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


Channel log from 3 December 2007   (all times are UTC)

00:27jes3 has joined #ltsp
00:28
<jes3>
compaq t5000?
00:29
:) I'm in the wrong room.
00:29
bye.
00:29jes3 has left #ltsp
00:50
<jgbillings>
this boot process looks so happy, it gets to ltsp login:
00:51
and then the screen flickrs and some garbled text is dispalyed on the screen
00:52
thin client is an old compaq deskpro booting from an edubuntu gutsy server
00:52
where do I go to start troubleshooting this?
00:53
says stuff about failed to detect any sound hardware
00:53
module-detect... failed to load
01:12MasterOne has joined #ltsp
01:27johnny_ has quit IRC
01:40cyberorg has quit IRC
01:41edgarin has quit IRC
01:54joebob777as7 has quit IRC
02:06alekibango has quit IRC
02:19cyberorg has joined #ltsp
02:38auctmore has joined #ltsp
02:39
<auctmore>
join kubuntu
02:39
sorry
02:43auctmore has left #ltsp
02:44auctmore has joined #ltsp
02:45auctmore has left #ltsp
03:08Daggett has joined #ltsp
03:33Q-FUNK has joined #ltsp
03:44Q-FUNK has quit IRC
03:52putty_thing has joined #ltsp
04:23ogra has joined #ltsp
04:25Nubae has joined #ltsp
04:26
<Nubae>
hi, I'm trying to autologin with the LDM_USERNAME and LDM_PASSWORD variables in lts.conf
04:26
but I get an X on a black screen
04:26
and it goes no further... any ideas?
04:30
<johnny>
known
04:31
<Nubae>
and, is there a solution?
04:32
I see from irc logs this has been discussed before... just not sure if there is a solution or not
04:32
<johnny>
sorta, but it requires manual patching of the patckage
04:32
somebody built a package, but it was built wrong
04:32
<Nubae>
ok, are there instructions somewhere?
04:33
<johnny>
if you want to learn how to build ubuntu packages in general i suppose there are
04:33
or you could wait til this guy generates a better one
04:33
<Nubae>
well, what exactly is wrong with the package?
04:33
<johnny>
it won't install on ubuntu
04:34
oh.. maybe you're not using ubuntu?
04:34
<Nubae>
ok was built for somethign else?
04:34
<johnny>
debian
04:34
<Nubae>
I am
04:35
is there a plan to port it to ubuntu (the package) soon?
04:35
<johnny>
who knows
04:35
it was meant to be built for ubuntu
04:35
so hopefully he'll fix it
04:35
but from my understanding it will never make gutsy
04:36
it sounds like it works for some people, like the author :)
04:36
<Nubae>
hmmm... ok well, I'm on Gutsy so...
04:36
<johnny>
the original feature
04:36
that is
04:36
<Nubae>
yeah there are a lot of those
04:37
<johnny>
ogra, if a patch won't make gutsy, it should be removed from the features list
04:38
<ogra>
johnny, the problem is that the patch is extremly intrusive, it changes a lot of code ... stable releases only get either security updates or very small non intrusive patches ...
04:38
the pob with the big ones is that you cant predict it breaks anything else
04:39
*prob
04:39
<Nubae>
is there a work around for the autologin, like maybe create a kiosk chroot?
04:40
<ogra>
additionally that ptch is a bit messy, it doesnt only tackle autologin, he has changed a lot more ... if i get a clean patch that only fixes autologin in an easy way like in a few lines so the security reviewers can uderstand it on first sight without knowing the ldm code much then it could go in
04:41
one good thing is that we have the code split now in hardy ... that makes it possible to use the fixed hardy version in gusy-backports
04:41
*gutsy
04:43
your best hope for now is that francois gives proper instructions for his ppa package so you dont end up with broen dependencies ... to bridge the tim where a hardy version is ready for backporting
04:44
<Nubae>
https://wiki.edubuntu.org/HowtoWriteLTSP5Plugins -- In there, there are some instructions on doing autologin, Cant I use that on a seperate chroot?
04:44
<johnny>
but then local devices won't work
04:44
<Nubae>
dont care about that
04:45
<johnny>
or sound?
04:45
<Nubae>
sound I need :-)
04:45
<johnny>
then no :)
04:46
<Nubae>
damn... ok... well stickers on the monitors its going to have to be
04:49
<ogra>
johnny, sound and localdevice will work ... since it doesnt even connect to the server, gdm and firefox are the only apps installed and they both run locally in that plugin
04:50
<johnny>
aha
04:50
<ogra>
the kios plugin only starts a local FF
04:50
*kiosk
04:50
<johnny>
that's differnet than the startx for ldm
04:50
aha
04:50
<ogra>
really, talk to francios to give clean instructions
04:50
the fix is out there, its packaged and apprently runs well for him
04:50
<johnny>
yeah.. i'll send an email to him later
04:51
<ogra>
i'll care for a backport to gutsy-backports if the time is there ...
04:51
<johnny>
i'll be upgrading to hardy when it comes out anyways
04:51
<ogra>
that wont happen before april 21st
04:52
<johnny>
so backports happen before the distro is even released?
04:52
<ogra>
they can, yes
04:53
its a decision of the devloper ... if he considers the code stable enough to backport thats fine
04:53
currently i dont change the code at all though as we're stil in the repackaging phase ....
04:53
and i wont change much before fedora has merged their stuff
04:54
<johnny>
if it requires a backported package, then it prolly shouldn't be in the features list
04:54
<ogra>
(since i expect that to change a lot)
04:54
well, hard to take back the release announcement now :)
04:55
its a known bug people get pointed to ... thats all i can do atm
04:55Daggett has quit IRC
04:58* ogra tests francios' package in a hardy image ...
05:00
<Nubae>
hmmm... the other option would be to use gdm instead of ldm to use autologin
05:00
<ogra>
well, that breaks in a lot of ways, you will need to do lots of extra stuff
05:01johnny has quit IRC
05:01
<ogra>
like disabling the ability to shut down/suspend/ hibernate the server for every user etc
05:01
<Nubae>
yeah, so I've read... there were some old docs somewhere in the wiki about doing that
05:01
<ogra>
sound and local devcies wont work either
05:01
if you can live with both, go ahead :)
05:01
<Nubae>
I need sound
05:01
its from primary kids
05:02
stickers on the monitors it is...
05:02meduxa has joined #ltsp
05:04meduxa is now known as toscalix
05:04
<ogra>
yay, it works fine in hardy
05:05
well ... there is a flaw in the implementation though ...
05:05* ogra wonders how to shut down the client without ldm gui to do so
05:06
<Nubae>
power button :-)
05:07johnny has joined #ltsp
05:10
<ogra>
Nubae, that will leave stale processes around
05:10
you need to be properly logged out to leave a clean environment behind
05:14Guaraldo has joined #ltsp
05:34
<Nubae>
heh... that never happens
05:35
getting people to log out properly is an almost impossible task
05:35
<ogra>
not really :) you just need to make the powerbutton a no-op as long as a session is running ;)
05:36
<Nubae>
then people will use other users sessions
05:36
<ogra>
if they cant kill the machine while logged in they have to log out
05:36
<Nubae>
and how do you stop someone hitting the power switch?
05:37
or unplugging the computer
05:37
<ogra>
you set the bios to ignore it and have some acpi code in ldm that switches on or off based on the session running or not
05:37
you cant prevent unplugging indeed
05:38
<Nubae>
the ideal situation is to teach good terminal ettiquette
05:38
I have a weird thing happening with my vmare... when I power on a virtual machine, it blinks black and then jumps out again... refuses to start up...
05:48Nubae has left #ltsp
06:30toscalix has quit IRC
06:31toscalix has joined #ltsp
06:31subir has quit IRC
06:56joebob777as7 has joined #ltsp
06:58tiagovaz has joined #ltsp
07:02martoya has joined #ltsp
07:03mikkel has joined #ltsp
07:22tiagovaz_ has joined #ltsp
07:23tiagovaz has quit IRC
07:33joebob777as7 has quit IRC
07:50wanie has joined #ltsp
07:51Greesmunky has joined #ltsp
07:52
<wanie>
hi! I use opensuse 10.2 and ltsp 4.2. After a kernel update i can´t connect via xdmcp with the thin client. wher can be the error?
07:53
<Greesmunky>
I have got ltsp up and running but am hitting a snag or two. Does nbd-server need to be running and what do I need in the nbd config file apart from the fairly empty default? NBD errors out saying nothing to do...I can boot the thinclient and get dumped to a busybox shell, by the way
07:56
actually, is it using nbd or nfs to run or mount?
07:57
<wanie>
nfs
07:57spectra has quit IRC
07:59
<Greesmunky>
in the pxelinux.cfg I have (default install of ltsp 5.0.38 in debian sid) ip=dhcp. Is that right or do I need to change it? DHCP server is running
08:07slidesinger has joined #ltsp
08:08martoya has quit IRC
08:12daya has joined #ltsp
08:13ogra has quit IRC
08:13ogra has joined #ltsp
08:14
<ogra>
Greesmunky, nbd server is run by inetd ... if yu use the ubuntu packages all should be set up for you
08:15
Greesmunky, oh, i didnt read the full backlog, if you are using debian your mileage may vary ... unionfs got stable for nbd mounts only recently there
08:15
<Greesmunky>
ogra: but if I try to start nbd by hand I get an error. Is that OK?
08:15
<ogra>
no
08:15
you shouldnt start it at all
08:16
nbdrootd does that
08:16
and the ltsp-update-image script shuld set up inetd for you
08:16
in ubuntu we run it as a plugin of ltsp-build-client i suspect you need a manual run in debian
08:16
<cliebow_>
wanie:simply a kernel update or could other processes have been changed...like gdm?
08:17
<Greesmunky>
ogra: inetd appears to be set up ok
08:17
<ogra>
good
08:17
<Greesmunky>
2 entries regarding ldminfod and nbdswap
08:17
nbdswapd
08:18
<ogra>
thats wrong, for nbd root you need nbdrootd as well
08:18
nbdswapd and ldminfod are also used by the old nfs mounts
08:18
<Greesmunky>
I get dumped to a busybox shell with mount and init errors if thats any help
08:18
<ogra>
nbdrootd is what you want for the squashed filesystem
08:19
indeed you do
08:19
since there is no nbdrootd :)
08:19wanie has quit IRC
08:19
<ogra>
did you run ltsp-update-image to have an image at all ?
08:20grithe has quit IRC
08:20
<Greesmunky>
I was following various wikis and I guess they missed this step..?
08:21
<ogra>
well, as i said, in ubuntu (which is the only distro with default nbdroot yet) its done automatically so its indeed not documented there (apart from maintenance docs etc)
08:21
run that command and check /opt/ltsp/images and your inetd.conf afterwards
08:22
you should have an .img file in the dir and inetd.conf should have any entry for it
08:22
<Greesmunky>
ok, am trying the command now...
08:23
<ogra>
get a coffee, it takes a moment :)
08:24
<Greesmunky>
The man page for nbdswapd contains that little nugget I have been searching for for hours: Add the following line to /etc/inetd.conf(5) <snip> Note that the ltsp-update-image(8) program will do this for you automatically.
08:24
<ogra>
yeah
08:24
<Greesmunky>
ogra: indeed, it does take some time
08:25
<cyberorg>
i hope warren figured out nbdroot/swap with xinetd, i havent got it to work, so we start rcnbd-server
08:26
<ogra>
just use a ditributable inetd :P
08:26
*distributable
08:26
xinetd has some license issues ... that why we dont ship it
08:27
<cyberorg>
nope, we checked that out, no issues at all
08:28
<ogra>
http://www.gnu.org/philosophy/license-list.html
08:28
License of xinetd
08:28
This is a copyleft free software license, incompatible with the GPL. It is incompatible because it places extra restrictions on redistribution of modified versions that contradict the redistribution requirements in the GPL.
08:29
<cyberorg>
ogra, ah, ok it is still distributable in original form
08:29
<ogra>
ubuntu only ships software you can actually change in the free parts of the distro :P
08:29
(i.e. in the "main" repository)
08:30
so it would never make it on the CD
08:30
<cyberorg>
ogra, ok, we ship acrobat reader, java, flash, mp3 support, realplayer and many other in non-oss repo and DVD media
08:30
<ogra>
CSS ?
08:31
<cyberorg>
css, whats that, the dvd playback thing? no we dont have video playback support at all out of box
08:31
just 1-click away
08:31
<Greesmunky>
ogra: well, I seem to be closer but I still have the Busybox shell. The first error listed is: mounting /dev/nbd0 on /rofs failed. No such device
08:31
<ogra>
yeah, we do tha same
08:32
Greesmunky, hmm, check if you can modprobe nbd in busybox
08:32
and also suqashfs and unionfs
08:32
*squashfs
08:32
these three are needed
08:33
<cyberorg>
nbdroot works fine even when running standalone mode, so not sure if running via xinetd would make more difference
08:33
<ogra>
debian ships a "broken" udev its likely it doesnt create the devices properly fo you
08:33
it might need a mknod call as well
08:33
<Greesmunky>
modprobe gives no error
08:34
<ogra>
cat /proc/modules contains it ?
08:34
<Greesmunky>
ogra: no, only nbd is listed
08:34
<ogra>
right thats what i meant :)
08:35
ls /dev/nbd* ?
08:35
<cyberorg>
it might also be /dev/nb*
08:35
<Greesmunky>
nbd's are 0 tthru 15
08:36
<ogra>
hmm
08:36
ten it should work
08:36
*then
08:36
echo $NBD_ROOT_SERVER
08:36MasterOne has quit IRC
08:36
<ogra>
echo $NBD_ROOT_PORT
08:36
<Greesmunky>
I can see on server log that client is connected to nbd
08:36
<ogra>
you need both
08:37burnerx has joined #ltsp
08:37
<ogra>
even though *_PORT defaults to 2000 which should be set in inetd.conf
08:38
<Greesmunky>
I have neither. Could this be the empty config file from /etc/nbd-server/config?
08:38
<ogra>
(unles you have multiple images, then it counts up for each )
08:38
no
08:38
/etc/nbd-server/config is not used at all
08:38
ignore it
08:38
<Greesmunky>
inetd lists port 2000 and the i386.img file
08:38
ok
08:38jammcq has joined #ltsp
08:38
<jammcq>
hey all
08:39
<ogra>
NBD_ROOT_SERVER defaults to $ROOTSERVER which should come form your dhcpd.conf
08:39
(from rootpath)
08:39
hey jammcq
08:40
Greesmunky, what you can do is modify your pxelinux.cfg/default file and add nbdroot="<serverip>:2000"
08:40
then try to reboot
08:40
(the client)
08:40
<Greesmunky>
ogra: ok, thanks will try that
08:41
ogra: the version in /opt/ltsp/i386 or /var/lib/tftpd/ltsp?
08:42
<ogra>
in /var/lib/tftpboot/ltsp
08:43
<Greesmunky>
do I need the quotes?
08:44
<ogra>
might help
08:44
i never tried without though :)
08:44
<Greesmunky>
ok ;-)
08:47
ogra: thanks for all your help but still no go here. I will lurk and google some more
08:48
<ogra>
i suspect google wont help much :)
08:49
whats the error you get
08:49
<Greesmunky>
No, I have discovered that already
08:49
<ogra>
well, most ubuntu users just dont have any probs with it ... and debian hasnt implemented it yet ....
08:49
i wrote that code so i guess i'm your best bet
08:50
is it still the same error ?
08:50
<Greesmunky>
Yeah, can mount /dev/nbd0 on /rofs
08:50
can't
08:50
<ogra>
but /dev/nbd0 exists
08:51
<Greesmunky>
yes
08:51
<ogra>
oh, wait, is the loop module loaded ?
08:52
/opt/ltsp/i386/usr/share/initramfs-tools/scripts/ltsp-nbd is the script btw
08:52
<Greesmunky>
ok, thanks for the location. I was getting a bit lost between /va/lib/tftpd/ opt/ltsp/i386 and the dhcpd stuff
08:53
<ogra>
ignore /opt/ltsp/i386 :)
08:53
<Greesmunky>
only nbd and hardware related modules are loaded, no loop
08:53
<cliebow_>
jammcq!!!!!1
08:53
<Greesmunky>
according to /proc/modules
08:54
<ogra>
all that counts for nbd is the image in /opt/ltsp/images and /var/lib/tftpboot/ltsp/
08:54
i just remembered that loop isnt needed for nbd ... only if you manually mount the squashfs on the server for editing etc, sorry, wrong hint
08:55* ogra is just working on a squashfs shell to edit the image in a GUI ... so i'm a bit on a different path atm :)
08:55
<Greesmunky>
no worries, a wrong hint is at least a direction to go in, which is something I don't really have now
08:56
<jammcq>
cliebow_: hey
09:02Egyptian[Home] has quit IRC
09:03
<cliebow_>
snowing there?
09:03Egyptian[Home] has joined #ltsp
09:04
<jammcq>
cliebow_: we had snow on saturday. rain yesterday. not it's just icy
09:05
<cliebow_>
supposed to get 14 inches..
09:06
i brought a pile of laptops to work on so im busy..
09:18burnerx has quit IRC
09:21Greesmunky has quit IRC
09:34
<slidesinger>
hu jammcq
09:34
*hi*
09:40
<jammcq>
hey slidesinger
09:40
how goes it?
09:52mcfloppy_ has joined #ltsp
09:53flopperl has quit IRC
10:10Gadi has joined #ltsp
10:10
<slidesinger>
jammcq: Quite well, thanks.
10:11
And with you?
10:13bkurt78 has joined #ltsp
10:13
<bkurt78>
anyone know why an etherboot client would seem to ignore the dhcpoffer
10:24
anyone??
10:26
<Gadi>
bkurt78: what makes you suspect it is ignoring it?
10:27
<bkurt78>
Gadi: in my logs I get a dhcpdisover from the client and the server responds with a dhcpoffer. It just keeps going in this loop and on the station itself it just says searching for server.
10:29slidesinger has quit IRC
10:29
<Gadi>
bkurt78: hmm... could be the switch? can you test it with a crossover cable?
10:29
<bkurt78>
no unfortunately I'm not on site with these devices right now.
10:30
<Gadi>
are other devices booting?
10:30
<bkurt78>
I have pxe based stations on the same lan and those are booting just fine, just none of the etherboot stations.
10:30
just updated their ltsp server from centos4 to centos 5
10:31MasterOne has joined #ltsp
10:31
<Gadi>
bkurt78: check your dhcpd.conf that it is set up to respond to Etherboot clients
10:32
<bkurt78>
is there something specific that is now needed for etherboot clients?
10:33* Gadi does not know how CentOS is config'd by default
10:33
<Gadi>
but should be the same
10:38
<bkurt78>
kind of what I thought, but it appears there is something different between dhcp on centos 4 as compared to centos 5.
10:39
jammcq: you there?
10:40
<MasterOne>
hi guys
10:40
gadi
10:40
I found an annoying problem on Edubuntu
10:40
has to do with flash (again)
10:40otavio has joined #ltsp
10:42
<MasterOne>
doing a fresh Edubuntu Gutsy installation (32bit), installing flash-nonfree, then pressing "play" on the flashmovieplayer at tv.com, firefox crashes. This seems to have to do with the fact, that no sounddevice or no local sound support on the LTSP5 server is installed
10:42staffencasa has joined #ltsp
10:42
<Gadi>
guys, sorry for loop delay ;) : bouncing between a bunch of things
10:43
MasterOne: install libflashsupport
10:43
<MasterOne>
the same procedure works just fine, if I install Edubuntu Gutsy 64bit (most likely due to the nspluginwrapper), and a desktop installation (like xubuntu, which I just tried on that server without soundcard)
10:44
gadi, libflashsupport does not solve that issue!
10:44
because without libflashsupport the player nevertheless has to function without causing firefox to crash, means it has to play the content just fine, but without sound
10:45bkurt78 has left #ltsp
10:45
<MasterOne>
with edubuntu gutsy 32bit something for local sound is missing, and that's why firefox crashes as soon as you click the play-button in that flash player
10:45bkurt78 has joined #Ltsp
10:45
<Gadi>
MasterOne: ls -ld /dev/shm
10:45
make sure that is chmod 1777
10:45
<MasterOne>
installing libflashsupport did not change that issue
10:47
/dev/shm is not the problme
10:48
edubuntu gutsy 32bit --> firefox crashes about every time that play button on tv.com is clicked
10:49
edubuntu gutsy 64bit --> no crashes, but no sound (due to the 64bit libflashsupport issue)
10:49
<Gadi>
and you are sure it is flash?
10:49* Gadi never used tv.com
10:49
<MasterOne>
xubuntu gutsy (desktop installation) --> no crashes
10:50K_O-Gnom has joined #ltsp
10:50
<ogra>
MasterOne, works flawless here
10:51
<MasterOne>
ogra, you try it on a thin client connected to a LTSP5 server which has no audio device installed?
10:51
<ogra>
no, actually my server has a soundcard
10:52
but that shouldnt matter at all since all sound should go to the virtual alsa device
10:52
<MasterOne>
I was testing this matter over the weekend, and I came to the conclusion, that flash needs a local sounddevice under certain circumstances
10:52
nevertheless the libflashsupport workaround
10:54
ok, I think I am unable to explain that properly, I dont't mean sounddevice as in soundcard, but something that's set up during installation for sound support
10:55
<ogra>
??
10:56
<MasterOne>
I think there is a reason, why installing ltsp-standalone-server on ubuntu recommends the installation of pluseaudio on the server, which is not installed on edubuntu
10:56
<ogra>
either you have a soundcard in the server then alsa will set that up during boot of the server or you dont
10:56
running pulse locally might break stuff
10:56
uninstall it and test again
10:57
<MasterOne>
I really suck in trying to explain that exact issue, but fact is, firefox + flash-nonfree crashes firefox on edubuntu gutsy 32bit, but does work (without sound) on edubuntu gutsy 64bit and xubuntu gutsy (desktop installation), so there is something wrong with the way edubuntu gutsy 32bit is setup
10:58
I don't have setup pluseaudio on the server
10:58
<ogra>
ignore the desktop installs please
10:58
ltsp sund works totally different, you cant even remotely compare one to the other
10:59
so saying "it works on desktop" doesnt help
10:59
<MasterOne>
I just did the desktop installation on that machine to check if the problem is also present, which was not the case
10:59
<ogra>
and beyonf that it works for lots of people with the 32bit setup
11:00
anything in the users .xsession-errors ?
11:00
<MasterOne>
but that problem is surely 100% reproduceable, because it is about a stock edubuntu gutsy 32bit LTSP5 installation, with just flash-nonfree added, and firefox can be seen to crash every time you press the play button in that flashplayer on tv.com
11:01
ogra, I can not tell details, because I wiped that edubuntu installation to try out the xubuntu desktop install on that machine
11:01bkurt78 has left #Ltsp
11:01mcfloppy_ has quit IRC
11:01bkurt78 has joined #ltsp
11:02
<ogra>
...
11:02
<MasterOne>
firefox gave output of snd_pcm_hw_open and core dumped
11:03
<ogra>
well, the correspnding errors in .xsession-errors would be intresting here
11:03
as well as dmesg etc
11:03
<MasterOne>
hmpf
11:03
<ogra>
i also dont understand why you installed pulse on the server ... recommends dont get installed by default from apt
11:04
<MasterOne>
the point is, I am quite sure if you install stock ubuntu gutsy, and then add the ltsp-standalone-server, it will work just fine, and I am also quite sure that that particular problem is an edubuntu gutsy 32bit LTSP5 server installation problem
11:05
<ogra>
well, it works on all my installs here and i guess we'd have at least a bug about it if it wouldnt work for any of the many schools out there using edubuntu gutsy 32bit
11:06
<MasterOne>
I mean, there is no way, that this is normal, that it works on edubuntu 64bit with the nspluginwrapper (so 32bit flash9 plugin in 64bit firefox), but not in edubuntu gutsy 32bit with 32bit flash9 in 32bit firefox
11:06
<ogra>
and i guarantee you that pulse isnt installed on the server with edubuntu gutsy
11:07
<MasterOne>
ogra, I also did not install pulse on the server, I just mentioned that, because I remember is was mentioned by apt-get as recommended package, when trying to install ltsp-standalone-server on a normal ubuntu desktop installation
11:08
<ogra>
well, get the logs and w can talk :)
11:08
<MasterOne>
ogra, how long are you around today? I am going to reinstall edubuntu gutsy 32bit server on that machine again, then I can provide the necessary data for further analyses
11:08
<ogra>
i know it works for others as well as for me
11:09
cant promise anything its late here and i started work early today ...
11:09
how about filing a bug and attaching the logs :)
11:09mcfloppy_ has joined #ltsp
11:10
<MasterOne>
takes too long, because I finally have to decide which setup I'll stick with (if I can not get this to work in edubuntu gutsy 32 bit, I most likely will go for ubuntu desktop 32bit + ltsp-standalone-server setup
11:10
<ogra>
well, there s no difference in them
11:11
apart from different artwork and desktop apps
11:11
<MasterOne>
that's what I tried to tell, there is a difference, because that issue is just not there is a stock ubuntu / xubuntu desktop install
11:11
<ogra>
on the tech level they are 100% the same
11:11
please dont tak about desktop
11:12
*talk
11:12
you cant compare that at all as i said above
11:13
(no pulse or virtual alsa devices are involved in desktop installs, flash works witout libflashsupport etc)
11:14
<MasterOne>
ogra, it really is difficult to tell, I talk about "desktop" in that context, because if you install ubuntu desktop, and that flash issue is not there, then you add ltsp-standalone-server, and the problem is still not there, there has to be a difference to a stock edubuntu installation, otherwise that problem wouldn't be there in edubuntu as well
11:15
ok, I'll install edubuntu 32bit on that machine again, and will report back, once I'm finished
11:15
<ogra>
well, as i said there is no difference at all in ubuntu+ltsp or edubuntu+ltsp without logs etc there is no way to tell
11:16BGomes has joined #ltsp
11:16
<MasterOne>
it shouldn't take all to long ;)
11:16
<ogra>
1h or so
11:28scrapbunny has joined #ltsp
11:31
<daya>
How do i setup bootsplasy in ltsp etch, does it setup autoamtically when I install bootsplash in chrooted
11:38Q-FUNK has joined #ltsp
11:40cyberorg has quit IRC
11:45bkurt78 has left #ltsp
11:48* kaminski-ltsp-br invite jammcq to join ltsp-br ;) lol
11:49twinprism_ has left #ltsp
11:54Stevecar has joined #ltsp
12:02
<Q-FUNK>
re
12:05mcfloppy_ has quit IRC
12:09sepski has joined #ltsp
12:12edgarin has joined #ltsp
12:14toscalix has quit IRC
12:30ogra1 has joined #ltsp
12:30ogra has quit IRC
12:30ogra1 is now known as ogra
12:34staffencasa has quit IRC
12:34staffencasa has joined #ltsp
12:42Q-FUNK has quit IRC
13:00MasterOne has quit IRC
13:05MasterOne_ has joined #ltsp
13:05MasterOne_ is now known as MasterOne
13:05
<MasterOne>
ogra: ping
13:06
I am back on Edubuntu Gutsy 32bit now, and I am already a bit further now
13:07
<ogra>
so you have a plain install without flashh and libflashsupport now ?
13:07lns has joined #ltsp
13:08
<MasterOne>
When playing that tv.com flashplayer logged in directly on the LTSP5 machine, it works (except these messages about snd_hw_something_open in terminal, but nevertheless, firefox does not crash, and the video plays without sound)
13:08
when I do the same on the thin client, firefox crashes instantly once I click the play button on tv.com
13:09
<ogra>
hmm
13:09
so you already added all the flash stuff
13:09
<MasterOne>
fresh installation, just flash-nonfree installed, and updated to the latest package versions
13:09
<ogra>
which libflashsupport did you install exactly ?
13:09
the one linked from https://help.ubuntu.com/community/EdubuntuFAQ ?`
13:10
(this one is reported to work fine)
13:10
<MasterOne>
again, no libflashsupport installed, these crashes of firefox happen independently of libflashsupport installed, means OK no sound without libflashsupport, but firefox is not supposed to crash with or without libflashsupport
13:10
this is what I get in a terminal:
13:11
$ firefox
13:11
firefox-bin: pcm_params.c:2351: sndrv_pcm_hw_params: Assertion `err >= 0' failed.
13:11
Aborted (core dumped)
13:11
<ogra>
whats the content of your ~/.asoundconf.asoundrc ?
13:11
<MasterOne>
this is what I find in .xsession-errors:
13:11
Error: could not open /proc/asound/cards
13:12
<ogra>
intresting
13:12* ogra has never seen that yet
13:12
<MasterOne>
~$ cat ~/.asoundconf.asoundrc
13:12
cat: /home/sysadmin/.asoundconf.asoundrc: No such file or directory
13:12
<ogra>
i wonder why it tries to access /proc adt all
13:12
thats broken ...
13:12
that file *must* exist if you are logged in through a thin client
13:13
ldm sets it up
13:13
<MasterOne>
ogra, that problem appears especially with the flashplayer on tv.com, the player on youtube.com works as expected
13:13
<scrapbunny>
hi there, quick question for ogra. I installed firefox 3, should it now open when i click the firefox link or is there something special i need to do?
13:13
<ogra>
sounds rather like a flsh bug to me
13:13
<MasterOne>
I have here .asoundrc and .asoundrc.asoundconf
13:14
<ogra>
right, the atter is what counts
13:14
*latter
13:14
it should have the setup for a virtual pulse/alsa device
13:14
<MasterOne>
quite possible that its a flash bug, but as mentioned, that problem is just not there on edubuntu 64bit
13:14
<ogra>
(while you are logged in from a TC)
13:15
<MasterOne>
$ cat .asoundrc
13:15
# ALSA library configuration file
13:15
# Include settings that are under the control of asoundconf(1).
13:15
# (To disable these settings, comment out this line.)
13:15
</home/sysadmin/.asoundrc.asoundconf>
13:15
~$ cat .asoundrc.asoundconf
13:15
# ALSA library configuration file managed by asoundconf(1).
13:15
#
13:15
# MANUAL CHANGES TO THIS FILE WILL BE OVERWRITTEN!
13:15
#
13:15
# Manual changes to the ALSA library configuration should be implemented
13:15
# by editing the ~/.asoundrc file, not by editing this file.
13:15
pcm.!default { type pulse }
13:15
ctl.!default { type pulse }
13:15
<ogra>
ergh
13:15
please use a the pastebot for more than 3 lines
13:15
!pastebot
13:15
<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.
13:16
<ogra>
thats right then, your sound setup is ok
13:16
<MasterOne>
no idea what else to try
13:17
you should be able to reproduce that error, if you visit tv.com from a thin client, and just press the play button in the flashplayer on the right side
13:17
<ogra>
me neither since i dont see such issues
13:17* ogra has to leave now ...
13:18
<ogra>
12h of work and my GF gets grumpy
13:18
<MasterOne>
hmpf, that issue is killing my head
13:18
please try it on tv.com yourself, as soon as you find the time
13:19
<ogra>
will do first thing tomorrow, sorry, but i'm risking my relationship if i stay on longer
13:19scrapbunny has left #ltsp
13:19
<MasterOne>
ok, have to leave as well
13:19
have a nice evening
13:19
<ogra>
you too
13:19Greesmunky has joined #ltsp
13:19MasterOne has quit IRC
13:32staffencasa has quit IRC
13:33staffencasa has joined #ltsp
13:33bobby_C has joined #ltsp
13:34Guaraldo has left #ltsp
13:35
<Greesmunky>
I have problems with nbd (I think) and wonder if its possible to test it from another box with nbd-client. Would "nbd-client server1 2000 /dev/nbd0" be the right command to test it?
14:02Greesmunky has quit IRC
14:06elisboa has quit IRC
14:06elisboa has joined #ltsp
14:07sepski has quit IRC
14:09tiagovaz_ has quit IRC
14:21Greesmunky has joined #ltsp
14:24bkurt78 has joined #ltsp
14:24
<bkurt78>
Gadi: you still there?
14:29daya has quit IRC
14:29
<Gadi>
in and out
14:30
<bkurt78>
you remember when you were assisting me with the issue with dhcp and etherboot earlier?
14:38* Gadi nods
14:38
<bkurt78>
found my answer: http://www.etherboot.org/wiki/troubleshooting
14:39
as of isc dhcp 3.0.3 they don't set a default next-server so you have to explicitly set it. CentOS 5 happens to be urnning 3.0.5 so that was the culprit.
14:39wanie has joined #ltsp
14:40
<wanie>
hi! can someone help me with a little sound problem under ltsp 4.2, pleas?
14:42
<johnny>
ogra, i'm looking at francis's patch , it might not be perfect, but i don't think it can be organized much better, it looks like the options had to move
14:42
so there could only be just a lil bit extra in there
14:42
<wanie>
when i start my thin client, in the boot messages i can read "starting sound server, accepting connection on port 16001" but how can i use the soundcard under gentoo?
14:43
<Gadi>
bkurt78: ah, good work
14:43
thx for letting me know
14:43
<bkurt78>
no problem. Thanks for taking the time to help out.
14:44
<wanie>
my sound server is esd but amarok or so say no soundcard avalible
14:44
<Gadi>
wanie: ur apps need to use esd and connect to the thin client's esd server
14:44bkurt78 has left #ltsp
14:44
<Gadi>
wanie: is that the setting in amarok?
14:44
amarok has its own settings beyond the system settings
14:47edgarin has quit IRC
14:47edgarin has joined #ltsp
14:47
<johnny>
ogra, really tho.. sbalneav should look it
14:47
he would certainly know the best way
14:48
<wanie>
gadi: it say "xine can't initializing audio drivers"
14:49
is the error in xine or in amarok?
14:50
xine is the only output that i can select, there is no esd
14:51
<mistik1>
hey guys
14:51
Gadi: btw, thanks man
14:51
<Gadi>
mistik1: ah, you got it? gr8
14:52
<mistik1>
yes sir
14:52
<Gadi>
wanie: try testing sound with esdplay first
14:52
then work on the harder app
14:52* mistik1 now making linux perform new tricks
14:53
<mistik1>
Its really quite amazing just how many levels of complexity is in this OS of ours
14:54
Gadi: use sfdisk much?
14:54
<Gadi>
mistik1: indeed I do
14:55
very nice for scripting
14:55
<mistik1>
oh man, its my new best friend
14:55
<johnny>
mistik1, how about looking at the diagram of possible sound systems, it'll make your head explode
14:56
<mistik1>
johnny: I dont doubt it at all
14:56
I went looking there before
14:57
<johnny>
the combinations are insane
14:57
you could loop yourself throught about 5 levels
14:57
i think
15:01BGomes has left #ltsp
15:07
<wanie>
gadi: ALSA lib confmisc.c:670:(snd_func_card_driver) cannot find card '0'
15:08
<Gadi>
wanie: that is to be expected
15:08
<wanie>
i haven't configured my card with alsaconf
15:08
<Gadi>
as you prolly have no sound card in the server
15:08
do not worry about that
15:08
can you open a terminal and type: esdplay <some wav file>
15:09
if not, my guess is that you do not have the ESPEAKER env var set properly
15:09
you should have a login script that sets ESPEAKER=<thin client ip>:16001
15:10
<wanie>
yes i did this an became this error message
15:10
<Gadi>
in fact, you can try: export ESPEAKER=<your_thin_client_ip>:16001 && esdplay <some wav file>
15:11
ah - then it sounds like esdplay was trying to connect to the *server* rather than the client and you got the error from esd on the server because it had no sound card
15:11
:)
15:15
<wanie>
ok, sorry my english is not so good, wher must be saved the login script
15:15cliebow_ has quit IRC
15:19K_O-Gnom has quit IRC
15:23
<Gadi>
wanie: not sure on gentoo - probably make it launch from /etc/profile
15:24mikkel has quit IRC
15:25spectra has joined #ltsp
15:26
<wanie>
ok now i type'd ESPEAKER=<your_thin_client_ip>:16001 && esdplay <some wav file> with a really ip and a existing file
15:27
now i don't becam any error's but i don't hear a sound :(
15:28
<Gadi>
Q: How can I use the soundcard of my workstation? A: There is a ltsp-sound package in gentoo, for more instructions read the included README file.
15:28
got that from: http://www.gentoo.org/doc/en/ltsp.xml
15:28
have you seen that?
15:31
<wanie>
no sorry
15:32
<Gadi>
wanie: it will probably help more than me - Im not much of a gentoo person
15:32
:)
15:33
<wanie>
ok, thank you
15:33edgarin has quit IRC
15:43wanie has quit IRC
15:50wanie has joined #ltsp
15:57
<Greesmunky>
how do I determine the thin client hostnames? how do I know/change what they are?
16:01Egyptian[Home] has quit IRC
16:05prpplague has joined #ltsp
16:06
<prpplague>
jammcq: whats cookin boss man?
16:24
<lns>
Greesmunky, my installs don't incorporate hostnames - They're useful, but AFAIK you have to assign them by /etc/hosts
16:27herson has joined #ltsp
16:30
<Greesmunky>
lns: so mac address must be specified in dhcpd.conf?
16:30
I haven't done that yet. I think I'm getting stuck at the nfs mount part (can mount from another box the network) but am not quite sure...
16:30
<lns>
Greesmunky, if you want specific hostnames for specific clients, I think you do have to set up static leases
16:31
but i'm not completely sure
16:31
<Greesmunky>
lns: To be honest, I'm just trying to get it to work right now. Hostnames don't bother me unless I need them
16:31
<lns>
Greesmunky, just based on what i've set up, i wouldn't recommend using them unless you need them
16:32
kind of a PITA
16:34
<johnny>
it's useful if you use ltsp on fiesty with gadi's autologin patch
16:34
then you can do LDM_USERNAME="hostname|"
16:35
quite nice
16:35
i used dnsmasq for my dhcp/dns
16:35
tho
16:36martoya has joined #ltsp
16:39
<lns>
ahh
16:40
Yes, for things like that, and sharing printers off of a thin-client, it's nice..
16:56spectra has quit IRC
17:01wanie has quit IRC
17:13jammcq has quit IRC
17:20Greesmunky has quit IRC
17:23bobby_C has quit IRC
17:25prpplague has quit IRC
17:56herson has quit IRC
18:03martoya has quit IRC
18:06joebaker has quit IRC
18:08Gadi has left #ltsp
18:14joebaker has joined #ltsp
18:56vagrantc has joined #ltsp
18:57staffencasa has quit IRC
19:00
<vagrantc>
ogra, sbalneav, warren: i started working on an ltspfs branch to upload to launchpad ...
19:00
<warren>
vagrantc, OK
19:00
vagrantc, did you see the mkdst I uploaded? any comments?
19:00
<vagrantc>
warren: yes, i saw the upload ...
19:01
<warren>
vagrantc, are the current contents OK?
19:01
<vagrantc>
i think ogra was saying it should remove some of the autogenerated stuff from the tarball ... automate.cache or something like that
19:01
warren: the other thing i was thinking is putting the version in release.conf seems ... silly.
19:02
warren: then we're maintaining version information in release.conf that will be outdated ... i think a --version flag or some other mechanism would be better
19:02
<warren>
vagrantc, why?
19:03
vagrantc, well, the nice thing is, since release.conf is sourced, release.conf itself could read VERSION from elsewhere too.
19:03
<vagrantc>
the version will both be in bzr and in release.conf, also ..
19:03
<warren>
vagrantc, both in bzr and release.conf? Not exactly. tags don't always equal version.
19:03
<vagrantc>
when does release.conf's version get updated?
19:03
<warren>
vagrantc, do you mean tags right?
19:03
bzr tag NAME-VERSION?
19:04
<vagrantc>
well, the bzr tag when releases should actually be the version
19:04
<warren>
vagrantc, where does bzr tag get the version from then?
19:04
<vagrantc>
warren: commandline?
19:04
<warren>
um
19:04
vagrantc, then we're talking an entirely different idea of "mkdst tag"
19:04
<vagrantc>
warren: where does release.conf get the information from? hard-coding it just means that it will usually be in incorrect.
19:05
<warren>
vagrantc, we have to hard code it SOMEWHERE
19:05
vagrantc, is it really hard to bump that number when we decide to make a release?
19:05* vagrantc doesn't see why it needs to be hard-coded anywhere
19:05
<vagrantc>
warren: and what about the commit immediately after the release?
19:05
warren: what version is it then?
19:06
<warren>
vagrantc, last version number + commit id.
19:06
vagrantc, --release cannot be done without a corresponding tag.
19:06
vagrantc, --test can make a tarball but it wont be official
19:07
<vagrantc>
i guess the tag function needs to error out if bzr tag fails
19:08
and hopefully bzr tag properly handles when the same version already exists
19:08
<warren>
yeah
19:08
<vagrantc>
er, same tag
19:08
<warren>
that'll be a good improvement
19:08
<vagrantc>
otherwise you'll end up with a tarball created that has the same version but a different bzr history
19:09
<warren>
vagrantc, here's an idea, I was trying to add some place in mkdst itself to print out the mkdst version whenever it runs. It would be handy to see during buildsystem logs, bug reports, etc.
19:09
<vagrantc>
warren: so, is it really that hard to require a version commandline argument?
19:09
<warren>
oh wait
19:09
no, it isn't
19:09
becaus mkdst is used long before that
19:09
vagrantc, let's make sure we agree on exactly how it would work first.
19:09
<vagrantc>
warren: sure. well, that's my comment for now.
19:10
<warren>
vagrantc, --tag=NAME-VERSION manually?
19:10
<vagrantc>
i've got intermittant internet access, so i need to catch up on a few things first...
19:10
warren: i kind of like release requiring an existing tag ...
19:10
<warren>
vagrantc, the trouble with this is suddenly your mkdst commands are more complex. The whole point of mkdst is to do things quickly without thinking about what you are doing.
19:10
--release cannot work with --tag=requird
19:11
unless you want --release=something
19:11
then mkdst is no longer easy to use
19:11* vagrantc still finds all the other steps easy to use
19:11
<warren>
vagrantc, I'm OK with making --release fail if the current stuff is not matching the latest tag.
19:11
<vagrantc>
warren: yeah, that would be a good first step
19:12
getting more than that may not be necessary...
19:12
<warren>
if (NAME-VERSION from release.conf doesn't match current tag) then fail;
19:12
vagrantc, another improvement I want to add is to make the --test tarballs differently named.
19:12
vagrantc, rather than NAME-VERSION that is certainly wrong, NAME-VERSION-test or something.
19:13
<vagrantc>
warren: sounds good
19:13
<warren>
NAME-version-commitid might work, except --test allows you to tarball things not yet checked in
19:13
which is REAL helpful to avoid tons of little checkins that are broken
19:13
<vagrantc>
well, you can always uncommit
19:13
<warren>
vagrantc, if I knew about it! =)
19:13
<vagrantc>
bzr uncommit
19:14
<warren>
vagrantc, it is not easy to know because I'm using several different distributed SCM's all with different commands for different projects.
19:14
<vagrantc>
just make sure not to uncommit published changes
19:14
warren: sure. well, now you know :)
19:14
<warren>
vagrantc, I might not implement this right way, but it might be like this.
19:15
vagrantc, --test creates NAME-VERSION-commitid if stuff is committed but not tagged since VERSION
19:15
vagrantc, --test creates NAME-VERSION-test if changes are not checked in
19:16
vagrantc, but I'm not sure how to do the former yet, so I'll just have it say NAME-VERSION-test until I do.
19:16
vagrantc, do you like the optional EXPORT_CHANGELOG option?
19:16
<vagrantc>
NAME-VERSION-DATESTRING-test?
19:17
warren: well, i generally think a changelog and vcs commits target a different audience, so i prefer to keep them separate
19:17
<warren>
vagrantc, hmm, maybe NAME-VERSION-post-DATESTRING is easiest.
19:17
<vagrantc>
warren: but those are always nice features
19:17
<warren>
vagrantc, yeah, not all projects can use EXPORT_CHANGELOG
19:17
especially not the LTSP projects split off from the original bzr repo
19:17
since it has tons of irrelevant changes
19:18
<vagrantc>
yeah, quite a long history of commits
19:18
many of which are ubuntu/debian specific
19:19
<warren>
vagrantc, NAME-VERSION-post-DATESTRING because VERSION would be the previous version, post indicates these are changes after VERSION
19:19
<vagrantc>
also, bzr log often produces a different log if your branch histories aren't identical ...
19:19
<warren>
yeah...
19:19
<vagrantc>
warren: so, how's fedora's ltsp coming?
19:21
<warren>
vagrantc, I've been fighting our mkinitrd unfortunately
19:21
vagrantc, I have the necessary features to boot LTSP written, but that means nothing if I cna't merge it.
19:21
vagrantc, it is a bit of a political battle
19:21
<vagrantc>
eeyk.
19:26Drakeweb has quit IRC
19:27
<vagrantc>
hmmm.. there's the ltspfs project ...
19:27
<warren>
vagrantc, using mkdst?
19:28
<vagrantc>
warren: well, not yet ... but ltspfs has it's own project on launchpad ...
19:28* vagrantc finds it funny that vagrantc is the top contributer to ltspfs without even being able to commit to it
19:30
<warren>
vagrantc, who has commit access to the ltspfs project?
19:30
<vagrantc>
warren: i can't tell
19:31
<warren>
vagrantc, looks like sbalneav didn't setup the commiters properly
19:31
vagrantc, I didn't want to create a new team for mkdst so I just assigned committers to ltsp-upstream team
19:31
<vagrantc>
well, it predates ltsp-upstream, ltsp-committers and ltsp-drivers, i think
19:34basanta has joined #ltsp
19:54underdog5004 has joined #ltsp
19:56jammcq has joined #ltsp
19:56
<jammcq>
hey all
19:56
<underdog5004>
howdy folks. I finally figured out how to get my thin client to boot from the server...but my X is all messed up...there are two of everything on the screen, and it's really shimmery. It almost looks like there are two X servers going...
19:56
<warren>
ogra, can we consolidate the the ltsp upstream committers to a single team to reduce confusion?
19:56
<jammcq>
sounds like the wrong xserver or the wrong amount of video memory was detected
19:57
<underdog5004>
jammcq: ok, I'll try manually setting it...
19:57
I'm reading about lts.conf right now...good stuff!
19:58
<warren>
jammcq, the T170 is proving to be very frustrating.
19:59
<jammcq>
hmm
19:59
<warren>
jammcq, it has a rare hardware bug (also seeing it on rare AMD's) that causes it to kernel panic maybe 50% of the time at startup. Began with upstream 2.6.23 kernel. All the kernel developers are really confused. The hardware seems to be doing the impossible.
19:59
<jammcq>
what's the problem with it?
19:59
hmm, but it worked fine with earlier kernels?
19:59
<underdog5004>
probably dividing by 0, :)
19:59
<warren>
jammcq, yes. this has something to do with a major rewrite for 2.6.23.
19:59
jammcq, so this is just a preview of what you'll see on other distros as they upgrade to 2.6.23+
20:00* jammcq remembers when major rewrites were odd version numbers
20:00
<warren>
jammcq, kernel developers have almost given up on it, blaming the hardware manufacturer because this is so rare.
20:00
<jammcq>
is there a kernel guy that should have one of the 170's to play with?
20:00
<warren>
jammcq, the kernel hackers in my office already had access, and they are perplexed.
20:01
jammcq, the debugging details are absolutely bizarre.
20:01
<jammcq>
hmm
20:01
<warren>
jammcq, "divide error" panic when the instructions have no division
20:01
<jammcq>
hmm, there's a guy I've talked to from Via who might be able to help
20:01
<warren>
jammcq, here's a question, boot a T170 on a kernel that you know works and save the dmesg, sent it to me via e-mail attachment.
20:02
<jammcq>
I don't have any 170's here
20:02
etyack is "the man" for that sort of stuff
20:02
<warren>
jammcq, https://bugzilla.redhat.com/show_bug.cgi?id=394981 grisly details here
20:03
jammcq, there is definitely a hardware bug involved here, triggering a very rare kernel behavior only in 2.6.23+
20:04
<underdog5004>
hmm, could it be a problem if my server and thin clients are using different video card drivers?
20:04
<jammcq>
underdog5004: no
20:04
underdog5004: the server doesn't even need a video card, ltsp doesn't care
20:04
<underdog5004>
ok...
20:04
even to run an Xserver instance?
20:04
huh...
20:05
<jammcq>
underdog5004: well, if you want to run X on the server console, you'll obviously need a video card, but that has nothing to do with ltsp
20:05
<underdog5004>
ok, gotcha.
20:07
I'm also having a problem logging into a thinclient...I've chrooted /opt/lib/tftpboot/ltsp/i386/ and did adduser matthew, but I still can't login...
20:12
<vagrantc>
underdog5004: you don't usually add users to the chroot
20:12
<underdog5004>
what's the standard way?
20:12
<vagrantc>
standard way is to not have users in the chroot
20:12
underdog5004: why did you try to add a user?
20:13
<underdog5004>
Am I supposed to be able to use the standard username/password that I use on the server?
20:13J45p3r has joined #ltsp
20:13
<vagrantc>
underdog5004: no
20:14
underdog5004: normally, you do not log into the thin client itself ...
20:14
<underdog5004>
...
20:14
<vagrantc>
underdog5004: are you trying to log in to troubleshoot your X problems ?
20:15
<underdog5004>
ok...I guess there's another problem I've got...when I boot the thinclient, it goes to the login screen, not the desktop...
20:16edgarin has joined #ltsp
20:17
<underdog5004>
is it usual for the thinclient to boot directly to the desktop?
20:19
<vagrantc>
it requires configuration, and is broken with LDM
20:19
<underdog5004>
beautiful
20:20
<vagrantc>
default behavior is to present an LDM login screen (with ltsp5)
20:20
<underdog5004>
can I set that in lts.conf ?
20:21
<vagrantc>
autologin requires patches to python-based LDM and is broken in C-based LDM (as far as i know)
20:21
you could also do autologin with GDM or KDM
20:21
<underdog5004>
sounds like a plan!
20:21
<vagrantc>
but then you loose simple configuration of sound and local devices
20:21
<underdog5004>
argh...what if I just go back to 4.2?
20:22
<vagrantc>
that's probably about the same as using GDM or KDM with ltsp5
20:22
<underdog5004>
darnit
20:23
ok, so I don't do autologin, that's fine...but how can I make a new user/password combo?
20:23
<vagrantc>
however you would normally add users to your server
20:24
<underdog5004>
vagrantc: I can't login with my usual username/password that I use to login to my server?
20:24
<vagrantc>
underdog5004: what distro?
20:25
<underdog5004>
ubuntu gutsy, but this is just a test server/thinclient, I will probably migrate to gentoo...
20:26
<vagrantc>
well, i don't think gentoo reall yhas an ltsp5 implementation yet ...
20:26
underdog5004: can you ssh to your server?
20:26
<underdog5004>
well, my ltsp server is the same as my desktop...
20:26
so, I have terminal access to it...
20:26
<vagrantc>
dpkg -l ltsp* | egrep ^ii
20:27
<johnny>
there is a patch for autologin
20:27
<underdog5004>
ii ltsp-server 5.0.39 Basic LTSP server environment
20:27
ii ltsp-server-standalone 5.0.39 Complete LTSP server environment
20:27
ii ltspfs 0.5-0ubuntu2 Fuse based remote filesystem for LTSP thin c
20:27
<johnny>
if you know how to build ubuntu packages...
20:27
<vagrantc>
underdog5004: ok, so, i'm asking troubleshooting information ... can you ssh to your server using the username and password?
20:28
<underdog5004>
yep, I've got sshd running.
20:28
<vagrantc>
underdog5004: that doesn't actually answer the question ...
20:28
<underdog5004>
yes, I can ssh to it
20:29
just did it.
20:31
<vagrantc>
ok, because all LDM does is: ssh -X username@server
20:31
basically
20:33
<underdog5004>
how can I check which dm is being used?
20:35
<vagrantc>
grep SCREEN $(find /opt/ltsp/*/etc /var/lib/tftpboot/ltsp/* -name lts.conf)
20:35
if no SCREEN_NN are defined, it should default to LDM
20:36
<underdog5004>
SCREEN_NN = gdm
20:36
look good?
20:36
<vagrantc>
no at all
20:36
<underdog5004>
I mean, that's what it should be, right?
20:36
<vagrantc>
no
20:36
<underdog5004>
I didn't get any output from that command string
20:36
<vagrantc>
then it should be using LDM
20:37
if you want to use GDM, you'll have to configure GDM to allow XDMCP, and set SCREEN_07=startx in your lts.conf
20:37
<underdog5004>
gotcha
20:37
I really don't care which dm I use, just so long as it works, you know?
20:38
<vagrantc>
i would recommend using the default, as that's the most tested and most supported
20:38
<underdog5004>
sounds good to me
20:38
<vagrantc>
figuring out why it's not working is probably better than just trying something else
20:38
<underdog5004>
lol
20:40
<vagrantc>
underdog5004: so, you get a login screen, you enter your username and password ... and then what happens?
20:41
<underdog5004>
It says verifying username...verifying password, please wait...then X restarts
20:41
<vagrantc>
underdog5004: anything in ~username/.xsession-errors ?
20:41
on the server
20:42
<underdog5004>
and when I do it from a terminal, it says Login incorrect
20:42
h/o
20:42
<vagrantc>
underdog5004: is that same user logged in on the server some other way?
20:42
i.e. from the local login ?
20:43
<underdog5004>
4 or 5 lines of : Xsession: X session started for matthew at Mon Dec 3 18:40:59 PST 2007
20:43
X connection to localhost:11.0 broken (explicit kill or server shutdown).
20:43alekibango has joined #ltsp
20:44
<underdog5004>
and : ** (nm-applet:5794): WARNING **: <WARN> nma_dbus_init(): could not acquire its service. dbus_bus_acquire_service() says: 'Connection ":1.17" is not allowed to own the service "org.freedesktop.NetworkManagerInfo" due to security policies in the configura
20:44
<vagrantc>
underdog5004: is that same user logged in on the server some other way?
20:44
<underdog5004>
nope
20:44
just me
20:45
<vagrantc>
underdog5004: so this is a different account that the one you're currently using?
20:45
<underdog5004>
nope
20:46
<vagrantc>
underdog5004: so, you just gave me two directly conflicting answers
20:46
<underdog5004>
what do you mean?
20:47
<vagrantc>
underdog5004: are you logged in as "matthew" (or whatever username it is) right now, and is that the same user you're trying to log in through LDM ?
20:47
<underdog5004>
yes
20:47
<vagrantc>
an you try with a second user?
20:47
can
20:47
<underdog5004>
hmm, lemme make a new user...
20:47
<vagrantc>
many desktops don't work very well with multiple users logged in simultaneously
20:48
multiple identical users, that is
20:48
<underdog5004>
gotcha...
20:50
testing...
20:50
fudge! it didn't work!
20:50
<vagrantc>
hrm.
20:50
<underdog5004>
made a user named matthew2, and no go
20:50
<vagrantc>
underdog5004: try setting SCREEN_08=shell in lts.conf
20:51
<underdog5004>
don't I want a gui though?
20:51
<vagrantc>
underdog5004: well ... both SCREEN_07=ldm and SCREEN_08=shell
20:51
underdog5004: then you'll be able to log in via ldm and then look at the log files and hopefully find some useful information
20:52
<underdog5004>
awesome, rebooting thinclient now...
20:54twinprism_ has joined #ltsp
20:57edgarin has quit IRC
20:57
<underdog5004>
ah HAH!
20:57
I love you!
20:57
in the most non-gay way ever!
20:58
erm, I was able to get to a desktop...btw
20:58
<vagrantc>
ok
20:59* vagrantc wonders what wasn't working
21:00basanta has quit IRC
21:01
<underdog5004>
now to figure out what the problem is with X...
21:03
thank you so much for your help!
21:25J45p3r has quit IRC
21:57vagrantc has quit IRC
21:59joebaker has quit IRC
22:10BGomes has joined #ltsp
22:11elisboa has quit IRC
22:11elisboa has joined #ltsp
22:33Stevecar has left #ltsp
22:38elisboa has quit IRC
22:55subir has joined #ltsp
23:18twinprism_ has quit IRC
23:41cyberorg has joined #ltsp