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


Channel log from 15 January 2009   (all times are UTC)

00:09alkisg has joined #ltsp
00:12try2free has joined #ltsp
00:21
<kc8pxy>
johnny: openrc 0.4.1-r1 has the following blockers.
00:21
[blocks B ] <sys-fs/udev-133 ("<sys-fs/udev-133" is blocking sys-apps/openrc-0.4.1-r1)
00:21
[blocks B ] <sys-apps/sysvinit-2.86-r11 ("<sys-apps/sysvinit-2.86-r11" is blocking sys-apps/openrc-0.4.1-r1)
00:21
<johnny>
go for it
00:21
install em
00:21
<kc8pxy>
and johnny -c them and try it?
00:21
<johnny>
-c them?
00:22
add em to your keywords in the chroot
00:22
and install it all
00:22hanthana has quit IRC
00:23
<kc8pxy>
since when in package.keywords, a directory?
00:23
2 files, ltsp and temp
00:28
johnny: not sure you understood what i said.. they are installed, and blocking openrc-0.4.1-r1
00:29
adding them in the only sane way i saw did nothign to the emerge -vp.. if i remove the -vp, it will just die on me.
00:31
<johnny>
since when what?
00:31
could it work as a directory? since years
00:32
you need to keyword them.. get a newer version?
00:32
<kc8pxy>
if i try and remove the blockers(at least udev), i get the 10 second, "do you REALIZE this will hose your system?!?" emerge countdown, which i halted.
00:32
<johnny>
don't remove them
00:32
upgrade them
00:32
i never said anything about removing them.. ever
00:33hanthana has joined #ltsp
00:33* kc8pxy must have an archaic view of how to fix blockers.
00:34try2free has quit IRC
00:36
<kc8pxy>
johnny: do we not need to specify ~arch anymore? it assumes that you are adding ~arch if it's in keywords?
00:37
baselayout 2.0? huge changes?
00:37
yeah, i know I'm probably pestering you.
00:39
<johnny>
it's not baselayout 2 by default?
00:39
did i screw that up?
00:40
hmm.. i was pretty sure that my setup would take care of getting baselayout
00:40
i was told that the default portage had the blocker dependency resolution that should take care of making it install
00:40
the stable portage
00:40
can you tell me which version of portage is stable in there?
01:01
<cyberorg>
how do we benchmark network bandwidht usage? i want to compare LDM_DIRECTX, normal LDM, xrdp, nx-agent etc
01:05F-GT has quit IRC
01:07F-GT has joined #ltsp
01:10
<kc8pxy>
johnny: openrc pulled in baselayout 2.0
01:11
johnny: portage 2.1.6.4 is installed
01:15
<johnny>
hmm.. i thought that one was new enough.. i was forcing 2.2 until i found that the changes had been backported
01:15
what other versions are there kc8pxy ?
01:32
<kc8pxy>
installing eix in the chroot to make this simpler
01:35* kc8pxy is watching a vbox boot ltsp
01:36
<kc8pxy>
johnny: i REALY like this poenrc.
01:36
openrc
01:37
unlike the standard init, i got only 2 warnings, and the WERE warnings, not errors, before i hit runlevel 3 :)
01:38
/var/lib/xkb doesn't exist, and /var/cache/localapps doesn't exist.
01:38
<johnny>
yeah.. i need to do something about those ..
01:39
<kc8pxy>
nice clean, sophisticated boot though, otherwise.
01:39* kc8pxy still thinks it would be cool to have a nice bootsplash :)
01:39
<johnny>
kc8pxy, try to generate one with genkernel.. see if it works.. we can change the flags..
01:40
as soon as i know what works.. i just didn't have time to investigate
01:40
<kc8pxy>
kk
01:40
np
01:40
<johnny>
and there are a few too many bootsplash situations going on right now..
01:40
<kc8pxy>
working is more important than pretty :)
01:40
ltsp client barfed.
01:41
warnings, setrlimit(RLIMIT_NICE, (31,31)): Operation not permitted
01:42
same barf on RLIMIT+RTPRIO, (9,9)
01:42
those are in main.c
01:43
module-detect.c barfed that it didn't find any soundcard(i have sound disabled in vbox).
01:44
<johnny>
kc8pxy, that's pulseaudio..
01:44
those aren't errors..
01:44
hopefully they will disappear from pulseaudio
01:44
i hope it is fixed in newer version :)
01:44
<kc8pxy>
3 earnings, 3 errors.
01:45
<johnny>
and then?
01:46
<kc8pxy>
after ltsp-clients, throes, network filesystems worked fine, syslog.ng and local, all fine.
01:47
but the ltps client finally bafed, saying it couldn't initialize daemon
01:47
dropped me to a CLI login
01:48cyberorg_ has joined #ltsp
01:49
<kc8pxy>
johnny: i need to crash soon.. recommendations?
01:49
<johnny>
ldm should start
01:49
it is possible you don't have ldm..
01:52ltspbot has joined #ltsp
01:52cyberorg_ is now known as cyberorg
01:52
<johnny>
kc8pxy, is it using the overlay in the chroot?
01:53
<kc8pxy>
johnny: i know i said some of this earlier, but i really appreciate the work you put into gentoo's ltsp implimentation, and putting up with my gripping to help me out :)
01:54ltspbot` has quit IRC
01:54
<kc8pxy>
johnny: it barfed on that path not being there. you mounted /usr/portage, and the new layman tree home is usr/local/portage,, right?
01:57
failed to source '/usr/local/portage/layman/make.conf'
01:57
<johnny>
that's why you should use that script i posted for you
01:57
it'll do all that stuff for ya
01:58
and then unmount it when you exit
01:58
that way.. if you rm -rf /opt/ltsp/x86 afterwords.. you won't destroy your portage directories and overlays
02:00
<kc8pxy>
johnny: i AM using your script
02:00
at least to fix things.
02:03
<johnny>
it should have mounted that for ya
02:03
<kc8pxy>
johnny: yes, should have.
02:03
<johnny>
altho that was just a quick rewrite
02:03
<kc8pxy>
ahh
02:03
<johnny>
of what is actually working
02:03
on my broken computer
02:03
im going to fix it soon
02:03
then i can be more help
02:04
altho.. i still have to update it to the new stuff.. which will suck
02:04
it'll have to use the backup on the other hard drive
02:04
<kc8pxy>
doh.
02:04
one look and i found a goof.
02:04
in yoru cript
02:05
you have it mounting /usr/local/portage/layman to $LTSP_CHROOT/usr/local/portage
02:06
i fixed that, and it should work :)
02:08
how do i manually fire ldm?
02:08
via xdm?
02:10generalsnus has joined #ltsp
02:11
<kc8pxy>
johnny: ideas?
02:11Chidder has joined #ltsp
02:12
<johnny>
no
02:12
so.. do you have ldm or not?
02:12
<generalsnus>
hi, how can i see what clients are currently connected to the thin client server? like a list of IP's
02:14
<kc8pxy>
johnny: yeah, tried to fire it manually, and it barfed on gtk+ then i hunt, and see that xorg is not installed.
02:15
johnny: mebbe it's just me, but doesn't a regular gui on a modern gentoo box need zorg-server?
02:15
xorg-server?
02:15
generalsnus: netstat on the server?
02:15
<johnny>
kc8pxy, yes.. alot of stuff didn't seem to happen
02:16
it died after the first try of extra_packages.. due to the missing blocker resolution
02:16
can you tell me the portage versions eix says you have?
02:18cyberorg has quit IRC
02:29ltspbot has joined #ltsp
02:29
<johnny>
kc8pxy, you might need to unmask it tho
02:29
unmask portage that is
02:30alkisg has quit IRC
02:32
<kc8pxy>
johnny: where do i do that in teh script?
02:33
johnny: nuke /opt/ltsp/x86 ?
02:34
night
02:36
<johnny>
i just told you
02:36
see if it is masked too
02:36
and make a file file in the similiar style
02:49Barbosa_ has joined #ltsp
02:52
<Appiah>
I'm using X_VERTREFRSH and X_HORZSYNC and it seams to get ignored in my lts.conf
02:53
anywhere I can check on the client that it really got these settings?
02:59kc8pxy has quit IRC
02:59tarbo has quit IRC
02:59v-pro has quit IRC
02:59ltsplogbot has quit IRC
02:59jcastro has quit IRC
02:59kc8pxy has joined #ltsp
02:59jcastro has joined #ltsp
02:59v-pro has joined #ltsp
03:00tarbo has joined #ltsp
03:00ltsplogbot has joined #ltsp
03:00Barbosa has quit IRC
03:05tarbo has quit IRC
03:05evilx has quit IRC
03:05tarbo has joined #ltsp
03:07
<polytan>
hi
03:10
kc8pxy, I'm also trying to get ltsp working on gentoo system
03:11
currently (client built yesterday with the last version of all programs) I can boot the client without any problem
03:11
I can specify a lts.conf
03:11tarbo has quit IRC
03:11* ogra waits for the but :)
03:11
<polytan>
but the options (keyboard, etc) in lts.conf aren't used :(
03:11
and ldm doesn't start
03:11tarbo has joined #ltsp
03:12
<johnny>
which keyboard?
03:12
polytan, i never had anything to test with
03:13
polytan, none of the X options are implemented afaik
03:13* ogra envys johnny for being able to use voise input instead of keyboards :)
03:13
<ogra>
*voice
03:13
<polytan>
I live in France, so my computers have a french jeyboards
03:13
<johnny>
polytan, sure.. what are you trying to set?
03:13
<polytan>
and doing "loadkeys fr" all the time is quite boring
03:13
<johnny>
what variable are you using in lts.conf ?
03:14
<polytan>
so I put that in the lts.conf (which my client are able to retrieve using tftf (I have a complete lts.conf file in /etc/lts.conf of the client))
03:14
<Appiah>
i put XSERVER = via and XSERVER = vesa in my lts.conf but it ends up using openchrome driver anyways
03:14
<johnny>
polytan, what are you putting in lts.conf?
03:15
<ogra>
XKBLAYOUT="fr" should be the one
03:16
<polytan>
http://dpaste.com/109278/
03:17
I also tried changing "azerty" by "fr" (right now actually)
03:17tarbo has quit IRC
03:18
<polytan>
and a manual "loadkeys fr" works
03:18tarbo has joined #ltsp
03:22
<ogra>
well, ldm calls setxkbmap internally
03:22
<johnny>
polytan,
03:22
configure_console() {
03:22
if [ -n "$CONSOLE_KEYMAP" ]; then
03:22
sed -i 's/keymap="us"/keymap="${CONSOLE_KEYMAP}"/' /etc/conf.d/keymaps
03:22
fi
03:22
}
03:22
that's what should happen..
03:23
is keymaps started?
03:23
<polytan>
it is started
03:24
but I have " KEYMAP = "us" " in /etc/conf.d/keymap
03:24tarbo has quit IRC
03:24
<polytan>
uit look like the lts.conf isn't read
03:24tarbo has joined #ltsp
03:24
<polytan>
johnny, did you see the bug https://bugs.launchpad.net/ubuntu/+source/ltsp/+bug/317286 ?
03:26
<ogra>
if you have KEYMAP = "us" the sed command wont work ... there are spaces and the case is wrong
03:26cyberorg has quit IRC
03:26
<polytan>
it is written in the file
03:26
KEYMAP="us"
03:27
is sed case sensitive ?
03:27
<johnny>
hmm.. i'm wondering if you have the wrong version of openrc?
03:28
in the version you have, it should be keymap not KEYMAP
03:28
ogra, what's this about the case being wrong?
03:29
<ogra>
sed -i 's/keymap="us"/keymap="${CONSOLE_KEYMAP}"/' wont match KEYMAP
03:29
<johnny>
KEYMAP is wrong
03:29
keymap is right
03:29
<ogra>
ah
03:29
<johnny>
polytan, it might be that you need to run etc-update in the chroot
03:29
and merge in all the openrc files
03:29* johnny is really really hoping it gets into stable portage soon
03:30
<polytan>
johnny, pfff
03:30
<Appiah>
any ideas on why my lts.conf X_MODE and X_HORZSTBC settings dont seam to affect my clients?
03:30
<polytan>
what happens with the new way of building clients ?
03:30
<johnny>
huh?
03:30
<polytan>
in the client I built yesterday
03:30
<ogra>
Appiah, release distro ?
03:30tarbo has quit IRC
03:30
<polytan>
I have a baselayout 1 and no openrc
03:31tarbo has joined #ltsp
03:31
<polytan>
it was working before critsmass, wasn't it ?
03:31
<johnny>
you shouldn't have
03:31
baselayout 2 is the only way i personally was able to get it to work :)
03:31
<polytan>
yes
03:31
So I am
03:32
it look like there wasn't emerge -uvDNa world done
03:32
<johnny>
yes.. that's part of the problem i had with kc8pxy
03:32
<polytan>
because I have baselayout 2 unmasked and openrc unmask
03:32
<Appiah>
ogra: Ubuntu 8.10
03:32
<johnny>
what you don't have .. is a version of portage that can resolve the blockers
03:32
<ogra>
Appiah, did you specify CONFIGURE_X=True ?
03:32
<Appiah>
...
03:32
no
03:32
<ogra>
Appiah, thats needed in 8.10, it defaults to not create an xorg.conf
03:33
<johnny>
polytan, i was told that the current stable version of portage had the dependency resolution, but it seems that it sitll only works in 2.2
03:33* Appiah facepalms
03:33
<Appiah>
thanks
03:33
<ogra>
since xorg moves away from that
03:33
<polytan>
johnny, sysvinit is missing in the unmask file
03:33
I have a problem (a simple one) with that)
03:33
<johnny>
polytan, oh.. they must have bumped it behind my back
03:33
<Appiah>
wonder if it says in the docs..
03:33
<ogra>
it should
03:33
<Appiah>
and I just missed it?
03:34
<polytan>
johnny, yes, I hope we will be able to have something working together ;)
03:34
<johnny>
polytan, tell them to hurry up and make openrc stable :)
03:35
<polytan>
johnny, you have unmask the version -r11 and it doesn't exists anymore
03:35
now, it is -r12 ;)
03:36
<johnny>
i wish the scdript would just error out instead of continuing..
03:36tarbo has quit IRC
03:36
<johnny>
i'll look into that
03:36tarbo has joined #ltsp
03:36
<johnny>
maybe that's the real problem.. and not portage.. :(
03:37
<Appiah>
Looks like the doc in 8.10 does not say that
03:37
(lts-parameter.txt.gz)
03:37
<polytan>
johnny, which version of udev can I unmask ?
03:37
I need one better that 133
03:37
<ogra>
well, the wiki should talk about it
03:37
<Appiah>
How can I suggest that this important command gets in ?
03:38
<ogra>
not sure lts-parameter.txt.gz will talk about it, especially since its obsolete in the recent upstream ...
03:38
<Appiah>
O_o
03:38
<ogra>
CONFIGURE_X isnt used at all anymore ...
03:38
<Appiah>
I understand
03:38
<ogra>
and before intreid it defaulted to True ...
03:38
*intrepid
03:39
<Appiah>
it's only used if you really need to force some settings
03:39
<ogra>
right
03:39
<Appiah>
which is what we needed to do here
03:39
(some crappy monitors)
03:39
that wont be replaced in about .. 1-2 years
03:39
<ogra>
since xorg's autodetection improved quite a lot we started to realy on it
03:39
*rely
03:40
and with future releases of it you will be able to do less and less in the config file
03:40
<Appiah>
how about changing the lts-paramater.txt.gz to say "THIS IS NOT REALLY NEED ONLY USE WHEN .." blah blah ?
03:40
<ogra>
feel free to file a bug :)
03:40
<Appiah>
oh ;)
03:41
<polytan>
johnny, it is the same on my machine
03:41
openrc has raised from 0.3 to 0.4 and that change everything
03:41
udev
03:41
sysvinit
03:41
hal
03:41
cryptsetup
03:42try2free has joined #ltsp
03:42tarbo has quit IRC
03:43tarbo has joined #ltsp
03:43try2free has left #ltsp
03:45
<polytan>
I hope my server will restart after all changes ...
03:46GandalfXx_ has joined #ltsp
03:47
<johnny>
hal? i wonder what was pulling that in.. it shouldn't have been openrc..
03:48
<polytan>
hal is because of udev going from <125 to >=135
03:49tarbo has quit IRC
03:49tarbo has joined #ltsp
03:51
<polytan>
I've got ldm !
03:51
<Appiah>
yaay \o/
03:51
<polytan>
but it freezes when I swith back to CLI
03:52
Appiah, thanks :)
03:52
<johnny>
polytan, as soon as i get my computer back running
03:52
i'll integrate the changes
03:52
unless you wanna make a patch for me..
03:52
to your quickstart profile
03:53
and see if it gets this far from scratch
03:54
<polytan>
johnny, I can do something
03:54
once it will work
03:55tarbo has quit IRC
03:55tarbo has joined #ltsp
04:00
<polytan>
johnny, how not to start ldm after boot ?
04:01
I have a lot of errors I would like to see
04:01
but once ldm has been started, I can't read them
04:01tarbo has quit IRC
04:01alkisg has joined #ltsp
04:01tarbo has joined #ltsp
04:01
<johnny>
you can comment out the screen stuff in /etc/init.d/ltsp-client
04:01
or actually wait
04:02
you could try SCREEN_07=shell
04:02
<polytan>
I can start a ldm session
04:02
<johnny>
polytan, it's possible that many of those errors.. are probably not really errors
04:02
like pulseaudio spew
04:02
<polytan>
johnny, there is one interesting :
04:03
/ is read only
04:03
<johnny>
yes / is read only
04:03
<polytan>
so lts.conf can't be downloaded to /etc
04:03
<johnny>
it should..
04:03
i set it to bind mount /etc/lts.conf
04:03
tmpfs bindmount
04:03
so that it is possible to copy files to it
04:03
<polytan>
what does that mean to "bind miunt /etc/lts.conf" ?
04:04
<ogra>
it creates a tmpfs bindmount for that file in ram
04:04
<johnny>
it means that /etc/lts.conf thinks it is writable.. even if the / isn't
04:04
<polytan>
ok, I see
04:04cyberorg has joined #ltsp
04:04
<polytan>
but I don't have lts.conf in mount
04:05
<ogra>
that works differently :)
04:05
<johnny>
you can see the list of mounts in /etc/conf.d/ltsp-client-setup
04:05
<polytan>
If I do "# mount" on the client, I have a lot of tmpfs, but no lts.conf
04:05
<johnny>
is it not in there?
04:05
oh you're right
04:05
it got lost in the shuffle somehow
04:06
<polytan>
no johnny, it isn't
04:06
<johnny>
polytan, add it to the last entry in that file
04:06
bind files
04:06
<ogra>
your /var/lib/ltsp-client-setup should be in mount
04:06
thats the tmpdir where files get used from
04:07
<polytan>
johnny, is it normal that I have to do "exportfs -va" after each modification of the files on the server ?
04:07
<ogra>
unless that mechanism changed in the past (/me doesnt know, ubuntu didnt use the slow bind methot since ages)
04:07
<johnny>
polytan, no..
04:07tarbo has quit IRC
04:07
<johnny>
polytan, any files? or just /etc/conf.d/ /etc/init.d/ files ?
04:08tarbo has joined #ltsp
04:08
<johnny>
polytan, you should always run rc-status after touching the files in those directories
04:08
<polytan>
all the folder / on the server
04:08
but it doesn't really matter
04:08
<johnny>
no idea..
04:08
<polytan>
it half works
04:08
<johnny>
i'm not all that familiar with nfs
04:08
<polytan>
lts.conf is well download now
04:08
<johnny>
the config options are the same as every other distro
04:09
polytan, that fix is committed
04:09
as of right now
04:10
<polytan>
but in /etc/keymaps, instead of keymaps="fr" or "us", I've $CONSOLE_KEYMAPS
04:10
-S
04:10
<johnny>
/etc/keymaps ?
04:10
i wonder where that is from
04:11
<polytan>
/etc/conf.d/keymaps
04:11
<johnny>
i've never had a chance to test that bit
04:11
seems pretty obvious how i got it wrong now
04:11
<polytan>
ok
04:12
I will see
04:12
ldm also doesn't check for all the session existing on the server
04:13
so if I have a new user, there is no default X session chosen and it fails to start
04:14
<johnny>
and you have ldminfod enabled in /etc/xinetd.d ?
04:14tarbo has quit IRC
04:14tarbo has joined #ltsp
04:14
<polytan>
I've never used xinted :)
04:15
in which file do I have to change something ?
04:15
rexec , rsh or rlogin ?
04:15
<johnny>
no.. ldminfod
04:16
/etc/xinetd.d/ldminfod, change disable=yes to disable=no
04:16
<polytan>
I don't have ldminfo in that folder
04:17
<johnny>
hmm.. i added a dep to recent ebuilds to fix that problem
04:17
it got moved upstream
04:17
add it to your ~arch
04:17
and emerge it
04:17
net-misc/ldminfod
04:17
it's also in the deps of ltsp-server
04:18
<polytan>
I've sync ltsp repo and I don't have ldminfod
04:18
eix ldm => ldm
04:20tarbo has quit IRC
04:21tarbo has joined #ltsp
04:23
<polytan>
johnny, ?
04:24xmagixx has quit IRC
04:24
<johnny>
it's in mine..
04:24
kc8pxy even got it
04:24
you layman -s ltsp ?
04:24
<polytan>
yes
04:24
layman -S
04:25
<johnny>
http://git.overlays.gentoo.org/gitweb/?p=proj/ltsp.git
04:25
its' right there
04:25
so.. update-eix or something :)
04:25
and then ~arch it
04:26cyberorg has quit IRC
04:27tarbo has quit IRC
04:27
<polytan>
* QA Notice: file does not exist:
04:27
*
04:27
* dosbin: server/ldminfod does not exist
04:27tarbo has joined #ltsp
04:27
<johnny>
tha'ts old
04:27
you do not hae the newest ebuilds
04:28
<polytan>
src = "git://git.overlays.gentoo.org/proj/ltsp.git"
04:29
<johnny>
perhaps your layman directories are screwed up.. ?
04:29
that you're sycing to a directory that isn't actually in PORTDIR_OVERLAY ?
04:33
polytan, try changing the line in /etc/ltsp/ltsp-client-setup to
04:33
sed -i "s/us/${CONSOLE_KEYMAP}" /etc/conf.d/keymaps
04:33
makes more sense anyways..
04:33tarbo has quit IRC
04:33tarbo has joined #ltsp
04:34
<johnny>
polytan, i'm going to have to go to bed
04:34
bbl
04:35
<polytan>
good night !
04:35tarbo has quit IRC
04:35cyberorg has joined #ltsp
04:41tarbo has joined #ltsp
04:42
<polytan>
johnny, I was playing with an extremly alod version of ltsp
04:47tarbo has quit IRC
04:47tarbo has joined #ltsp
04:49
<alkisg>
If I try to unmount a local device such as a usb or a cdrom, I get "...device... is not in fstab, and you're not root". Any ideas?
04:49
(intrepid)
04:50
(and of course I'm talking about the local devices in the thin clients)
04:53tarbo has quit IRC
04:53tarbo has joined #ltsp
04:57
<generalsnus>
i have a problem here. the students are using "proxy" websites to gain access to websites they are not allowed while on school.. ive tried to block these pages whitin host file.. but it wont work.. becouse we "need" to use a proxy server. we dont maintain this proxy, and its seldom updated.. users cant connect to the net if we dont use it.
04:58
what can i do to make firefox, check with host file first? or some other solution
04:59
<alkisg>
I think you could install squid, then tell squid to use your proxy and tell all the clients to use squid
05:00
Then you could cut anything you want from squid.conf
05:00cyberorg_ has joined #ltsp
05:01cyberorg has quit IRC
05:01cyberorg_ is now known as cyberorg
05:11
<generalsnus>
alkisg: sigh.. cant figure out how to use squid.. installed the gui for squid aswell.. gadmin-squid.. but there is like 1000 settings to play with
05:11
<alkisg>
generalsnus: you don't need most settings, all I had to change for my school needs were 2-3 lines in squid.conf
05:12
Then you just go to the System menu > preferences > proxy, put 127.0.0.1:3128 there and click on "set as system default" (I'm not sure about the button names)
05:33bobby_C has joined #ltsp
05:38|Paradox| has quit IRC
05:39|Paradox| has joined #ltsp
05:40jond3rd has joined #ltsp
05:45hanthana has left #ltsp
05:50jond3rd has left #ltsp
05:51polytan has quit IRC
05:53polytan has joined #ltsp
05:53
<polytan>
re
05:56polytan has quit IRC
05:58polytan has joined #ltsp
06:00
<generalsnus>
alkisg: ok, do you have a example for me? what do i need to write in the config to block... block.com..for instance?
06:02alkisg1 has joined #ltsp
06:06
<generalsnus>
alkisg1: ok, do you have a example for me? what do i need to write in the config to block... block.com..for instance?
06:07
<alkisg1>
generalsnus: first hit in google: http://www.freesoftwaremagazine.com/articles/web_blocking_squid/
06:07
It's really only a few lines, nothing difficult
06:07
For anything difficult, you may also ask at #squid channel
06:08tarbo has joined #ltsp
06:09alkisg has quit IRC
06:10alkisg1 is now known as alkisg
06:10
<generalsnus>
thanks
06:15xmagixx has joined #ltsp
06:22GandalfXx_ has quit IRC
06:23
<alkisg>
!localdev
06:23
<ltspbot>
alkisg: "localdev" is (#1) Information regarding using pluggable devices and cdroms with LTSP can be found at http://wiki.ltsp.org/twiki/bin/view/Ltsp/LTSP-42-LocalDev, or (#2) for ubuntu: https://wiki.ubuntu.com/EnableLTSP5LocalDevices, or (#3) for debian-edu: http://wiki.debian.org/DebianEdu/HowTo/LocalDeviceLtspfs, or (#4) http://wiki.ubuntu.com/DebugLocalDev
06:29GandalfXx_ has joined #ltsp
06:35GandalfXx_ has left #ltsp
06:46pmatulis has joined #ltsp
06:50
<polytan>
johnny, you must add the program "setxkbmap"
06:50
it is needed by lts
07:01
<Appiah>
What do you people say , 2 nic setup vs 1 nic setup , pros and cons?
07:01
2nic setup = 1 nic for external acess - 1 nic for clients
07:02
<cyberorg>
Appiah, we recommend 2 nic setup
07:09
<Appiah>
can you tell me alittle bit why cyberorg ?
07:09
is it about bw?
07:10
<cyberorg>
for the reason you mentioned or to isolate ltsp network from normal network
07:11
<Appiah>
ok , but this current setup , is seperated from the normal network in the switch
07:11
1nic setup
07:13
<cyberorg>
that is also ok, having one external facing card allows us to leave firewall up there and disabled on ltsp facing card
07:24evilx has joined #LTSP
07:28
<alkisg>
I prefer 2, 3 or even 4 nic setup on the same switch (bonding) for speed reasons... :)
07:30evilx_ has joined #LTSP
07:33
<generalsnus>
ack.. i cant seem to find out how to use squid together with our proxy..
07:34
<quickslip>
does anyone know how to get vnc working on LTSP ?
07:37
<alkisg>
generalsnus: something like this, maybe: http://www.linuxquestions.org/questions/linux-newbie-8/need-help-to-connect-a-squid-proxy-to-connect-to-another-squid-proxy-server-302067/
07:38
quickslip: have you seen italc.sourceforge.net? It's easier to install than vnc...
07:38
Just sudo apt-get install italc-master and it's ready...
07:42likuidkewl has joined #ltsp
07:42likuidkewl is now known as dmaran
07:46
<quickslip>
no - I have not - let me check it out - thnx
07:47
<alkisg>
quickslip: here are some instructions for installing italc on ubuntu: https://help.ubuntu.com/community/UbuntuLTSP/iTalc - they may be of some use (I don't know your distro)...
07:55alkisg has quit IRC
07:58alkisg has joined #ltsp
08:04six2one has joined #ltsp
08:06F-GT has quit IRC
08:08F-GT has joined #ltsp
08:24cliebow has joined #ltsp
08:32Gadi has joined #ltsp
08:34jammcq has joined #ltsp
08:34
<jammcq>
good morning #ltsp
08:35pdjbarber has joined #ltsp
08:36
<cliebow>
j!
08:37
<jammcq>
hey chuck
08:38johnny has left #ltsp
08:38johnny has joined #ltsp
08:38johnny is now known as Guest2043
08:52
<Appiah>
alkisg: may I ask how many clients do you have when you run a 4 nic setup? :P
08:53vvinet has joined #ltsp
08:53brendan0powers has joined #ltsp
08:54brendan0powers has left #ltsp
09:23RobertLaptop has quit IRC
09:23CAN-o-SPAM has joined #ltsp
09:28
<alkisg>
Appiah: 8 clients!!! Well, the point is that I have 400 Mbps instead of 100 Mbps. ;)
09:30
<Appiah>
:D
09:30
why not 1 gbit interface?
09:31
<alkisg>
No money... We'd need ~200 € to upgrade to gigabit
09:31
<Appiah>
ah
09:31
I see
09:31
<alkisg>
But even with gigabit, I'd use 4 NICs
09:31
More to decrease the latency than to increase the bandwidth
09:31
<Appiah>
What does these 8 clients use that uses to much bw?
09:32
<alkisg>
Well, firefox + flash OR videos OR paint OR wine applications ...
09:33
I think each client should have at least 70 Mbps (for a peak)
09:50
<Gadi>
alkisg: are you using ubuntu?
09:50
<alkisg>
Gadi: yes, intrepid mostly
09:51
<Gadi>
you may want to put /opt/ltsp/images in a ramdisk
09:51
that will boost your throughput even more
09:52
<alkisg>
Hm... I don't think my students access the images so often... is there any way to benchmark how much they access them?
09:52
<Gadi>
(depending on the quality of your drives and drive IO and such)
09:52
well, if you do localapps more it could help
09:52
<alkisg>
Nah, I only have 128 RAM in one lab, and 64 MB RAM in the other one
09:52
<Gadi>
ah
09:52
<alkisg>
And I don't really care about boot times
09:52
<Gadi>
then, yeah not so much
09:52
:)
09:54
<alkisg>
In my gigabit lab, each client is only able to fetch ~150 Mbps (with netperf) before using 100% CPU, so I'm never able to use all the bandwidth
09:55
I'd like to put 2-3 more gigabit nics in the server to reduce the latency, though
10:00
<Appiah>
alkisg: what wine applications do the students use?
10:01
<alkisg>
Some Greek educational applications... And also some translated ones like http://www.microworlds.com/
10:06
<Appiah>
I see
10:06
what cpu in the gigabit lab?
10:07
<alkisg>
quad core, 2 point something GHz
10:09
...and the clients Celerons @450 MHz
10:24staffencasa has joined #ltsp
10:24spectra has joined #ltsp
10:26staffencasa_ has joined #ltsp
10:28staffencasa__ has joined #ltsp
10:29bobby_C has quit IRC
10:44staffencasa has quit IRC
10:44staffencasa_ has quit IRC
10:45staffencasa has joined #ltsp
10:48staffencasa_ has joined #ltsp
10:52pdjbarber has quit IRC
10:59staffencasa__ has quit IRC
11:03quickslip has quit IRC
11:06staffencasa has quit IRC
11:12six2one has quit IRC
11:15staffencasa has joined #ltsp
11:15staffencasa_ has quit IRC
11:25RobertLaptop has joined #ltsp
11:27hanthana has joined #ltsp
11:37generalsnus has quit IRC
11:38
<pmatulis>
my screen display sucks on LTSP5/Ubuntu810 with a koolu thin client. where does one begin?
11:38
<laga>
you'd begin by describing your issue
11:39vvinet has quit IRC
11:42
<pmatulis>
the color depth is low (no smooth gradients) and there are some anomolies on web pages (little black squares)
11:43
<alkisg>
pmatulis: run xwininfo and click on a window to see the actual color depth
11:45
<Gadi>
iirc, koolu is geode based
11:45
you may need: XSERVER=geode in lts.conf
11:46
<pmatulis>
Gadi: i'll try that
12:03Guest2043 is now known as johnny
12:04johnny is now known as Guest14076
12:04Guest14076 is now known as johnny
12:08dmaran has quit IRC
12:17
<warren>
pmatulis: is the geode even capable of that many colors at higher resolution? it isn't here.
12:18alkisg has quit IRC
12:19
<pmatulis>
warren: one moment
12:20alkisg has joined #ltsp
12:21* pmatulis is having other troubles
12:22sepski has joined #ltsp
12:22
<pmatulis>
alkisg: it reports a depth of 16
12:26
warren: ^^
12:27
<warren>
pmatulis: that's what I see here too, is the geode capable of more at a high resolution?
12:27
geode is OLD
12:27
<pmatulis>
not sure but i don't remember this issue with my previous ltsp feisty install
12:28
will check that and report back soon
12:28
<warren>
pmatulis: OLPC using Geode LX is 1200x900x16bit
12:37
<alkisg>
pmatulis: so the color depth is fine... I don't know anything about Geode though, so listen to the other people in here! ;)
12:38
<pmatulis>
weird, just logged in to my old ltsp4 rig and i don't get the same bad display
12:38
could it be a network throughput issue? new rig is routed through a laptop
12:38
at the moment
12:48
<re-G>
CAN-o-SPAM: now i have two wlan-ltsp client up. I think 802.11g is too slow for more than two. Can you test multiple clients over 802.11n?
12:49
<johnny>
we have wimax in baltimore .. i wonder how it would work with that :)
12:50
<stgraber>
warren: I reproduced the GX2 bug with F10 so I'm pushing the fix now
12:50
<CAN-o-SPAM>
re-G: thats what i had thought would happen, any success right now seems to be 1 AP / client, i wonder how we can navigate that issue
12:50
i don't have .n
12:51
<warren>
stgraber: ok
12:51
stgraber: be sure to file an upstream bug too
12:51
<stgraber>
warren: I filed a LP bug and Q-FUNK poked upstream (upstream is subscribed to the bug)
12:52
<warren>
stgraber: what is the exact xorg.conf option needed?
12:52
<stgraber>
AccelMethod "XAA" in Deivce
12:52
<warren>
stgraber: xorg 1.5+ probably?
12:52
<stgraber>
*Device
12:52
that options existed with < 1.5 and XAA was the default back then
12:52
so it won't be a problem with earlier versions of Xorg
12:54
<warren>
stgraber: Ubuntu 8.04 or 8.10?
12:54
<re-G>
CAN-o-SPAM: how many clients will work ok with 100Mbit lan? Maybe I will buy .n stuff...
12:55
<stgraber>
warren: both probably but IIRC 8.04 had a few more issues with GX2
12:55
I'm testing with 8.10 running the LTSP from 9.04
12:56
<CAN-o-SPAM>
re-G: how ever many your server will support if thats the only traffic on that network
12:59Lns has joined #ltsp
13:08sbalneav has joined #ltsp
13:08
<sbalneav>
Afternoon all
13:08
<jammcq>
Scotty !!!!!!!!!!!!!!!!!!!!!!!!
13:09
<sbalneav>
Hey jammcq!
13:09
<Lns>
Morning =p
13:09vagrantc has joined #ltsp
13:09
<Gadi>
stgraber: you may want to add a check for an existing X_DEVICE_OPTION_01 to that hack
13:10
just in case someone sets one
13:18
<stgraber>
Gadi: right, we should actually have a function to add a new X_DEVICE_OPTION that'd then just guess the right index to use
13:19
Gadi: in that case I want to apply that settings anyway as otherwise X will just crash
13:20
Gadi: I don't care if the use has already set a X_DEVICE_OPTION_01 as without mine he'd just not get X to start
13:20
so for now I won't add that check, though having a function to append a new X option would be great and avoid that kind of issue
13:21
<re-G>
CAN-o-SPAM: how many wlan clients you could test same time over .g -network?
13:22
atheros works best this time
13:23
<Lns>
a MILLION! *pinky in mouth*
13:25
<CAN-o-SPAM>
re-G: many, but none would be atheros unless i went and got those cards
13:28
<warren>
stgraber: my thought exactly
13:44
<re-G>
CAN-o-SPAM: which chipsets you have
13:47Eghie has joined #ltsp
13:47
<Eghie>
hello
13:47
I have an LTSP on boot timeout
13:47
right before it loads the rootfs (i guess)
13:47
fully at random though
13:47
the same client could be working and after reboot, timeout
13:48irc has joined #ltsp
13:48bobby_C has joined #ltsp
13:48
<Eghie>
server is connected via 1Gb/s and the clients with 100Mb/s
13:48
does ndb have maximum connections or something?
13:48irc is now known as Guest48777
13:50pmatulis has quit IRC
13:52pmatulis has joined #ltsp
13:55spectra has quit IRC
13:57Guest48777 is now known as yaun
13:57yaun is now known as yanu
13:59
<Eghie>
or does someone else have experience with boot timeouts on LTSP?
13:59
<Lns>
Eghie: I heard something just the other day about this, and it turned out to be some sort of xinetd max connection thing
13:59
not sure if its the same issue but it sounds similar
14:01
<Eghie>
ok, but that could be a clue
14:01
maybe NBD has some max connections
14:01
inetd does not have that option
14:02
not that i know of anyway
14:02
<Lns>
Eghie: I wouldnt think nbd does itself...
14:02
<warren>
NBD itself doesn't
14:02
<Eghie>
I thoughed so already yes
14:02sepski has quit IRC
14:02
<Eghie>
could a switch be a potential harmfull creature?
14:03
<Lns>
Eghie: what's the exact error?
14:03
<Eghie>
after DHCP boot info: error: Connection: connection timed out
14:03
<Lns>
Eghie: whats your distro?
14:03
<Eghie>
mounting /rofs on /root/rofs failed: Invalid argument
14:03
ubuntu intrepid
14:03
64 bits, with 32 bits client images
14:04
<Lns>
Eghie: you're not doing remote syslog are you?
14:05
<Eghie>
I did try it yes
14:05
hmm, now lets kill that option first
14:05
<Lns>
Eghie: thats probably your prob
14:05
I had that same issue.
14:06
Gadi can confirm that its really not a good idea to implement it. Thus the warnings on the Ubuntu wiki page about it (if thats what you read to implement it)
14:07
<Eghie>
I just removed it from the lts.conf
14:08xet7 has joined #ltsp
14:08
<Lns>
Eghie: Remember to revert your /etc/default/syslogd.conf or whatever too
14:09
<Eghie>
on the client chroot?
14:09
or the server?
14:09
<Lns>
Eghie: server
14:10hanthana has quit IRC
14:10
<Eghie>
even after disabling syslog, it still does that
14:11
at least, I think I disabled remote syslog loggin in LTSP (all lts.conf I could find in the TFTP folder)
14:12
i looks like he cannot find rootfs
14:12
from the network information i got anyway
14:13
<Lns>
Eghie: why not pastebin your errors so we can take a closer look
14:13
<Eghie>
that would be very difficult I guess
14:13
or I need to do a lot of typing
14:14
in booted initramfs, but not going further I get some info
14:15
<Lns>
ah
14:15
<Eghie>
in /tmp/net-eth0.conf
14:15
<alkisg>
Eghie, a little before this error the client obtains ip, netmask etc. Can you see if it gets a correct IP for the server?
14:15
<Eghie>
ROOTSERVER is set
14:15
but ROOTPATH is empty
14:15
<alkisg>
Do you have another dhcp server running?
14:16
<Eghie>
crap, and there you have the jackpot
14:16
!
14:16
stupid me
14:16
<Lns>
haha!
14:17Blinny has joined #ltsp
14:17
<Lns>
alkisg: how the hell did you figure that out from ROOTPATH being empty?
14:17
<Blinny>
Are there any LTSP-specific considerations for using NIS as an auth scheme?
14:17
<alkisg>
I get it all the time at schools with routers acting as a second dhcp server.. :)
14:17
<Eghie>
working for a whole week, last night 02:30 in bed (gambian time) from work
14:17
<Lns>
hehe
14:18
<alkisg>
...so now I always put IPAPPEND 3 in /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default, even if it isn't needed!
14:18
<Eghie>
Lns: after a while you get an initramfs prompt
14:18
Lns: cat /tmp/net-eth0.conf will give you a textfile with filled in network variables
14:21
Lns and also alkisg thanks for the great help
14:21
<Lns>
Eghie: =) I didnt do nothin, alkisg is the genious here
14:22
<alkisg>
Nah... just someone who got into trouble too often! :(
14:22
<Eghie>
hehe
14:23
<alkisg>
Eghie: you don't have to turn off the second dhcp server, IPAPPEND 3 is a fine workaround for this case
14:23
(if you need it and it's not by mistake, I mean)
14:24
<Eghie>
well its useful when the server for some sort of reason crashes
14:24
then the router will pick up
14:24
<alkisg>
OK, just split the range in half so that the offers don't overlap.
14:24
<kc8pxy>
johnny: heya.. i have portage 2.2 in the chroot
14:25
<Eghie>
range is already split
14:25
<kc8pxy>
alkisg: dunbify the routers :)
14:26
<alkisg>
Eghie, so, IPAPPEND 3 in /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default and you're done
14:26
<Eghie>
yeah, it takes the ip settings from PXE boot, right?
14:26
<alkisg>
kc8pxy: sorry, I didn't get it (dunbify) - google doesn't know dubify and neither my dictionary.... :)
14:26
Yup, but only works for pxe, not etherboot
14:27
<Lns>
alkisg: does that workaround basically stop a second DHCP request from happening, and re-use the initial one from PXE DHCP?
14:27
oh
14:27
lol
14:27
<Gadi>
alkisg: what is IPAPPEND 3?
14:27
<kc8pxy>
alkisg: if you have routers working needlessly as dhcp servers, dumbify them and turn them into hubs :)
14:27
<alkisg>
Lns, yes - Gadi, what Lns said
14:27
<Lns>
So, unless the second (rogue?) DHCP server is configured with NEXTSERVER or whatever, its going to still work randomly ?
14:27
<alkisg>
kc8pxy: ah, got it
14:28
<Gadi>
alkisg: its a kernel argument?
14:28
<alkisg>
Gadi, it converts the ip from pxe into ip=xxx kernel argument
14:28
It's a pxelinux thing
14:28
<Gadi>
really?
14:28* Gadi investigates
14:28
<Lns>
alkisg: have you run into the case where the initial PXE DHCP req. is picked up by the "other" DHCP server, therefore killing the boot process?
14:29
<alkisg>
Lns, multiple servers are allowed by the dhcp protocol, as long as they're configured properly
14:29
Lns, routers typically don't offer a boot filename, so the clients ignore the offers
14:29
<Lns>
alkisg: right...which wont help if someone plugs in a linksys AP with DHCP server unknowingly enabled
14:29
alkisg: they ignore it if it doesnt pass the correct parms for pxe boot?
14:29
<alkisg>
It can be configured by dhcpd.conf, no need for anything special in the router
14:30
Yes, pxe enabled cards completely ignore dhcp offers that don't have a boot filename
14:30
<Lns>
oh wow, didnt know that
14:30
crafty
14:30
<Eghie>
ah, that is good to know
14:30
<alkisg>
Even gpxe supports something similar with a dhcp option called gpxe.priority
14:32
<ogra>
sbalneav, http://www.notebooks.com/2009/01/07/new-generation-of-netbooks-199-and-299-eight-hour-battery-sexy-design/
14:33
<stgraber>
anyone about to upload something to ltsp-trunk or can I tag it now ?
14:34yanu has quit IRC
14:35yanu has joined #ltsp
14:37Blinny has quit IRC
14:38
<sbalneav>
ogra: Cool
14:38
When that arm based one comes out, I'll pick one of them up, too.
14:39
I am REALLY liking the little aspire one.
14:39
<jammcq>
I'm waiting for the Leg based one
14:39
<Lns>
I want one that the dimensions are based on the golden ratio
14:40
<rjune>
ogra, that the one you're working on?
14:40* sbalneav contemplates what other body part we could mention, and decides to stay silent, due to the family based nature of this channel.
14:40
<Lns>
...and can make me pancakes
14:40
sbalneav: you obviously weren't here yesterday, haha
14:40
<sbalneav>
Well, heck, *I* can make pancakes!
14:40
<jammcq>
mmmm, pancakes
14:41
<sbalneav>
Just ask anyone at the last hackfest!
14:41
<rjune>
howdy jammcq
14:41
<jammcq>
hey rjune
14:41
<Lns>
http://www.oswald.us/wp-content/uploads/2007/05/bunny-pancake.jpg
14:41
<rjune>
how's life?
14:41
<jammcq>
damn, now i'm hungry for breakfast food
14:41
life is good
14:41
<Lns>
jammcq: sorry
14:41
<jammcq>
busy as can be
14:42
<sbalneav>
Lns: That's oolong, the rabbit:
14:43
<rjune>
glad you're staying busy
14:43
<sbalneav>
http://en.wikipedia.org/wiki/Oolong_(rabbit)
14:43
<rjune>
idle hands are the devil's playthings and all that
14:43
<Lns>
sbalneav: hahaha!!! http://en.wikipedia.org/wiki/Oolong_(rabbit)
14:43
err crap
14:43
i need to read more =p
14:44
<sbalneav>
http://www.fsinet.or.jp/~sokaisha/rabbit/rabbit.htm
14:44
That was one of my alltime favourite websites
14:44
He used to take that rabbit for trips, walks in the country, etc.
14:46
<Lns>
That is so cool... I have 5 rabbits, but they just stay in my back yard and dig tunnels =p
14:47Eghie has quit IRC
14:53
<Gadi>
alkisg: it seems to me that IPAPPEND bypasses the dhcp mechanism for the OS-present address
14:53
which means a dhcp client never really runs OS-present.
14:53
is that right?
14:54* Gadi supposes that is prolly a good thing in an LTSP env
14:54
<Gadi>
just need to make sure I understand the consequences
14:56
<alkisg>
Gadi: yes, exactly
14:57
But ipconfig doesn't respect the leases anyway, it's a one-time thing, so it doesn't make any difference
14:57
A normal dhcp client like dhclient should be called for things to be done properly, but that would cost in RAM
14:58
(...and ipconfig is somewhat broken, so I try to avoid calling it wherever I can...)
14:58
<Gadi>
u bet - but changing IP midstream on an LTSP client would be catastrophic anyway
14:58
<vagrantc>
right.
14:59
<Gadi>
any reason to avoid IPAPPEND 3?
14:59
<alkisg>
Why would changing ip be catastrophic?
14:59
<vagrantc>
recent dhcp3-server checks if an existing machine is using the ip address anyways
14:59
<Gadi>
alkisg: you have a constant nbd mount
14:59
which would break
14:59
<vagrantc>
or NFS
14:59
<Gadi>
or nfs
14:59
<alkisg>
But that's before nbd mounting starts, isn't it?
14:59
and/or nfs...
15:00
Ah, you mean the leases thing...
15:00
<Gadi>
i meant if you had a dhcp client
15:00
running
15:00
and it got a lease change
15:00
<alkisg>
I thought about using dhclient. Well, dhclient would ask for the same ip, just for a renewal
15:00
<Gadi>
that would suck
15:00
<alkisg>
Renewals are respected, the IP wouldn't have to change
15:01
<Gadi>
right - but it is catastrophic if it does
15:01
so, ipconfig is only one time?
15:02
<alkisg>
Then DHCP_INFORM would be the way to go...
15:02* Gadi wonders why the dhcp server holds the lease
15:02
<alkisg>
Yes, ipconfig is only called once from the initramfs
15:02
Leases have limited life time because addresses are limited
15:02cliebow has quit IRC
15:03
<Gadi>
right
15:03
but, we dont see the dhcp server handing out duplicate leases after the lease expires
15:03
<alkisg>
So if 1.000 people with laptops pass from an airport in a day, leases would have to be very small
15:03
<Gadi>
so, it must ask the client for a DHCP_INFORM
15:04
<alkisg>
If the lease expires, and the same client boots and asks for an IP, the server may give it a different ip
15:04
<Gadi>
thats not what I mean
15:04
<alkisg>
But if it was within the lease time, it would give it the same ip
15:04
<Gadi>
if the lease expires, and a different thin client boots, it should not get the same IP
15:05
<alkisg>
Why?
15:05
<Gadi>
because 2 thin clients with same IP = IP conflict
15:05
<alkisg>
If someone has 50 addresses and 100 clients (in different days), it'll give the same IP
15:05
<Gadi>
I am talking about this:
15:05
1. Lease time = 1 hour
15:06
<johnny>
kc8pxy, but did you get baselayout2?
15:06
<Gadi>
2. Boot client A
15:06
<alkisg>
Ah, you mean about pinging to see if the address exists
15:06
<Gadi>
3. wait an hour
15:06
4. keep it on
15:06
5 boot client B
15:06
6. client B should not get same IP
15:06
but how does dhcp server know?
15:06
<alkisg>
OK, theoretically, it *could* get the same IP, if the dhcp server didn't check to see if the IP existed
15:06
<Gadi>
right
15:06
so, it must
15:06
<alkisg>
Let me find the dhcpd.conf option for this...
15:07
<Gadi>
just as if the machine had static IP
15:07
sorry
15:07
gotta go get the kids
15:07CAN-o-SPAM has quit IRC
15:07
<Gadi>
ciao
15:08Gadi has left #ltsp
15:08
<johnny>
usually it caches by mac addresses does'nt it?
15:08
so you should never run into that situation
15:08
<alkisg>
Gadi: ping-check. If the administrator set it to off, then the scenario you described could happen, resulting in duplicate IPs
15:09pmatulis has quit IRC
15:10
<alkisg>
It defaults to 1 sec, meaning that the dhcp server sends an ICMP request and waits for 1 sec for an answer, to see if another PC uses the IP to be assigned.
15:11
So normally, it doesn't happen, but with "ping-check off" it may happen.
15:11
(if the dhcp client doesn't respect the leases, like ipconfig)
15:11
<johnny>
dnsmasq must have more agressive caching by default
15:11
as i never ever get the same IP while the machine is running
15:11
it stores the mac against the ip
15:12
for as long as the server is running
15:27six2one has joined #ltsp
15:33kernelfreak has joined #ltsp
15:33
<kernelfreak>
Hi..
15:34
<Lns>
hi kernelfreak
15:34
<kernelfreak>
Does anybody knows how to improve performance over a nfs server thats is a server for 80 clients ? or any alternative to nfs ?
15:36
<Lns>
nbd?
15:37
<sbalneav>
kernelfreak: How many nfsd's do you have running?
15:37
What OS are you running?
15:37
<kernelfreak>
there is 16 runnng at one ubuntu-server
15:38
<sbalneav>
For 80 terminals, that's probably not enough.
15:38
I'd bump to at least 32
15:38
You're using nfs-kernel-server?
15:38
<kernelfreak>
Ive tried that. but didnt notice many improvements
15:38
yes
15:39
<sbalneav>
bump up RPCNFSDCOUNT=32 in /etc/default/nfs-kernel-server
15:39
<vagrantc>
kernelfreak: raided filesystem?
15:39
<johnny>
which os?
15:39
<kernelfreak>
yes raid 1
15:39
<sbalneav>
kernelfreak: OK, well, instead of us playing the "guess what you've already done" game, why dont you tell us:
15:40
1) Why you think it's an NFS problem, and 2) What steps you've taken to fix it already?
15:40
<kernelfreak>
right
15:40
let me say
15:40
<vagrantc>
!release
15:40
<ltspbot>
vagrantc: "release" is please mention the linux distro and release you're using :)
15:41
<kernelfreak>
There is a Atlhon X2 with 2gb ddr2 running with 2 7500rpm SATA disks on RAID 1
15:41
on ubuntu server 8.4
15:42
Ive already tried increase the number of threads
15:42
Ive noticed theres a high level of IO
15:42
<sbalneav>
You've got 80 clients, only with 2 gig of ram?
15:42
<vagrantc>
kernelfreak: mount with "noatime" ... might help some. otherwise all reads are effectively writes, which is slow on raid1.
15:42
<johnny>
ubuntu doesn't use nfs by default kc8pxy
15:42
<stgraber>
hey scottie !!!
15:42
<johnny>
err kernelfreak
15:42
<sbalneav>
Hey stgraber!
15:43
<johnny>
why did you switch to nfs kernelfreak
15:43
<kernelfreak>
vagrantc: by the moment its mounted with relatime
15:43
<sbalneav>
kernelfreak: Check the "free" command, or "top". Are you swapping?
15:43
with 80 clients, and only 2 gb of ram, I suspect you're ram starved.
15:43
<kernelfreak>
sbalneav: no there isnt any swap
15:43
<sbalneav>
lol
15:44sepski has joined #ltsp
15:44
<Lns>
yikes!
15:44
<kc8pxy>
johnny: what do they use?
15:44
<johnny>
nbd
15:44
kc8pxy, that wasn't to you .. :)
15:44
it was accident
15:44
<Lns>
kernelfreak: are you using a GUI for the TCs??
15:45
<kc8pxy>
johnny: I'm gonna try and get my production setup working later.
15:45
<johnny>
kc8pxy, make sure baselayout2 is installed after ltsp-build-client
15:45
<kernelfreak>
Lns yes it run a Gnome
15:46
<sbalneav>
You have nowhere near enough ram.
15:46
<kernelfreak>
its for a developers network using eclipse all the time
15:46
<johnny>
eclipse!
15:46
lol
15:46
adding insult to injury..
15:46
kernelfreak, why are you using nfs????
15:46
nbd comes default on ubuntu
15:46
it does not use nfs at all
15:46
unless you manually fiddle with settings
15:47
<kernelfreak>
hmm
15:48
<sbalneav>
kernelfreak: You should have something like (at least) 256 + (128 * users)
15:48
<kernelfreak>
I was thinking about combine two server with GFS and "mount" or use it in a kind of load balance server
15:48
<sbalneav>
kernelfreak: Or about 10 gigs of ram.
15:49
<kernelfreak>
sbalneav: but theeres is no application running on the server just the dhcp and nfs services all the applications on the users machine
15:49
<Lns>
how mature is ltsp-cluster stuff?
15:49
kernelfreak: do you mean localapps?
15:49
<sbalneav>
kernelfreak: so these arenet LTSP thin clients?
15:50* Lns puts a pancake on his head
15:50
<sbalneav>
kernelfreak: Are you running LTSP diskless thin clients?
15:50
<kernelfreak>
Lns yes.. I just use the nfs to prove the / and /home
15:51
<johnny>
that is not ltsp..
15:51
<Lns>
kernelfreak: then what is LTSP part for?
15:51* sbalneav backs away slowly
15:51
<sbalneav>
Going home for the day. Be on later tonight.
15:52
<Lns>
bye sbalneav
15:52
<kernelfreak>
Lns to make the machine boot
15:52
<Lns>
kernelfreak: ok ... so, you're using "localapps" or "fat client" setup here?
15:53
<kernelfreak>
fat client
15:54
<kc8pxy>
johnny: baselayout 2.0 and openrc 0.4.1-r1 are a pair. also installed portage 2.2_rc20
15:54
<Lns>
kernelfreak: ok
15:54
<kc8pxy>
johnny: talk to you next, from FGCMH
15:55
<Lns>
kernelfreak: so does 'top' or 'free' tell you your RAM is still used up?
16:01six2one has quit IRC
16:03vvinet has joined #ltsp
16:08Egyptian[Home] has quit IRC
16:09Egyptian[Home] has joined #ltsp
16:14RobertLaptop has quit IRC
16:36dirigeant has joined #ltsp
16:40sepski has quit IRC
16:48evilx has quit IRC
16:48twinprism has quit IRC
16:54twinprism has joined #ltsp
17:09bobby_C has quit IRC
17:11johnny has left #ltsp
17:17carlos__ has joined #ltsp
17:17carlos__ is now known as irule
17:17irule has quit IRC
17:17irule has joined #ltsp
17:19johnny has joined #ltsp
17:20jammcq has quit IRC
17:30alkisg has quit IRC
17:32xet7 has quit IRC
17:41dirigeant has quit IRC
17:41dirigeant has joined #ltsp
18:01mistik1_ has joined #ltsp
18:01mistik1 has quit IRC
18:02mistik1_ is now known as mistik1
18:12japerry has joined #ltsp
18:45RobertLaptop has joined #ltsp
18:54johnny has left #ltsp
18:56staffencasa has quit IRC
19:08chrisinajar has quit IRC
19:08brendan0powers has joined #ltsp
19:31Lns has quit IRC
19:36japerry has quit IRC
19:36johnny has joined #ltsp
19:37japerry has joined #ltsp
19:39dirigeant has quit IRC
19:47vagrantc has quit IRC
19:50alekibango has quit IRC
19:54japerry has quit IRC
20:05chrisinajar has joined #ltsp
20:06japerry has joined #ltsp
20:28irule has quit IRC
21:04warren has quit IRC
21:13warren has joined #ltsp
21:33nobios has joined #ltsp
21:40warren has quit IRC
21:46hanthana has joined #ltsp
22:01CaScAdE^FarAway has joined #ltsp
22:16nobios has quit IRC
22:18shamino has quit IRC
22:18johnny has quit IRC
22:18xmagixx has quit IRC
22:18CaScAdE^1arAway has quit IRC
22:18primeministerp has quit IRC
22:18rjune_ has quit IRC
22:18wwx has quit IRC
22:18Patina has quit IRC
22:18sutula has quit IRC
22:18sep has quit IRC
22:18loather has quit IRC
22:18johnny has joined #ltsp
22:18xmagixx has joined #ltsp
22:18CaScAdE^1arAway has joined #ltsp
22:18primeministerp has joined #ltsp
22:18rjune_ has joined #ltsp
22:18wwx has joined #ltsp
22:18Patina has joined #ltsp
22:18loather has joined #ltsp
22:18sep has joined #ltsp
22:18sutula has joined #ltsp
22:18CaScAdE^1arAway has quit IRC
22:19shamino has joined #ltsp
22:22wwx has quit IRC
22:22Patina has quit IRC
22:22sutula has quit IRC
22:22johnny has quit IRC
22:22sep has quit IRC
22:22rjune_ has quit IRC
22:22primeministerp has quit IRC
22:22loather has quit IRC
22:22xmagixx has quit IRC
22:24shamino has quit IRC
22:25shamino has joined #ltsp
22:25johnny has joined #ltsp
22:25xmagixx has joined #ltsp
22:25primeministerp has joined #ltsp
22:25rjune_ has joined #ltsp
22:25wwx has joined #ltsp
22:25Patina has joined #ltsp
22:25loather has joined #ltsp
22:25sep has joined #ltsp
22:25sutula has joined #ltsp
22:48
<kc8pxy>
heya all :)
22:48
johnny: other than portage 2.2_rc20 and baselayout 2.0, what else do i need to get ldm going?
22:49
<johnny>
so.. what happens now?
22:49* kc8pxy grabs his mounse, which he took to FG with him.
22:56
<kc8pxy>
ltsp client booting
22:57
boots to a CLI login
22:58
mostly succesfully
22:58
same 2 files don't exist
22:59
johnny: that's it.
22:59
xorg-server?
22:59
or something else?
22:59
<johnny>
rc-status | grep 'ltsp-client$'
23:00
that should work anyways..
23:00chrisinajar has quit IRC
23:02
<kc8pxy>
ltsp-client is started
23:02
it flopped though
23:02
initialization failed,
23:06
johnny: with sound enabled, it does less flopping, but still has cli-only.
23:07
<johnny>
ok.. try startx
23:08
<kc8pxy>
johnny: failed to connect to an x server
23:09
<johnny>
hmm.. weird
23:09
and /var/log/Xorg.0.log says ?
23:09
<kc8pxy>
johnny: does it matter that I'm using xorg on the server?
23:09
<johnny>
no
23:10
<kc8pxy>
johnny: ther eis no xorg.0.log on the client
23:12
<johnny>
you are logged in how exactly?
23:12alkisg has joined #ltsp
23:14
<Ryan52>
grrr. my nbi.img downloads really fast on my test network, but when I siwtch to my main network, it takes forever (to the point where I give up). the only difference is that instead of my ltsp server serving dhcp, my normal system is...any ideas?
23:16
oh, well, it's on a different interface of my ltsp server...I wonder if that network card has something wrong with it...
23:17
<alkisg>
Ryan52: are the network cards on different subnets / switches?
23:17
<Ryan52>
yes.
23:18
tho I'm pretty sure the switch works fine.
23:19
<alkisg>
no, I wasn't worried about the switch; I've seen delays because e.g. the traffic to eth1 was passing through eth0...
23:19
...and the rest of the traffic (in the same, problematic card) is fast? It's only nbi.img that is slow?
23:20
<Ryan52>
oh, heh, nevermind.
23:21
etherboot makes it appear things are downloading, when they aren't, I guess...
23:21* Ryan52 tried with gpxe and it errored out right away
23:21
<alkisg>
....and why do you use nbi.img, anyway?!!! Won't your clients (even if they're etherboot) accept pxelinux.0?
23:22
<Ryan52>
I dunno :p
23:22
that's the way the "working" network has it, which was a default setup, I think...
23:22* Ryan52 will figure that one out later.
23:24
<Ryan52>
there. fixed the tftp problem, and switched to pxelinux. and everything works. yay! :)
23:26
<alkisg>
Ryan52, is "next-server" correct in your dhcpd.conf?
23:27
AH, never mind, I didn't see the last line! :)
23:29alkisg has quit IRC
23:41japerry_cat has joined #ltsp
23:55warren has joined #ltsp
23:58japerry has quit IRC