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


Channel log from 16 February 2010   (all times are UTC)

00:04alexqwesa_ has joined #ltsp
00:04alexqwesa has quit IRC
00:10vagrantc has joined #ltsp
00:11frederickjh has joined #ltsp
00:21Damianos_ has quit IRC
00:32alkisg has joined #ltsp
00:47pts has joined #ltsp
01:24klausade has joined #ltsp
02:16mikkel has joined #ltsp
02:35alkisg has quit IRC
02:36Kicer86 has joined #ltsp
02:38alkisg has joined #ltsp
02:48alkisg has quit IRC
02:50alkisg has joined #ltsp
03:06sene has joined #ltsp
03:41vagrantc has quit IRC
03:45monteslu has quit IRC
03:48monteslu has joined #ltsp
04:05ogra has quit IRC
04:05ogra has joined #ltsp
04:19bobby_C has joined #ltsp
04:55bobby_C has quit IRC
05:45Faithful has joined #ltsp
05:57bobby_C has joined #ltsp
06:03bobby_C has quit IRC
06:16vvinet has quit IRC
06:19bobby_C has joined #ltsp
06:20pmatulis has joined #ltsp
06:23
<alkisg>
Gadi_eeepc: Updating /var/lib/tftpboot directories for chroot: /opt/ltsp/i386
06:23
Port already in use. Changing to 2000.
06:24
(works fine, just the message is misleading)
06:35bobby_C has quit IRC
06:47Blinny has joined #ltsp
07:22Gadi_eeepc has left #ltsp
07:26shawnp0wers has joined #ltsp
07:36lucascoala has joined #ltsp
07:37vvinet has joined #ltsp
07:39johnny has quit IRC
07:54mikkel has quit IRC
07:59alkisg has quit IRC
08:06etyack has joined #ltsp
08:07mikkel has joined #ltsp
08:12lucascoala has quit IRC
08:21zeltron has joined #ltsp
08:22
<zeltron>
Hi. I'm testing setup of ltsp-cluster, and I've got a problem when my terminal tries to start ldm
08:22Gadi has joined #ltsp
08:23
<zeltron>
Error message : syntax error in /var/cache/getltscfg-cluster/lts.conf
08:23
<ogra>
fix your lts.conf then :)
08:24
<zeltron>
ogra : OK, but there is no lts.conf in /var/cache/getltscfg-cluster
08:24
(neither on terminal, nor on servers
08:24
In fact, I can't find an lts.conf file
08:25
I guess it should be created by control-center (the web interface in which I put some parameters),
08:26
I guess too (maybe I'm wrong) it puts these parameters into postgresql db
08:26
but I can't find no lts.conf file
08:27
<Blinny>
Sounds like a Rolling Stones song.
08:28
<ogra>
i might be wrong but i think even in ltsp-cluster lts.conf lives somewhere under /var/lib/tftpboot on the server
08:29
<alincoln>
Blinny: hah!
08:30
<zeltron>
ogra : I thought so too, but no trace there !
08:31
<Gadi>
guys, in ltsp-cluster, lts.conf is created on the fly from a wget call to the control center
08:32
if you have none, it means your client did not successfully contact the control center
08:32
<ogra>
ah
08:32
it shouldnt fail to boot without one though
08:32
<Gadi>
have a look at /opt/ltsp/i386/usr/bin/getltscfg-cluster
08:33
no, if you get a syntax error in any lts.conf, it will use defaults
08:33
<ogra>
right, else its a serious bug
08:33
<Gadi>
which may or may not produce a useable system (ie video might not be right)
08:33
<ogra>
ldm should still come up
08:34
<zeltron>
Gadi : I've got a file /opt/ltsp/i386/usr/bin/getltscfg-cluster
08:34
but it's a little too complicated for me !
08:34
<Gadi>
zeltron: right - thats a shell script that is what ltsp-cluster runs to get the lts.conf file
08:35
if you get to a shell on the client (either by installing ssh and/or setting a root password, you can try running getltscfg-cluster -a at a client shell
08:36
otherwise, be sure to follow the ltsp-cluster instructions carefully
08:36
all of those lts.conf vars that it suggests to add to the control center, you should add
08:36
there are some that are needed just to get to ldm
08:37
<zeltron>
Gadi : I tried getltscfg-cluster -a on the client, and got the error message
08:37
(the same one)
08:40
<Gadi>
zeltron: is the client's /etc/ltsp/getltscfg-cluster.conf correct?
08:42
<zeltron>
It seems. The only difference between mine and the doc's one is I put the IP from the server instead of his dns name
08:42
(The server is not in dns at this time)
08:42
Should it be the source of my problem ?
08:43
<Gadi>
no, IP should be better
08:44pts has quit IRC
08:48frederickjh has quit IRC
08:49frederickjh has joined #ltsp
08:51
<zeltron>
I just noticed something really strange :
08:52
I'm logged as root on the client console and I try : ls /var/cache/getltscfg-cluster
08:52
The answer is changing every second !
08:53
the first time, there are 3 files (and an lts.conf), the second time, 2 files, then one, then no directory at all, then 3 filess...
08:54rjune has joined #ltsp
08:56
<zeltron>
Ok : I found the solution !
08:56
The doc on the net gave the following value for LDM_SERVER :
08:57
%LOADBALANCER%
08:57
I thought it was a constant ltsp knew, but it seems it is not
08:58
I replaced it with the IP of the server and it works... better
08:58
Only better, because now, I've got a screen on the client that tells me the client conf (MAC ADDRESS, IP, CPU...), but no way to get further !
09:00
<alincoln>
isn't there something funny about that window not responding to one or the other of mouse/keyboard?
09:00
try the keyboard if the mouse doesn't work.
09:01
<zeltron>
alincoln : I'm beginner with LTSP-cluster, but not with computer ;-)... I already tried it, of course
09:01
<Gadi>
zeltron: you can only use %LOADBALANCER% if you are using ltsp-cluster's loadbalancer
09:01
<alincoln>
zeltron: heh, fair enough
09:01
<zeltron>
(return key, alt-o, altgr-o...), nothing works
09:02
Gadi : I am using it
09:02
<sbalneav>
Morning al
09:02
all
09:04
<zeltron>
alincoln : You were (already) right : I had to move the mouse, there was no cursor, but at a time, I saw the 'OK' bouton highlight.
09:04johnny has joined #ltsp
09:05
<zeltron>
But click didn't work, so I had to press the enter key !
09:05
<alincoln>
zeltron: ah!
09:05
i sat at that screen once and wondered what to do.
09:07etyack has quit IRC
09:14CAN-o-SPAM has joined #ltsp
09:14cliebow has joined #ltsp
09:26lucascoala has joined #ltsp
09:37alkisg has joined #ltsp
09:38lucascoala has quit IRC
09:44etyack has joined #ltsp
09:45cliebow has quit IRC
09:45cliebow has joined #ltsp
09:54nosedrum has joined #ltsp
09:54
<nosedrum>
hi
09:55
I experiment a Firefox issue running on a ltsp client
09:56
On a buisness jsp site the result of a form normally appear on local firefox but not on the ltsp client
09:56
I have no idea why
09:56
nobody have some experience with that ?
09:57
<sbalneav>
Soryy, what does "result of a form" mean?
09:57
when you're typing, it doesn't appear?
09:57
or the java applet's showing blank, or what?
09:57
We'd need more info.
09:58
<nosedrum>
I'm typing a reference in a form and submit for result
09:58
Typing OK on local and client
09:58
but the result of submit (a list of two links) appear on local firefox (server) but not on client side
09:59
<sbalneav>
First off, is this an embedded java applet? Or is the jsp just returning a "standard" web page?
10:00
<nosedrum>
it's return standard web code
10:00
js and html
10:00
<sbalneav>
ok
10:00
<nosedrum>
i trace it with firebug
10:01
<sbalneav>
so, when you do a "show page source" on the returned page, is it the same for both the local and thin client?
10:04
<nosedrum>
sorry was on phone (work)
10:04
good idea
10:04
i'll check that
10:04
i'll be back soon
10:06lucascoala has joined #ltsp
10:09
<nosedrum>
ok
10:09
so some diff
10:09
diff local.srv client.srv
10:09
like:
10:09
< src="FC97A38A351269D186E84FB6842724B57.srv?dlg=CtrDefault&action=Enter"
10:09
---
10:09
> src="E47305932E8DAEECB9F65C78C41E3C447.srv?dlg=CtrDefault&action=Enter"
10:10
just on id
10:10
just normally things
10:11
i think
10:12
the rest of page source is the same
10:15
<sbalneav>
So, the code for the links is there, but the links simply aren't displayed?
10:15
<nosedrum>
right
10:16
exactly
10:16
<sbalneav>
What's there in their place?
10:16
just blanks?
10:16
<nosedrum>
yes just blank
10:17
<sbalneav>
If you highlight the page, are the links there? I'm just wondering if they're in white for some reason?
10:17staffencasa has joined #ltsp
10:17
<nosedrum>
no, i have try... normally the result is a table with two icons on bottom and the links on table cells
10:18
<cliebow>
Canadians Rock!
10:18
<nosedrum>
on the client is ther just table headers with title and the two iconw on botton.. no cells, no links...
10:18
<sbalneav>
Hm, never seen that before.
10:18Kicer86 has quit IRC
10:18
<sbalneav>
What are you running?
10:18
<nosedrum>
ubuntu 9.04
10:19
<sbalneav>
is your chroot up-to-date?
10:19
<nosedrum>
firefox shiretoko
10:19
iep
10:19
<sbalneav>
firefox which?
10:19
<nosedrum>
ff 3.5.7 = shiretoko
10:19
<sbalneav>
So a different firefox than ubuntu comes with?
10:20
<nosedrum>
Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1.7) Gecko/20100106 Ubuntu/9.04 (jaunty) Shiretoko/3.5.7
10:20
<cliebow>
64 bit?
10:20
<nosedrum>
iep
10:20lucascoala has quit IRC
10:20
<sbalneav>
iep?
10:21
<nosedrum>
yes
10:21
<cliebow>
like yep?
10:21
<nosedrum>
sorry i'm french
10:21
<cliebow>
8~)
10:21
<nosedrum>
:D
10:22
<sbalneav>
ok, then I'm at a loss. Only thing I could suggest is that with 9.04, I beleive the libxcb bug wasn't fixed yet.
10:22
Have a look at stgraber's ppa for a fixed libxcb, try installing that in both the chroot and the server.
10:22
!stgraber
10:22
<ltspbot>
sbalneav: "stgraber" :: rules
10:22
<sbalneav>
umm
10:22
!stgraber-ppa
10:22
<ltspbot>
sbalneav: "stgraber-ppa" :: https://launchpad.net/~stgraber/+archive/ppa
10:22
<sbalneav>
^^
10:22zeltron has quit IRC
10:22
<nosedrum>
ok and up to 9.10 on local and chroot not a solution ?
10:23
for this fix ?
10:23
<sbalneav>
that would be a solution too.
10:23
Or something you could try, anyway
10:23
<nosedrum>
ok. i'll try that bu it's very strange.
10:23
<sbalneav>
I've never seen web pages not render correctly
10:23
<nosedrum>
i didn't specify you this appening at random
10:24
some time result appear sometimes not
10:24
<sbalneav>
If the results don't appear, is the firefox spinner still spinning?
10:24
<nosedrum>
firefox spinner ?
10:24
<sbalneav>
i.e. is this a case that the page isn't fully loading
10:25
throbber, whatever they call it. The "still loading" indicator :)
10:25
<nosedrum>
ok... no spinner always finish to search
10:25
<sbalneav>
ok
10:25
Was just wondering if this was maybe a networking problem.
10:25
<nosedrum>
no network problem
10:25
sure
10:25xet7 has joined #ltsp
10:25
<sbalneav>
Honestly, I've never seen this before.
10:26
I'd try an upgrade to 9.10
10:26
<nosedrum>
Ok i'll try
10:26
<sbalneav>
after full backups, of course :)
10:26
<nosedrum>
a last question
10:26
<sbalneav>
Sure
10:26
<nosedrum>
my local server is x86_64 and my lts image is build with -a i86
10:26
is it a problem ?
10:26
i386*
10:27
<johnny>
that's what most people do..
10:28
<nosedrum>
ok so 64 application lauch from thin client are not "unstable"
10:28
<sbalneav>
No, that should be fine.
10:28
Lots of people deploy ltsp that way.
10:29
<nosedrum>
(Good question after 80 production sites installed... lol)
10:30
<cliebow>
heheh!
10:30
<sbalneav>
heh
10:30
<nosedrum>
Ok thanks for help and info... And for all what you do on LTSP project ! see you
10:30
<sbalneav>
No problem
10:34
<stgraber>
pushing a new nbd-proxy to ltsp-trunk
10:34
it'll need some more testing to make sure it's stable though
10:34
<cliebow>
cool! 80 independent sites?
10:37
<stgraber>
new ltsp-trunk snapshot on its way
10:37
alkisg: I'd appreciate if you can test it. It's supposed to be the same as what we gave you but we never know what's happening on LP :)
10:38
<alkisg>
stgraber: thanks, I'll test it
10:42otavio has quit IRC
10:43otavio has joined #ltsp
10:55Faithful has quit IRC
11:31litlebuda has joined #ltsp
12:13Gadi has left #ltsp
12:18Lns has joined #ltsp
12:20nosedrum has quit IRC
12:20
<knipwim>
i have a weird xorg issue on my client
12:20
ldm starts
12:21
but a startx fails
12:21
when logging on to the server, the login process also fails at starting X
12:22
any ideas where the problem might be?
12:22
<alkisg>
compiz?
12:22
Ah, that's gentoo?
12:22
<knipwim>
gentoo id, and no compiz
12:23
<alkisg>
~/.xsession-errors?
12:23
<knipwim>
all hal stuff is compiled in on the client
12:29
hmmm, /usr/lib/X11/xdm/Xsession: line 71: exec: xsm: not found
12:29
and it's not installed
12:29
that could be it :)
12:34
yeah! it workes
12:35
alkisg: thanks a lot, i've been trying things for hours
12:35
thinking it was a kernel/module problem
12:37
<rjune>
howdy all
12:39
<alkisg>
knipwim, you're welcome
12:58litlebuda has quit IRC
12:59litlebuda has joined #ltsp
13:12litlebuda has quit IRC
13:16litlebuda has joined #ltsp
13:18dro has joined #ltsp
13:21ogra has quit IRC
13:21ogra has joined #ltsp
13:27ogra has quit IRC
13:28evilx_ has quit IRC
13:29ogra has joined #ltsp
13:36litlebuda has quit IRC
13:51GodFather has joined #ltsp
13:53Sarten-X2 has quit IRC
14:01
<knipwim>
johnny: you there?
14:01
<johnny>
hi
14:03
<knipwim>
i'd like to discuss two issues with you
14:03
first: http://bazaar.launchpad.net/~wimmuskee/ltsp/ltsp-gentoo-dev/revision/1497
14:03Sarten-X has joined #ltsp
14:03
<knipwim>
which includes hal support for the client xorg-server
14:06mikkel has quit IRC
14:07xet7 has quit IRC
14:10cliebow has quit IRC
14:11Blinny has quit IRC
14:12shawnp0wers has quit IRC
14:18cliebow has joined #ltsp
14:18evilx has joined #ltsp
14:23
<johnny>
knipwim, add todos for those, so we can search
14:24
and switch the ordering of hal and dbus, so the comment only applies to hal
14:25
knipwim, you might also want to look into the changed to the ubuntu/debian initscripts to see if we can remove some of the stuff we had there
14:25
there were some things .. i can't recall
14:25
also.. ltsp cluster stuff at some point :)
14:29
<alkisg>
Wow... "Gentoo is getting very close" ;)
14:29
Well done guys
14:29
<johnny>
it's been there for ages
14:30
it can be changed anytime
14:30dro has quit IRC
14:30
<johnny>
actually.. there's still one major stop for awesome integration tho :(
14:30
alkisg, our initramfs generator is not very good
14:30
and still hastn' gotten better
14:31
i really really wanted to use what gentoo puts in their own guide
14:31
genkernel
14:31
BUT......... it does not work for us at all with ltsp.. not enough network drivers by default, and you must edit a file manually to add more
14:32
so.. we probably need to switch to dracut..
14:33
<knipwim>
johnny: the other thing i wanted to discuss
14:33
was that ltsp-client won't login unless a wm is present
14:33
<johnny>
of course
14:33vagrantc has joined #ltsp
14:33
<johnny>
what do you expect
14:33kebax has joined #ltsp
14:33
<knipwim>
but the default wm, tvm, is not installed by default
14:33
<kebax>
hello
14:34
<johnny>
hmm.. did that change?
14:34* vagrantc expects good food and drink
14:34
<knipwim>
in the xinit package, the minimal use flag is enabled by default now
14:34
<johnny>
ah!
14:34
bastards..
14:34
<knipwim>
leaving out xsm, xterm, tvm and such
14:35
<johnny>
so.. should ltsp-server dep on some virtual ?
14:35
is there such a virtual?
14:35
<knipwim>
so we could add xinit[-minimal] to ltsp-server
14:35
<johnny>
virtual/vm ?
14:35
err virtual/wm
14:35
btw.. i can't believe portage 2.2 is not stable yet
14:35
wtf
14:35
how long has it been?
14:36
<knipwim>
they're proably working on it
14:36
<johnny>
that's why i stopped using gentoo on my main desktop.. crap like that
14:36
<knipwim>
they do work on openrc
14:36
<johnny>
sure.. but it's been 2 years now
14:36
<knipwim>
i see bug deps being removed in bugzilla
14:36
<johnny>
that's good..
14:36
<knipwim>
i know, but there is hope :)
14:36
<johnny>
but would be better to just switch to upstart or somethin..
14:37
it stopped being worth it to compile software that wasn't bleeding edge and new
14:37
fedora has newer stuff
14:37
<knipwim>
or make the ebuild yourself
14:37
<johnny>
had to do that too many times
14:38jstewart_ has joined #ltsp
14:38jstewart_ is now known as fudoreaper
14:38pmatulis has quit IRC
14:38
<knipwim>
no virtual/wm though
14:39
<cliebow>
vagrantc:you going somewhere with the Jammman?
14:39
<knipwim>
so we could include the xinit[-minimal] dependency
14:39
<vagrantc>
cliebow: no
14:39
<knipwim>
or do nothing and put a msg on the wiki
14:39
<vagrantc>
i just prioritize good food on a daily basis :)
14:39
<knipwim>
or a message ibn the ebuild
14:39
<cliebow>
you said goood foood.. first thought was jammcq 8~
14:39
)
14:41
<fudoreaper>
i hate to be such a newb, but. I have a Ubuntu 9.10 LTSP server, and i want to know how to get a client to boot to it. i don't understand the etherboot/gPXE steps exactly
14:44
is it all about setting the correct DHCP options?
14:44
<kebax>
stgraber here?
14:45
<cliebow>
fudoreaper, yes yes
14:45
fudoreaper, but i am so rusty i have to leave it at that
14:46
<fudoreaper>
cliebow: i see.. unfortunately i currently have an MS DHCP server on all my networks...
14:46
<cliebow>
well that i can help with..i think..
14:47
<vagrantc>
stgraber: i'm a little pressed for time regarding nbd-proxy ... i can check tonight.
14:47
hopefully
14:48
<fudoreaper>
cliebow, any help would be appreciated
14:49
all the howtos i have found assume the LTSP server is the DHCP server as well, on an isolated network
14:49
<cliebow>
i had a bunch o stuff on the old ltsp wiki..ooking for it now
14:49
<fudoreaper>
i am not really in that situation
14:49
<johnny>
knipwim, putting a post install message would do it
14:49
or perhaps ltsp-build-client should warn?
14:50
<alkisg>
fudoreaper: there are 2 ways, one is to adjust your windows dhcp server, but if you don't succeed with that, there's also the possibility to use a proxydhcp on your linux server: https://help.ubuntu.com/community/UbuntuLTSP/ProxyDHCP
14:51
<knipwim>
johnny: how could it check for installed wms?
14:51
<johnny>
didn't say it should
14:51
<cliebow>
fudoreaper, i have to go..but this link has a little dhcp stuff..
14:52
<johnny>
but.. ltsp-build-client could warn , or the package post install could warn
14:52
that you need to install one
14:52
whether you have one or not
14:52
<cliebow>
basically option 17 and 67
14:52
<stgraber>
vagrantc: would be great, ideally, I'd release 5.2 tomorrow just in time for Ubuntu's feature freeze
14:53
<vagrantc>
stgraber: early or late in the day?
14:53
<stgraber>
alkisg: did you have the chance to do a quick test
14:53
vagrantc: late
14:53
<vagrantc>
stgraber: good, i might get some time to test in the afternoon.
14:53
<fudoreaper>
cliebow, thanks for your help
14:53
<vagrantc>
worst case is it doesn't work for debian, but it's non-default anyways.
14:53
and we fix it later.
14:54
<alkisg>
stgraber: I was waiting for upower to be available in the archives, because it breaks new fat client chroots, but I'll test now with an existing thin chroot
14:54
<vagrantc>
plain NBD works fine with debian, and gets most of the advantages.
14:54
<stgraber>
alkisg: great, thanks. Also if you can get the user who had it happen on real hardware to update and test with nbd-proxy, that'd be awesome.
14:55
<cliebow>
oohh..no longr even need rootpath..
14:55
<alkisg>
stgraber: hmm I'll do that, but not before FF, it'll take some time for him to try it.
14:55
cliebow: yup, that's for nfs. Some weird PXE clients need it though, to accept the lease, even if it isn't used later on.
14:55cliebow has quit IRC
14:58
<knipwim>
johnny: a message in ltsp-server seems better because it's package related
14:58
and it's server related
15:02
<alkisg>
stgraber: nbd-proxy from 5.2~bzr1668-0ubuntu1~ppa1~lucid1 works fine in my vbox. ;)
15:09leio_ has joined #ltsp
15:12leio has quit IRC
15:16rjune has quit IRC
15:17
<alincoln>
hey folks - i wrote a script called ltsp-chroot-tool that's intended to be part of ltsp-server, and i'm looking for feedback on features, style, bugs, etc. here's a link to the file in my personal branch: http://ur1.ca/miua
15:18
<johnny>
alincoln, please don't use a shortener unless you have to..
15:18
/me assumes you're not posting from twitter or sms to here
15:18
shorteners can die.. and then so does your link
15:18
<alincoln>
johnny: ok. http://bazaar.launchpad.net/%7Eadamjlincoln/%2Bjunk/my-ltsp-trunk/annotate/head%3A/server/ltsp-chroot-tool
15:19
<johnny>
alincoln, yeah.. you should add support for plugins to that
15:19
distro plugins i mean
15:20
<alkisg>
alincoln: why two parameters, update and don't update image?
15:20
<alincoln>
alkisg: if you don't specify either, it asks
15:20
johnny: i can do that, yep.
15:20
<johnny>
knipwim, check it out.. would be neat
15:20
<alincoln>
johnny: in fact, uh, it's silly if i don't do plugins, eh?
15:21
<johnny>
if your goal is to get it upstreamed. then i suppose so :)
15:21
<alincoln>
hehe
15:21
yes, you are right.
15:21
<johnny>
if your goal is ti get it included in just debian/ubuntu.. then not so much
15:21
alincoln, also.. vagrantc started some ltsp info tool you might want to share infra with
15:22
in fact.. we should probably do some braining in how ltsp-build-client, ltsp-chroot-tool, and that ltsp info tool fit together
15:22
see how much more can be shared
15:22
<alkisg>
alincoln: """if ! $(mount | grep -F "on ${CHROOT}/proc"); then""" => "on" won't work with localized mount - I think the current code checks for proc/cpuinfo to see if it's mounted
15:22
<johnny>
localized mount?
15:22
oh.. the mount command?
15:22
<alincoln>
alkisg: ah, ok. i stole that from somewhere... build-client plugin, maybe?
15:23
<johnny>
ah.. see.. we should share more!!
15:23
between these neat commands
15:23
make more of ltsp-build-client into a library
15:23litlebuda has joined #ltsp
15:23
<alkisg>
alincoln: also, I think it's not adviced to mount /dev/pts.... maybe make a parameter for that, with default=off?
15:24
<alincoln>
alkisg: dunno. build-client does it
15:24
but i'm open to understanding better why we might want to.
15:24
i thought it may be a dpkg thing
15:24
<alkisg>
MOUNT_PACKAGE_DIR => MOUNT_PACKAGE_CACHE
15:24
(dir is if the user provides a dir)
15:25
<alincoln>
alkisg: i plan to add user specified dir, so i used the same convention as build-client, where mount-package-cache uses MOUNT_PACKAGE_DIR but with a hardcoded default
15:25
<alkisg>
alincoln: and I think you can reuse some ltsp-common-functions, e.g. boolean_is_true...
15:25
<alincoln>
ah, yes, you're right
15:25
<johnny>
so.. a suggestion for ltsp 5.3
15:26
a more unified ltsp-build-client/ltsp-chroot-tool/ltsp-info
15:26
we should use launchpad blueprints
15:26
or somethin
15:26
<alincoln>
in general, would something like this be useful to people?
15:27frederickjh has left #ltsp
15:27
<johnny>
sure.. why not
15:27
alincoln, perhaps it could work with that easy ltsp gui thing.. or somethin
15:27
<alkisg>
alincoln: also check if you want: cp /etc/resolv.conf "$ROOT/etc/", and export LTSP_HANDLE_DAEMONS=false
15:27
<johnny>
alincoln, or.. perhaps.. is there any way we can use packagekit ?
15:27
not sure if packagekit handles chroots.. but that woul be ultrasweet
15:27
<alincoln>
johnny: yeah, i kind of had in mind a broader chroot management thing.
15:28
<johnny>
so we don't have to manually handle the abstraction
15:28
alincoln, perhaps we should try to get upstream to help us here
15:28
if packagekit does enough for us.. that would be great..
15:28
packagekit now supports all the distros ltsp supports i think..
15:29
<knipwim>
johnny: dpkg --print-architecture is not available on gentoo
15:29
also, the dpkg command prints out i386 while gentoo uses x86
15:29
<johnny>
knipwim, that's why it needs a plugin.. system
15:29
<knipwim>
also an issue in ltsp-update-image btw
15:29
<johnny>
?
15:29
yes
15:30
we need to spend some time unifying these things is exactly what i was saying
15:30
<alkisg>
Will gento be able to use nbd?
15:30
<johnny>
alkisg, why not?
15:30
<alkisg>
Just asking :)
15:30
<johnny>
it won't come default, since we can't rely on the server kernel including it
15:30
but it's certainly possible if we replace our initramfs crapola
15:30
which needs to be done anyways..
15:30
alkisg, i kept holding out for the developers to care..
15:31slidesinger has quit IRC
15:31
<knipwim>
haven;t got it working also ...
15:31
<johnny>
too long now tho..
15:31
knipwim, dracut?
15:31
well.. we can work with upstream
15:31
<knipwim>
dracut and nbd
15:31
haven't put to much time in both tough
15:31
not working out of the box anyway
15:31
<johnny>
just using dracut in general would be a big deal
15:32
and it would make gentoo chroot work out of the box for nfs..
15:32
<knipwim>
but on the unification, i would like a generic ltsp function to call and get my arch
15:32
<johnny>
we have that.. but only in the plugin
15:32
<knipwim>
depending on the distro i'm using
15:32
<johnny>
err ltsp-build-client plugin
15:32
knipwim, also.. ltsp-info .. that vagrantc wrote
15:32
in fact.. ltsp-chroot-tool, and ltsp-build-client should use that
15:33
<Lns>
stgraber: do you know about a possible bug in lucid daily/ldm-server that hangs during "Setting up ldm-server (2:2.0.54-0ubuntu1) ..." ?
15:33
<stgraber>
Lns: haven't seen that yet
15:33
<Lns>
stgraber: just ran into it
15:33
anything I can do to troubleshoot?
15:34
<alincoln>
Lns: i've seen that. you can ctrl-c, then dpkg will go back at the end and reconfigure it
15:34
and it works fine
15:34
some process runs away or something
15:34
<Lns>
alincoln: hrmmm
15:35
wondering if i can pinpoint anything for a bug report or something
15:35
<stgraber>
Lns: ps aux would help
15:36
<alkisg>
Lns: open another terminal
15:36
in that, write: sudo service openbsd-inetd restart
15:36
Then ldm-server.postinst will continue... ;)
15:37
<Lns>
stgraber: http://pastebin.com/m797e063d
15:37
alkisg: is ldm-server trying to contact the (failed to start) dhcp3-server?
15:37
err
15:37frederickjh has joined #ltsp
15:37
<Lns>
n/m ;)
15:38etyack has quit IRC
15:38
<stgraber>
ok, so it's update-inetd that's failing ...
15:39
<Lns>
alkisg: stgraber: what alkisg said above works
15:39
<alkisg>
I've seen it in all installations I've done in the last month...
15:40
Lns, also, tftpd-hpa might not start on boot. If you reboot and see it not started, ping me to tell you a workaround
15:41
<Lns>
alkisg: thx. I might have to go get a small switch and get a test env. going so i don't have to swap dhcp servers all the time..would make full testing much easier ;)
15:42
<stgraber>
alkisg: any idea why we need to restart it ? sounds like update-inetd doesn't do what it's supposed to do ...
15:42
<alkisg>
stgraber: no, I forgot to look into it... :(
15:43
<_UsUrPeR_>
hey all. Having problems getting 10.04 Alpha working with Etherboot...
15:43
<knipwim>
the ltsp-info looks good, but it can only be applied if all distros support the same functions
15:45
<_UsUrPeR_>
I used mknbi the same as what fixed 9.10, but it's not really booting this time. Just stops after ".............done"
15:46
<alkisg>
_UsUrPeR_: aren't those workstations flashable to gpxe?
15:47
<_UsUrPeR_>
alkisg: haha.
15:47
<alkisg>
...so I'm guessing "nah"? :D
15:47* _UsUrPeR_ flies to brazil to flash some dude's T-150E
15:48
<alkisg>
Would flashing require hardware? /me doesn't know about real thin clients...
15:50
<Lns>
crossover cable ftw!
15:51
<_UsUrPeR_>
I have never flashed one either, but it depends on the mobo revision of the thin client. Some are flashed, some have sockets for new eeprom chips.
15:51
but yeah, to flash a new chip, you'd need the hardware for it
15:53
<alkisg>
Ah ok, I thought it would be as easy as netbooting it with an e.g. award flasher...
15:55
<_UsUrPeR_>
alkisg: yeah, it's dependent on the model
15:55
waaaay aside though, any ideas about etherboot on 10.04? :P
15:56* alkisg saw that the 9.10 etherboot bug was solved in a few ...months, that's why he asked about flashing :D
15:57artista_frustrad has quit IRC
15:58
<_UsUrPeR_>
seriously though, we have a lot of clients in a lot of places, and flashing them separately isn't really an option :/
15:58
the last fix for 9.10 was simply installing mknbi on the client image, and updating ltsp's kernel list
16:06
<Lns>
alkisg: Ok, booting my client gives "TFTP Error - File Not found" - is this what you're talking about? ;)
16:08
I have "/usr/sbin/in.tftpd --listen --user tftp --address 0.0.0.0:69 --secure /var/lib/tftpboot" running
16:08leio_ has quit IRC
16:10leio_ has joined #ltsp
16:10artista_frustrad has joined #ltsp
16:11frederickjh has left #ltsp
16:16leio_ has joined #ltsp
16:20GodFather has quit IRC
16:20
<alkisg>
Lns: you have that running?
16:20
<Lns>
alkisg: yeah
16:20
<alkisg>
My problem was that it didn't run at boot...
16:21
The symptom is the same, tftp: file not found
16:21leio_ has quit IRC
16:21
<alkisg>
So you rebooted the server, and without manually starting tftpd-hpa, it was running? weird.. :(
16:22
<Lns>
alkisg: yeah.. but still no love to the client
16:22
oh jeez
16:22
i'm an idiot
16:22
i haven't ltsp-build-client yet ;)
16:22
<alkisg>
Heh
16:22* Lns smacks himself on head
16:22* alkisg curses race conditions :(
16:22leio_ has joined #ltsp
16:23
<alkisg>
Lns, do you want to try sch-scripts?
16:23
<Lns>
alkisg: i'd love to
16:23
<alkisg>
It builds thin clients from a gui :D
16:23
***thin chroots
16:23
<Lns>
oh
16:23
maybe i can test it to build some fat client chroots? =) even though i'm not going to want to run a fat client out of my koolu amd geode tc ;)
16:24
<alkisg>
Sure, but not today, the ubuntu archive is broken today because of upower
16:24
(a new package)
16:24
For fat clients, the package selection is done in .conf files
16:24
<Lns>
ikes
16:25
<alkisg>
OK if you want we may go over it tomorrow or some day after that (maybe some messages are greek only, so you'll need my help, and I'll need yours, if we're to translate them :D)
16:27
<fudoreaper>
i'm setting DHCP options to get my workstation to boot...
16:27
i've set option 17 and 67... but don't i need to specify the IP address of the boot server?
16:28
<Lns>
alkisg: no worries, that would be cool. =)
16:28
<alkisg>
67=boot filename, 17==? is that next-server?
16:28
fudoreaper: if 17=tftp server, then that can also be used as the boot server
16:29
<Lns>
fudoreaper: that's a windows dhcp server?
16:29
<fudoreaper>
yeah, MS DHCP server i'm working with
16:29* Lns has only seen numerical dhcp option references w/ms)
16:29
<Lns>
fudoreaper: try https://help.ubuntu.com/community/UbuntuLTSP/LTSPWindowsDHCP
16:29
<fudoreaper>
well they're ALL numerical, LNS
16:29
and have a name
16:29
in the IANA spec, i mean
16:29
but, okay
16:29
<Lns>
oh gotcha
16:30
<alkisg>
066 Boot Server Host Name: <ip address>
16:30
<Lns>
yikes..that page could use some cleanup
16:30
<alkisg>
Also it should say "install ltsp-server, not ltsp-server-standalone" instead of telling people to remove dhcp3-server
16:31
<fudoreaper>
okay, yes, set option 66
16:31
i'm trying to use gPXE to boot the client...
16:32
<alkisg>
With gpxe you don't even need a dhcp server at all :D
16:32
<vagrantc>
you and your crazy gpxe.
16:32
<fudoreaper>
ho ho, getting somewhere
16:33
i see all the right options returned
16:33
<alkisg>
Well, if BIOSes had gpxe, we could even have ssh keys on the TCs :D
16:33
<fudoreaper>
the IP addresses all seem right... trying to load tftp now...
16:33
alkisg, um... don't need a DHCP server at all?
16:33
i want to be able to netboot off a CD, at the moment
16:34
<alkisg>
fudoreaper: yeah, you can hardcode all the values in an embedded gpxe script
16:34
<fudoreaper>
auto-netbooting is something for the future
16:34* Lns wonders why the wiki page says you *have to* specify MACs in the MS dhcp server to boot them..i've never had a problem with simply putting them in the global scope...well, unless all your machines try to network boot by default =p
16:34
<alkisg>
And download that from rom-o-matic, in a CD image
16:34
<fudoreaper>
huh... i could try etherboot
16:34
<Lns>
isn't gpxe == etherboot?
16:34
<alkisg>
etherboot == very old gpxe, sucks
16:34
gpxe is waaaaaaaay better
16:34
<highvoltage>
vagrantc: what's so crazy about gpxe :p
16:35
<vagrantc>
highvoltage: nothing ... hoping it hits debian soon. :)
16:35
<highvoltage>
awesome
16:35
<alkisg>
vagrantc: I added a grub2 /etc/grub.d/40_gpxe script in your gpxe package
16:36
If you're still updating it, maybe you want to include it?
16:36
<vagrantc>
i put some work into it, and panthera was putting some work into it ... and otavio was maybe going to help too
16:36
<alkisg>
It adds a "boot from network" entry to grub2
16:36
<vagrantc>
alkisg: ooooh. nice. :)
16:36
<panthera>
and it's almost finished
16:36* vagrantc dances
16:36
<alkisg>
Yey!
16:36
<panthera>
eta is saturday evening
16:36
<alkisg>
Goi go panthera!
16:37
<fudoreaper>
now my TFTP server isn't working!
16:37
console@proxy:~$ atftp 10.1.100.13
16:37
tftp> get ltsp/i386/pxelinux.0
16:37
timeout: retrying...
16:37
:(
16:37
<panthera>
.oO(use tftp-hpa; aftp sucks)
16:37
<alkisg>
fudoreaper: ls -l /var/lib/tftpboot/ltsp/i386/pxelinux.0 ?
16:38
<fudoreaper>
onsole@bal-linux2:~$ ls -l /var/lib/tftpboot/ltsp/i386/pxelinux.0
16:38
-rw-r--r-- 1 root root 14776 2010-02-16 13:34 /var/lib/tftpboot/ltsp/i386/pxelinux.0
16:38
<alkisg>
fudoreaper: grep tftp /etc/inetd.conf ?
16:39
<vagrantc>
panthera: i've been meaning to grumble about tftpd-hpa changes... but haven't had the chance to do so properly. :)
16:39
<fudoreaper>
console@bal-linux2:~$ grep tftp /etc/inetd.conf
16:39
tftp dgram udp wait root /usr/sbin/in.tftpd /usr/sbin/in.tftpd -s /var/lib/tftpboot
16:39
<alkisg>
vagrantc: Ubuntu chose to keep using /var/lib/tftpboot for now
16:39
fudoreaper: sudo service openbsd-inetd restart
16:40
And try sudo apt-get install tftp-hpa / tftp 10.1.100.13 / bin / get /ltsp/i386/pxelinux.0 while sitting on the server.
16:40
<fudoreaper>
... bin, eh?
16:41
<alkisg>
I think there were some problems with the default netascii mode, but I don't remember details
16:41
<panthera>
fudoreaper: tftp-hpa in stable isn't a good package, the one from testing/unstable is much better
16:42leio_ is now known as leio
16:42
<panthera>
fudoreaper: and runs standalone and asks through debconf any optional stuff, so the setup is just 'apt-get install tftpd-hpa' and it works out-of-the-box.
16:42
much nicer user experience ;)
16:43
<fudoreaper>
console@bal-linux2:~$ tftp 10.1.100.13
16:43
tftp> bin
16:43
tftp> get ltsp/i386/pxelinux.0
16:43
Transfer timed out.
16:43
so that's what's happening on the LTSP server now
16:43
<alkisg>
netstat -nap|grep :69
16:43CAN-o-SPAM has quit IRC
16:44
<Lns>
woohoo! boots to ldm =) i just need to get a small usb keyboard, it doesn't like my ps2->usb converter ( -> kvm switch) ;)
16:45
<fudoreaper>
alkisg, nothing... but i was sure i saw something...
16:45
<alkisg>
fudoreaper: see the logs for any related messages, e.g. sudo grep tftp /var/log/*
16:49
<fudoreaper>
alkisg, okay, got it
16:49
tftpd-hpa is disabled by default in /etc/default
16:49
i was confused and saw tftpd running on a different server
16:49
<alkisg>
fudoreaper: that's because it starts from inetd
16:49
<fudoreaper>
ha
16:49
<alkisg>
Don't enable it from there...
16:54
<fudoreaper>
alkisg, oh right...
16:54
<alkisg>
Try again looking at inetd.conf and restarting openbsd-inetd and looking at the logs
16:57leio has quit IRC
16:57
<fudoreaper>
ah... it's called openbsd-inetd
16:57
that's the trick
16:58
k, think i got it now...
16:58leio has joined #ltsp
16:59
<fudoreaper>
oh man it's booting!
17:00* alkisg remembers the joy of the first network boot... :D
17:00
<fudoreaper>
yeah!
17:00
i've got a login pompt
17:00
alkisg, thanks for all your advice here
17:00
<alkisg>
np
17:01
<fudoreaper>
this is excellent
17:01
i'm at a school now
17:01
all MS environment
17:01
heh
17:01
TILL NOW!
17:04* alkisg calls it a day. Goodnight all...
17:05
<fudoreaper>
awwwww
17:05
i can't login!
17:05
it reboots X on me
17:05
<alkisg>
if you changed the ip, you need to run ltsp-update-sshkeys and ltsp-update-image
17:05alkisg has quit IRC
17:06
<Lns>
fudoreaper: congrats! Oh, I remember when my first client booted...I was so ecstatic =)
17:07
<fudoreaper>
this is amazing
17:07
yeah
17:07
but i can't login
17:07
Feb 16 17:00:27 bal-linux2 sshd[22827]: Accepted password for bhs from 10.1.100.126 port 47396 ssh2
17:07
Feb 16 17:00:27 bal-linux2 sshd[22827]: pam_unix(sshd:session): session opened for user bhs by (uid=0)
17:07
Feb 16 17:00:29 bal-linux2 sshd[22891]: subsystem request for sftp
17:07
Feb 16 17:00:30 bal-linux2 sshd[22827]: pam_unix(sshd:session): session closed for user bhs
17:07
i see that in the auth.log
17:08
so it seems i've got the right password...
17:08
<Lns>
fudoreaper: did you ltsp-update-sshkeys and ltsp-update-image?
17:08
<fudoreaper>
i'm doing that right now
17:08
the update-sshkeys did nothing
17:08
<Lns>
fudoreaper: it doesn't normally give output
17:08
<fudoreaper>
i installed ltsp packages today...
17:09
but i'm redoing the i386 arch now
17:13
i saw a message this time 'no reponse from server, restarting'
17:13
it's like X reboots...
17:14
actually it says no response when i put in the wrong pass
17:14
on right pass, it just restarts X
17:17
<johnny>
why does it say subsystem request for sftp ..
17:17
is that normal ?
17:17
<fudoreaper>
johnny: i dunno
17:17
<johnny>
doesn't seem normal..
17:18
<fudoreaper>
some kind of encrypted FS mount?
17:18
i tried to login on console1, like non-gui
17:18
didn't work, though
17:18
and that doesn't seem to hit the server...
17:18
<Lns>
johnny: to mount the user's home?
17:19
fudoreaper: tty1 on thin client is the client chroot. you have to set up a user in the chroot for that to work
17:19
<fudoreaper>
Lns: okay, that makes sense
17:19
same thing with the X gui?
17:20
no, must not be, SSH sees and accepts the user/pass...
17:20
<johnny>
Lns, ah.. if you have local apps..
17:20
<Lns>
fudoreaper: your login gui is LDM which is run on the client, and authenticates to the server via ssh to get your user session (on the server)
17:21
johnny: and possibly localdev stuff too maybe..?
17:22
<fudoreaper>
Lns, okay, that makes sense
17:22
why isn't it working?
17:22
:P
17:22
i just created a user with 'sudo adduser' on the server (not chrooted)
17:23
<Lns>
is it adduser or useradd that doesn't create a homedir...?
17:23
fudoreaper: check to make sure you have a homedir for your user
17:23
<fudoreaper>
yeah, i have a /home/bhs
17:23
adduser is the wizard
17:24
useradd makes you specify that stuff
17:24
<Lns>
fudoreaper: after you ltsp-update-image you rebooted the client right?
17:24
<fudoreaper>
Xsession: X session started for bhs at Tue Feb 16 17:13:11 CST 2010
17:24
Xsession: warning: xrdb command not found; X resources not merged.
17:24
Xsession: unable to start X session --- no "/home/bhs/.xsession" file, no
17:24
"/home/bhs/.Xsession" file, no session managers, no window managers, and no
17:24
terminal emulators found; aborting.
17:25
yes, i rebooted the client
17:25
actually, it TOLD me that
17:25
and rebooted
17:25
heh
17:25
this is a snippet from /home/bhs/.xsession-errors
17:25
<johnny>
what OS is this?
17:25
<fudoreaper>
it is true, there is no .xsession file, or .Xsession
17:26
johnny: Ubuntu 9.10
17:26
<johnny>
do you not have gnome/kde/something installed?
17:26
<fudoreaper>
johnny: uh... possibly?
17:26
this is a server, so
17:26
i haven't installed any gui stuff
17:26
<Lns>
fudoreaper: you need to have a desktop environment installed on the server weather you use it *on* the server or not
17:26
<fudoreaper>
just LTSP packages
17:26
hahah
17:26
okay
17:26
<Lns>
sudo apt-get install ubuntu-desktop
17:27
<fudoreaper>
so, apt-get install
17:27
okay
17:27
<Lns>
;)
17:27* Lns sees fudoreaper's brain getting bigger by the minute
17:29
<fudoreaper>
ha ha
17:29
well that's pretty funny
17:30
i'm installing it now
17:30
will i have to copy a .xauthority file, or will that be automatic?
17:30
and, the server is AMD64... will that be a problem?
17:30
<Lns>
fudoreaper: you should be golden after the install
17:30
<fudoreaper>
i guess not, cause it's executing server side
17:30
<Lns>
nah, maybe just with flash
17:30
<fudoreaper>
oh flash
17:31
i hate it
17:31
but even our school website depends on it
17:31
:(
17:31
<Lns>
yeah..join the club ;)
17:34
flash should work fine really with the newer versions..just still a complete resource hog
17:36
<fudoreaper>
right
17:42
<Lns>
the world needs an OSS answer to flash/silverlight..and not gnash (or anything else that tries to *be* flash)
17:48vagrantc has quit IRC
17:48
<johnny>
Lns, well. silverlight is open ..
17:48
mostly anyways
17:49
Lns, the part that won't be open.. is the drm
17:49
which is not something OSS can truly provide an answer to..
17:49
<Lns>
johnny: is it really?
17:49
<johnny>
thus the problem isn't in the software, but content providers
17:49
Lns, sure.. the spec is open, and there is a free implementation via moonlight
17:50
<Lns>
is it open source?
17:50
<johnny>
Lns, but really.. we're moving in that direction via html5 and canvas
17:50
moonlight is
17:50
<Lns>
but we're still chasing someone's tail as far as updates/upgrades right...?
17:50Faithful has joined #ltsp
17:50* Lns googles html5 canvas
17:50
<johnny>
not as much as flash is
17:51
Lns, also.. theora/vorbis for audio/video is already feasible in newer browsers
17:51
the real problem is google using x264
17:51
for html5 video for youtube
17:51
<Lns>
right but i meant besides the video tag
17:51
that's a whole other deal it seems..
17:51
<johnny>
sure.. canvas will let you do stuff like you can do in flash
17:51
<Lns>
at least as far as html5 is concerned
17:51
<johnny>
altho the apis aren't as well developed
17:51
and it's still a bit too slow
17:51
<Lns>
gotta start somewhere!
17:51
<johnny>
but people have written games in them
17:52
i even saw a partial wolfenstein3d clone
17:52
Lns, part of the problem is the spec might be too open..
17:52vvinet has quit IRC
17:52
<johnny>
thus opened up for too many people to stop soemthing from being good enough from getting out..
17:52
and the perfect never comes
17:52
thus the chasing..
17:53bobby_C has joined #ltsp
17:53
<Lns>
perfect never comes, period
17:53
we know this from flash ;)
17:53
<johnny>
the real problem is social imo
17:53
<Lns>
well that's much better than technical problems..at least we can solve that in an open manner
17:53
<johnny>
and.. we've often been hampered by stuff sticking around for too long..
17:54
because of peope using old browsers and other software in their quest for "stability"
17:54
so.. you're part of the problem too :)
17:55
<Lns>
so sue me for not wanting to reboot a linux server daily ;)
17:55
<johnny>
Lns, but now that more apps arei n the browser
17:55
people like you can't stop the evolution
17:55
only thing stopping that is the browser version itself :)
17:56
<Lns>
i'm far from wanting to stop technological evolution johnny.. but in the context of having a fruitful technology platform for thousands of students, I will go for stability over bleeding edge (and unstable) software. I'll wait to catch up with the newest stuff when everyone else has hammered the crap out of it and fixed all the bugs. :)
17:57
in the meantime i'm happy to TEST the latest and greatest and report bugs/etc
17:57
<johnny>
that seems not to be working too well for ubuntu
17:57
imo at least..
17:57Sarten-X has quit IRC
17:57
<Lns>
why do you say that?
17:57
<johnny>
bugs stick around too long that affect tons of people
17:58
big regressions, etc
17:58
that don't get fixed during the cycle
17:58
because X or Y component can't be changed
17:58
because it's "stable"
17:58Sarten-X has joined #ltsp
17:59
<johnny>
anyways.. it'll be a big longer before it's safe to stay more up to date tho
17:59
so.. it's not like we can always do latest and greatest now ..
17:59
i guess i see the problem as finding the right balance
17:59
and the stack itself still has growing pains causing some of those issues
17:59
<Lns>
imho, stability, like security, is a process, not a product
17:59
<johnny>
so it's not like i'm blaming ubuntu..
18:00
completely anyways
18:00
it's just a good example of the right balance not being found yet
18:00
<Lns>
there have been a lot of growing pains in the FOSS world in the past few years, FF 3.x, LTSP5, etc...that affect us all
18:01
<johnny>
thos are minimal imo
18:01
<Lns>
=p
18:01
<johnny>
a big example i can think of is when hal stuff started happening
18:01
or that intel gfx card issue
18:01
<fudoreaper>
oh Lns, things are working great now
18:01
amazingly fast
18:01
i'm impressed, actually
18:01
i can't even tell it's not local
18:02
<kebax>
well I can
18:03
although I'm runnin pixie inside virtualöbox wlan away from the ltsp
18:03
<Lns>
fudoreaper: =) That's the beauty of ltsp =)
18:03
johnny: agreed
18:03
kebax: haha! nice.
18:04
<kebax>
=)
18:04
<johnny>
Lns, and then in other areas.. we have things holding us back
18:04
like lack of free drivers with 3d..
18:04
<Lns>
well that's a hardware manufacturer holdback
18:04
<johnny>
at least nouveau is going to have some 3d in fedora13
18:05
<Lns>
seems that wherever there is profit to be made from software alone,there are problems in the OSS world
18:05
we don't think in that context (for the most part)
18:05
<fudoreaper>
Lns, i have some problem
18:05
ltsp-client wants to install
18:05
but is broken..
18:05
The following packages have unmet dependencies:
18:05
ltsp-client: Depends: ltsp-client-core but it is not going to be installed
18:05
E: Unmet dependencies. Try 'apt-get -f install' with no packages (or specify a solution).
18:06
sound works
18:06
video playback works
18:06
incredible
18:06
do i need ltsp-client installed, or...
18:06
<johnny>
you're not trying to install that on the server areyou?
18:06
<Lns>
fudoreaper: that should only be installed in the chroot i believe
18:06
<fudoreaper>
it says i should not install it on a regular machine...
18:06
<johnny>
truly
18:06
so don't do it
18:06
<fudoreaper>
johnny, i'm not trying... it seemed to come automatically
18:06
as i was installing gstreamer codecs with aptitude
18:07
<Lns>
you borked it! =p
18:07
<johnny>
weird
18:08
you don't want ltsp-client-* anything on the server
18:08
<fudoreaper>
okay
18:08
purged
18:08
no problems
18:08
man, 50 mbit of traffic for streaming vid
18:09
<johnny>
that's why some people run local apps
18:09
that and server cpu usage..
18:25Gadi_eeepc has joined #ltsp
18:27sene has quit IRC
18:28sene has joined #ltsp
18:33Lns has quit IRC
18:38bobby_C has quit IRC
19:07
<kebax>
ltsp-localapps /usr/bin/xterm is pretty interesting
19:08sene has quit IRC
19:14
<kebax>
=)22:10 < pscheie> kebax, the Revolution Linux folks have taken netbooks, put a kernel, small X, and nxclient on them which then connect to the ltsp server, achieving a similar effect.
19:14
22:11 < pscheie> over wireless
19:15
er, wrong window, sorry =)
19:23mgariepy has joined #ltsp
19:24
<mgariepy>
good evening everyone
19:26
<stgraber>
hey mgariepy
19:26
long time no see ! :)
19:28
<mgariepy>
yeah it's been an hour and a half :P
19:34
<fudoreaper>
johnny
19:34
local apps? you can do that?
19:34
like have a hybrid?
19:34
<johnny>
yes
19:34
you install the apps in the chroot
19:34
<fudoreaper>
i was just playing around, right?
19:34
<johnny>
but you need to have a decently powerful lclient..
19:34
<fudoreaper>
on different lab machines
19:35
flash was *rough*
19:35
i should say: painfully slow
19:35
<johnny>
flash is one of the biggest reasons
19:35litlebuda has quit IRC
19:35
<johnny>
but you should probably have at least 256mb ram in the client
19:35
<fudoreaper>
everything else we tried performed fine
19:35
<johnny>
before trying
19:35
yes.. flash sucks on linux.. and even worse over ltsp
19:35
<fudoreaper>
even full-screen video
19:35
<johnny>
blame adobe..
19:35
<fudoreaper>
oh, i do, man, i do
19:35
blame adobe
19:36
but yeah, some of the clients
19:36
<johnny>
so.. yes.. it's common to install firefox+flash locally
19:36
if you have enough ram to make it usable
19:36
<fudoreaper>
live 2GHz 1.5 GB
19:36
<johnny>
sure..
19:36
<fudoreaper>
so you would install the software on the chroot
19:36
<johnny>
jusyes
19:36
yes*
19:36
<fudoreaper>
and then it would be loaded over... whatever network I/O i have going on
19:36
<johnny>
and then there are some lts.conf options you set to make it use the local ones vs remotes
19:36
and it dynamically rewrites the .desktop files
19:37
<fudoreaper>
even flash is running in a 32 bit wrapper, on the server
19:37
man flash sucks
19:37
but yes, this LSTP is pretty amazing
19:37
once you figure it out
19:37
:P
19:37
<johnny>
so.. yeah.. you might wanna do the local apps thing
19:37
<fudoreaper>
but you guys have been a huge help
19:38
<johnny>
and you can also do fat clients..
19:38
with some apps remote too
19:38
but that's not in a released version on a distro yet
19:38
<fudoreaper>
hey is all the server-client communication in an SSH tunnel?
19:38
<johnny>
yes
19:38
<fudoreaper>
wow
19:38
<johnny>
you can set LDM_DIRECTX=Y
19:38
<fudoreaper>
isn't that CPU-expensive?
19:38
<johnny>
if you wanna try it without
19:38
sure.. but it's for security..
19:38
<fudoreaper>
well, heh
19:38
right
19:39
<johnny>
so.. if you don't need it.. yo ucan get better performance with LDM_DIRECTX=Y
19:39
<fudoreaper>
yeah, essentially i don't think we have anyone listening on the network
19:39
<johnny>
well give it a try
19:39
<fudoreaper>
i will
19:39
but man, it's 19:30 here
19:39
i officially get off at 16:00
19:40
hey also
19:40
i found that the fancy video effects resulted in a lot of bad video rendering
19:40
even to the point that some things don't appear
19:41
<johnny>
well try with DIRECTX and whatnot
19:41
and disabling destkop effects
19:41
<fudoreaper>
right, disabling worked great
19:41
but how could i make that the default?
19:42
oh and my plan is to try to link the server to the active directory domain
19:42
so that the AD users are now LTSP users
19:43
<johnny>
it's possible.. but i don't know anything about it
19:43
<fudoreaper>
yeah, i've been playing with it a bit
19:43
i feel confident it's possible
19:44
<Gadi_eeepc>
fudoreaper: you may find this to be a good reference: https://help.ubuntu.com/community/UbuntuLTSP
19:45
it has links like: https://help.ubuntu.com/community/UbuntuLTSP/ActiveDirectoryIntegration
19:45
<fudoreaper>
Gadi_eeepc, i've seen it.
19:45
ooooh
19:45
good
19:45
i will look tomorrow
19:45
i didn't notice that link
20:02wuffi600 has joined #ltsp
20:02
<wuffi600>
hello.
20:06
what ways are there to run ltsp performance clustered? (I've read http://openmosix.sourceforge.net/ltsp-omr4-1.html , are there other ways?)
20:09
<johnny>
!ltsp-cluster
20:09
<ltspbot>
johnny: Error: "ltsp-cluster" is not a valid command.
20:09
<johnny>
err guess not
20:09
there's some ltsp cluster thing , but it is currently only packaged for ubuntu
20:09
it's on the ubuntu wiki
20:09
i don't know much about it tho
20:12
<wuffi600>
I guess the graphical output show on a client-maschine is transport via X, right? Are there ways to replace it by something else (with the ability to reconnect, lower bandwidth, ...), e.g. neatx, nx, ...
20:26
<stgraber>
https://www.ltsp-cluster.org
20:26
we should make an alias
20:26
sbalneav: ^ !ltsp-cluster should point to https://www.ltsp-cluster.org
20:28
<johnny>
why it is a seperate page stgraber ?
20:29
shouldn't it be all itnegrated into ltsp.org ?
20:30
<stgraber>
johnny: we got a funding to work on a project website for ltsp-cluster. For now it's separate but we'd love to use that as a base for ltsp.org at some point and merge the two.
20:36dro has joined #ltsp
21:09dro has quit IRC
21:27dmarkey_ has joined #ltsp
21:27alincoln_ has joined #ltsp
21:28dmarkey has quit IRC
21:28alincoln has quit IRC
21:29wuffi600 has quit IRC
21:30vvinet has joined #ltsp
22:03vagrantc has joined #ltsp
22:21
<vagrantc>
stgraber: on second thought, i'm totally fine with incrementing ltspfs and ldm's ... i.e. ltspfs 0.6.x and ldm 2.1.x
22:22
we just did some fairly significant changes to ltspfs ... and ldm's certainly seen some significant changes since 2.0.0 :)
22:23
<stgraber>
great, that'll make a good reference point then
22:23
<vagrantc>
it'll be easier to see strange things out of sync. that's a good enough reason.
22:24
while i'd prefer to maintain compatibility, and not have hard versioned dependencies... it's sometimes good to see that everything is "in sync" in case we missed something.
22:36
initial testing... no sound :(
22:37
though doesn't seem like ltsp's fault ... no sound modules.
22:38
<mgariepy>
gnite everyone
22:38mgariepy has quit IRC
22:48johnny has left #ltsp
22:56
<vagrantc>
hrm.
22:57
the changes to ltsp-update-image make it harder to switch between NFS and ltsp_nbd
23:00
i suppose there's commandline flags to avoid that ...
23:01johnny has joined #ltsp
23:07
<vagrantc>
well, cdrom mount is getting treated as a usbdevice...
23:08vvinet has quit IRC
23:08vvinet has joined #ltsp
23:09
<vagrantc>
nevermind. sound works fine.
23:11gentgeen__ has quit IRC
23:12japerry has joined #ltsp
23:13
<vagrantc>
ltspfs is a little wonky ...
23:14
but no idea if it's any worse than it was.
23:14
ldm seems to be working...
23:19
stgraber: nbd-proxy is working!
23:23
other than ltspfs seeming a little wonky, everything's working ok. which is no regression in debian/testing...
23:23* vagrantc tries the backports features...
23:23japerry has quit IRC
23:29Faithful has quit IRC
23:35johnny has left #ltsp
23:38cyberorg has quit IRC
23:38johnny has joined #ltsp
23:45cyberorg has joined #ltsp
23:56alkisg has joined #ltsp