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


Channel log from 6 July 2009   (all times are UTC)

00:26jammcq has quit IRC
00:48intelliant has joined #ltsp
01:46hanthana_ has joined #ltsp
01:55hanthana has quit IRC
02:19l0C0 has joined #ltsp
02:22OvaKill has joined #ltsp
02:22F-GT has quit IRC
02:26mikkel has joined #ltsp
02:45seppy has joined #ltsp
02:47ogra has quit IRC
03:01ogra has joined #ltsp
03:06alkisg has quit IRC
03:08alkisg has joined #ltsp
03:08ogra has quit IRC
03:12ogra has joined #ltsp
03:28OvaKill has quit IRC
03:29F-GT has joined #ltsp
03:37ogra has quit IRC
03:39ogra has joined #ltsp
03:43ogra has quit IRC
03:45ogra has joined #ltsp
03:56ogra has quit IRC
04:00ogra has joined #ltsp
04:04ogra has quit IRC
04:11ogra has joined #ltsp
04:14ogra has quit IRC
04:18ogra has joined #ltsp
04:43yoshi___ has joined #ltsp
04:44ogra has quit IRC
04:47ogra has joined #ltsp
04:50yoshi__ has quit IRC
04:54hanthana_ is now known as hanthana
05:27ogra has quit IRC
05:29Faithful has quit IRC
05:31ogra has joined #ltsp
05:48erieslabnx has joined #ltsp
06:02ogra has quit IRC
06:14tjikkun_work has joined #ltsp
06:18agike_ has joined #ltsp
06:28ogra has joined #ltsp
06:28pmatulis has joined #ltsp
06:36seppy has quit IRC
06:46sene has joined #ltsp
06:57mikkel has quit IRC
06:57
<_UsUrPeR_>
morning all :)
07:00
<erieslabnx>
morning sir
07:09Faithful has joined #ltsp
07:11mesias has joined #ltsp
07:17erieslabnx has quit IRC
07:24mesias has quit IRC
07:29pmatulis has quit IRC
07:45elisboa has joined #ltsp
07:52pmatulis has joined #ltsp
08:04alkisg1 has joined #ltsp
08:04alkisg has quit IRC
08:05alkisg1 has quit IRC
08:10hanthana has quit IRC
08:14alkisg has joined #ltsp
08:17alkisg has quit IRC
08:17CAN-o-SPAM has joined #ltsp
08:21gentgeen__ has quit IRC
08:22gentgeen__ has joined #ltsp
08:23Gadi has joined #ltsp
08:33ogra_ has joined #ltsp
08:34ogra has quit IRC
08:34
<_UsUrPeR_>
gadi: ping
08:53
<sbalneav>
Morning all
08:57erieslabnx has joined #ltsp
08:58
<_UsUrPeR_>
hey sbalneav
08:58
!s
08:58
<ltspbot>
_UsUrPeR_: "s" is Scotty!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
08:58
<erieslabnx>
Is there any simple diagnostic tool for servers? For no apparent reason only 2 of my 10 clients can login, they took almost 5 minutes to do so and I don't know why.
08:59
I don't know what could have changed outside of the new ubuntu updates that were on my screen this morning
09:00
server is running slow as *&^ now too, Firefox taking minutes to respond
09:01
after updates do I need to rebuild the boot image?
09:04
<_UsUrPeR_>
erieslabnx: you only need to rebuild an image if you have updated the image's chroot
09:04
i.e. made changes in /opt/ltsp/i386/*\
09:04
err... ignore the last backslash
09:04
<erieslabnx>
ok then it must not be that :)
09:04
<_UsUrPeR_>
yeah :/
09:05
are you noticing any strange stuff in top/system monitor? What linux derivative are you using?
09:05
<erieslabnx>
januty, nothing strange in system monitor, but even system monitor can take minutes to start
09:06
jaunty*
09:06
<_UsUrPeR_>
and this slowness is new?
09:06
<erieslabnx>
yes today it is new, reboot does nothing to help
09:07
ran fine when I left it on Thursday
09:07
all clients were online
09:07
<_UsUrPeR_>
Have you checked your drive capacities to ensure you're not pushing 100% on something?
09:07
I have had that creep up on me before
09:08
("df -h" in terminal)
09:08
<erieslabnx>
I am at about 70% free on drives, 2ghz Athlon with 2GB of RAM
09:09
<_UsUrPeR_>
hmm~~~ and your RAM is not being fully used (i.e. swapping)
09:09
<erieslabnx>
INSANITY!!! LOL
09:09
ok terminal finally lit
09:10
15% in use
09:10
<_UsUrPeR_>
:/
09:10
getting your clients working seems like the least of your issues compared to your server at this point. At least they still work
09:11
<erieslabnx>
what drives me nuts is I can't figure out what could have changed :/
09:11
or more importantly why a reboot doesn't help :P
09:12
<_UsUrPeR_>
what type of video card are you running?
09:12
<erieslabnx>
Nvidia I think
09:12
yes Nvidia
09:13
GeForce 8200
09:13
<_UsUrPeR_>
hmm. you ran top as root to make sure you weren't noticing some crazy process right?
09:13* _UsUrPeR_ will brb
09:14
<erieslabnx>
no I just run the system monitor, and gnome-system-monitor is the highest CPU using process
09:15dorimar has quit IRC
09:23
<_UsUrPeR_>
erieslabnx: hmm. You're sure the AMD chip isn't kicking in to a over-temp protected mode? I am running out of guesses
09:23
<erieslabnx>
no worries, when day is done I will reinstall all :)
09:24
thanks for the efforts
09:25alkisg has joined #ltsp
09:30
<_UsUrPeR_>
no problem. Good luck.
09:34lucascoala has quit IRC
09:36dan_young has joined #ltsp
09:50intelliant has quit IRC
09:55dan_young has quit IRC
09:56Q-FUNK has joined #ltsp
09:57
<Q-FUNK>
:)
10:02bieb has joined #ltsp
10:03bobby_C has joined #ltsp
10:12staffencasa has joined #ltsp
10:16mikkel has joined #ltsp
10:16dan_young has joined #ltsp
10:19cyberorg has quit IRC
10:40l0C0 has quit IRC
10:47alkisg has quit IRC
10:47alkisg has joined #ltsp
10:57Q-FUNK has left #ltsp
11:04WhorkMan has joined #ltsp
11:04
<WhorkMan>
good morning everyone :)
11:09
<sbalneav>
Morning
11:11
<WhorkMan>
hows your day going?
11:11
<Lumiere>
!s
11:11
<ltspbot>
Lumiere: "s" is Scotty!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
11:12
<Lumiere>
hi sbalneav, WhorkMan
11:12
<WhorkMan>
Lumiere: good day :)
11:12
<Lumiere>
do either of you know of how people fix entropy on larger ltsp installations?
11:13
<WhorkMan>
Lumiere: i dont :/
11:13
<sbalneav>
We're talking about firefox's random number generator problem here?
11:13
<Lumiere>
sbalneav: we're not even getting logged in
11:13
it seems
11:14
but firefox may be eating entropy too
11:14
<sbalneav>
Odd. How many workstations?
11:14
<Lumiere>
20-30
11:14
<sbalneav>
I have 40 workstations, but never have problem with entropy.
11:14
<Lumiere>
ubuntu?
11:15
<sbalneav>
However, I don't run the generic kernel, I run the -server kernel.
11:15
yes, hardy.
11:15
<Lumiere>
hmm
11:15
<sbalneav>
Are you running with the generic kernel?
11:15
<WhorkMan>
my ubuntu and gentoo servers dont play nice together :/
11:15
<Lumiere>
not sure
11:15
I have to look at it this afternoon
11:16
<sbalneav>
balneav@sputnik:~$ uname -a
11:16
Linux sputnik 2.6.24-19-server #1 SMP Wed Aug 20 23:54:28 UTC 2008 i686 GNU/Linux
11:16
-server's the key there.
11:17
<Lumiere>
k
11:18
<WhorkMan>
hrm. i have a quick question about lts.conf in ltsp5... i assume it is pulled via tftp correct? at what point does it get pulled?
11:18
<ogra_>
initramfs
11:18ogra_ is now known as ogra
11:19
<Lumiere>
yea... prior to the mountpoint swap from initramfs to nbd/nfs
11:19
<WhorkMan>
thats what i was thinking, though i dont see it trying to pull the file at all
11:20
thanks, i am going to go digging around in my images a bit. :)
11:21
<ogra>
it determines the tftp path from the filename direcive in the dhsp reply
11:21
*dhcp
11:23
<WhorkMan>
ogra: i had read that, everything else tftp related seems to be working ok
11:24
probably doesnt help that the dhcp and tftp server isnt on the same box as the ltsp enviroment. heh
11:26
<ogra>
hmm, it might be a problem with next-server not being used
11:29fotanus has quit IRC
11:29
<WhorkMan>
i have next-server set to 10.0.0.1, ltsp is setup on 10.0.0.2
11:32IBC_jkenney has joined #ltsp
11:32alkisg1 has joined #ltsp
11:40
<alkisg1>
WhorkMan: (echo "get ${tftpath:-/ltsp/i386}/lts.conf"|tftp ${ROOTSERVER} >/dev/null 2>&1) ==> the LTSP server is considered to be the tftp server as well...
11:42vagrantc has joined #ltsp
11:54alkisg has quit IRC
11:55alkisg1 is now known as alkisg
12:03
<WhorkMan>
alkisg thanks :)
12:04
<alkisg>
Np... it would be easy to patch it if you want, or well put lts.conf in the ltsp server :)
12:12
<erieslabnx>
any quick guess off hand as to why the sound will play locally on clients half the time and the other half it plays the sound on the server itself (not playing on the thin client)?
12:12
FarmTown on Facebook has annoying music lol
12:17rjune_ has joined #ltsp
12:18
<vagrantc>
erieslabnx: distro and version?
12:18
<erieslabnx>
Ubuntu Jaunty
12:19
I can search it, I just figured if it was a common thing someone would know off the top of their head :)
12:19
<vagrantc>
erieslabnx: and by half the time, do you mean sometimes login plays on the thin client, and on logging in again, it might end up playing on the server?
12:20
or that any given application might play on the server one time, and without logging out, when started again it plays on the thin client?
12:20
<erieslabnx>
yes
12:20
very random
12:21
<vagrantc>
erieslabnx: which is it? i described two different scenarios...
12:21
maybe i didn't make it clear enough...
12:21
<erieslabnx>
the first
12:22
<vagrantc>
ok, so any given login will stick with server or thin client for sound?
12:22
<erieslabnx>
correct
12:22
it seems to stick like that, but it seems hit and miss
12:22
like I never know when or why it chooses to do so
12:23
<vagrantc>
from the same thin client?
12:23fotanus has joined #ltsp
12:23
<erieslabnx>
from more than one but yes it seems to be 2-3 particular machines
12:23
some clients have never done it
12:23
<vagrantc>
particular users?
12:24
<erieslabnx>
yes
12:24
seems to be just 2 user accounts
12:25
could it be not loading drivers properly?
12:25
<vagrantc>
probably some setting stuck into a . file in their home directory...
12:25
or maybe a running daemon that doesn't close on logout
12:25
<erieslabnx>
from that assumption in theory would recreating the user account fix it?
12:25
<vagrantc>
if you wipe the home directory
12:26
typically the home directory isn't deleted when the account is
12:26
i'm just guessing, though.
12:26
<bieb>
vagrantc: are home directories required? I will have our clients authenticating against Active Directory, and the students don't curently have home dirs
12:26
<vagrantc>
never seen any problem like that
12:27
bieb: yes
12:27
<erieslabnx>
ok :) I will make sure to take note more when it happens in case my memory isn't working right
12:27
<bieb>
vagrantc: any recommended or minimum size for the HomeDir?
12:27
<vagrantc>
bieb: it's probably not impossible to do without them, but it's likely to lead to tears
12:28
!serversize
12:28
<ltspbot>
vagrantc: "serversize" is http://wiki.ltsp.org/twiki/bin/view/Ltsp/ServerSizing
12:28
<vagrantc>
bieb: try that ^^
12:28
<bieb>
thanks
12:29
<Gadi>
bieb: https://help.ubuntu.com/community/UbuntuLTSP/ActiveDirectoryIntegration
12:29
if you are on ubuntu
12:30
<bieb>
yeah I will be using Ubuntu 9
12:30
<vagrantc>
Gadi: so, i tried NBD+squashfs+aufs, just like ubuntu, and still /media/root/* stuff just plain didn't work.
12:30
Gadi: works for your setups?
12:30
<Gadi>
vagrantc: after whats-his-name had issues last week, I did a fresh jaunty install and tried it and it worked fine
12:31
no fancy footwork
12:31
I think your ltspfssd is not compiled with the new patch
12:31
<vagrantc>
which new patch?
12:31
something outside of ltspfs-trunk?
12:31
<Gadi>
from the troubleshooting I did with him, it didn't have the no auth when connecting from localhost bit
12:31
that is in upstream
12:31
are you compiling from upstream?
12:32
<vagrantc>
0.5.12
12:32
<Gadi>
yeah
12:32
but is it compiling fresh?
12:32
<vagrantc>
i'm missing my spelling typo, which is the only patch above and beyond that
12:32
Gadi: what do you mean by compiling fresh?
12:32
<Gadi>
I mean, is it perhaps not make cleaning
12:32
and using the old ltspfsd binary
12:33
<vagrantc>
Gadi: no
12:33
Gadi: i do all my builds in a clean setup
12:33
<Gadi>
can you try installing ltspfsd from jaunty?
12:33
or from stgraber's ppa
12:33
<vagrantc>
i don't think that'll work, as the udev rules end up in a different place
12:33* ogra would take a deep look at udev
12:33
<vagrantc>
i guess i could manually move them
12:34
<Gadi>
or at least just grab his ltspfsd binary
12:34
<WhorkMan>
hrm. when booting my tc's, it shows IP-Config: eth0 complete with the various info from DHCP, then shows Negotiation: and hangs. any suggestions on what to look for?
12:34
<ogra>
ubuntu doesnt use the debian packages anymore since jaunty iirc ... but packages upstream unmodified
12:34
<vagrantc>
basically, it mounts /media/root/* if it's there are boot, but doesn't handle device removal or insertion post-boot
12:35
ogra: ubuntu and debian are synced with ltspfs
12:35
<ogra>
right but not with udev
12:35
<Gadi>
vagrantc: yeah, I dont have that issue on jaunty
12:35
<vagrantc>
ogra: there's a rule in debian/rules that puts the udev rules in a different spot, is all.
12:35
<ogra>
we have a completely different udev than you
12:35
<vagrantc>
ah, got it
12:36
<Gadi>
WhorkMan: sounds like Lumiere's issue
12:36
WhorkMan: try: sudo pkill -9 nbd-server
12:36
WhorkMan: and reboot a client
12:36
<ogra>
Gadi, so you mean these two should form a club ?
12:37
<Gadi>
ogra: yes, and whack you with it :P
12:37
<bieb>
vagrantc: I didnt see anything on the per user home dir size. I did appreciate the RAM calculation though.. I was looking for that earlier I was just figuring 100mb per client
12:37
<vagrantc>
ogra: what's confusing, is there aren't any different udev rules for local mounts vs. remote ltspfs mounts ... and the local mounts work fine
12:37
<ogra>
right, but udevd might behave differently
12:38
<vagrantc>
sure ...
12:38
<WhorkMan>
Gadi: same thing. :/
12:38
<vagrantc>
perplexing, though
12:38
<Gadi>
WhorkMan: ps -efw|grep nbd
12:38
<ogra>
its the only idea i have
12:38
apart from different kernels indeed
12:38
<Gadi>
WhorkMan: are there a ton of nbd processes?
12:38
<WhorkMan>
just one
12:39
<Gadi>
WhorkMan: do you have any firewalls blocking port 2000?
12:39
<WhorkMan>
the LAN side of the router has all ports allowed
12:40
maybe i should try a clean install. heh
12:41
<IBC_jkenney>
hello i am trying to setup ltsp (thinclient) to use LDAP for authentication and goup policy is that even possible or am i spinning my wheels
12:46
<WhorkMan>
Gadi: i figured out the problem, it was showing one IP (the correct one) on the thinclient and trying to mount another (the wrong one)
12:49
<Gadi>
IBC_jkenney: sure - just follow any LDAP howto on the server
12:49
<IBC_jkenney>
Gadi am i following the howto on the ldap server then do i have to rebuild the client?
12:49
sorry ldap=ltsp
12:50
the ldap server is all ready running they are seperate VM'S
12:50
<Gadi>
IBC_jkenney: no, the thin clients use SSH
12:50
so it is the server's SSH auth that auth's the client
12:50
as long as ssh auths using LDAP on the server, ur good
12:51
<IBC_jkenney>
how about group polices for access to certain application's.
12:51
<Gadi>
totally different subject
12:51
:)
12:51
you can lock that down with your desktop environment's lockdown tools
12:52
<IBC_jkenney>
can you point me in a direction for that subject matter so i can read then come back with questions
12:52
so i'm not wasting anyone's time
12:54
<ltsppbot>
"Gadi" pasted "for _UsUrPeR_" (34 lines) at http://pastebot.ltsp.org/404
12:55
<Gadi>
IBC_jkenney: https://help.ubuntu.com/community/LDAPClientAuthentication
12:55
if ur on ubuntu
12:55
<_UsUrPeR_>
gadi: :D
12:55
<IBC_jkenney>
Thanks Gadi
12:56
<_UsUrPeR_>
gadi: one question: is it alright not to include video stuff in xorg.conf?
12:57
<Gadi>
_UsUrPeR_: I think so - it should take things piece-meal and modify the bits you supply
12:57
<_UsUrPeR_>
hmm
12:57
<Gadi>
but, I wasn't sure how that applied to input devices
12:57
:)
12:57
so, ur guess is as good as mine
12:58
if you set CONFIGURE_X=True, it will add skeleton bits for the rest
13:00alkisg has quit IRC
13:01
<_UsUrPeR_>
ahh
13:06Faithful has quit IRC
13:12tjikkun_work has quit IRC
13:13tjikkun_work has joined #ltsp
13:15alkisg has joined #ltsp
13:20sene has quit IRC
13:20
<_UsUrPeR_>
gadi: where will the changes show up on the client filesystem once it's booted?
13:21
<Gadi>
/var/run/ltsp-xorg.conf
13:22
<_UsUrPeR_>
gadi: it has correctly put in a section "inputdevice" with /dev/ttyS0 as the device
13:22
however, no mouse movement :/
13:22
it does look correct though
13:22
driver is with elotouch installed
13:22
is there anything you would like me to pastebin?
13:22
<Gadi>
elotouch, not elographics?
13:23
<klausade>
vagrantc: fyi: i just created a squeeze chroot on a lenny machine, and ltsp-localapps "just works"! I need to look at homedirs via ldap, and test some more.but, for sure a good start.
13:24
<vagrantc>
klausade: yeah!
13:24
<_UsUrPeR_>
:3 my bad, elographics it is.
13:25
<vagrantc>
klausade: there are some issues with locales (localapps get the same locale as LDM, rather than the user's session), but overall, it works :)
13:25
<Gadi>
_UsUrPeR_: can you pastebot the ltsp-xorg.conf?
13:25
<_UsUrPeR_>
gadi: np
13:26
<vagrantc>
Gadi: so i tried ltspfs local mount stuff again, and this time, it actually registered a CD eject, but not an insertion...
13:26
<klausade>
vagrantc: iceweasel+flash running localy is a sure winner, with youtube and the likes.
13:26
<Gadi>
vagrantc: can you try a USB device?
13:26
hehe - I love it when they call it iceweasel
13:27
<ltsppbot>
"_UsUrPeR_" pasted "For Gadi" (6 lines) at http://pastebot.ltsp.org/405
13:27
<vagrantc>
Gadi: you mean using real hardware? :)
13:27
<_UsUrPeR_>
gadi: yeah
13:27
<Gadi>
_UsUrPeR_: wheres the rest of it?
13:27
<_UsUrPeR_>
I can make the changes to your script for a USB device
13:27
?!?
13:27* vagrantc rummages around for a physical thin client
13:28
<Gadi>
the mods I pastebot'd should have a ServerLayout section
13:28
did you update the image?
13:28
<klausade>
Gadi: don̈́'t hit me! i just learnt how to spell iceweasle properly.
13:28
<Gadi>
klausade: no, its just funny because "firefox" is a real animal name
13:28
it is not a "fire" "fox"
13:29
so, "ice" "weasel" is just all the more ridiculous
13:29
<_UsUrPeR_>
it's true. ice and weasels typically are never combined
13:30
gadi: gimme a moment. lemme make sure my lts.conf is setup correctly
13:30
<Gadi>
np
13:30
<_UsUrPeR_>
was there supposed to be options called in lts.conf for minx and maxx and whatnot, or were those going to be put in by the script
13:30
?
13:33
<klausade>
vagrantc: how do I best controll the sound volume? using pulseaudio(pavucontroll)? or alsamixer/kmix/... running localy?
13:34johnny has joined #ltsp
13:34
<Gadi>
_UsUrPeR_: lts.conf
13:35
<_UsUrPeR_>
gadi: rgr. just re-read your script
13:35
<Gadi>
the driver should have sane defaults
13:37
<vagrantc>
klausade: no idea
13:38
<ltsppbot>
"_UsUrPeR_" pasted "For Gadi" (12 lines) at http://pastebot.ltsp.org/406
13:39WhorkMan has quit IRC
13:39
<_UsUrPeR_>
gadi: I forgot the important parts :)
13:39
gadi: unfortunately, there is no response to mousing :(
13:39
<Gadi>
that still is not all of it
13:39
I don't think you updated the file I pasted
13:40
http://pastebot.ltsp.org/404
13:40
that should replace the: /opt/ltsp/i386/usr/share/ltsp/screen-session.d/XS95-touchscreen file
13:40
and update the image
13:40
<_UsUrPeR_>
:O oh. I thought this was another .sh file :)
13:40
I have been using it as such
13:40erieslabnx has left #ltsp
13:41
<Gadi>
ah, sorry
13:41
<_UsUrPeR_>
it appears to have worked properly thus far :)
13:42
<klausade>
vagrantc: ok. and now the "ignorant" question. must all my browserplugins be in the chroot? as long as the browser runs as localapp?
13:42
<vagrantc>
klausade: yes
13:43
<klausade>
vagrantc: i thought so.
13:43
<vagrantc>
klausade: well, users can install plugins in their homedirs, but for system-wide, yes, in the chroot
13:44
<sbalneav>
!docs
13:44
<ltspbot>
sbalneav: "docs" is For the most current documentation, see http://wiki.ltsp.org/twiki/bin/view/Ltsp/LtspDocumentationUpstream
13:45akSeya has joined #ltsp
13:45
<knipwim>
akSeya: you had a comment on the gentoo wiki a few days ago?
13:45
<akSeya>
yeap
13:46
<vagrantc>
Gadi: usb sticks just as unreliable...
13:46
<knipwim>
i wrote most of it
13:46
<akSeya>
hum.. nice
13:46
so
13:46
i had some problems
13:46
<Gadi>
vagrantc: /me wonders what is so different about the debian environment
13:46
<vagrantc>
Gadi: works if booted with the device, and registers device removals (although sometimes veryu slowly) ... but insertion doesn't kick in
13:46
<akSeya>
lemme check
13:47
<vagrantc>
Gadi: thankfully, it doesn't break the remote mounts like previous incarnations did
13:47
<sbalneav>
New version of docs posted. Lots of cleanup, commenting out of "non-ltsp" related info, explanation of the LIKE keyword, more consistent layout, etc.
13:47
<ltsppbot>
"_UsUrPeR_" pasted "For Gadi again :)" (16 lines) at http://pastebot.ltsp.org/407
13:48
<vagrantc>
klausade: actually, you can use the ltsp* and ldm* packages from squeeze on lenny, if you want a little more stability.
13:48
<_UsUrPeR_>
gadi: there's an "EndSection" on the end of that
13:48
I missed it in my copy-paste
13:48
<vagrantc>
Gadi: yeah, it's a mystery ...
13:49
<Gadi>
_UsUrPeR_: where's the AlwaysCore option?
13:49* _UsUrPeR_ shrugs
13:49
<_UsUrPeR_>
is that something that I am supposed to be specifying in lts.conf?
13:50
<Gadi>
no
13:50
it was in the script I pasted
13:50
it should have put it in the xorg.conf file
13:50
weird...
13:50
did you overwrite what I pasted with what you had?
13:50
<_UsUrPeR_>
I see alwayscore in the script
13:51
directly below "DeviceName"
13:51
I overwrote it and updated the server yes
13:51
err client image I mean
13:51pmatulis has quit IRC
13:51
<Gadi>
somethings weird
13:52
also, the S in TouchScreen is capitalized in 2 places
13:52* _UsUrPeR_ copies the lts.conf he's using:
13:52
<Gadi>
and I uncaps'd it everywhere
13:52
<johnny>
knipwim, you wrote waaay too much
13:52
<akSeya>
i had 2 problems with ltsp-build-client on gentoo, but I was able to fix it changing somethings on quickstart/profile.qs
13:52
<johnny>
we don't need unionfs at all
13:52
it is completely unnecessary
13:53
the only manual modification needed (other than out of chroot setup like tftp, and dhcp) is modifying genkernel to include more nic modules
13:53Egyptian[Home] has quit IRC
13:53
<johnny>
knipwim, in fact.. i consider it completely unsupported
13:53
<vagrantc>
Gadi: found the issue: XAUTHORITY=$(find /tmp -maxdepth 1 -mindepth 1 -type f -name xauth-$DISPLAY-*)
13:54
Gadi: there's no xauth-* files in /tmp
13:54
<johnny>
there is a bit of funionfs-fuse support in genkernel git now.. but union mounts looks to be making traction, so i think i'll stick with tmpfs bindmounts until they occur in a released kernel
13:54
<vagrantc>
Gadi: are there on ubuntu?
13:54
<johnny>
knipwim, i wish you would have spoken to me before leading people down such a long path
13:55
<Gadi>
vagrantc: shouldn't be - and anyway, we shouldn't need XAUTHORITY at all
13:55Egyptian[Home] has joined #ltsp
13:55
<Gadi>
you should be able to comment that out
13:55
since connections from localhost are not auth'd
13:55
<vagrantc>
do we need DISPLAY then, too?
13:55sene has joined #ltsp
13:56
<akSeya>
johnny, knipwim well.. i'm here to help to
13:56
johnny, remember that error I was getting with ltsp-built-client?
13:56
<johnny>
yes
13:56
<Gadi>
vagrantc: not 100% sure - I forget what ltspfsd or ltspfsmounter complain about
13:56
<ltsppbot>
"_UsUrPeR_" pasted "For Gadi -- script + current lts.conf" (80 lines) at http://pastebot.ltsp.org/408
13:56
<Gadi>
vagrantc: try without
13:56
<johnny>
akSeya, we neeed a place to host a chroot targz.. so we don't need peeps tobuild all the time
13:56
<vagrantc>
Gadi: and now... i switch back to NFS... :)
13:57
<Gadi>
hehe
13:57
_UsUrPeR_: ttyUSB0 is only for a USB-to-serial converter
13:57
<akSeya>
I had to add iso-codes on extra_packages
13:57
<Gadi>
is that what you are using?
13:57
<akSeya>
without it ldm was not building
13:58
and svg to USE flag
13:58
<klausade>
vagrantc: yes. my approach is to start with the latest-and-greatest,and from there move over to more stable stuff.
13:58
<akSeya>
after that ltsp-build-client was successfull
13:59
<knipwim>
reading ...
13:59
<akSeya>
now it tells me to append a correct root= boot option
14:00
<knipwim>
johnny: did you read my reply in the wiki discussion?
14:00
<vagrantc>
klausade: well, localapps, localdev, sound all work with lenny and the latest ltsp*/ldm* packages :)
14:01
klausade: in fact, i can't get sound to work on squeeze
14:02
<_UsUrPeR_>
gadi: just changed /dev/ttyUSB0 to /dev/ttyS0. still no changes :/
14:02
<vagrantc>
Gadi: appears to be just as broken with it commented out
14:02
<Gadi>
_UsUrPeR_: can you add: CONFIGURE_X=True and pastebot ltsp-xorg.conf?
14:03
oh, actually, before that
14:03
<_UsUrPeR_>
gadi: yes, but no gurantees the client will boot afterwards :)
14:03
<Gadi>
can you paste the Xorg log file from the client?
14:03
and also note, I said: CONFIGURE_X=True NOT CONFIGURE_X_COMMAND
14:04
<_UsUrPeR_>
gadi: rgr
14:04
<Gadi>
we don't want to use a sh script
14:04
<_UsUrPeR_>
:)
14:04
<akSeya>
knipwim, johnny, where does the client read the boot options.. like kernel name and etc?
14:04
<Gadi>
_UsUrPeR_: we may need to specify the screen in the server layout
14:05
<knipwim>
akSeya: at the tftboot
14:06
<akSeya>
client does not have grub nor lilo ... it's says it could not find kernel image: linux
14:06
<_UsUrPeR_>
gadi: there's no change to ltsp-xorg.log on the client when using CONFIGURE_X = True
14:07
<johnny>
akSeya, /var/lib/tftpboot/ltsp/x86/pxelinux.cfg/default
14:07
<knipwim>
akSeya: your tft server is like your grub or lilo, networkbased
14:09
<johnny>
akSeya, so what fixes did you make to the build?
14:09
knipwim, the real problem we have to solve is genkernel initramfs, but i'm thinkng about switching to dracut
14:09
depending if they can get their own module inclusion stuff in order
14:09
<akSeya>
johnny, I added svg to USE flag and iso-codes to extra_packages
14:10
on /etc/ltsp/quickstart/profile.qs
14:10
<ltsppbot>
"_UsUrPeR_" pasted "for Gadi" (44 lines) at http://pastebot.ltsp.org/409
14:10
<johnny>
hmm.. that shouldn't be in extra_packages
14:10
it should be a dep of the requiring package
14:10
i actually want to avoid using extra_packages except for the ltsp packages themselves
14:10
and let htme handle their own deps :)
14:10
i thought i had somebody commit the fix to ldm to require iso-codes
14:11
at least.. i think ldm is what requires iso-codes
14:11
<knipwim>
yes
14:11
<johnny>
too bad i can't run the digest :(
14:11
knipwim, do you have a git checkout of the ltsp overlay?
14:11litlebuda has joined #ltsp
14:11
<knipwim>
nope
14:11
i thought you had one
14:11
<johnny>
can you?
14:11
i do, but no gentoo box to run ebuild digest :)
14:12
<knipwim>
o yes
14:12
<johnny>
i need somebody to hook me up with a decent vm or something
14:12
<klausade>
vagrantc: sound works just fine here in my squeeze chroot. both playback and record.
14:12
<akSeya>
johnny, yeap.. it was ldm which required iso-codes
14:12
<johnny>
my laptop just can't handle gentoo
14:12
<Gadi>
_UsUrPeR_: out of curiosity, there is no /etc/X11/xorg.conf in the chroot, right?
14:13
I mean on the runtime thin client
14:13
only /var/run/ltsp-xorg.conf
14:13
<johnny>
i coulda swore i got that committed before
14:13
<alkisg>
johnny: virtualbox?
14:13
<johnny>
buti will commit the svg useflat
14:13
alkisg, that won't help!
14:13
it's the hardware that can't handle it
14:14
putting a vm on the hardware that isn't good enough won't magically make it be better hardware :)
14:14
<alkisg>
Why, gentoo can't be installed as a guest os in vbox?
14:14
<johnny>
of course it can
14:14
but the underlying hardware isn't good enough
14:14
<alkisg>
Ah, just throw it away then :)
14:14
<johnny>
it's my main laptop
14:14
so.. no :)
14:14
<knipwim>
get some virtual memory and proc :)
14:14
<johnny>
well.. it is actually my only computer atm
14:14pmatulis has joined #ltsp
14:14
<johnny>
it's a pentium m 1.5ghz
14:15
<_UsUrPeR_>
gadi: correct. I have used etc/X11/xorg_1420.conf (I am using a dlw 1420 workstation)
14:15litlebuda has quit IRC
14:15
<akSeya>
johnny, i don't have "default" on that folder.. I have it on /tftpboot/pxelinux.conf
14:15
<johnny>
plus.. why throw away hardware that works perfectly well with fedora 11
14:15
<akSeya>
but looks like this is not being read
14:15
<johnny>
alkisg, that isn't right
14:15
err akSeya that isn't right
14:15
<alkisg>
johnny: heh... I gave my celeron 1.6 laptop to my daughter, and got a new one
14:16
<vagrantc>
klausade: weird. i've not seen remote soudn working in squeeze as long as i can remember.
14:16
<johnny>
sure.. when i can afford it i will
14:16
you need to set your tftp server to look into /var/lib/tftpboot/ltsp/
14:16
<akSeya>
if I type a kernel filename that is on /tftpboot/ltsp/x86 it starts to boot (but crashed)
14:16
<Gadi>
_UsUrPeR_: and, you are connecting by serial port?
14:16
<johnny>
akSeya, we are standardizing on /var/lib/tftpboot, please use that
14:17
<akSeya>
hum..
14:17
<vagrantc>
Gadi: so, running the whole su ${USER} -c "/usr/sbin/ltspfsmounter ${MOUNTPOINT} $action"
14:17
<akSeya>
ok
14:17
<Gadi>
_UsUrPeR_: can u paste ur: /var/log/Xorg.*.log
14:17
<johnny>
akSeya, the ltsp-update-kernels should be looking there.. iirc
14:17
<vagrantc>
Gadi: seems to be where it's failing ... or at least, when i run it manually, i get a few warnings and errors
14:17
<_UsUrPeR_>
gadi: correct. serial port. Not using a usb adapter
14:17
<vagrantc>
Gadi: but where does the root mounting happen from?
14:18
<akSeya>
johnny, it was ltsp-build-client fault.. :D
14:18
<vagrantc>
Gadi: "Cannot open X display" and "Error: /tmp/.root-ltspfs/usbdisk-sda1 is not mounted"
14:18
<Gadi>
vagrantc: if USER is unset, ltspfsmounter should run as root, and with a connection from localhost, ltspfsd should not require auth
14:19
vagrantc: then, something needs DISPLAY set
14:19alkisg has quit IRC
14:19
<Gadi>
even if it is not used
14:19
<ltsppbot>
"_UsUrPeR_" pasted "for Gadi -- Xorg.7.log" (1025 lines) at http://pastebot.ltsp.org/410
14:19
<_UsUrPeR_>
gadi: there you go
14:20
<johnny>
akSeya, yes.. as i said.. i'm going to tar it up and ask new users to just use the chroot instead of building it off unreliable sources all the time
14:20
<akSeya>
hum.. it's going to be interesting
14:21
<vagrantc>
Gadi: on device removals, if someone is logged in, it will run as that user rather than as root.
14:22
<ltsppbot>
"Gadi" pasted "for _UsUrPeR_ - take 2" (38 lines) at http://pastebot.ltsp.org/411
14:22
<vagrantc>
Gadi: so you get into the scenario where it is initially mounted as /media/root/cdrom, but if the cdrom is then removed while the user is logged in, it tries to unmount /media/$user/cdrom ...
14:23
<_UsUrPeR_>
gadi: will change that out
14:23
<Gadi>
vagrantc: true
14:23
<vagrantc>
oh hell... OOM errors.
14:23
<Gadi>
vagrantc: it seems to be hard to satisfy both local_apps and RDP (other screen scripts)
14:24
all because of lbmount
14:24
<vagrantc>
i've mostly been testing with thin clients with 64MB of ram... wonder if i've been getting OOM errors all along
14:24alkisg has joined #ltsp
14:24
<Gadi>
perhaps we should make a separate ltspfsmounter script for localhost mounts
14:24
<_UsUrPeR_>
gadi: updating image now
14:24
<Gadi>
that does not use lbmount
14:25
and that can just mount in /media
14:25
of course, we would still need to handle permissions on the mounted device
14:29alkisg has quit IRC
14:29
<Gadi>
honestly, for local apps, the easiest thing is to create a symlink on login to the server's /media/$USER dir in the user's home directory
14:29
<ltsppbot>
"_UsUrPeR_" pasted "for Gadi -- Xorg.7.log part2" (1025 lines) at http://pastebot.ltsp.org/412
14:30
<vagrantc>
Gadi: and remote-mount /media?
14:30
<Gadi>
no
14:30
<ltsppbot>
"_UsUrPeR_" pasted "for Gadi -- ltsp.xorg part 85 million ;)" (21 lines) at http://pastebot.ltsp.org/413
14:30
<vagrantc>
how does a symlink work across computers, then?
14:30
<Gadi>
the symlink should lead the local app to the remote /media
14:30
in sshfs
14:30
<vagrantc>
ah.
14:31
<Gadi>
then, we can handle RDP+localdev solely
14:31
with the local ltspfsmounter
14:31
<vagrantc>
actually, it seems like we need two events...
14:32
i mean, a minor tweak to the current code would be to always run as root, and sometimes run as the user.
14:32alkisg has joined #ltsp
14:32
<vagrantc>
for the su ${USER} -c "/usr/sbin/ltspfsmounter ${MOUNTPOINT} $action"
14:32
part
14:33
<ltsppbot>
"Gadi" pasted "for _UsUrPeR_ - take 3" (38 lines) at http://pastebot.ltsp.org/414
14:33
<Gadi>
sorry, _UsUrPeR_ - forgot something the first time
14:33
:)
14:33
<_UsUrPeR_>
Gadi: which line?
14:34* _UsUrPeR_ tries to read closely
14:34
<Gadi>
line 2
14:35
vagrantc: right - its the sometimes thats tricky
14:35
<_UsUrPeR_>
:D
14:35
<akSeya>
johnny, now I'm not able to start tftp :/
14:35
<_UsUrPeR_>
Gadi: what /dev/ should I be using if I want to try USB?
14:35
<akSeya>
permissions maybe?
14:35spectra has joined #ltsp
14:37
<Gadi>
_UsUrPeR_: not sure
14:37
_UsUrPeR_: it would be one of the /dev/input/event*
14:37
I guess
14:37
<_UsUrPeR_>
gadi: ahh :/
14:37
<Gadi>
but, get serial going first
14:37
;)
14:37
<knipwim>
akSeya: the tftp server is started?
14:38
<akSeya>
no.. I'm trying to find it's log to see whats going on
14:39
<ltsppbot>
"_UsUrPeR_" pasted "for Gadi -- ltsp.xorg part 4" (20 lines) at http://pastebot.ltsp.org/415
14:39
<akSeya>
be back soon.. going to try something..
14:40
<ltsppbot>
"_UsUrPeR_" pasted "for Gadi -- Xorg.7.log part 4" (1025 lines) at http://pastebot.ltsp.org/416
14:40
<Gadi>
_UsUrPeR_: it seems to be missing something still
14:40
<_UsUrPeR_>
gadi: same. no results on serial
14:41
<Gadi>
duh
14:41
oh no
14:41
wait
14:41
hmm...
14:41
<_UsUrPeR_>
did you still want CONFIGURE_X = True on?
14:43
<Gadi>
you can try with it off
14:43
I am just wondering why the write_screen 0 >> $XCONF didn't do anything
14:47
<bieb>
Gadi: Do you know if these instructions should still work with ubuntu 9?? https://help.ubuntu.com/community/UbuntuLTSP/LTSPWindowsDHCP
14:49
<Gadi>
bieb: you need to add an nbdserver kernel parameter to the pexlinux.cfg/default file, as well
14:49
otherwise, the OS will most likely not find the nbd server
14:50
bieb: sorry - "nbdroot" not "nbdserver"
14:51
<bieb>
ok.. looking for ndb root docs
14:52
<johnny>
ldm now is updated akSeya and knipwim
14:53
<Gadi>
bieb: just add: nbdroot=<LTSPserverIP>:2000
14:53
<bieb>
ok
14:53
<Gadi>
after "quiet" in /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default
14:54
and if you change the IP of the LTSP server, update it
14:54
<johnny>
also added the svg flag
14:54
<bieb>
gotcha
14:55
and nothing needs to be done for nbd root swap right?
14:56
<Gadi>
you may need: nbdswap=<LTSPserverIP>:9572
14:56
as well
14:56
<vagrantc>
Gadi: ok, so, it definitely needs XAUTHORITY...
14:56
<Gadi>
for nbdswap to work correctly
14:56
<vagrantc>
Gadi: it also needs to always needs to process /media/root ...
14:57
Gadi: can't get it to do /media/$user locallly, though.
14:57
<Gadi>
vagrantc: there is an xauth thing for non-ldm sessions in xinitrc.d
14:57
<bieb>
Gadi: so whether it needs either of the NBD's is trial and error?
14:57
<Gadi>
I forget where it places it
14:57
bieb: no, nbdroot is for the rootfs, nbdswap is for swap over nbd
14:57
use both
14:58
personally, I don't use swap
14:58
<bieb>
ok..
14:58
<Gadi>
so, I never care to do the second
14:58
:)
14:58
<bieb>
:)
14:58
<vagrantc>
Gadi: yes, but that only handles delayed-mount stuff, not udev insertion/removal.
14:59
Gadi: if the udev rules are working properly, the delayed mount stuff shouldn't be needed
14:59
Gadi: for /media/root/*
14:59
<Gadi>
vagrantc: right, but put back that XAUTHORITY line we had, and s/tmp/\/var\/run/
14:59
<vagrantc>
Gadi: yes, that's what i did (although the syntax keeps changing...
15:00
<alkisg>
bieb: where's your tftp server?
15:00
<vagrantc>
Gadi: i still can't get it to work for $USER, though.
15:00pmatulis has quit IRC
15:00
<bieb>
it will be a linux server in the rack..
15:00
<_UsUrPeR_>
gadi: FYI "cat /dev/ttyS0" does output stuff
15:00
<Gadi>
vagrantc: but does root work better
15:00
<_UsUrPeR_>
so the connection between monitor and client does work
15:01
<Gadi>
_UsUrPeR_: did you add the ">> $XCONF" to the write_screen line?
15:01
<alkisg>
bieb: ah, ok, then I don't think nbdroot is needed at all...
15:01
<bieb>
Gadi: we are running AD and the DHCP server is a windows box. That's why I was asking about the https://help.ubuntu.com/community/UbuntuLTSP/LTSPWindowsDHCP doc
15:02
<_UsUrPeR_>
Gadi: yeah, just copy-pasted your last creation in. the results showed no changes though :/
15:02
<vagrantc>
Gadi: yes, /media/root is working great for me now.
15:02
<Gadi>
vagrantc: ah, thats good
15:02
<bieb>
the LTSp server is currently sitting on my desk. I have to make a hardware change from the original configuration.. then Ubuntu will be installed and then inserted in the rack
15:03otavio has quit IRC
15:04
<Gadi>
_UsUrPeR_: from a shell on the client, can you run: sh /usr/share/ltsp/screen_session
15:04
and then look at /var/run/ltsp-xorg.conf
15:05
and make sure that line 2 of that file is coorect
15:05
on the client
15:05
<_UsUrPeR_>
when running "sh /usr/share/ltsp/screen_session", I am getting the repeating error "openvt: --: illegal vt number
15:05
"
15:05
from this screen, I cannot look at /var/run/ltsp-xorg.conf
15:06
<Gadi>
crap
15:06
ctrl-c
15:06
hehe
15:06
<_UsUrPeR_>
k
15:08otavio has joined #ltsp
15:10
<ltsppbot>
"Gadi" pasted "for _UsUrPeR_ - take 4" (43 lines) at http://pastebot.ltsp.org/417
15:10
<Gadi>
ok, here
15:10
you should definitely get a Screen section this time
15:10
which is what we need
15:11
I gotta go in 5 mins
15:15
vagrantc: can you push your changes upstream?
15:15
to ltspfs-trunk
15:15
<vagrantc>
Gadi: working on it
15:15rjune_wrk has quit IRC
15:15
<Gadi>
great, thx
15:15
gotta run
15:15
<ltsppbot>
"_UsUrPeR_" pasted "for Gadi -- Xorg.7.log part 5" (1025 lines) at http://pastebot.ltsp.org/418
15:16
"_UsUrPeR_" pasted "for Gadi -- ltsp-xorg.conf part 5" (25 lines) at http://pastebot.ltsp.org/419
15:17
<Gadi>
_UsUrPeR_: on the client console, can you run: X :2 -conf /var/run/ltsp-xorg.conf
15:17
<_UsUrPeR_>
Gadik
15:17
<Gadi>
and see if the cursor moves when you touch it
15:18
the log file looks like it isn't picking up the conf file at all
15:18
try adding the CONFIGURE_X = True back in, too
15:18
I need to go
15:18
will touch base in a bit
15:18Gadi has left #ltsp
15:20
<ltsppbot>
"_UsUrPeR_" pasted "for Gadi -- Xorg.0.log part 1" (966 lines) at http://pastebot.ltsp.org/420
15:21Ahmuck has joined #ltsp
15:24
<bieb>
vagrantc: how many clients are you running?
15:24
<akSeya>
johnny, it doesn't make sense.. I am not able to change pxelinux.cfg.default ...
15:25
I do the changes, save it, restart xinetd .. but changes does not take effect
15:25
I searched for pxelinux.cfg/default on other folders.. there is no other...
15:27
<vagrantc>
bieb: in what context?
15:29
<bieb>
vagrantc: context of how many thin clients are you runnning on your ltsp server.. is it a lab? or do you have them rolled out to people's desks?
15:30
<vagrantc>
bieb: i have many ltsp servers, and most of them are fragmented into different pieces of ltsp server (i.e. nfs+tftp on one, dhcp on another, applications on several) ..
15:31
bieb: some lab, some desks.
15:31
and of course, development, which is usually virtual thin clients on my laptop, or occasionally an external thin client
15:32
bieb: at freegeek, we've got probably 15-25 thin clients, and about 3-4 application servers.
15:32rjune_ has quit IRC
15:32
<bieb>
vagrantc: what type of apps do you provide to users?
15:32rjune_ has joined #ltsp
15:33
<vagrantc>
bieb: iceweasel/firefox, openoffice ... a handfull of others
15:33Lns has joined #ltsp
15:33
<bieb>
vagrantc: what size home dir do you provide your users?
15:34
<vagrantc>
bieb: it's not a per-user thing...
15:34
some users use 4-6GB, some users use 150MB
15:34
some less that tha
15:35
than that
15:35ogra_ has joined #ltsp
15:35
<bieb>
vagrantc: I see.. thats a pretty big split
15:35
<vagrantc>
bieb: it just depends on what they're doing
15:35
and how lazy they are at cleaning up old junk
15:36
<ogra_>
junk ? where ?
15:36* ogra_ grabs the sweeper
15:36
<Lns>
I hear the linux defrag utility automatically cleans up stuff you don't use. I heard a microsoft employee who wrote the "disk cleanup utility" actually wrote it
15:37
<ogra_>
bah, just use ext4
15:37
<johnny>
akSeya, huh.. what filename ? what you just wrote wasn't right
15:37* vagrantc wonders what count's as "don't use"
15:37
<_UsUrPeR_>
Lns: the real kicker: does it warn you if you have not used a desktop icon in a while?
15:37
<akSeya>
johnny, pxelinux.cfg/default
15:37
<vagrantc>
frag defrag.
15:37
<Lns>
_UsUrPeR_: yes, as well as hiding all the inactive applets on your panel
15:37
<bieb>
vagrantc: thanks... I am trying to determine the hard drive size for the server (OS and Apps) and then a separate /usr drive for home Dirs
15:37
<ogra_>
heh
15:38
<akSeya>
johnny, but I find out that it's loading pxelinux.0 instead
15:38
<Lns>
bieb: /home for users
15:38
<_UsUrPeR_>
ogra/lns/anyone else. Gadi and I have been working on touchscreen stuff for 9.04
15:38
<ogra_>
cool
15:38
<bieb>
Lns: your right.. my bad.. reading the wrong stuff as I am typing here.. :D
15:38
<johnny>
akSeya, it loads pxelinux.0 and then the config file
15:39
<_UsUrPeR_>
and it's gotten sort-of-almost there
15:39
<bieb>
damn multi tasking
15:39
<Lns>
_UsUrPeR_: nice =)
15:39
<_UsUrPeR_>
but it's not quite there :(
15:39
<akSeya>
johnny, hum.. but it is ignoring the config file :(
15:39
<_UsUrPeR_>
and it's driving me insane
15:39
<ogra_>
_UsUrPeR_, note that the touchscreen handling will change a lot in the near future
15:39
<johnny>
are you sure?
15:39
or is the config file empty?
15:39
<_UsUrPeR_>
ogra_: not what I wanted to hear ;)
15:39
<ogra_>
well, thats why i didnt touch it in jaunty
15:40
<_UsUrPeR_>
yeah :P
15:40
<ogra_>
i wanted to give xorg and the kernel time to settle
15:40
<johnny>
akSeya, if i had a gentoo box available, i'd be able to fix some of the issues like automatically symlinking pxelinux.0 and writing a base config file
15:40
<_UsUrPeR_>
we're just trying to use the X_TOUCH_<command> stuff
15:40
<ogra_>
in karmic the plan is to add all usb touchscreens to the kernel driver
15:40
<johnny>
mmm.. food
15:40
<_UsUrPeR_>
man, that would be great
15:41
<ogra_>
if thats done you should be able to easily configure them at runtime
15:41
<_UsUrPeR_>
the problem I am having right now: I can see in /var/run/ltsp-xorg.conf that everything is inserted in the xorg.conf
15:41
<ogra_>
though the ltsp solution probably takes until karmic+1
15:41
<_UsUrPeR_>
yet nothing is showing up in Xorg.7.log
15:41
<ogra_>
right
15:41
use hal instead
15:41
<_UsUrPeR_>
however, if I cat /dev/ttyS0, I will get output from the monitor
15:41
<ogra_>
i gave Gadi the right pointers
15:42
oh, its a serial touchscreen ?
15:42
thats very tricky
15:42
<_UsUrPeR_>
it's got both. Gadi is saying serial will be easier for the moment
15:42
and USB later
15:42
<ogra_>
serial will always need configs
15:42
<_UsUrPeR_>
which is fine. We have the configs in order
15:42
<ogra_>
right
15:42
<_UsUrPeR_>
in the lts.conf
15:42
<ogra_>
but i cant help you much
15:42
<_UsUrPeR_>
ok
15:43
<ogra_>
serial is something you cant probe ...
15:43
i usually work on automation
15:43
<_UsUrPeR_>
well if you ever want to start testing out USB on an elotouch with ltsp, let me know :)
15:43
<ogra_>
if on a serial port your config doesnt work, it doesnt work
15:44
no chance to fix it if you dont know the right config parameters
15:44
<_UsUrPeR_>
(FYI: right now, it is detected properly, but the mouse input is @ 800x600, and it's upside-down)
15:46
^^ Referring to USB input
15:46
<ogra_>
right, as i said ... know the right config and you are golden ... if not ... feel the serial pain
15:46
oh
15:46
well, there should be a swapxy setting or so
15:46
<_UsUrPeR_>
according to etyack, I have the exact minX, maxX, minY, maxY settings. He has told me he's gotten it to work in ltsp 4.2
15:47
<ogra_>
right, but the driver might need s swap command
15:47
does it use a special kernel module ?
15:47
if so you can usually change direction with a module parameter
15:48
<_UsUrPeR_>
nothing that's not pre-installed...
15:48
how do I find out what module it's using?
15:48
<ogra_>
well, lsmod on a client and look for suspicious modules
15:50* vagrantc is starting to suspect "su" is the culprit with local ltspfsmounter
15:50
<ogra_>
su ?!?
15:51
where the heck is that used ?
15:53
<vagrantc>
su ${USER} "/usr/sbin/ltspfsmounter ...
15:53
<ogra_>
hrm
15:53
<vagrantc>
both for delayed mounter and in the udev scripts
15:53
to support localapps with ltspfs devices
15:54
<ogra_>
thats not my code
15:54
<vagrantc>
yes, it's not really my code either
15:54* vagrantc cringes
15:54
<ogra_>
where would it get ${USER} ?
15:54
<vagrantc>
i've made some edits to it
15:54
ogra_: it is the epitome of hack: grepping /proc/mounts for an sshfs mount
15:55
<ogra_>
ugh
15:55* ogra_ is shocked
15:55* ogra_ grabs the sweeper
15:55
<ogra_>
:)
15:55
<vagrantc>
ok, so, now i seee ... fusermount: Permission denied
15:56
<_UsUrPeR_>
ogra_: it looks like it loads joydev
15:56
when the usb screen is plugged in
15:56
joydev = module
15:56
<ogra_>
hrm
15:56
yeah
15:56
likely wrong
15:56
<vagrantc>
user is in fuse group, file looks like fuse group, /dev/fuse has writeable group permissions on fuse ...
15:56
<ogra_>
i guess thats something usbtouchscreen will fix
15:56
<vagrantc>
what's fusermount got to complain about?
15:57
<ogra_>
is user in fuse group locally ?
15:57
or only on the server
15:57
fusewrmount is executed setgid
15:57
at least it was like that until recently
15:58
i think it was just dropped
16:00* ogra_ grins reading the debian mono stuff
16:00agike_ has quit IRC
16:01
<vagrantc>
ogra_: yeah, the user is in the fuse group locally...
16:01
fusermount is still setgid here.
16:02
<ogra_>
gid might not match what you get from the server though
16:02
<vagrantc>
er, setuid, apparently.
16:02
gid is the same.
16:02
<akSeya>
knipwim, i was able to fix the pxelinux problem.. now I'm getting mount: RPC: Unable to Send; errno = Network is unreachable !! NFS Mounting failed. Is the pash correct ... I took another computer on the network and I was able to mount the NFS normally
16:03
<vagrantc>
well, this baffles my mind.
16:05* vagrantc chmod +x fusermount
16:05* ogra_ needs to go to eat, else susie gets angry ... later
16:06
<vagrantc>
well hell, that works.
16:08akSeya has quit IRC
16:08akSeya has joined #ltsp
16:13pmatulis has joined #ltsp
16:24Pulga has joined #ltsp
16:26sene has quit IRC
16:35Gadi has joined #ltsp
16:35
<Gadi>
_UsUrPeR_: ping
16:39
<akSeya>
well.. my client in unable to mount NFS.. is this something I should care about: in.tftpd[7100]: tftp: client does not accept options
16:39spectra has quit IRC
16:40
<johnny>
not really
16:40
that error is unrelated
16:40
do you have the proper network drivers in your genkernel initramfs?
16:40
that is the one problem with the current setup.. still waiting on genkernel devs to fix
16:40
i'm probably going to end up switching to dracut because of it
16:40
<akSeya>
hum
16:41
<johnny>
you need to chroot in and edit /usr/share/genkernel/x86/modules_load
16:41
and add your nic modules
16:41
<akSeya>
ok ;)
16:41
<ltsppbot>
"Gadi" pasted "_UsUrPeR_ : move the file to XS80-touchscreen and make it look like this" (38 lines) at http://pastebot.ltsp.org/421
16:41
<johnny>
and then genkernel initrd or genkernel ramdisk
16:41
depending on the version
16:42
<vagrantc>
Gadi: well, for reasons i can't explain, the problem all along has something to do with group permissions on /usr/bin/fusermount ...
16:42
<akSeya>
huuum.. now i'm understanding this better
16:44
<Gadi>
vagrantc: hmm... do tell
16:44
<akSeya>
johnny, so, I'll need to add all network cards my lan has on modules_load.. right?
16:45
<vagrantc>
Gadi: for some reason, even though the user is locally in the fuse group, and fusermount is executable by the fuse group, it still refuses to let the user run fusermount ... adding it to be world-executable, and it works
16:45
Gadi: though it doesn't unmount properly.
16:46
although the commits i just made were required to get that much...
16:46
getting XAUTHORITY properly set and such
16:47sindfil has quit IRC
16:48
<vagrantc>
Gadi: *seemed* to work with nbd+squashfs+aufs now that i got that other code updated... but i'm just now double-checking for sure.
16:48
<Gadi>
vagrantc: umount as well as mount?
16:49
<vagrantc>
Gadi: maybe different GIDs on the server cause problems for NFS somehow ... though i don't understand why, as locally it uses the appropriate gid.
16:49sindfil has joined #ltsp
16:49
<vagrantc>
Gadi: umount doesn't appear to work ... my guess is /etc/passwd and freinds get reset too early?
16:49
<Gadi>
if you chown it to the server's gid does it work? :P
16:50
<vagrantc>
that would seem all kinds of wrong...
16:50
but worth trying
16:50ioizzgd has joined #ltsp
16:50lnchginos has joined #ltsp
16:50
<Gadi>
we are at the fringes of what is normal
16:50
:)
16:52
<vagrantc>
Gadi: that might be too kind
16:55
well, that test went all kinds of wonky.
16:55
now remote mounts didn't kick off.
16:58
<akSeya>
johnny, ok.. already have the correct module.. it is loading.. but still not able to mount nfs
17:03mikkel has quit IRC
17:03dan_young has quit IRC
17:04
<akSeya>
johnny, on a normal machine i can mount it with no problems...
17:05
gonna try on another NIC
17:10
hum.. on some machines I get the same error, on others (when I use a floppy) I get "segment [a number]:[other number] does not fit in any memory region"
17:16fotanus has left #ltsp
17:16alkisg has quit IRC
17:22bieb has left #ltsp
17:24J45p3r__ has joined #ltsp
17:24J45p3r__ has left #ltsp
17:46lucascoala has joined #ltsp
17:47
<akSeya>
johnny.. knipwim ... I guess I'll stick with ltsp-4.2 for now.. anyway, thanks a lot for your help and attention
17:48
<johnny>
ltsp 4.2 has security holes
17:49
and is not updated
17:49
i don't know why nfs is not working,
17:49
you haven't even shown me what your pxelinux.cfg/default looks like
17:49
perhaps it is not correct
17:49
<akSeya>
hum..
17:50
one second
17:50
<johnny>
akSeya, the real problem here is that i just don't have a gentoo machine around :(
17:51
<akSeya>
well.. i'm able to boot now.. it loads the kernel, the nic modules, but is unable to mount root
17:51
also says that network is unreachable
17:52
johnny, http://pastebin.com/m70919c2b
18:01
johnny, my problem is that I have only tomorrow to get it working
18:04
<johnny>
that's not what i had in my config
18:04
mine was somewhat differently
18:04
i wonder if i still even have it :(
18:06bobby_C has quit IRC
18:08
<johnny>
dberkholz, are you here?
18:18
<akSeya>
johnny, I have good news
18:19
I added debug to the append line, an ifconfig shows me only lo
18:19
then I did ifconfig eth0 10.0.0.250 up
18:19
<johnny>
is your nic module actually loaded?
18:19
oh.. so it is
18:19
<akSeya>
exit the console
18:19
and it loaded
18:19
<johnny>
ok so.. that means it's not getting dhcp
18:19
are you sure your dhcp server is loading then?
18:19
<akSeya>
hum...
18:20
<johnny>
well i guess it must if you used pxe
18:20
<akSeya>
yeah...
18:20
hum.. i have and idea.. one second..
18:21
huum
18:21
looks like ltsp-build-client does not changed /etc/conf.d/net
18:21
<johnny>
are you gonna be around for a little while ? because i must travel 20 minutes awa
18:21
but then will be back
18:23
<akSeya>
hum.. i'll wait
18:23
;)
18:23
<vagrantc>
Gadi: ok, with my last ltspfs patches, it works fine with nbd+ext2 and nbd+squashfs ... so the bug seems to have something to do with the way NFS handled GIDs.
18:23
Gadi: and group-executable files...
18:25
<akSeya>
johnny, I added config_eth0=( "noop" ) to /etc/conf.d/net on the chroot
18:25
lets see if it works :P
18:25
nope
18:27
<johnny>
you don't need to add that , i didn't
18:27
ok bbiaf
18:27johnny has left #ltsp
18:30
<Lns>
bbiaf = be back, i am farting
18:30* Lns needs to go home for the day =p
18:37staffencasa has quit IRC
18:38vagrantc has quit IRC
18:39
<akSeya>
LTSP 5 tooks a lot longer to boot ... :-/
19:04johnny has joined #ltsp
19:09
<akSeya>
johnny, welcome back :D
19:10
<johnny>
akSeya, i think your default pxe file isn't correct
19:10
and now i just remember where it is .. but i am not near it now
19:11
<akSeya>
yeah.. i think something is missing there.. so the eth0 does not activate
19:13agike has joined #ltsp
19:14
<akSeya>
johnny, the pxelinux/dega
19:15
ops.. default is not distro dependant.. right?!
19:17Lns has quit IRC
19:17
<johnny>
distro dependant.. no
19:23pmatulis has quit IRC
19:23fotanus has joined #ltsp
19:24
<akSeya>
well.. i'm going home.. will try tomorrow a but more
19:24
c ya
19:25akSeya has quit IRC
19:54CAN-o-SPAM has quit IRC
19:59agike has quit IRC
20:03johnny has left #ltsp
20:14jammcq has joined #ltsp
20:14
<jammcq>
hey all
20:25
<stgraber>
hi jammcq
20:35Pulga has quit IRC
21:37hanthana has joined #ltsp
22:02vagrantc has joined #ltsp
22:09try2free has joined #ltsp
22:36
<sbalneav>
Gadi: ping
22:38
<vagrantc>
yay. i done gone figured out what was borked with ltspfs and local mounting on debian
22:38
<jammcq>
Scotty !!!!!!!!!!!!!!!!!!!!!!!!!!
22:38
vagrantc!!!!!!!!!!!!!!!!!!!!!
22:38
<sbalneav>
Hey jammcq
22:38
<vagrantc>
the delayed mounter hooks were running when the user wasn't in /etc/passwd
22:38
<jammcq>
how's things going?
22:38
<sbalneav>
vagrantc: What was causing the problem?
22:38
<vagrantc>
jammcq, sbalneav: how was brazil?
22:38
<jammcq>
twas great
22:38
<sbalneav>
Excellent.
22:39
vagrantc: Got any good tutorials on how to i18n docbook documents? We have people who want to translate the docs into spanish and portugese
22:47
<vagrantc>
sbalneav: alas, not much.
22:48
sbalneav: maybe po4a ?
22:49
the other ltspfs local mounting problem is that when using NFS, even if the client says it's group fuse, if the server has a different GID for the fuse group, that somehow still messes things up. even though the client thinks it's the fuse group ... at least for purposes of group-executability ... it refused to run fusermount, essentially.
22:50
<sbalneav>
nfs gid mapping goo needed?
22:50
<vagrantc>
perhaps. but i don't really understand why ... as the NFS server should just be serving the the numeric ID, and the client thinks it's fuse. and it's running on the client ... so i don't get why it's borked.
22:52
but it is.
22:53
<sbalneav>
is the fuse gid different on the server and the client?
22:53
<vagrantc>
yes.
22:53
in this case, fuse is 107 on the server, and 106 on the thin-client/chroot.
22:54
the file is gid 106.
22:59
works fine using NBD with ext2 or squashfs.
23:07
<sbalneav>
yeah, with nfs, you HAVE to have gid/uid's consistent between client and server
23:07
or bad mojo happens.
23:08
<vagrantc>
it just seems inexplicable why there is even a problem.
23:08
if it was just sticking with the GIDs and not using the group NAMEs, everything would be fine.
23:09
somehow it's actually using the group names or something
23:11
this local ltspfs mounting requires so many ugly hacks ... we really need to re-think ltspfs a little for localapps and such.
23:12
<sbalneav>
? Aren't we using sshfs for localapps?
23:13
<vagrantc>
sbalneav: for the /media mounts, we're using ltspfs locally.
23:13
sbalneav: otherwise that's a lot of round-trips to get a localapp to talk to the local media... but gah. it's a mess.
23:14johnny has joined #ltsp
23:26johnny has left #ltsp
23:28johnny has joined #ltsp
23:39Sarten-X has quit IRC
23:44cyberorg has joined #ltsp
23:49Ahmuck-Jr has joined #ltsp
23:58alkisg has joined #ltsp