
14:47
<vagrantc>
it seemed fairly simple to manually backport the current localapps implementation
14:47
<warren>
vagrantc: except for the sshd part...
14:47
<vagrantc>
it's pretty much just a single plugin to ldm
14:48
warren: what part do you mean?
14:48
<warren>
vagrantc: openssh-5.1
14:49
<vagrantc>
well, doing whatever it is they're doing now, they could still backport the xatoms and group handling stuff... home directory mounting might be trickier if they actually need the cleaner sshfs
14:49
but no matter what, they need to mount home directories, so that doesn't really matter.
14:49Egyptian[Home] has quit IRC
14:50Egyptian[Home] has joined #ltsp
14:54
<warren>
ooh
14:54
found the cause of my nbd0 being confused
14:54
nbd0 is mounted root
14:54
but ltsp-init-common has:
14:54
nbd-client $SWAP_SERVER $NBD_PORT /dev/nbd0 && \
14:54
swap_devices="$swap_devices /dev/nbd0"
14:54
is nobody else using ltsp-init-common or something?
14:55* ogra uses the same initscript he always used ... from his /debian dir in the package
14:55
<ogra>
i think vagrantc uses the upstream initscript
14:55
thoguh they shouldnt differ in that area
14:56
<vagrantc>
hmmm...
14:56
i don't think i've tested the NBD swap stuff in ages.
14:56
definitely not with both nbd root and nbd swap ...
14:56
<ogra>
th ehardcoded nbd0 is worrying
14:56
<warren>
vagrantc: I recently found that it was broken in fedora because somebody changed the paths to /bin/nbd-*
14:57
should I add a loop that tries nbd-client -c /dev/nbdX
14:57
?
14:57
<ogra>
to connect ? ... hmm
14:57
<warren>
vagrantc: oh, so logging out of KDE after adding your ldm patch... gets stuck
14:57
<vagrantc>
nbd-client -c would require a versioned dependency on a version that implements -c
14:57
<ogra>
well, all recent ones have it
14:57
you must have a very old nbd to not have -c
14:58
<warren>
vagrantc: well, if you upgrade to ltsp-latest upstream you better have recent versions of stuff
14:58
<vagrantc>
yes, let's make sure nothing works for longer than 6 months.
14:58
<ogra>
yeah, that also applies to you being cautious about openssh-server 5.x
14:58
we never really supported different versions of client and server
14:59
<vagrantc>
you never supported it
14:59
it's worked for me for ages.
14:59
<ogra>
you support etch (roiginal etch) chroots with leanny servers and the other way round ?
14:59
<vagrantc>
warren: debugging the ldm patch ... what processes are still running as that user on the server?
15:00* ogra bets on dbus
15:00
<warren>
hmm, I killed a few and still stuck
15:00
<vagrantc>
ogra: i've done it on many occasions, and rarely have problems. some features don't work on etch (nbdroot) ... but in general, ther'es not really many issues.
15:00
for me, the process that always gets stuck seems to be dbus-launch
15:01
<warren>
I killed everything except these, and it is still stuck
15:01
test 4522 4518 0 15:45 ? 00:00:00 sshd: test@pts/5
15:01
test 4523 4522 0 15:45 pts/5 00:00:00 bash -c echo LTSPROCKS; /bin/sh -
15:01
test 4525 4523 0 15:45 pts/5 00:00:00 /bin/sh -
15:01
<vagrantc>
that's very different symptoms from me
15:02
<ogra>
dbus-launch properly run by gnome session for me ... so it dies with gnome ... kde and other WMs are not handling it from the session binary
15:02K_O-Gnom has joined #ltsp
15:02
<ogra>
as soon as dbus-launch is handled by a wrapper script around the session manager it gets wonky
15:03
(thats why gnome changed it to be spawned by gnome-session)
15:04
<warren>
argh, I need a nap, huge headache
15:04
<vagrantc>
ogra: are you saying: "dbus-launch --exit-with-session " is borked?
15:04
<warren>
i'm going to add a nbd-client -c loop to ltsp-init-common later today
15:04
<ogra>
no. tats what gnome-session calls as well
15:05_UsUrPeR_ has quit IRC
15:05
<warren>
vagrantc: go ahead and move ldminfod, i'll adjust later today
15:05
<ogra>
but spawning it from a binary seems to make a difference
15:05
<warren>
maybe I"ll have time to implement the LDM_SESSION plan
15:05
nap
15:06indradg has quit IRC
15:06indradg has joined #ltsp
15:07
<vagrantc>
warren: i'll give a little more time for others to comment before moving it
15:11
<warren>
ok
15:11
hurting too much for nap
15:13
at least I figured out why nbd was failing
15:16_UsUrPeR_ has joined #ltsp
15:17spective has quit IRC
15:19spective has joined #ltsp
15:22Q-FUNK has quit IRC
15:43jonnor has quit IRC
15:55alekibango has joined #ltsp
15:57jonnor has joined #ltsp
15:58GodFather has quit IRC
16:01Q-FUNK has joined #ltsp
16:01mhterres has quit IRC
16:01btil has quit IRC
16:02_UsUrPeR_ has quit IRC
16:03btil has joined #ltsp
16:03spectra has quit IRC
16:04bobby_C has joined #ltsp
16:04btil has left #ltsp
16:06spective has quit IRC
16:13jammcq has quit IRC
16:13nicoAMG has quit IRC
16:14Lns has joined #ltsp
16:15
<warren>
is /usr/bin/seq installed in your chroot?
16:16
<vagrantc>
it's part of coreutils on debian, which is required.
16:16
so yes
16:16
i'm guessing ubuntu has it too.
16:19
<warren>
# Detect first unused nbd device, skip nbd0
16:19
for num in `seq 1 15`; do
16:19
nbd-client -c /dev/nbd${num} > /dev/null
16:19
[ $? -eq 0 ] && break
16:19
done
16:19
nbd-client $SWAP_SERVER $NBD_PORT /dev/nbd${num} && \
16:19
sane?
16:19
skipping nbd0 since it is likely to be used
16:20gbolte has joined #ltsp
16:24
<vagrantc>
i swear i implemented that before.
16:24
well, not exactly that ... but not hard-coding /dev/nbd
16:25* vagrantc wonders if it got reverted somewhere along the way
16:25
<warren>
perhaps I'm the only one using this function
16:26jonnor has quit IRC
16:29
<vagrantc>
warren: no, we're using it in debian ...
16:29indradg has quit IRC
16:29
<vagrantc>
it was probably in one of my gazillion branches of unmerged features...
16:29
unmerged and unreleased, that is
16:30
<warren>
I have zero unmerged branches
16:31
if I work on something that goes unmerged it will be forgotten
16:32
<vagrantc>
well, that's what happens to me too
16:32
sometimes it's not merged for a reason, though.
16:33
<johnny>
hmm.. you don't have to use seq
16:33
<vagrantc>
you could just hard-code 0 1 2 3 4 ... 15
16:33
<warren>
right, ok
16:33
<vagrantc>
but seq sure is simple
16:35
already found issues with using "env" for the tcsh stuff.
16:35
localdev breaks with tcsh...
16:35
well, tcsh and LDM_DIRECTX
16:36
i guess i should still try the /bin/sh -c 'blah blah blah' approach ...
16:36
but if the env approach seemed non-evil, i figured i might try and get it into the next debian release ... *sigh*
16:36
<johnny>
uggh.. burn them
16:37
<vagrantc>
if i can find a relatively simple way to fix it, i don't want to force people to use a particular shell. i'm just not that way.
16:38
<warren>
vagrantc: right, tcsh is a fucking different language. env is not going to fix that.
16:38
<vagrantc>
heh.
16:38
warren: the issue was env only applied to the first command.
16:39
i.e. our chaing of foo ; bar ; baz ... bar and baz didn't get it
16:39
or soemthing like that.
16:39
or maybe not ...
16:40
ah, i think the issue was we'd have to put env in the S10-delayed-mounter too ...
16:40
<warren>
My essay is written in English. I'm going to s/the/le/ and make it work in French.
16:40
It is a DIFFERENT LANGUAGE.
16:40
<vagrantc>
yes, it's a pain to support tcsh.
16:41
<warren>
Make it use /bin/sh as a script interpreter, don't try to change the syntax itself.
16:41
/bin/sh as interpreter is the legitimate fix
16:41
<vagrantc>
we'd have to hard-code that in all of our calls to ssh ...
16:41
i'm starting to lean towards not supporting non-bourne compatible shells at all.
16:42
<johnny>
vagrantc, i have a fix for you..
16:42
<warren>
+1
16:42
<vagrantc>
handling the environment differently is just too danged complicated, and we make extensive use of that.
16:42
<johnny>
tell them the upstream maintainers of the project are forcing you to not support tcsh
16:42
<warren>
it isn't just the environment that is a problem
16:42
<johnny>
and ther'es nothing you can do about it
16:42
<warren>
it is a DIFFERNT LANGUAGE
16:43
<vagrantc>
warren: well, so far, the environment is the only problem i've encountered.
16:43
<warren>
johnny: appeal to authority? =
16:43
=)
16:43
<johnny>
warren, yes :)
16:43
you know.. get the flames off his back
16:43
<vagrantc>
i don't need to appeal to authority. i can handle it myself.
16:43
<warren>
csh needs a different syntax to set variables
16:43
<vagrantc>
but i only have so much time.
16:43
<warren>
conditionals use endif instead of fi
16:44
<johnny>
fi is kinda silly tho
16:44
<warren>
if ($TERM) exit
16:44
this is a different language!
16:44* vagrantc never debated that it wasn't a different language.
16:44
<vagrantc>
but if there were simple patches to bridge the divide, i'd be willing to try them. there just aren't.
16:48indradg has joined #ltsp
17:03btil1 has joined #ltsp
17:03robbie has joined #ltsp
17:04btil1 has quit IRC
17:10bobby_C has quit IRC
17:10johnny has left #ltsp
17:10K_O-Gnom has quit IRC
17:12btil1 has joined #ltsp
17:13
<btil1>
anyone on?
17:14
<Q-FUNK>
nope
17:14
<btil1>
that sucks
17:15
i shouldn't be here either, but i have no life at all
17:17* Lns 's life currently consists of trying to teach linux filesystem basics to lab techs that don't want to learn new things.
17:17* vagrantc marked the ldm/tcsh bug as wontfix
17:18
<btil1>
don't know what's worse, mine or Lns 's
17:18
<Lns>
heh
17:19
<btil1>
teaching unwilling pupils is not fun, but neither is writing documentation
17:19
<Lns>
"But in Windows it's so easy to put permissions on files and folders!" "...Oh, and why can't you have group ownership of a file that doesn't exist?" *bangs head on wall*
17:20* Lns prefers documentation as you can hand it to people and walk away
17:20
<btil1>
don't know what i'd do if i was stuck developing on windows
17:20
working without chmod/chown -R or grep
17:20
or screen?
17:20
true, but the writing of it is mind-numbing
17:21
<Lns>
I actually kind of enjoy it
17:21
<btil1>
sometimes. seems i only enjoy it when i don't have to do it.
17:21
right now it's been forced upon me, so it sucks
17:21
i liked doing the ltsp docbook stuff, probably because i chose to.
17:24
<warren>
vagrantc: ogra: I just found a security hole in <unspecified part of your LTSP>
17:24
I haven't been using that part yet.
17:31Q-FUNK has quit IRC
17:37japerry has joined #ltsp
17:54
<vagrantc>
warren: hmmmm?
17:54adeel has left #ltsp
18:11shogunx has quit IRC
18:20shogunx has joined #ltsp
18:28
<rjune>
Lns: You must work at a school
18:35barbosa has quit IRC
18:39
<Lns>
rjune: yes.
18:39
well..i work FOR a school.
18:39
they are my client.
18:40
I am hopeful that they will change their attitudes, and I think they will - it's just frustrating when you're trying to get people excited about something they hate by default.
18:40
just because it's different than what they know.
18:41
<warren>
vagrantc: false alarm
18:41robbie has quit IRC
18:52
<gbolte>
lns thats kinda an issue with us here people dont like new things but have eventually settled down and accepted the fact they are not getting their way
18:53
<Lns>
gbolte: that's good news to hear - i think it's just a matter of pushing it through their short-term memory. If they don't like it at first they'll fight it but if they see it's still there after a while they'll start to adapt
18:54
<gbolte>
yup
18:55staffencasa has quit IRC
19:02Lns has quit IRC
19:32gbolte has quit IRC
19:35etyack has joined #ltsp
19:42japerry has quit IRC
20:06shogunx has quit IRC
20:07johnny has joined #ltsp
20:10indradg has quit IRC
20:19btil1 has quit IRC
20:24rcy has quit IRC
20:28shogunx has joined #ltsp
20:40japerry has joined #ltsp
20:50japerry has quit IRC
20:54lejo has quit IRC
20:54otavio has quit IRC
20:54C7 has quit IRC
20:54Loto_ has quit IRC
20:54MacIver has quit IRC
20:54ninkendo has quit IRC
20:54otavio has joined #ltsp
20:54C7 has joined #ltsp
20:54ninkendo has joined #ltsp
20:55MacIver has joined #ltsp
20:55Loto_ has joined #ltsp
20:59lejo has joined #ltsp
21:01shogunx has quit IRC
21:09shogunx has joined #ltsp
21:19
<vagrantc>
hmmmm... my gnome fails to logout issue seems to be http://bugs.debian.org/486007
21:20shogunx has quit IRC
21:30shogunx has joined #ltsp
21:33petre has joined #ltsp
21:33
<petre>
warren, do you know the url for reporting bugs for Fedora 10?
21:34
I can find lots of webpages about bur reporting, just can find the bloody page to actually log the bugs
21:34
<warren>
petre: bugzilla.redhat.com
21:34
<petre>
or search for them
21:34
<warren>
petre: bugzilla calls fedora 10 "devel" until it is released, then it becomes 10
21:34
<petre>
hmm, should my login from the k12ltsp page work there?
21:34
k12linux , i mean
21:35
<warren>
what do you mean?
21:35
petre: i'm using bugzilla for k12linux bug tracking
21:35
petre: not trac
21:40
<petre>
I've got a fedora login, but apparently that doesn't work at bugzilla.redhat.com
21:40
<warren>
bugzilla.redhat.com is a different account system
21:42
<etyack>
hi warren
21:42
<petre>
ok, I created a new account on bugzilla.redhat.com, still waiting for the email
21:42
<warren>
hi
21:43
<petre>
BTW, nautilus on F10 does the same thing if launched from a terminal by root: mounts all the partitions on the local hard disk
21:44
tried some switches with it, too, but to no avail
21:44
<warren>
might need to ask upstream for a switch that wont do that
21:44
<petre>
gnome project?
21:44
<warren>
yes
21:45
<etyack>
did the core dump of the ldm error UsrP.. emailed shed any light about what caused it?
21:45
<petre>
so, you don't think it's actually a bug?
21:45
<warren>
etyack: I posted the details of the traceback to the list
21:45
etyack: It really only crashes on that one client and not others? that's curious
21:45
<etyack>
they have 12 of the same hardware
21:46
which list did you post it to?
21:47
<warren>
etyack: ltsp-devel
21:56petre has quit IRC
22:01
<sbalneav>
Evening all
22:01
<etyack>
hey sbalneav
22:03
<warren>
etyack: sbalneav is best to ask about that ldm segfault
22:03
etyack: he did the majority of the C code
22:03
<etyack>
sbalneav: did you happen to look at the segfault warren posted to the devel list?
22:03
sbalneav: if so, any thoughts or questions?
22:04
<sbalneav>
Briefly. Lemme have a look again.
22:07
Well, it's dying in a fprintf.
22:07
in the die routine. :)
22:07
so it's already dead, for other reasons.
22:08
<etyack>
isn't that special
22:09
the clients are ~10 yr old 800 Mhz celerons w/128mb ram
22:09
<sbalneav>
hold on, I'll look at the code.
22:09
do they get logged in at all? Is it intermittent?
22:10
<etyack>
no and no. they always fail with a segfault.
22:10
they have 12 and all behave in the same way.
22:11
<sbalneav>
Other ones they have work?
22:12
<etyack>
they have numerous different hardware configurations and approx 150 clients. this group are the only ones experiencing problems.
22:12
<sbalneav>
ah.
22:12
I see the problem.
22:13
<etyack>
do tell
22:13
<sbalneav>
hm.
22:14
one of these days, I should really write myself some notes. :) This whole bzr thing still has me flummoxed.
22:14
We need to move the fclose and the close to just before the exit(1).
22:14
We close the logfd, but then ldm_wait() tries to write to it.
22:15
<etyack>
why do you think this hardware combo exposed the error?
22:16
<sbalneav>
Incedental. The real problem here is probably that the X server isn't autodetecting that video card, or something. Does X even light up the display, or does it just stay on the blinking cursor?
22:16
Eitherway, the early close is the problem for the segfault.
22:16
That WONT fix their problem, but it WILL eliminate the segfault.
22:17
So, what's the easiest way to create a patch here, Warren?
22:17
<warren>
huh
22:17
<etyack>
we are supporting them remotely. this is a old sis 6236 based video card and I would not be surprised if X does not properly detect the card.
22:17
<warren>
i'm doing a few things at once
22:18
sbalneav: why is it segfaulting only for these users?
22:18
oh I see
22:18
<sbalneav>
Dunno. Maybe the ldm_wait write simply silently fails on other boxes.
22:19
Not sure.
22:19
<warren>
I don't know exactly what you mean
22:19
so I can't fix it
22:19
<sbalneav>
Hold on, lemme create a regular patch file. Uno momento.
22:19
<etyack>
possibly tomorrow, but Friday for sure, we will have a chance to look at the xorg log files.
22:20
<warren>
etyack: on ebay I see lots of thin clients for sale for like $5-20 each, but they're all video cards that upstream X no longer supports.
22:21
pre-GX2 geode's are plentiful
22:21
<johnny>
i wish i had some real thin clients.
22:22
i also need some monitors..
22:22
<etyack>
warren: i also saw ~50 of our thin clients damaged during shipping on ebay.
22:22
<warren>
etyack: eh?
22:22
<ltsppbot>
"sbalneav" pasted "diff -Naur for ldm.c fix." (19 lines) at http://pastebot.ltsp.org/47
22:23
<sbalneav>
et, voila.
22:23
<warren>
etyack: you want a F9 i386 build of that?
22:23
<etyack>
warren, we made an insurance claim for damaged thin clients and a salvage company put them on ebay.
22:23
<warren>
etyack: that's very common
22:23indradg has joined #ltsp
22:23
<etyack>
warren: yes. we will test it tomorrow or Friday.
22:24
<warren>
etyack: sometimes you can take advantage of that though. for example a few years ago my dad's crane company was lifting a giant 120ton chiller on top of a building, and the crane toppled over.
22:24
etyack: million+ insurance payout...
22:24
<etyack>
warren: i'm sure it is, but i the pictures used in the auction were taken in our warehouse!
22:25
warren, by the claims adjuster
22:25
<warren>
etyack: the chiller was put for auction. my dad bought it, fixed it up, and was renting it out ever since at a profit. it would have been impossible to buy a chiller for rental that is profitable otherwise.
22:25
now my dad wasn't involved in the accident
22:25
he wasn't at the site
22:26
<etyack>
warren: wow. we could by out thin clients cheaper on ebay than our vendor.
22:26
<warren>
etyack: exactly
22:26
etyack: we paid about $15k to buy a chiller that would have been $300k
22:26
etyack: about $20k in repairs, and it made $40k in the first 2 month rental job.
22:26
<etyack>
warren: nice
22:27
that's the way to make money
22:27
<warren>
not a normal way
22:27
you don't drop million dollar cranes....
22:27
<etyack>
true
22:27
<warren>
etyack: oh, and it helps that this is in hawaii, where it takes months to get a new chiller
22:27
<etyack>
that must have been one hell of a scary moment
22:27
<warren>
etyack: so we have a monopoly on emergency chiller rental there
22:28
etyack: if a building (like a hospital)'s chiller fails, either they rent our emergency chiller or shut down the hospital
22:29
<etyack>
warren: there's an advantage to being the "monopoly"
22:29
<warren>
yeah...
22:29
although we went all last year without renting it out
22:29
<etyack>
sbalneav: thanks for the help
22:30
warren: let me know where to access the new ldm version and we will coordinate testing with the school.
22:33
<sbalneav>
etyack: NP. Remember: this will fix the coredump, but not the LDM not working, that's due to X not liking the SiS. I suspect either 1) vesa 2) custom lts.conf/xorg.conf fiddling or 3) 12 cheap ati Rage 128 cards will solve the problem :)
22:34
<warren>
etyack: fedora ltsp does not support specifying an X driver in lts.conf
22:34
etyack: yet
22:34
etyack: nobody asked for it yet
22:35
<etyack>
sbalneav: understood. purchasing 12 shiny LTSP Term 1220's will fix their problem also. I read some posts about this chipset causing x problems.
22:35
warren: i was just about to ask...
22:35
<warren>
it has been lower on my priority list
22:35
i'm really trying to avoid using xorg.conf at all
22:36
etyack: you can test it specifying a driver in /etc/X11/xorg.conf in the chroot. if you cna get it working with vesa, then I guess I'l hack up the lts.conf thing to make it work.
22:36
etyack: OTOH i suspect even vesa wont work
22:37
etyack: oh, and get me their xorg log (without any vesa)
22:37
<etyack>
warren: i'm not very confident either
22:37
warren: will do.
22:39
<warren>
etyack: http://people.redhat.com/wtogami/temp/ldm-2.0.11-4.20080820.23.fc9.i386.rpm
22:39
<etyack>
warren: thanks
22:39
<warren>
etyack: seems to work in cursory tests
22:40
<etyack>
warren: we'll test it here first in the morning.
22:41
<vagrantc>
something is very, very strange. the m68k build daemon was the first to pick up my latest upload.
22:41
<warren>
m68k is making a comeback
22:41
over alcohol the hackers were talking about m68k fedora
22:42
<vagrantc>
m68k is barely making a come-back with debian ... let alone distros that never supported it ..
22:43
it was dropped for the last official debian stable release, and never really recovered.
22:43
<warren>
some things are better dropped
22:43
<etyack>
thanks guys
22:44
time to go home.
22:45etyack has left #ltsp
22:53gonzaloaf_laptop has joined #ltsp
22:57
<warren>
heh, I got my EEEPC to work as a LTSP server
22:59
<johnny>
neat
22:59
<warren>
of course it sucks
23:00
<johnny>
hmm.. would be neat to carry around an ltsp server tho
23:00
something without a screen
23:01
dberkholz, so what's goin on?
23:06
<sbalneav>
johnny: Laptop with a dead screen? :)
23:06
<johnny>
sbalneav, i meant something designed without one :)
23:06
hey scott
23:11
<sbalneav>
heh, hey
23:12
But (devils advocate) EEEPC = $(~300) Laptop with dead screen = Free (from friends, neighbors, work, etc) :)
23:12
<johnny>
sure.. but the eepc has too much stuff.. speakers, screen,keyboard
23:13
<sbalneav>
Plus you get to say: "Look how COOL this LTSP stuff is! I can even make a DEAD LAPTOP worth something!!!"
23:13
<johnny>
guess you could build a custom case for the dead laptop, but it'd still be too big
23:13
<sbalneav>
It'd be a total babe magnet, trust me :)
23:13
<johnny>
well.. it could be..
23:14
<sbalneav>
As an LTSP developer, supermodels throw themselves at my feet constantly.
23:14
<johnny>
if that "babe" is setting up a something where they need to provide a bunch of computers cheaply and in a jiffy
23:14
and you show up with a magic box that makes it happen
23:14
you'd be her hero
23:15
<sbalneav>
Exactly.
23:15
So, now all you need to do is wait around while Anna Kournikova needs to set up a school lab in a hurry, and YOU'D TOTALLY BE SET!!!!111oneone.
23:16
<johnny>
well i doubt it will be her..
23:16
more like somebody who is actually doing something of value for society even
23:17
<sbalneav>
Are you QUESTIONING the value that SUPERMODELS and CELEBRITIES provide to society?!?! Heaven forfend! :)
23:18
Think of all the employment that Paris whatsername generates :)
23:18
<johnny>
well. you know me..
23:18
did i ever show you where i use ltsp?
23:18
<sbalneav>
No, don't think you ever did. Do tell!
23:19
<johnny>
www.redemmas.org
23:19
<sbalneav>
Oh, cool!
23:20
Looks like one of the places I hang out in Winnipeg, at Mondragon. Hold on, I'll find the urlllllll.
23:20
http://mondragon.ca/
23:21
<johnny>
http://flickr.com/photos/45891628@N00/351914359/
23:21
there's a picture of two of em :)
23:21
yes.. that is us :)
23:21
lol
23:22
we aren't exlusively vegan tho
23:22
sbalneav, while in portland, i went to a place using ltsp
23:22
red and black cafe
23:22
thanks to free geek
23:22
freegeek is really a great project..
23:22
<sbalneav>
Yeah, I was really impressed
23:23
For an old soldering-iron wrangler such as myself, I was in heaven.
23:23
<johnny>
we should do that again
23:23
<sbalneav>
I'd go back.
23:23
<johnny>
altho.. i would like you to come down to baltimore
23:23
i have a much bigger room than we had at freegeek
23:24
it's in a church
23:24
<sbalneav>
I'd be all in favor of spring in Portland and Fall in Maine for the two annual LTSP hackfests.
23:24
<johnny>
check it
23:24
http://redemmas.org/2640/images/
23:24
<sbalneav>
vagrantc: Hey, I got my art home BTW. It's hanging out at my cottage.
23:25
<johnny>
altho.. we won't need that big room.. there's a smaller one next to it
23:25
<sbalneav>
Love the bulding. When was it built? Looks like 1890'
23:25
s
23:26
<johnny>
1889 i think?
23:26
or maybe 1879 .. i forget
23:26
definitely around that time tho
23:26
<sbalneav>
Yeah, the posts look very pre 1900's
23:26
<johnny>
our group is involved with that space
23:26
<sbalneav>
Cool.
23:27
<johnny>
started as a simple kitchen rental.. as we don't have enough space at red emma's to do the baking we like to do
23:27
<sbalneav>
Hey, I heard something cool, about longterm thinking. Ever heard of "The Long now" project? Anyway...
23:27
<johnny>
and we needed to bake in a food certified kitcen
23:27
<vagrantc>
sbalneav: glad you found somethin to purty up yer cottage :)
23:28
<sbalneav>
I loved portland.
23:28
<johnny>
and then we started renting out that big room so the church could stay open
23:28
<vagrantc>
sbalneav: i'll work on scrounging up a clicky keyboard for ya
23:28
<sbalneav>
When you find it, save it for next year.
23:28
<vagrantc>
sbalneav: if you give me a whole year, i could probably even wrangle up two or three :)
23:28
<sbalneav>
johnny: How long's RE been open?
23:29
<johnny>
coming up on 4
23:29
in november
23:29
<sbalneav>
Nice.
23:29
I'd go back out to freegeek for another hackfest in a heartbeat.
23:29* vagrantc tries to sneak a few more patches into lenny
23:29
<vagrantc>
sbalneav: glad y'all liked it :)
23:29
<johnny>
i've only been a member of the collective since last november.. in fact.. the same day as the birthday :)
23:30
<sbalneav>
sneak that patch I posted earlier in. It's a definite bug
23:30
<johnny>
been in baltimore since last june, not too long
23:30
<vagrantc>
sbalneav: which patch?
23:30
<sbalneav>
http://pastebot.ltsp.org/47
23:31
<johnny>
sbalneav, i'd love to go to portland again, i just hope i'll be able to afford it
23:31
<vagrantc>
sbalneav: still using ldm 2.0.6 in lenny
23:31
<johnny>
and next time.. i'll be able to contribute more effectively, as all my stuff will have made it into gentoo itself
23:31
<vagrantc>
just getting into the bronze age or something
23:32
sbalneav: you think that patch is relevent for the older version? how ugly of a bug is it?
23:32indradg has quit IRC
23:35
<warren>
vagrantc: that only prevents an ldm segfault right before ldm would have stopped anyway
23:35
vagrantc: because X failed to start
23:35
vagrantc: I suppose the benefit of that patch is people don't blame ldm for an X failure
23:35
<vagrantc>
ah.
23:36
<warren>
I just pushed it up to trunk
23:37indradg has joined #ltsp
23:39
<vagrantc>
warren: do you know how to configure rsyslog for recieving remote logs ?
23:40
<sbalneav>
I do, on Debian/Ubuntu
23:40
<warren>
sorry no
23:40
<vagrantc>
sbalneav: how? :)
23:40
<warren>
I don't have that working in fedora
23:40
vagrantc: if you figure out how...
23:40
<vagrantc>
i figured out how to send remote syslogs, but i've been switching to sysklogd to test it
23:40
<sbalneav>
edit /etc/default/syslogd
23:41
<vagrantc>
sbalneav: /etc/default/rsyslog ?
23:41
<sbalneav>
chenge the "SYSLOGD="" " stuff to SYSLOGD="-r"
23:41
no, should be syslogd
23:41
<warren>
[warren@newcaprica ltsp-trunk]$ bzr pull
23:41
Using saved location: bzr+ssh://wtogami@bazaar.launchpad.net/~ltsp-upstream/ltsp/ltsp-trunk/
23:41
ssh: connect to host bazaar.launchpad.net port 22: Connection refused
23:41
<sbalneav>
Lemme check on a debian box
23:41
<vagrantc>
sbalneav: debian switched to rsyslog
23:41
sbalneav: in lenny
23:42
older lenny installs still probably use sysklogd
23:42
<warren>
brb shower
23:42
<johnny>
metalog provided my favorite config out of the box..
23:42
<sbalneav>
ah
23:42
ok, for lenny, dunno
23:44
<vagrantc>
apparently, i could try to switch to an older compatibility mode ... but i haven't managed to get it to work
23:50
<warren>
/opt/ltsp/images/i386.img
23:50
is this the default location of your image?
23:50
location and name
23:50
<vagrantc>
yes
23:50
<warren>
k
23:57juanpaul has joined #ltsp