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


Channel log from 31 March 2008   (all times are UTC)

00:19
<johnny>
laga, are you about?
00:19
uggh.. guess not..
00:22AzMoo has joined #ltsp
00:27
<AzMoo>
Hi, I have my ltsp client booting from across the network, but I can't seem to login to it with a username and password that works fine on a different client. Any ideas why?
00:37
<johnny>
so.. does every other one work?
00:38
<AzMoo>
Well there's only one other one, and it does work.
00:38
Another client I mean. It's an Xubuntu LTSP so I can't log in as root.
00:42
If I chroot to the ltsp image and add a user, I have to do ltsp-update-image before it will recognize it, right?
00:43
<cyberorg>
AzMoo, users are created on the server, not in chroot
00:44
<AzMoo>
cyberorg: I see. Well, why can I log in to the server with my user and not the client? Is there somewhere I have to define which users are allowed to log in? Sorry for asking such fundamental questions, it's just that I can't really find any comprehensive docs.
00:45
<cyberorg>
AzMoo, use the account you can ssh into the server with
00:45
<AzMoo>
cyberorg: I have, it's just denying me.
00:46
<cyberorg>
AzMoo, can you ssh into the server from some other machine?
00:46
<AzMoo>
cyberorg: Yep.
00:47
<cyberorg>
AzMoo, then check logs on the server when client tries to log in
00:50
<AzMoo>
cyberorg: Which logs? I'm looking at /var/log/auth.log and messages and syslog, but I'm not seeing anything.
00:50
<cyberorg>
AzMoo, syslog should show some requests from the cleint
00:51
/var/log/messages
00:51
<AzMoo>
Nah, it's not showing anything, oddly enough.
00:52
The last entry in that is: Mar 31 13:16:28 kiosk -- MARK --
00:52
And it's 16:21 now.
00:52
<cyberorg>
AzMoo, is syslog service running?
00:53
<AzMoo>
Certainly is.
00:55
<cyberorg>
sorry, no idea then
00:55
<AzMoo>
Can you point me towards some decent documentation?
00:56
I think part of my problem is that I really don't understand how this is all working.
00:56
<cyberorg>
AzMoo, the client run "ssh -X user@serverIP /path/to/Xsession"
00:57
that is all there is to it :)
00:57
<AzMoo>
That can't be right. X is failing to load on the client.
00:57
<cyberorg>
if you can do that normally, you should be able to do via ldm
00:57
<AzMoo>
But I've got a vty login prompt.
00:58
<cyberorg>
oh, your issue is different then, most likely X configuration issue
00:58
<AzMoo>
Yeah, I think so, but I can't see what's failing because I can't log in to the client.
00:58
<cyberorg>
you cannot login to tty using server username
00:59
<AzMoo>
correct.
00:59
<cyberorg>
check out the docs about xeserver config that goes into lts.conf
01:00
not sure if you have to rebuild the image if you change lts.conf, best wait for ogra or someone else from ubuntu team
01:00
<AzMoo>
What time are they usually around?
01:00* AzMoo is at work.
01:01
<cyberorg>
they should be in few hours time
01:01
<AzMoo>
OK, thanks mate.
01:08
<johnny>
you can set a root password on the chroot tho, if you wanna poke around in there
01:09
then you can log into the local tty
01:11Shingoshi has quit IRC
01:14
<AzMoo>
OK, I got into it. Set a root password, rebuilt the chroot and logged in. Seems the X driver I'm using doesn't support 32bpp. Where do I change the X config? Within the chroot or on the server normally?
01:15Shingoshi has joined #ltsp
01:15
<AzMoo>
It would have to be on the server normally, because there is no xorg.conf in the chroot.
01:16praveer_cool has joined #ltsp
01:19
<cyberorg>
AzMoo, X_COLOR_DEPTH=whatyouneed in lts.conf
01:20
under the client section[MAC]
01:21
for all variables that you can set, see configure-x.sh
01:22
<AzMoo>
Ah, wicked, thanks.
01:28
<dberkholz>
what weirdo driver is that?
01:29
<AzMoo>
i810
01:29
<dberkholz>
Note: depth means the number of bits in a pixel that are actually used to determine the pixel colour. 32 is not a valid depth value. Most hardware that uses 32 bits per pixel only uses 24 of them to hold the colour information, which means that the colour depth is 24, not 32.
01:31
<AzMoo>
Yeah, I got that. It was just the default settings. I've put it down to 16 and it's running X now.
01:31
<dberkholz>
i810 should definitely work at 24
01:31
i'm using it right now
01:35
<AzMoo>
Do I have to run ltsp-update-image each time I change a config value?
01:37AzMoo has quit IRC
01:52deavid has joined #ltsp
02:02Pascal_1 has joined #ltsp
02:03Pascal_1 has quit IRC
02:04Pascal_1 has joined #ltsp
02:21
<Pascal_1>
Bonjour !
02:35K_O-Gnom has joined #ltsp
02:39subir has quit IRC
02:53ogra_cmpc has joined #ltsp
02:58dberkholz has quit IRC
02:58jammcq has quit IRC
02:58klausade has quit IRC
02:58F-GT has quit IRC
03:01jammcq has joined #ltsp
03:06dberkholz has joined #ltsp
03:11klausade has joined #ltsp
03:11F-GT has joined #ltsp
03:13makghosh has joined #ltsp
03:20klausade has quit IRC
03:20F-GT has quit IRC
03:21klausade has joined #ltsp
03:21F-GT has joined #ltsp
03:41tjikkun_work has joined #ltsp
03:41xp_prg has joined #ltsp
03:42
<tjikkun_work>
hi, i get the message that the client is not authorized to authenticate, any idea what is wrong? this is on a fedora install with modified network settings
03:42xp_prg has left #ltsp
03:45
<cyberorg>
tjikkun_work, run ltsp-update-sshkeys
03:46
<tjikkun_work>
cyberorg: thanks, that helps a bit.. although it still doesn' t log in, i just get the login screen back
03:47
<cyberorg>
tjikkun_work, try ssh user@serverip from another machine on the network
03:47
<tjikkun_work>
cyberorg: that works without problems
03:48
<cyberorg>
hmm, you need to wait for warren for more :)
03:49
<tjikkun_work>
cyberorg: ok, will do that, do you know what timezone he is from?
03:49
<cyberorg>
nope, somewhere in asia, japan?
03:50
<tjikkun_work>
ok. i'll just wait and see then
03:50
thanks :)
03:50pdjbarber has joined #ltsp
03:50Q-FUNK has joined #ltsp
04:08
<ogra_cmpc>
tjikkun_work, does "ssh -Y user@serverip xterm" get you an xterm on another machine ? you probably need to modify the sshd config to allow that in fedora
04:08
(an dindeed teh server needs a proper desktop installed to work at all)
04:10
cyberorg, warren lives on US eastcoast
04:10
<cyberorg>
ogra_cmpc, ok :)
04:11mikkel has joined #ltsp
04:23
<tjikkun_work>
ogra_cmpc: xterm comes up without problems
04:24
<ogra_cmpc>
and you got a desktop installed so the Xsession script can run properly ?
04:25
(thats what ldm executes after login, not sure where that sits on redhat systems)
04:31makghosh has quit IRC
04:32Q-FUNK has quit IRC
04:39
<tjikkun_work>
yes, i have gnome installed
04:39
<ogra_cmpc>
then it shoud work ...
04:39
anything in ~/.xsession-errors ?
04:41
<tjikkun_work>
nope, that file doesn' t exist
04:41
<ogra_cmpc>
auth.log ?
04:42
<tjikkun_work>
also doesn' t exist
04:42
<ogra_cmpc>
in /var/log ?
04:43
your sshd should create it on start if it doesnt exists
04:44
<tjikkun_work>
ah, i think it is /var/log/secure
04:44
<ogra_cmpc>
aha ...
04:44* ogra_cmpc hsant touched any redhat based OS since RH 4.2
04:45
<tjikkun_work>
well, don' t see anything weird in the log
04:46
<ogra_cmpc>
and you are using the recent fdeora ltsp packages ?
04:46
they should DTRT ... weird that they dont
04:47
<tjikkun_work>
well, maybe something is just wrong with my fedora setup
04:47
it is rawhide, but i think that is supossed to work
04:47
<ogra_cmpc>
might be ... thats actuallys requiring warren i guess ... i'm out of suggestions from my ubuntu view ....
04:48
<tjikkun_work>
ok, thank you very much anyway
04:52praveer_fedora has joined #ltsp
04:58viking-ice has joined #ltsp
05:02subir has joined #ltsp
05:09praveer_cool has quit IRC
05:21Q-FUNK has joined #ltsp
06:04cyberorg has quit IRC
06:25TelnetManta has quit IRC
06:27
<daduke>
!seen vagrantc
06:27
<ltspbot>
daduke: vagrantc was last seen in #ltsp 19 hours, 37 minutes, and 8 seconds ago: <vagrantc> seems like with pack-0.92, it actually even works.
06:31toscalix has joined #ltsp
06:35
<ogra_cmpc>
daduke, he migth be on his way back to the us, not sure how long his event in spain goes on
06:38
<daduke>
ogra_cmpc: yeah I remembered something like this. thanks. You don't happen to be familiar with the ltsp process in debian?
06:38
<ogra_cmpc>
process ?
06:39
you mean package approval between testing/unstable ?
06:39
or what exactly ?
06:41
<ltsppbot>
"daduke" pasted "chroot package install failure" (3 lines) at http://pastebot.ltsp.org/487
06:41
<daduke>
ogra_cmpc: I wonder how the ltsp-client-core package gets installed. backround: I'm working on my debian-live boot image (you might remember), and in the live chroot ltsp-client-core is not being installed: pb/487
06:41subir has quit IRC
06:41
<ogra_cmpc>
(i'm in teh uploaders list and in teh pkg-ltsp devel group of debian, but wouldnt dare to make an upload there witrhout talikg to vagrant forst)
06:48
daduke, ltsp-client-core should be pulled in by ltsp-client
06:50spectra has joined #ltsp
06:50
<daduke>
ogra_cmpc: it is, but it fails to install with said error message
06:52
<ltsppbot>
"daduke" pasted "preinst script" (17 lines) at http://pastebot.ltsp.org/488
06:53
<daduke>
ogra_cmpc: and I cannot see why the preinst script (pb/488) should fail. what unknown argument?
06:53
<ogra_cmpc>
daduke, touch /etc/ltsp_chroot
06:54
it is supposed to fail in any environment that doesnt have that file
06:54
so people dont install it by accident on their normal systems
06:55
<daduke>
ogra_cmpc: ahh, now we're talking... good idea, lemme try...
06:55
ogra_cmpc: will take ~ 5 min...
06:59cliebow has joined #ltsp
07:00jammcq has quit IRC
07:02
<daduke>
ogra_cmpc: -rw-r--r-- 1 root root 0 Mar 31 13:54 chroot/etc/ltsp_chroot, but it still does not install...
07:02
<ogra_cmpc>
thats odd, it surely should
07:03
do you have your debconf frontend set to noninteractive ?
07:04
<daduke>
ogra_cmpc: I'm afraid I don't really know what that means :(
07:04
<ogra_cmpc>
export DEBCONF_FRONTEND=noninteractive
07:05
that needs to be at the top of your chroot builder script
07:05mhterres has joined #ltsp
07:05mhterres has quit IRC
07:06
<daduke>
ogra_cmpc: I'll try. I tried export DEBCONF_DEBUG='.*', but this did not change anything. I'm not sure what all this lh_config and lh_build magic does...
07:06mhterres has joined #ltsp
07:07
<ogra_cmpc>
s/lh_/dh_ ?
07:07
man debconf :)
07:07
and man debconf-deve
07:07
l
07:07
<daduke>
ogra_cmpc: nope, it's called lh_*, I guess it stands for live-helper
07:07
<ogra_cmpc>
ah
07:08
that would be odd though ... forking debconf just fro naming conventions
07:08
my live scripts in ubuntu just use the existing tools ... so i cant judge what lh_* does for you
07:08
<daduke>
ogra_cmpc: not guilty, your honor, I just installed and use it ;)
07:09
<ogra_cmpc>
i didnt blame you :)
07:09
<daduke>
ogra_cmpc: I know
07:09Guaraldo has joined #ltsp
07:11
<daduke>
ogra_cmpc: it fails... dunno know
07:11
<ogra_cmpc>
me neither, that would be a question to the debian-live folks i guess
07:12
<daduke>
ogra_cmpc: don't you think they'd blame it on you ltsp ppl since the other 200M install just fine?
07:12
<ogra_cmpc>
well, it works fine in a normal chroot env where we use dbehelper scripts
07:13
*debhelper
07:13
so if they forked debhelper to be something different, i'd rather blame their forked something
07:14
<daduke>
ogra_cmpc: I can try. however, I'm not sure lh_build is not 1 layer 'higher' than debhelper...
07:15
<ogra_cmpc>
might be, i have never heard of it
07:15
in any case the preinst should only fire if /etc/ltsp_chroot doesnt exist in the chroot
07:15
if you have that file and it still fires, something is wrong with debconf
07:16
if [ ! -f /etc/ltsp_chroot ]; then is a pretty clear statement :)
07:18
the debconf parts below only show a note to the user
07:19
<daduke>
ogra_cmpc: I agree, but touching it didn't change a thing. Could the error msg I got be a result of this bailing out? I just asked the debian-live ppl.
07:20
<ogra_cmpc>
note really, the code under the fi statement should be executed at all if the file is there
07:20
s/fi/if
07:21
<daduke>
ogra_cmpc: I guess then we have a fine example of the "2. Hauptsatz der Informatik": der Fehler liegt woanders.
07:23
<ogra_cmpc>
hmm
07:24
do you use ltsp-build-client to create the chroot ?
07:25
it sets a bunch of environment defaults yourpackages might need
07:27
(like DEBCONF_FRONTEND for example)
07:35
<daduke>
ogra_cmpc: nope, the chroot's done by lh_config. but this might be a lead indeed...
07:35
<ogra_cmpc>
oh
07:35
yeah
07:35
you should use ltsp-build-client and call that from whatever tool you want to useto build the iso
07:35
<daduke>
ogra_cmpc: let me take a look at ltsp-build-client (again)
07:36
<ogra_cmpc>
there are lots of settings the plugins make to make sure the packages install and find their defaults as needed
07:36
<daduke>
ogra_cmpc: yeah right, those plugins... hmmm
07:37
ogra_cmpc: thing is, I don't know what the overlap/difference is between the chroots of ltsp and debian-live (apart from ltsp* of course)
07:38TelnetManta has joined #ltsp
07:39slidesinger has joined #ltsp
07:41
<daduke>
ogra_cmpc: lh_config only creates a config/ dir, all the rest (which is a lot) is done by lh_build. I'm afraid that if I replace the lh chroot with the ltsp one, all the debian-live stuff is missing and it won't boot any longer...
07:42
<ogra_cmpc>
well, in the end you only need a modified initramfs for having the live stuff
07:43
(thats why i love casper so much, it just works with any possible chroot :) )
07:45slidesinger has quit IRC
07:47
<daduke>
ogra_cmpc: would you have any idea on how to create a live-capable initramfs for the ltsp chroot?
07:48
<ogra_cmpc>
i would leave that to casper in ubuntu and tweak some settings for it
07:48slidesinger has joined #ltsp
07:48* daduke has to talk to casper then...
07:49
<ogra_cmpc>
not sure thats still suported or regulary updated in debian though
07:49
they seem to have switched everything to debian-live
07:49
<daduke>
:-(
07:50
<ogra_cmpc>
casper is actually an initramfs script ... with tons of optional settings to tweak its behavior ...
07:50
it doesnt run mkisofs or something, thats something you would have to do with a wrapper script
07:51
<daduke>
ogra_cmpc: then it might be easier to try to run the important parts of ltsp-build-client and have lh_build do the rest
07:52
<ogra_cmpc>
yeah, thats waht i was suggesting with "you should use ltsp-build-client and call that from whatever tool you want to use to build the iso"
07:53deavidsedice has joined #ltsp
07:54
<cliebow>
anyone have a link to warren's tutorial from yesterday?
08:17cyberorg has joined #ltsp
08:19Gadi has joined #ltsp
08:20
<tjikkun_work>
cliebow: maybe https://fedorahosted.org/k12linux/wiki/InstallGuide ?
08:21mikkel has quit IRC
08:23
<timborn>
cliebow - http://deltacfax.dyndns.org/timborn/LTSP-warren-080330.txt
08:24
i'm sure he will post it somewhere sane soon
08:30timborn has quit IRC
08:30K_O-Gnom has quit IRC
08:37deavidsedice has quit IRC
08:41
<cliebow>
timborn:thanks..
08:47praveer_cool has joined #ltsp
08:49mikkel has joined #ltsp
08:56praveer_fedora has quit IRC
08:58jammcq has joined #ltsp
08:58
<jammcq>
hey homies
09:00FLOLO has joined #ltsp
09:02
<cliebow>
jammcq!
09:05
<jammcq>
hey chuckster
09:05
<cliebow>
8~)
09:11
so if i have done a co of ltsp-trunk.. which exists in /ltsp-trunk..do i bzr branch or pull to create my own playpen version?
09:11
<ogra_cmpc>
yeah
09:11
branch
09:12
bzr branch ltsp-trunk ltsp-cliebow
09:12
<cliebow>
ahhh..workewd that time..
09:12
686 revision
09:12
s
09:15
<ogra_cmpc>
sounds familiar
09:18
<cliebow>
how elaborate is building deb from the source tree?
09:18mccann has joined #ltsp
09:19
<ogra_cmpc>
not working anymore since we had to split out the packaging bits
09:20
pull the source package, roll a new orig.tar.gz, update the packaging bits to apply to that and roll a new source package is todays way
09:21
<cliebow>
cool!..that i can get my head around
09:21
<ogra_cmpc>
its a ton more complicated than it was before but was necessary if we want to have independent upstream
09:21
<Pascal_1>
hello
09:21
any news about the pam/ldm bug ?
09:21
<ogra_cmpc>
you might also need the mkdst tree
09:22
<cliebow>
ok..
09:22
<ogra_cmpc>
(to roll the tarball, even though just tarring up teh source should suffice, you will just have all teh bzr metadata in it then)
09:22
(that increases size by about 2M)
09:24
<cliebow>
i see..
09:24* cliebow cliebow looks at mkdst
09:25Q-FUNK has quit IRC
09:26
<cliebow>
not much more than a shell script then...
09:27toscalix has quit IRC
09:28
<ogra_cmpc>
right
09:28
just calling bzr export and tar ... and giving the right filemanes
09:32MasterLTSP has joined #ltsp
09:49vagrantc has joined #ltsp
09:58
<cliebow>
/plugins/ltsp-build-client/Ubuntu..wjere are these sourced from?
09:58
<ogra_cmpc>
you mean the target dir where they are installed on the system ?
09:59
or do you meqan from which script ?
09:59
the latter is ltsp-build-client, the dir should be /usr/share/ltsp/plugins/ltsp-build-client/
10:00
<cliebow>
ohhh..that was dumb..
10:00
<vagrantc>
i /me waves from madrid
10:00
er.
10:00* cliebow cliebow waves from snowbank
10:00
<ogra_cmpc>
vagrantc, airport ?
10:00
<vagrantc>
ogra_cmpc: hostel
10:00
<ogra_cmpc>
ah
10:00
not on your way back yet ?
10:01
<vagrantc>
way back?
10:02
whoah! mips built ldm!
10:02
it's a miracle
10:02
<cliebow>
so on install /usr/share/ltsp is populated..
10:03
must be that hostel environment
10:06
<ogra_cmpc>
vagrantc, well, i thought its a weekendish event
10:06
<vagrantc>
i think it's the 2nd-6th-ish ... and i got here a couple days early to actually be mentally conscious once we get to working :)
10:09
<ogra_cmpc>
ah
10:09
welcome to my timezone then :)
10:10
i just booked my prague flight
10:10sepski has joined #ltsp
10:17
<vagrantc>
hrm. so now it's only hppa that's holding back sid -> lenny progression for ldm and ltsp
10:17DonSilver has joined #ltsp
10:28staffencasa has joined #ltsp
10:42tjikkun_work has quit IRC
10:44FLOLO has quit IRC
10:46pdjbarber has quit IRC
10:53j_h has joined #ltsp
10:55j_h has left #ltsp
11:17makghosh has joined #ltsp
11:29makghosh has quit IRC
11:39MattyOv has joined #ltsp
11:39
<MattyOv>
Hi all - I am having some problems with LTSP on edubuntu - can anyone help at all?
11:47
<cliebow>
MattyOv, tell us what the prob is...
11:48
<MattyOv>
Hiya, well I have an intel quad core and the thin clients just hang - even if I build the clients with ltsp-build-client --arch i3867
11:48
(i386)
11:50
<cliebow>
where do they hang??
11:51rjune has joined #ltsp
11:51
<MattyOv>
on the ctrl-alt-f1 screen I just get Ubuntu ltsp tty
11:51
ltsp login:
11:52
<cliebow>
are you saying you drop to a busybox shell because the squashfs doent mount?
11:53
<MattyOv>
Erm I have no idea if that's what I am saying :) I have set this up before and it just worked - it seems the Intel quad core server is causing a prob with AMD64. I have followed lots of stuff in google for dealing with this but got nowhere
11:53
have now gone back to a 32 bit install on this machine to try to get the ltsp server working ok
11:54
<cliebow>
ogra may be listening in..
11:54
in /etc/inetd.conf you have a line for nbdroot right?
11:55
<MattyOv>
2000 stream tcp nowait nobody /usr/sbin/tcpd /usr/sbin/nbdrootd /opt/ltsp/images/i386.img
11:55
like that?
11:55
<Gadi>
MattyOv: have you changed lts.conf at all?
11:56
<cliebow>
Gadi is THE MAN
11:56
<MattyOv>
nope not that I know of - I did change dhcp.conf
11:56
<Gadi>
sounds like X is just not coming up for you
11:56
if you ctrl-alt-f7 are you at a graphical login?
11:57
<MattyOv>
nope - just hda_intel: azx_get_response timout, switching top single_cmd mode
11:57
<cliebow>
rmphh
11:58
<Gadi>
do this:
11:58
sudo chroot /opt/ltsp/i386 passwd
11:58
enter a password
11:58
then, sudo ltsp-update-image
11:58
and reboot the terminal
11:58
then, you can ctrl-alt-f1, log in as root on the terminal, and have a look at /tmp/ldm.log
11:59
<MattyOv>
ah that sounds like a good plan!
12:00
<cliebow>
i was just about to say that too..what Gadi said
12:00
8~)
12:00
<MattyOv>
:)
12:01
<Gadi>
MattyOv: you may also want to check the X log on the terminal
12:01
my guess is that X died
12:01
<MattyOv>
I just have a file called /tm/ldm/192.168.0.1
12:02
ah right ldm.log is /var/log/ldm.log
12:02
it just says starting Xorg
12:02
<cliebow>
check xorg log..
12:03
<MattyOv>
in Xorg.6.log I get 'fatal server error: no screens found'
12:03
<Gadi>
there ya go
12:03
<MattyOv>
above that 'Screen(s) found, but none have a usable configuration
12:03
<cliebow>
what video card on client?
12:03
hmmm
12:03
<Gadi>
have a look at your /etc/X11/xorg.conf
12:04
<MattyOv>
it's a laptop - but the same happens on the desktop I have too
12:04
<cliebow>
some nvidia chip?
12:04
<MattyOv>
maybe not too sure what the chip is (in either)
12:04
<Gadi>
grep Driver /etc/X11/xorg.conf
12:05
<cliebow>
he could force vesa driver in lts.conf..
12:05
?
12:05
see what happens/..
12:05
<MattyOv>
I have Driver "vesa"
12:05
<cliebow>
hee hee
12:05
maybe an intel choip?
12:05
<Gadi>
MattyOv: can you paste your xorg log?
12:05
!pastebot
12:05
<ltspbot>
Gadi: "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.
12:06
<Gadi>
you can scp it from the terminal to the server
12:08
<ltsppbot>
"MattyOv" pasted "Xorg.1.log" (652 lines) at http://pastebot.ltsp.org/489
12:09
"MattyOv" pasted "Xorg.6.log" (2025 lines) at http://pastebot.ltsp.org/490
12:09deavid has quit IRC
12:10
<Gadi>
its your monitor
12:10
do this:
12:10
create a file on the server called "/var/lib/tftpboot/ltsp/i386/lts.conf"
12:10
in there put:
12:10
[default]
12:10
X_VERTREFRESH = "60-85"
12:11
X_HORZSYNC = "30-100"
12:11
... then, reboot the thin client
12:13
<MattyOv>
wow - that totally makes it work
12:13chupacabra has quit IRC
12:13
<MattyOv>
:) I am so rubbish :(
12:13
<Gadi>
?
12:13
<MattyOv>
so it was basically that X could not drive the monitor - for both clients?
12:13
<Gadi>
no
12:13
<MattyOv>
:) oh ok
12:13
<Gadi>
it was that X could not derive the correct sync ranges from the monitor
12:13
<johnny>
vagrantc, so how did have the bind mounts method working?
12:14
<Gadi>
X asks the monitor what sync ranges to use
12:14
<vagrantc>
johnny: could you be more specific?
12:14
<Gadi>
and it was returning bad numbers
12:14
<johnny>
oh.. i forgot i had a more immediate question
12:14
<vagrantc>
johnny: or you just want a general overview?
12:14
<Gadi>
so, there were no valid video modes for the ranges specified
12:14
<johnny>
./autogen is failing in ltspfs
12:14
<MattyOv>
well thank you very much for that - it has been driving me nuts for days
12:15
<Gadi>
MattyOv: now, try this:
12:15
<johnny>
.autogen.sh that is, it's looking for a ./ChangeLog file
12:15
<Gadi>
MattyOv: comment out those two lines in lts.conf
12:15
<johnny>
but that was removed to use bzr commit log instead iirc
12:15
<Gadi>
and put: CONFIGURE_X=False
12:15
and see if you get the same result
12:16
<cliebow>
Gadi Rocks 8~)
12:16
<MattyOv>
so this is all irrelevant what hardware I am running on - it is just config thing for X?
12:16chupacabra has joined #ltsp
12:17
<Gadi>
MattyOv: not irrelevant at all
12:17
here's the deal:
12:17
<MattyOv>
(I get that no codecs installaed message)
12:17
<Gadi>
there is a configure-x script on the thin client that uses X -configure to create a xorg.conf file
12:17
if you turn CONFIGURE_X = False, the script does not execute
12:17
<vagrantc>
johnny: maybe we should include a single changelog entry that says "this file should be automatically generated using bzr log."
12:18
<Gadi>
in generating xorg.conf, configure-x will get the sync ranges from the monitor
12:18
<johnny>
i don't think that is necessary
12:18
<Gadi>
so, you have 2 choices:
12:18
turn off the script
12:18
or set the ranges manually
12:18
<johnny>
Makefile.am: required file `./ChangeLog' not found
12:18
<Gadi>
1: CONFIGURE_X = False
12:19
<vagrantc>
well, how do we ensure a ChangeLog gets created?
12:19
<Gadi>
2: X_VERTREFRESH.... X_HORZSYNC....
12:19mccann has quit IRC
12:19
<vagrantc>
./autogen.sh shouldn't assume you're using bzr
12:19
<MattyOv>
ok cool - well it appears turning it off also causes it to fail so I need the second option the I guess
12:19
<Gadi>
MattyOv: thats a better one anyway, imo
12:19
:)
12:19
<cliebow>
configure-x.sh ?
12:20
<MattyOv>
Ok cool - so what I should do now - is go back to a 64bit architecture of 8.04 on my server and build the client with the --arch i386 and make these changes again?
12:21
<johnny>
vagrantc, are you sure we should even care?
12:21
just make it not care about a ChangeLog file
12:21
btw.. this is pulling from a bzr branch
12:21
<vagrantc>
johnny: i am *sure* that we *should* care about having a ChangeLog of some form or fashion.
12:21
<johnny>
i think one web accessible is fine?
12:21
<vagrantc>
johnny: wether we actually do is another issue.
12:22
johnny: how do you get a web accessible changelog from a tarball?
12:22
<Gadi>
MattyOv: if that was all it was :)
12:22
<johnny>
it's already web accessible in loggerhead
12:22
<MattyOv>
Gadi - many thanks for that help - it has been days fiddling and getting no-where. You are indeed a star!!!
12:22
<vagrantc>
johnny: how do you know that the changelog included in your tarball matches the web accessible one?
12:23
<MattyOv>
Yeah it could be a mix of things but now I am a bit better armed I can't see why the LTSP stuff would be so different from 7.1 to 8.04 of ubuntu
12:23
(optimist)
12:23
<vagrantc>
7.10
12:23
<Gadi>
MattyOv: pop in anytime
12:24
<johnny>
vagrantc, why would it? especially if the tarball is out of date
12:25
<cliebow>
Gadi, configure-x.sh is sourced from ltsp-core?
12:25
<vagrantc>
johnny: i don't get it.. .how on earth can you get a changelog from the web and knnow that it corresponds to an arbitrary tarball?
12:25
<johnny>
it won't.. and i don't care that it does
12:25
<vagrantc>
and we *should*
12:25
that's what i'm saying.
12:25
<johnny>
either way.. i'm operating from a bzr repo :)
12:26
atm
12:26
and i can't make ltspfs
12:26
i think it should ignore ChangeLog if it doensn't exist
12:27MasterLTSP has quit IRC
12:27
<vagrantc>
it should create a dummy changelog if it doesnt't exist, i think...
12:27
<johnny>
i'm not that concerned about the argument of wheter we should have a changelog, i just want ./autogen.sh to work
12:27
<vagrantc>
sure.
12:28
<johnny>
http://bazaar.launchpad.net/~ltsp-upstream/ltspfs/ltspfs-trunk/revision/wtogami%40redhat.com-20080326204915-dibhtmhkgiehlt6l?start_revid=wtogami%40redhat.com-20080326204915-dibhtmhkgiehlt6l
12:28
<Gadi>
cliebow: ltsp-client-core
12:29
<cliebow>
ahhh
12:29
<johnny>
it looks like the intention was to generate the changelog from bzr, but it's not doing that atm
12:29
<vagrantc>
johnny: it does it from mkdst
12:30
<johnny>
can't autogen ignore this missing changelog file?
12:30
the odd part.. is i can't figure out how its looking for the ChangeLog grep -r ChangeLog shows nothing
12:30* vagrantc won't restate vagrantc's opinion
12:30
<dberkholz>
i love how i can't get to any ubuntu-related websites for work.
12:30
from work, that is
12:30
<johnny>
that's fantastic..
12:30
<cliebow>
Gadi:ltsp-client-core is generated from ltsp-trunk?
12:31
<Gadi>
cliebow: ltsp-trunk may have moved things around
12:31* Gadi hasn't been keeping up with the saga that is ltsp-upstream of late
12:31
<Gadi>
other than the spicy emails :)
12:31
<cliebow>
i am trying to get a handle on the moment
12:32
<Gadi>
heh
12:32vagrantc has quit IRC
12:32
<cliebow>
find /ltsp-trunk -name *client* reurns nada
12:33
<dberkholz>
in xorg, we autogenerate ChangeLog from git at `make dist` time
12:33
<johnny>
my autofoo must be weak
12:33
~/projects/ltsp/ltspfs-trunk $ grep -r ChangeLog *
12:33
how does it even know to look for a ChangeLog if that returns nothing?
12:34MattyOv has quit IRC
12:36
<johnny>
this is before ./autogen.sh
12:37
after that, i see it being used in Makefile.in
12:41
<dberkholz>
johnny: did you see where vagrantc mentioned mkdst
12:42
<johnny>
sure
12:42
<dberkholz>
that's where the magic happens
12:43
<johnny>
so.. if i was doing a bzr live ebuild, how would that work?
12:43
ldm atm doesn't require mkdst.. it just works
12:46
<dberkholz>
you would use mkdst if you need a changelog, otherwise don't
12:47rjune has quit IRC
12:49
<johnny>
i don't need one, but ./autogen.sh does
12:49
it generates an exit status >0
12:49
so BZR_BOOTSTRAP fails
12:50
and i don't think autogen.sh should hard require that file
12:50
but i can't figure out how it's doing it
12:52
in automake --help it has a section called "Files which are automatically distributed, if found:"
12:52
and ChangeLog is mentioned in there
12:53
oh.. maybe -a is doing it..
12:53chupacabra has quit IRC
12:57chupacabra has joined #ltsp
12:59rjune has joined #ltsp
13:00mikkel has quit IRC
13:03
<dberkholz>
johnny: why do you feel like you have to run autogen.sh?
13:04
autoreconf -v --install
13:04
(or inherit autotools and run eautoreconf in an ebuild environment)
13:04
<johnny>
as i said.. my autofoo is weak.. i just thought that was the proper method
13:04
i've only done ebuilds for scripts
13:05
i thought eautoreconf was like really slow :)
13:05
<dberkholz>
it's not any slower than autogen.sh, because it does the same stuff
13:06
<johnny>
that errors out the same
13:07MattOv has joined #ltsp
13:07
<MattOv>
Hey all
13:08
<dberkholz>
ok, run the one by hand then. probably eautoreconf doesn't do --install
13:08
<MattOv>
Gadi - are you about at all. I have a nice new problem now I have gone back to 64bit ;)
13:09
<johnny>
dberkholz, i was talkng about the one by hand :)
13:09
i ran that in the bzr checkout
13:10
there's something hard requiring the file, but i don't know what it is
13:10
grepping shows nothing .. so how is that requiring it???
13:11
<MattOv>
Can anyone help me with an issue where my thin client just hangs just after it says 'Negotiation: ..size=143508KN'
13:11
<dberkholz>
johnny: eh, gnu tools require a changelog file. i was hoping --install would touch an empty one
13:12
johnny: if you touch ChangeLog, it'll work around that
13:12
<johnny>
me too :)
13:12
yes.. i did work around that
13:12
i thought the add missing option would put it in there
13:13
i naively hoped that EBZR_BOOTSTRAP="touch ChangeLog && ./autogen.sh" would work
13:13
EBZR_BOOTSTRAP="autogen.sh works..
13:13
but it does eval..
13:17chupacabra has quit IRC
13:19chupacabra has joined #ltsp
13:20
<johnny>
i'm gonna get ready and talk to the leasing manager, and also deposit some money in the bank
13:20
bbiab
13:21stgraber_ has joined #ltsp
13:22sonjag has joined #ltsp
13:23DonSilver has quit IRC
13:25stgraber_ has left #ltsp
13:28Pascal_Debian has joined #ltsp
13:29
<MattOv>
Could it be something to do with nbd-server (I am really clutching at straws!)
13:30
<Gadi>
MattOv: I am in and out
13:30
whats up?
13:30
<MattOv>
ah hi - sorry
13:30
I went back to ubunutu 8.04 and now it's dead again
13:31
<Gadi>
ah, 8.04
13:31
<MattOv>
just hangs at Negotiation: ..size=143508KB
13:31
<Gadi>
more like 8.04 alpha
13:31
:)
13:31
<MattOv>
yeah it seems that way
13:31
<Gadi>
ogra_cmpc can help you more on that end
13:31
I have not played with the packages there
13:32
hmm...
13:32* Gadi wonders what point Negotiation is...
13:32
<MattOv>
cool... cheers fr all the help. It seems it is hard to get all the individual components to work
13:32
I was just reading about nbd-server but it all means very little to me really
13:32
<Gadi>
is that nbd-client?
13:33
ah
13:33
sounds like nbd
13:33
which sounds like it found the file
13:34
you built the chroot with --arch=i386
13:34
<MattOv>
yeah it def found the file now - I edited the dhcp.conf
13:34
yeah indeed
13:34
<Gadi>
edit: /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default
13:34indradg has joined #ltsp
13:34
<Gadi>
and remove the words: "quiet splash"
13:34
then, reboot
13:35
<cliebow>
is nbdserver compiled for 64 bit?
13:36
seems it must be
13:39
any good rreason not to edit sudoers with vi?
13:39
<MattOv>
sorry back - keyboard went crazy !
13:40
it hangs at: Begin: Running /scripts/nfs-bottom ...
13:40
Running /scripts/init-bottom
13:41Pascal_Debian has quit IRC
13:41
<MattOv>
(I am not sure what nbdserevr is complied for)
13:42pascal has joined #ltsp
13:48
<sonjag>
Hey Gadi... can you tell me how to restrict the use of Wanda the Fish? My students are using it to run vncviewer and causing mischief ;)
13:51
<cliebow>
dont suppose fish in pessulus can be disabled?
13:52
<sonjag>
cliebow, is pessulus in 7.04 or is that an add on? I don't see it in system->admin
13:52
<cliebow>
i think i installed with aptget
13:53
<sonjag>
cliebow, what's lockdown editor? I see that I can check off fish but I don't know what checking it off does?
13:53
<cliebow>
heh..i am not sure..i cant ebven find the fish
13:54
<Gadi>
uhm, does wanda the fish have an executable name?
13:54
(go to properties)
13:54
<cliebow>
isee fish_applet
13:54
<Gadi>
if so: chmod a-x <executable_name>
13:55
<johnny>
i'd really like to get pessulus up to speed
13:55
but i need a helper :)
13:55
i'll get back to that after i get my ltsp integration finished
13:56
<MattOv>
Has anyone ever got LTSP going with ubuntu 8.04 with an AMD64 server and i386 clients?
13:56
I am beginning to think it is not possible ;)
13:58
<johnny>
why are you using 8.04/
13:58
i have it working in 7.10 just fine
13:58
minus flash
13:58sepski has quit IRC
13:58
<MattOv>
well I am just doing a new install and would rather do it in 8.04 rather than have to upgrade it all later
13:59
<johnny>
well i haven't upgraded my gutsy yet, since hardy hasn't been released
14:00
i do have hardy working with i386 on both sides
14:00
so.. i don't see the diff
14:01
<MattOv>
My old server with is i386 works fine - it is just this new server which is 64bit doesn't
14:01
<cliebow>
MattOv, loks to me like init-bottom has one script..unionfs_cow
14:01
any reason why unionfs would bork in amd64?
14:02
<MattOv>
I can't even find the init-bottom thing
14:03viking-ice has quit IRC
14:03
<johnny>
did you try the new server with i386?
14:03
otherwise you're testing two different things?
14:03chupacabra has quit IRC
14:04
<MattOv>
I tried this box with 7.10 and i386 and after some tweaking with Gadi that worked
14:04
I have tried this box 8.04 i386 before but it didn't work and now back to AMD64 :)
14:05
<johnny>
so.. the problem has nothing to do with64 bit then
14:05deavid has joined #ltsp
14:07
<MattOv>
I don't know where the exact problem lays - my skills are lacking too much to determine the exact location of the problem (or rough location!). I have followed all the howtos about setting it up but all my clients hang
14:07chupacabra has joined #ltsp
14:07
<MattOv>
(I am using a laptop and a spare desktop to test on the thin client side)
14:08
<cliebow>
MacIver, i think you'll have to see ogra..he is on German time..
14:08
whoops
14:09
Mat
14:09
yikes
14:09
<MattOv>
ok cool - he seems to be the man in the know :)
14:09
<cliebow>
he wrote most of it
14:09
<MattOv>
unlike me - who is the dunce.
14:09
<cliebow>
with help from his firends
14:10
<MattOv>
well yes you are all obviously rather excellent ;)
14:10* cliebow cliebow preens cluelessly
14:10
<MattOv>
I have now got to the point where I just need to solve this to stop me going mad - I mean how hard can it be (or alternatively how thick must I be!)
14:11
man vs machine and machine winning about 20 nil
14:11
;)
14:11
<cliebow>
20000 nil
14:11
<MattOv>
:)
14:13
<cliebow>
i only got ltsp-trunk today..so your questions were great stimulus
14:14
<MattOv>
:) Glad to be of some help!!!!
14:15
I am sure answering my questions helps you to re-enforce the basics ;)
14:16
<cliebow>
ther is something Googly in unionfs/hardy..waiting for lonk to open
14:16
<warren>
lonk?
14:17
<cliebow>
wink..link schmink..8~)
14:20Shingoshi has quit IRC
14:21otavio has joined #ltsp
14:22bobby_C has joined #ltsp
14:22Shingoshi has joined #ltsp
14:25otavio_ has joined #ltsp
14:28
<MattOv>
yay - got it going
14:29
found a post from mr Oli :)
14:30
Solved the problem - it was a color depth thing
14:35otavio has quit IRC
14:35otavio_ has quit IRC
14:35otavio has joined #ltsp
14:46indradg has quit IRC
14:47Gadi has left #ltsp
14:49
<sonjag>
cliebow, looks like lockdown editor is not an admin tool but just for an indiv to lock down their acct. Bummer. I guess I need to look at Pessulus/Sabayon and see if I can make that work. Any idea if it will work if you're using AD for authentication?
14:51chupacabra has quit IRC
14:52chupacabra has joined #ltsp
14:52jimcooncat_ has joined #ltsp
14:52viking-ice has joined #ltsp
14:53jimcooncat_ has left #ltsp
14:59rjune has quit IRC
15:04jimcooncat_ has joined #ltsp
15:04rjune has joined #ltsp
15:04
<jimcooncat_>
I'd like to know what happens on a small ltsp setup with a long term power outage. When all the systems come back up, do light clients hang because the DHCP server isn't ready yet? How do you deal with this?
15:06MattOv has quit IRC
15:07
<dberkholz>
easy way might be to put the server on a different power switch that you could power up a few minutes earlier
15:07
<warren>
jimcooncat_, most clients repeatedly retry DHCP until it works.
15:08
<jimcooncat_>
thanks, my short tests made me think that they just hung, but I need to more fully investigate when I get a chance.
15:08
<dberkholz>
eh, i guess my comment might apply more to a downed nfs server
15:08
i have a horrible memory
15:08
<jimcooncat_>
dberkholz: it's still a good suggestion.
15:09
<cliebow>
sonjag: i wouldnt think authentication would matter...
15:10mhterres has quit IRC
15:10mhterres has joined #ltsp
15:13antocm has joined #ltsp
15:13
<antocm>
Hello. Does anyone knows about suspending a session in one LTSP terminal, and opening the same session again from another terminal?
15:13
Something like http://wwwu.edu.uni-klu.ac.at/mkropfbe/xray.html
15:15cliebow has quit IRC
15:15
<johnny>
nothing is integrated into ltsp to do that atm afaik
15:16
it would be a good addition
15:17
<dberkholz>
you could use gnome's session saving
15:17
<antocm>
I think the same. We should draw LTSP people attention to something like this.
15:17
<johnny>
antocm, they have enough to do already
15:17
<jimcooncat_>
antocm: I do that with vnc
15:17
<johnny>
we need somebody to implement it :)
15:18
dberkholz, that only reopens programs to where they were
15:18
i think he wants something like screen, but for X
15:18
<antocm>
vnc is not the same. Gnome is also not the same. Handling X protocol directly is the way.
15:18
<dberkholz>
johnny: that's generally what i think of when i think of session suspending
15:18
<johnny>
gnome session saving doesn't do that
15:18
afaik
15:18
<dberkholz>
being where you were
15:18
<warren>
jimcooncat_, due to power spiking dangers you will want things to not turn on after a power outage automatically anyway.
15:18
you don't want EVERYTHING to turn on at once
15:18
most computers use a lot more power at bootup than during normal operation which is variable
15:19
<jimcooncat_>
freenx is supposed to do that, but I've never been able to get it to restore after suspending
15:19indradg has joined #ltsp
15:20
<jimcooncat_>
warren: I think you're quite right, I'll probably want to set the client's BIOS to stay off after a power out
15:20DonSilver has joined #ltsp
15:21sonjag has quit IRC
15:22
<jimcooncat_>
antocm: there's also xmove, which I haven't tried yet
15:22
<antocm>
jimcooncat_, I will analyze your suggestions. Thanks.
15:22otavio has quit IRC
15:22chupacabra has quit IRC
15:23
<jimcooncat_>
antocm: yw. Wish you luck, I'd like to find out myself how to do this "right".
15:24
<antocm>
johnny: Wouldn't it be very nice, to see an implementation under LTSP with any tool that could provide session roaming?
15:24
<johnny>
it would be nice to have it for anything, not just for ltsp
15:25
<dberkholz>
someone worked on Xscreen a couple years ago for summer of code
15:25
feel free to pick it up =)
15:25
<johnny>
did it go very far? i only saw the spec, and i couldnt find any implementation
15:26
<antocm>
So, options you know can be x-ray, freenx and Xscreen right?
15:26
<dberkholz>
http://lists.freedesktop.org/archives/xorg/2007-April/023169.html
15:27jimcooncat_ has left #ltsp
15:34TelnetManta has quit IRC
15:36
<antocm>
dberkholz: i don't see anything other than specs. Do you know of something else?
15:37
<dberkholz>
did you read that email?
15:37
specifically, paragraph 2
15:38
<antocm>
OK, i see. I read http://code.google.com/soc/2006/xorg/appinfo.html?csaid=73A89F18E7770493
15:40
<dberkholz>
i have no idea what the status of that code is, however
15:40DonSilver has quit IRC
15:40
<dberkholz>
i just asked the author, to see if he wrote anything
15:41chupacabra has joined #ltsp
15:46
<johnny>
hmm.. gentoo-sources doesn't include fuse by default :(
15:59mccann has joined #ltsp
16:01
<antocm>
dberkholz: If he say something, would you keep me in the loop?
16:04
<dberkholz>
johnny: emerge fuse ...
16:04
antocm: sure.
16:05
<warren>
dberkholz, what means do you use to ensure that users have access to fuse?
16:05
dberkholz, fuse group? setuid binary?
16:05
<johnny>
we don't use a fuse group in gentoo by default :(
16:05
<dberkholz>
warren: this is kind of experimental, i haven't done much with fuse before, only used kernel unionfs
16:05
<johnny>
dberkholz, fuse is needed for ltspfs
16:06
which is needed whether we use unionfs or not
16:06
<dberkholz>
-rws--x--x 1 root root 26576 2008-03-16 22:42 /usr/bin/fusermount
16:06slipttees has joined #ltsp
16:07
<dberkholz>
we also allow nonroot users to specify allow_other or allow_root mount options through fuse.conf 'user_allow_other'
16:07
oh, fuse is in 2.6.25? didn't realize that. cool
16:08
<johnny>
fuse was in earlier i thought?
16:08
<dberkholz>
not according to our fuse ebuild, it checks for >= .25
16:08
<warren>
fuse was much earlier
16:08
<johnny>
look at an earlier ebuild dberkholz :)
16:08
<warren>
2.6.14
16:08
<dberkholz>
ah, i bet i know
16:09
the separate module just isn't compatible with .25
16:09
<johnny>
but it's not installing under 2.6.24 either..
16:09
<dberkholz>
the logic in there isn't really making sense to me.
16:10
<johnny>
the x86 ebuild doesn't work in 2.6.24 at all
16:10
<dberkholz>
well, the logic is, the warning messages aren't.
16:10
<johnny>
with a check
16:10
<dberkholz>
non-native english speaker, i guess
16:10
<johnny>
the ~x86 ebuild tries to work under 2.6.24, but isn't for me :)
16:11deavid has quit IRC
16:12
<dberkholz>
it should be obvious at this point that i don't know jack about fuse
16:13antocm has left #ltsp
16:13mhterres has quit IRC
16:14
<dberkholz>
and the ebuild's using an extremely poorly documented feature of CONFIG_CHECK. trying to figure out what it does
16:16
did the name or install location of the fuse kernel module change at some point?
16:24
<warren>
does it work for the root user?
16:25
dberkholz, which fuse filesystem are you trying to test with?
16:25
dberkholz, something more normal like sshfs or encfs might be a better test to see if it is working at all.
16:29chupacabra has quit IRC
16:33chupacabra has joined #ltsp
16:38jammcq has quit IRC
16:39slidesinger has quit IRC
16:43slipttees_ has joined #ltsp
16:43slipttees has quit IRC
16:46
<dberkholz>
warren: there's a pretty neat one that does ftp sites
16:46
<warren>
dberkholz, what is it called?
16:46* warren wants one that can do http sites
16:46
<dberkholz>
curlftpfs
16:46
<warren>
oh!!!
16:46
curl
16:46
that has potential to do http as well
16:52slipttees_ is now known as slipttees
16:52Guaraldo has left #ltsp
16:55vagrantc has joined #ltsp
16:58
<johnny>
dberkholz, i think we're going to need to provde our own custom kernel with patches at this rate :(
16:59
<vagrantc>
whoah.
16:59
<dberkholz>
johnny: could you provide a little more detail, please? i'm not following you.
17:00
wb vagrantc
17:00* vagrantc is getting a good share of cheese and wine in spain
17:01
<dberkholz>
lucky man
17:01
<vagrantc>
estoy baracho
17:05
<johnny>
dberkholz, since gentoo-sources doesn't include aufs/unionfs and fuse isn't compiled in by default
17:05
<dberkholz>
johnny: who cares about default? and aufs is available as a separate package
17:05
johnny: we can certainly provide a custom kernel config to genkernel
17:06
<johnny>
now if only genkernel would freakin let me cross compile the utils :(
17:07slipttees has quit IRC
17:07
<johnny>
vagrantc, have you seen this patch? for initramfs-tools conf? https://bugs.launchpad.net/ubuntu/+source/ltsp/+bug/190016
17:08
with aufs support
17:10
<vagrantc>
no, haven't
17:11* vagrantc would like to re-write the initramfs-tools stuff almost from scratch
17:14
<johnny>
so.. i have working ebuilds for ldm/ltspfs/ltsp-server/ltsp-client
17:14
oh.. not ltsp-server yet.. oops
17:14
only thing ltsp-client needs is working init scripts
17:15
i've been working on that
17:15chupacabra has quit IRC
17:18
<dberkholz>
johnny: excellent
17:19
<johnny>
ok.. i'm going back to fighting genkernel again
17:19
genkernel is the evil
17:19
i did find an initramfs-tools ebuild
17:19
but only in the forums oddly enough
17:19chupacabra has joined #ltsp
17:20* johnny wonders why genkernel does it's own initramfs thing instead of using initramfs-tools
17:20
<vagrantc>
if you can get initramfs-tools on gentoo, that will be a whole lot more code we can share.
17:20
presumably.
17:21
although not using the typical gentoo conventions probably has it's own costs.
17:21
<johnny>
sure, but genkernel is not very maintained atm, and won't be revamped until after 2008.1
17:21
err 2008.0
17:22
err. that's unfair
17:22
it is maintained, but only really for issues that concern the 2008.0 release
17:23
<dberkholz>
it's maintained, they just aren't adding new features they have no need for. i have no doubt that they will take patches
17:23
initramfs-tools probably didn't do some kind of fancy bootsplash they wanted
17:25
<vagrantc>
there's debian packages for fancy bootsplashery
17:25
<johnny>
it'd be nice to have it crosscompile, so i can run it outside of the chroot tho :(
17:25
<vagrantc>
it's very pluggable.
17:25
<johnny>
i'll get back to that later i guess
17:26
<dberkholz>
vagrantc: one of our devs wrote his own splash stuff, kinda doubt it made it to debian
17:26
<johnny>
now we have that v86d thing..
17:26
<dberkholz>
yeah, that'll make it upstream at some point
17:26
<vagrantc>
dberkholz: well, i'm not saying that gentoo's stuff is in debian, but that sort of thing is possible with initramfs-tools
17:26
dberkholz: and that initramfs-tools is modular enough that it might be easy to add.
17:27
<dberkholz>
vagrantc: maybe so. then we'd have to push for an overall move to initramfs-tools (to do this properly), and i'm not sure of the size of that commitment
17:28
<vagrantc>
dberkholz: right, that's the sort of thing i mean by the costs of not using typical gentoo conventions
17:28
<dberkholz>
basically we'd have to rip out the current initramfs code in genkernel and replace it with initramfs-tools
17:28
should be doable, not familiar enough to estimate the time
17:29
<johnny>
it shouldnt' be that big of a deal, the genkernel initramfs stuff isn't that much code
17:29
<dberkholz>
in general, i always like the idea of using other people's code
17:29
<johnny>
at least in genkernel3
17:29
maybe i should bug genkernel4 peeps to use initramfs..
17:30
<dberkholz>
gk3 is all we care about till gk4 (which needs a rename, anyway, it's a total rewrite) makes a release
17:30
use initramfs-tools, you mean?
17:30
<johnny>
yes
17:30
<dberkholz>
pretty sure they already use initramfs =)
17:30
<johnny>
vagrantc, how long are you going to be about?
17:31
i'm really thinking about getting food
17:31
tummy growling
17:31
<dberkholz>
johnny: i'd love to start looking at your code at some point. when do you think you might be able to post your current work?
17:31
<johnny>
which do you want to see?
17:31
<dberkholz>
johnny: everything you've done
17:31
=)
17:31
<johnny>
ok.. i'll put it up on my web
17:32
<dberkholz>
johnny: can just post the bzr repo somewhere and i'll pull it ... for ltsp overlay, just push to the upstream place
17:33
<johnny>
i want to show you before committing
17:33
<dberkholz>
johnny: it might be nice to have a history of changes
17:34
<johnny>
well i don't like committing my ebuilds when i've never done an offical one
17:34
i'd like to to show me where i screwed them up :)
17:34
you to show me*
17:35
<vagrantc>
johnny: i should really try and get some sleep. some debian developers fed me spanish cheese and got me drunk.
17:36
<johnny>
well get the hell out of here then
17:36
lol
17:36
get some sleep
17:37
<vagrantc>
yeah.
17:38
<johnny>
dberkholz, see how i'm doing here please :)
17:38
http://pastebin.ca/965332
17:39vagrantc has quit IRC
17:41
<johnny>
the dodoc for the example dir is fine, but not for NEWS and README
17:41
<dberkholz>
why not?
17:41
<johnny>
dberkholz, i've never done an ebuild that was going to be used by somebody other than me or a close friend
17:42
<dberkholz>
you can delete the src_compile, it's the default behavior
17:42
you aren't using any EAPI=1 behavior, get rid of that
17:43
the changelog touch could use comments, and i suspect glib also belongs in RDEPEND
17:43
<johnny>
it is?
17:43
seemed like a build dep to me
17:43
<dberkholz>
try to avoid mixing the nonstandard BZR variables in with the standard ones, put them above to me
17:43
glib is a library, chances are good that it gets linked to
17:44
<johnny>
hm.. i always get confused there..
17:44
ok. i'll give that a shot
17:44
<dberkholz>
basically DEPEND=glib, RDEPEND=$DEPEND fuse
17:44
<johnny>
oh.. also belongs..
17:45
i missed also :)
17:45
<dberkholz>
if libfuse or fuse headers aren't used at all during the build
17:45joebob777as7 has joined #ltsp
17:45
<johnny>
ok.. so did you want me to add the kinda hackery i've seen to make one ebuild work for releases and vcs?
17:46
like if version equals 9999 do this.. else do this?
17:46
<dberkholz>
sure, if you're planning to use the same ebuild for both
17:46
<johnny>
is that proper behaviour ?
17:46
aren't we?
17:46
at least for now?
17:46
<dberkholz>
you tell me =)
17:46
<johnny>
ok.. we are then :)
17:47
ok.. food time.. then i'll fix up this ebuild
17:47
and then show you the ltsp-build-client stuff
17:49
<dberkholz>
johnny: might be a bit later, i'm still at work
17:50
<johnny>
that's cool
17:50
you and your fancy west coast
17:58chupacabra has quit IRC
18:04bobby_C has quit IRC
18:15chupacabra has joined #ltsp
18:39lns has joined #ltsp
18:41score has quit IRC
18:43
<lns>
Where could I go to start troubleshooting thin clients hard-locking?
18:43
I'm getting reports at one of my sites where students' thin-clients are locking up completely during work (usually in Firefox).. I have XRAMPERC set to 75
18:44joebob777as7 has quit IRC
18:44TelnetManta has joined #ltsp
19:00Faithful has joined #ltsp
19:10staffencasa has quit IRC
19:35jammcq has joined #ltsp
19:59robfantini has joined #ltsp
19:59
<monteslu>
warren, does redhat still have a directory server distro type thing
19:59robfantini has left #ltsp
20:00
<monteslu>
i remeber hearing about it 3 or so years ago
20:00
<warren>
monteslu, yeah, we bought out Netscape Directory Server and open sourced the entire thing.
20:00
<monteslu>
i think smbldap+ltsp+load balancing needs to get easier
20:00
<warren>
monteslu, I don't think they have all the pieces packaged yet, it is a monumental task
20:00viking-ice has quit IRC
20:00
<warren>
you can install the entire thing from a single RPM from the project site
20:01
monteslu, http://directory.fedoraproject.org/
20:01
<monteslu>
i guess we just need a best practice doc on setting up a school with more than one ltsp server
20:01
<warren>
monteslu, I have no experience at all with it
20:02
<monteslu>
i've just been using the smbldap installer
20:02
<warren>
looks like many people are in #fedora-ds though.
20:02
<monteslu>
my only load balancing has been throught lts.conf
20:02
and its just splitting up the workstations
20:03johnny_ has joined #ltsp
20:08
<warren>
monteslu, I know nothing about directory servers in general
20:17
<monteslu>
i know just enough to know its a lot of work to get all the right pieces together
20:18
dtrask's and moquist's scritps help quite a bit though
20:20
<warren>
monteslu, former K12LTSP user?
20:20
<monteslu>
yeah several years
20:20
<warren>
monteslu, maybe you could contribute to Fedora by getting dtrask and moquist's scripts into an official RPM package
20:21
<monteslu>
cool. we need to have a path to keeping those up to date
20:22
<warren>
do they check it into a source repository to keep track of version numbers and changes?
20:22
<monteslu>
not sure what their currnet process is, i've only ever been an end user of their stuff
20:23
looks like the last build was for fc5
20:24
i think it would just need to be a meta package with an extra rpm of the setup script
20:24
<warren>
why not the RPM with shell script Requires all the server bits?
20:25
<monteslu>
well, the current script also supports ubuntu
20:25
<warren>
monteslu, the thing about RPMS in Fedora is you craft the RPM to work on Fedora
20:26
I don't know exactly how the scripts work
20:26
so this might be premature
20:27
<monteslu>
it checks the system for fedora/ubuntu, then pulls down the additional rpms/debs, installs them, then does some setup of openldap and samba
20:27
<warren>
ok, then part of that can be cut out
20:28
the RPM can satisfy the dependencies, and the script simply does the last part of what you described.
20:28
<monteslu>
yeah, pulling down of the missing rpms could just be done witha meta package right?
20:28
<warren>
foopackage Requires package1, package2, package3, package4, ...
20:28
foopackage contains script to setup openldap and samba
20:28
no metapackage needed
20:28
<monteslu>
oh, right
20:30
and the setup script could be kept somewhat distro agnostic since the packaged dependcy would be handled through yum or apt
20:32
config file locations called by the script would be the main difference I imagine, since samba is samba and openldap is openldap
20:32
unless the fedora one used netscapse directory
20:36
<warren>
well, we'll see
20:36
are you on k12linux-devel-list?
20:36
you could write up something there and get ideas from others
20:37
<monteslu>
yup
20:37
well, wait. I'm on the k12osn list
20:38
the k12linux-devel list new?
20:38
<warren>
yes
20:39
https://fedorahosted.org/k12linux/ see details
20:50johnny_ has quit IRC
20:52johnny_ has joined #ltsp
20:55praveer_cool has quit IRC
21:22johnny_ has quit IRC
22:00Faithful has quit IRC
22:43spectra has quit IRC
22:46captain_magnus has quit IRC
22:47captain_magnus has joined #ltsp
22:53Egyptian[Home1 has joined #ltsp
23:08Shingoshi has quit IRC
23:10Egyptian[Home] has quit IRC
23:11Shingoshi has joined #ltsp
23:13captain_magnus has quit IRC
23:13captain_magnus has joined #ltsp
23:16
<dberkholz>
johnny: i'm around now, fwiw
23:28
<johnny>
ok
23:28
i'm just cleaning up my ebuilds as i picked up a few more tricks
23:29* johnny <3 eclasses
23:34indradg_ has joined #ltsp
23:34Shingoshi has quit IRC
23:38subir has joined #ltsp
23:39Shingoshi has joined #ltsp
23:43
<dberkholz>
johnny: install eclass-manpages
23:47
<johnny>
ah!
23:47
good idea!
23:47
never noticed that
23:48
seems like that ebuild should be updated :)
23:52indradg has quit IRC
23:57mccann has quit IRC