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


Channel log from 21 August 2012   (all times are UTC)

00:09andygraybeal has joined IRC (andygraybeal!~andy@h206.212.213.151.dynamic.ip.windstream.net)
00:15andygraybeal has left IRC (andygraybeal!~andy@h206.212.213.151.dynamic.ip.windstream.net, Ping timeout: 268 seconds)
01:11warren has left IRC (warren!~warren@fedora/wombat/warren, Read error: Operation timed out)
01:43cl2 has joined IRC (cl2!~cl2@68-114-241-134.dhcp.fdul.wi.charter.com)
01:44
<cl2>
when using xfce with ltsp, removable drives are not being displayed. I'm assuming this is because thunar is looking for volumes in /media/, but ltsp is mounting them in /media/$USER/. how can I fix this?
02:05NeonLicht has left IRC (NeonLicht!~NeonLicht@darwin.ugr.es, Ping timeout: 240 seconds)
02:12warren has joined IRC (warren!~warren@fedora/wombat/warren)
02:23mordocai has joined IRC (mordocai!~mordocai@mail.mordocai.net)
02:37mordocai has left IRC (mordocai!~mordocai@mail.mordocai.net, Quit: Leaving)
02:42NeonLicht has joined IRC (NeonLicht!~NeonLicht@darwin.ugr.es)
03:02cl2 has left IRC (cl2!~cl2@68-114-241-134.dhcp.fdul.wi.charter.com, Quit: Leaving)
03:22vagrantc has joined IRC (vagrantc!~vagrant@c-98-232-129-196.hsd1.or.comcast.net)
03:22vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc)
03:22vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc)
03:24quiliro has joined IRC (quiliro!~quiliro@186.4.149.245)
03:24
<quiliro>
hi
03:25
I have installed ltsp server on a free version of ubuntu (FSF aproved)
03:25
it is based on ubuntu 11.04
03:26
i have ltsp-build-client --arch taranis
03:27
taranis is a previous version of trisquel (the free version of ubuntu) successfully
03:27
but when I boot clients i have a frozen splash
03:28
i removed splash and quiet to verify errors on the client
03:29adam__ has joined IRC (adam__!~adam@cpe-65-28-48-231.wi.res.rr.com)
03:29
<adam__>
does anyone here use ltsp with xfce? i'm having trouble with xfce recognizing removable drives...
03:30
<vagrantc>
look in /media/USERNAME
03:30
they won't appear on the desktop automatically
03:30
<stgraber>
vagrantc: I'll probably be releasing a new upstream release before the Ubuntu feature freeze to pick up these ARM changes I pushed to trunk and a couple more fixes I commited earlier.
03:30
vagrantc: so we won't be on the same release after all, but pretty close ;)
03:31
vagrantc: I'll also have a few packaging changes to clean up and send you (for the bits that would benefit Debian or can be merged without causing any change)
03:31
<vagrantc>
stgraber: ah well
03:31
stgraber: cool :)
03:32
<quiliro>
I get udhcpc: no lease failing
03:32
udhcpc: sending discover
03:32
<stgraber>
I believe I actually have commit rights to your packaging branch on alioth, but I think it's best to send patches as you may not agree with all the changes (not that I'm really aware of anything major) ;)
03:33
<quiliro>
when i add IPAPPEND 3
03:34
I get error: failed to connect to nbd server
03:35
<adam__>
vagrantc: is there a way to make them show up on the desktop automatically? nautilus recognizes the devices even though they're mounted in /media/$USER
03:35
<quiliro>
and then abusybox shell
03:35
<vagrantc>
adam__: see the scripts in /usr/share/doc/ltspfs/examples/
03:35
adam__: you might be able to adapt those to make it work on xfce
03:36
<quiliro>
On busybox: nbd-client 1.2.3.4 2000 /dev/nbd0 returns falure
03:37
error: connect: no oute to host
03:38
error: connect: no route to host
03:38
what do you think I should test?
03:38
i have been trying for a week
03:39
with different methods
03:39
<vagrantc>
quiliro: so, sinnce it's not actually exactly ubuntu, they may have needed to make modifications to the ltsp-build-client scripts ... did they?
03:40
<quiliro>
the thing is that trisquel works with ltsp for version 4.1 but won't with the latest version which works with a modern browser (html5)
03:40
<adam__>
vagrantc: I can look into that, but that won't really work in my setup because the script creates a .desktop file in the user's desktop. however, i'm using one global desktop with read-only launcher shortcuts for all ltsp users. there is no native way for xfce to recognize these mounted devices? isn't it just b/c they're not directly in /media but instead /media/$USER? there is no way to config xfce to look there instead?
03:40
<quiliro>
vagrantc: no they didn't because 5.5 is not lts
03:41
4.1 is
03:41
but i will do it if i can find out and send them to trisquel
03:43
<vagrantc>
adam__: i don't know XFCE, sorry.
03:44
<quiliro>
there is one problem with ltsp-build-client --arch brigantia on trisquel 5.5 which does not come up with ltsp-build-client --arch taranis (brigantia is 5.5 and taranis is 4.1)
03:44
or ltsp-build-client --arch taranis on trisquel 4.1
03:44
<vagrantc>
quiliro: might be worth reporting to the people who produce this distro...
03:45
<quiliro>
i did but they will not repair it for 5.5. they will for 6.0
03:45
which will not be released til mid september
03:46
and the college work is for this week
03:47
will you point me to documentation on how to solve this problem?
03:48
<adam__>
vagrantc: that's alright. i think i'm going to try to make xfce use nautilus to render desktop because removable drives under ltsp work great in nautilus / gnome
03:55sha has joined IRC (sha!~sha@g231162058.adsl.alicedsl.de)
03:58sha_ has left IRC (sha_!~sha@e177161099.adsl.alicedsl.de, Ping timeout: 252 seconds)
03:59alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
04:05Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Quit: Leaving.)
04:08staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Ping timeout: 244 seconds)
04:09Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas)
04:26
<quiliro>
how can I update the client to a new kernel that is not available on the sources.list?
04:27
can I just chroot to /opt/ltsp/i386 and dpkg -i the new deb linux-image?
04:28
<alkisg>
Which distro/version?
04:36jimjimovich has joined IRC (jimjimovich!~jim@77.239.224.238)
04:37quiliro has left IRC (quiliro!~quiliro@186.4.149.245, Ping timeout: 245 seconds)
04:39
<jimjimovich>
Good morning everyone
04:39quiliro has joined IRC (quiliro!~quiliro@186.4.149.245)
04:39
<quiliro>
alkisg: the client is ubuntu 10.04 and the server is ubuntu 11.04
04:40
<alkisg>
quiliro: there are backported versions of kernels in the repositories, in Ubuntu LTS versions
04:40
So you can just apt-get install any backported kernel you want
04:40
http://packages.ubuntu.com/search?keywords=lts-backport
04:41
<jimjimovich>
We're getting ready for the second attempt of the big upgrade to Ubuntu 12.04. We wanted to try the approach without the chroot, I forget what it's called. Anyone have a link?
04:41
<alkisg>
!ltsp-pnp
04:41
<ltsp>
alkisg: ltsp-pnp: ltsp-pnp is an alternative (upstream) method to maintain LTSP installations for thin and fat clients that doesn't involve chroots: https://help.ubuntu.com/community/UbuntuLTSP/ltsp-pnp
04:42
<jimjimovich>
thank you kindly
04:44
<quiliro>
alkisg: linux-image-3.0.0-15-genericis in the clients repos when in chroot
04:44
all i have to do is install it and run ltsp-update-image?
04:45
<alkisg>
quiliro: chroot /opt/ltsp/i386 apt-get install linux-image-generic-lts-backport-oneiric
04:45
...and sudo ltsp-update-image afterwards
04:46
Don't install a kernel directly, use the -lts-backport metapackage, so that it selects the right kernel for you, and then you'll also get updates automatically
04:47
Ah, don't forget to run ltsp-update-kernels as well
04:48
<quiliro>
i don't have any -lts-backport
04:48
and the repos I use do not contain non free packages (they are from trisquel)
04:48
<alkisg>
sudo chroot /opt/ltsp/i386 apt-get update
04:49
sudo chroot /opt/ltsp/i386 apt-cache policy linux-image-generic-lts-backport-onei
04:49
They're _not_ non free, they're in main
04:49
<quiliro>
i cannot confirm other packages do ar don't cantain nonfree packages
04:50
ubuntu contains nonfree kernels in main
04:50
but trisquel doesn't
04:50
<alkisg>
I'm not sure I understand... you're using an Ubuntu 10.04 chroot with ...trisquel repositories?
04:50
<quiliro>
trisquel contains no nonfree packages
04:50
<alkisg>
How is that possible?
04:51
<quiliro>
trisquel is a version of ubuntu
04:51adam__ has left IRC (adam__!~adam@cpe-65-28-48-231.wi.res.rr.com, Quit: Leaving)
04:51
<quiliro>
with no nonfree packages
04:51
<alkisg>
So, what's your /opt/ltsp/i386/etc/apt/sources.list like? Can you pastebin it?
04:52
<quiliro>
suro
04:52
sure
04:54
http://pastebin.com/jk771jAu
04:57
<alkisg>
quiliro: http://archive.trisquel.info/trisquel/pool/main/l/linux-lts-backport-oneiric/
04:57
So, you should have linux-lts-backport-oneiric available
04:58
<quiliro>
ok...thank you very much
04:59
it could not find the package!!!! let me check
05:02
<alkisg>
Anyways, if you don't find it, you can always install a specific kernel, and manually update to newer ones when you need to
05:03
The linux-image metapackages are there just to help with upgrades
05:04
<quiliro>
that package is virtual and there is only the header version
05:04
is it correct what i did before?
05:05
sudo chroot /opt/ltsp/i386 apt-get install linux-image-3.0.0-24-generic
05:05
<alkisg>
Sure
05:05
But you might want to report the problem to your distro so that they generate meta packages as well, if they don't
05:06
Otherwise you'll need to manually update when 3.0.0.-25 becomes backported
05:06
<quiliro>
ok...i did that
05:07
<alkisg>
dpkg -l '*lts-backport*' should tell you if the packages are available (or, apt-cache search)
05:07Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Quit: Leaving.)
05:09
<quiliro>
http://pastebin.com/NaRRdmBe
05:12
<alkisg>
Yeah, file a bug report about the lts-backport meta packages
05:12
<quiliro>
ok
05:17
https://trisquel.info/en/issues/6165
05:17bauerski has joined IRC (bauerski!~witekb@frodo.psp.opole.pl)
05:21vmlintu has joined IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi)
05:27
<quiliro>
something changed...the boot splash
05:27
but the error is almost the same
05:27
error:failed to connect to nbd server
05:28
nbd-client 192.168.0.1 2000 /dev/nbd0 gives an error
05:29
connection refused
05:29
when typed on the busybox prompt
05:29
<alkisg>
(08:27:24 πμ) quiliro: but the error is almost the same => do you mean that you had this error before the upgrade?
05:30
<quiliro>
yes before installing the new kernel
05:30* alkisg checks the irc logs...
05:32dtrask has joined IRC (dtrask!4a4b499b@gateway/web/freenode/ip.74.75.73.155)
05:32
<alkisg>
Hmmm ok, indeed, some ltsp-build-client plugins might not have been ran due to your distro not labelling itself "ubuntu" or at least generating a symlink for the Ubuntu dir
05:32
But anyway, is your nbd-server up and running on your server?
05:32
sudo netstat -nap|grep nbd
05:33
grep nbd /etc/inetd.conf
05:33
<dtrask>
alkisg: hey...thanks for the help, but it didn't work ;-)
05:33
<alkisg>
Hi dtrask
05:33
Let me read your bug reply...
05:33
<dtrask>
alkisg: I'm not at school....I only have access to my server via ssh at the moment
05:33
alkisg: It's very early in the a.m.
05:34
<alkisg>
dtrask: you can copy it with plain `cp`, but I'd really like to troubleshoot the problem when you have the chance
05:34
<dtrask>
alkisg: but I can try
05:34
<alkisg>
It's openssl related, not epoptes related, but it looks like it affects a few people
05:34
Unfortunately I've never been able to reproduce it....
05:34
<dtrask>
alkisg: was there an epoptes update in the past couple days?
05:34
<alkisg>
Nope
05:34
<dtrask>
alkisg: can't seem to figure what may have broken it
05:35
<quiliro>
alkisg: http://pastebin.com/Dbrt6sLt
05:35
<alkisg>
Maybe an openssl update
05:35
<quiliro>
alkisg: itis listening
05:35
<alkisg>
`sudo cp /etc/epoptes/server.crt /opt/ltsp/i386/etc/epoptes` should give you a working setup, but really, if you have the time sometime, come here again so that we share a screen so that I troubleshoot it
05:35
quiliro: 10809? wrong port
05:35
It should be 2000 for your version
05:36
<quiliro>
yes!
05:36
how can I change it?
05:36
<dtrask>
alkisg: could be....I will be back at school in 8 hours or so
05:36
<alkisg>
dtrask: ok, first fix the problem, and then, at your convenience, let me troubleshoot it, can do so even on a running system without breaking it
05:36
<dtrask>
alkisg: sure! Thx
05:37
<alkisg>
Ty too, bb
05:37
quiliro: did you manually modify your nbd-server configuration, maybe as part of the troubleshooting process?
05:37
E.g. the "cannot find /etc/nbd-server/config" message is actually what you're supposed to get, in that version...
05:38
<quiliro>
alkisg: possibly but i doubt it. i do not get that message either
05:38
<alkisg>
sudo mv /etc/nbd-server/config /etc/nbd-server/config.not-needed
05:39
and you need the inetd.conf line for nbd-server...
05:39
Check /var/lib/dpkg/info/ltsp-server.postinst, I think it's there
05:39
(got no access to 10.04 systems to find it for you)
05:40
You're looking for an `update-inetd` call
05:42
When you've added the line with port 2000 in your inetd,conf, then you need to stop nbd-server, restart openbsd-inetd, and reboot the client
05:42
<quiliro>
http://pastebin.com/NSWKnD9F
05:43ricotz has joined IRC (ricotz!~rico@p5B2ADBCC.dip.t-dialin.net)
05:43ricotz has joined IRC (ricotz!~rico@unaffiliated/ricotz)
05:43
<alkisg>
quiliro: hmm no, not that one
05:43
The other one with port 2000... I hope it was in that file and not somewhere else
05:44
Something like this: 2000 stream tcp nowait nobody /usr/sbin/tcpd /usr/sbin/nbdrootd /opt/ltsp/images/i386.img
05:44
<quiliro>
please check the updated paste
05:44
it has the whole file now
05:44dtrask has left IRC (dtrask!4a4b499b@gateway/web/freenode/ip.74.75.73.155, Quit: Page closed)
05:45
<alkisg>
quiliro: it might be in some other file, anyway, put that line above ^ manually in inetd.conf
05:46
<quiliro>
by the way: the distro is based on ubuntu 11.04 and the client is contructed with a version ased on 10.04
05:46
2000 stream tcp nowait nobody /usr/sbin/tcpd /usr/sbin/nbdrootd /opt/ltsp/images/i386.img
05:46
<alkisg>
Right
05:46
<quiliro>
on inetd.conf?
05:46
<alkisg>
Yup
05:46
<quiliro>
ok
05:47
<alkisg>
sudo service nbd-server stop; sudo service openbsd-inetd restart
05:47vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Quit: leaving)
05:47
<alkisg>
And reboot client
05:50
<quiliro>
no starting nbd-server?
05:50
<alkisg>
No, it's not supposed to be running
05:50
LTSP runs it from openbsd-inetd
05:51
(in those old versions)
05:52
<quiliro>
it's taking longer...good symptom
05:53
but the workstation is a 64 bit machine....it shouldn't take long
05:54Parker955 is now known as Parker955_Away
05:54
<alkisg>
If it drops you to a busybox shell, try again the nbd-client command
05:54
Also, it's possible that the missing plugins caused problems
05:56
<quiliro>
no busybox shell... just the splash
05:56
and on ctrl+alt+f1 no errors
05:56
ltsp-build-client said it was successful
05:57
doesn't that mean there were no errors?
05:57
<alkisg>
That doesn't matter, if the distro hasn't integrated the plugins the client might fail to boot because of them
05:57
!quiet-splash
05:57
<ltsp>
alkisg: quiet-splash: to disable the splash screen in Ubuntu, in order to see any boot error messages, run `sudo gedit /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default` and remove quiet splash .
05:58
<quiliro>
i did remove them...they reappeared!
05:59
i am rebooting the client with the changes
05:59
<alkisg>
When you run ltsp-update-kernels, pxelinux.cfg/default gets copied again
06:00
<quiliro>
kernel panic this time
06:00
oh
06:01
should i return to the original kernel?
06:02
<alkisg>
If nbd-server was your problem, yeah, you've no need for a newer kernel
06:02
quiliro: if you rely on ipappend 3, that got removed too
06:02
You need to add it again
06:03
<quiliro>
sudo chroot /opt/ltsp/i386 apt-get remove linux-image-3.0.0-24-generic
06:03
<alkisg>
Try adding it first, before removing the kernel
06:03
<quiliro>
i just removed it
06:03
<alkisg>
You didn't run ltsp-update-kernels/image though, so it doesn't matter
06:03
<quiliro>
ok
06:08
NOTHING
06:08
no kernel panic but no freeze
06:09
<alkisg>
Last message?
06:10
<quiliro>
more messages
06:11
<mealstrom>
good morning..
06:12
<quiliro>
info: task fstype:334 blocked for more than 120 seconds
06:12
mealstrom: it's 1am good morning
06:12
<alkisg>
Hi mealstrom
06:13
quiliro: did you see "nbd-client... negotiation... block-size=xxx" etc?
06:13
Before the one you mention now...
06:13
<mealstrom>
alkisg: hi alkisg
06:13
<quiliro>
yes
06:13
<alkisg>
quiliro: cat /opt/ltsp/i386/etc/network/interfaces
06:13
<mealstrom>
quiliro: so we are from different time zones ).
06:14
<alkisg>
And, sudo chroot /opt/ltsp/i386 dpkg -l '*network-manager*' | grep ^ii
06:14
<jimjimovich>
alkisg: in the ltsp-pnp setup, how does a client know from which server to download the image?
06:15
<quiliro>
http://pastebin.com/Wd3RwSdX
06:15
<alkisg>
jimjimovich: same as in normal ltsp, do you mean that you have multiple ltsp servers?
06:15
<jimjimovich>
alkisg: I just thought that it got that info from the DHCP server, and that the pnp setup doesn't have a DHCP server running.
06:15
<quiliro>
mealstrom: i'm from assange-land: ecuador!
06:16
<alkisg>
jimjimovich: it gets the BOOT server from the (proxy)dhcp server, and then that's also the ltsp server, unless overriden from lts.conf
06:17
<mealstrom>
quiliro: im from Ukraine (gmt +2)
06:17
<alkisg>
quiliro: I'm guessing that you're missing the plugin that prohibits the network service from re-creating a connection on boot, and thus your nbd connection is broken when the service loads
06:17
<jimjimovich>
alkisg: So, the LTSP server is the one giving out IPs, it's just doing it through a proxy? Does this mean that the real DHCP server needs to be inaccessible to the network where the clients are?
06:18
<alkisg>
quiliro: so from this point on, we're troubleshooting distro-specific problems, I don't think I want to go there... ask your distro maintainers or switch to an ltsp supported distro
06:18
!proxydhcp | echo jimjimovich:
06:18
<ltsp>
jimjimovich: proxydhcp: A proxy DHCP server is defined by the PXE specification as a server which sends auxiliary boot information to clients, like the boot filename, tftp server or rootpath, but leaves the task of IP leasing to the normal DHCP server. More info: https://help.ubuntu.com/community/UbuntuLTSP/ProxyDHCP
06:18
<quiliro>
negotiation size: 688588KB
06:18
<alkisg>
jimjimovich: so no, the dhcp server gives the ip leases, it needs to be accessible
06:19
jimjimovich: if you don't like that, you can still use normal dhcp with ltsp-pnp
06:19
<jimjimovich>
alkisg: okay. I'll read those docs.
06:19
<alkisg>
You just need to deviate from the wiki a bit
06:19
<quiliro>
ok...thank you for your help alkisg
06:20
<jimjimovich>
alkisg: so basically, if I have a regular old DHCP server somewhere in the network, the wiki instructions should just work, right?
06:20
<alkisg>
quiliro: nbd connects, loads the networked disk, and then the network service tries to make a network connection,
06:20
<quiliro>
mealstrom: 7 hours difference
06:20
<alkisg>
which breaks the existing connection to the nbd server, thus removing the networked disk, thus giving you the messages you saw
06:21
You need to remove the network service or add eth0 manual to /etc/network/interfaces
06:21
<quiliro>
on the client?
06:21
<alkisg>
On the chroot, yeah
06:21
<quiliro>
ok
06:21
thks
06:22
<alkisg>
jimjimovich: yup
06:22
<quiliro>
but that will not make it give different ips to the different clients
06:22
<jimjimovich>
alkisg: super!
06:22
<alkisg>
quiliro: eth0 manual means "don't touch my ip, the initramfs set it up from the dhcp info"
06:26Bootless has joined IRC (Bootless!~AnDyLap@yath1870.cronon.net)
06:26
<quiliro>
alkisg: i am testing with workstation i was testing before the kernel update now and i get a busybox
06:26
<alkisg>
Btw, after that, you might find more plugins missing + causing problems
06:26
We're out of the "troubleshooting an installation" step, and into the "porting ltsp to another distro" step...
06:26
...which I've no time to go to, sorry
06:27
<quiliro>
i understand
06:27
please point me to documentation so i can do it
06:28
<alkisg>
ltsp.org has some information about porting, but without diving at the code you won't be able to do it
06:28
<quiliro>
i'll do that then will you please suggest where to start?
06:29
<alkisg>
If your distro ever supported ltsp 5, look at the packages it provided
06:29
<quiliro>
i'll do that then. will you please suggest where to start?
06:30
ok...thanks
06:30
your help was very valuable
06:32
<alkisg>
You're welcome
06:33
The main idea is that you want to run ltsp-build-client again, but ensure that the Ubuntu plugins get to run correctly even if your distro uses a custom name
06:33
<quiliro>
ok
06:35alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
06:38jimjimovich has left IRC (jimjimovich!~jim@77.239.224.238, Ping timeout: 244 seconds)
06:41jimjimovich has joined IRC (jimjimovich!~jim@77.239.224.238)
06:43jimjimovich has left IRC (jimjimovich!~jim@77.239.224.238, Remote host closed the connection)
06:44
<muppis>
Is it a common issue that Firefox leaves it lock files in place during logoff/shutdown?
06:52stgraber has left IRC (stgraber!~stgraber@ubuntu/member/stgraber, Ping timeout: 272 seconds)
06:52stgraber has joined IRC (stgraber!~stgraber@ubuntu/member/stgraber)
06:56jimjimovich has joined IRC (jimjimovich!~jim@77.239.224.238)
06:57telex has left IRC (telex!~telex@freeshell.de, Ping timeout: 240 seconds)
06:58telex has joined IRC (telex!~telex@freeshell.de)
06:59lefteris has left IRC (lefteris!~user@pc8692.cti.gr, Read error: Connection reset by peer)
07:00lon3wol7 has joined IRC (lon3wol7!c426aa82@gateway/web/freenode/ip.196.38.170.130)
07:02
<lon3wol7>
Hi Guys
07:03
I have a LTSP server about 85 users we keep losing are panels and the machine keeps spiking the load any idea's
07:03jimjimovich has left IRC (jimjimovich!~jim@77.239.224.238, Ping timeout: 240 seconds)
07:04JustSomeGuy has joined IRC (JustSomeGuy!~JustSomeG@port-87-193-233-146.static.qsc.de)
07:04jimjimovich has joined IRC (jimjimovich!~jim@77.239.224.238)
07:06
<lon3wol7>
We are running an IBM IX3550 machine with 32 GB Memory
07:06jimjimovich has left IRC (jimjimovich!~jim@77.239.224.238, Remote host closed the connection)
07:06jimjimovich has joined IRC (jimjimovich!~jim@77.239.224.238)
07:07telex has left IRC (telex!~telex@freeshell.de, Remote host closed the connection)
07:07
<lon3wol7>
Anybody here ?
07:08telex has joined IRC (telex!~telex@freeshell.de)
07:17jimjimovich has left IRC (jimjimovich!~jim@77.239.224.238)
07:21lon3wol7 has left IRC (lon3wol7!c426aa82@gateway/web/freenode/ip.196.38.170.130, Quit: Page closed)
07:27ricotz has left IRC (ricotz!~rico@unaffiliated/ricotz, Quit: Ex-Chat)
07:42alkisg has joined IRC (alkisg!~alkisg@ppp089210202102.access.hol.gr)
07:42alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
07:46Lumiere has left IRC (Lumiere!~jstraw@unaffiliated/jstraw, Ping timeout: 240 seconds)
08:10quiliro has left IRC (quiliro!~quiliro@186.4.149.245, Quit: Leaving.)
08:12
<JustSomeGuy>
hi i have a problem with my ltsp setup im using ubuntu 12.04. when i shutdown or reboot the system i get a lot of squashfs errors (i mount the root with nbd over the network) and the computer just hangs up (i have to hard reset the computer) any advice?
08:13
<alkisg>
Does reboot work?
08:13
<JustSomeGuy>
nope
08:13
same errors
08:14
<muppis>
Sure you aren't rebooting server instead of client?
08:14
<JustSomeGuy>
yes
08:14
its a fat client
08:14
and i run the command on the client
08:15
the strange thing is that some hardware configurations work fine (they just shut down) but others hang
08:16
<Hyperbyte>
JustSomeGuy, interesting... I've noticed the same thing on my most recent setup of 12.04
08:16
Are you using ltsp-pnp or regular ltsp?
08:16
<JustSomeGuy>
i have a suspicion that the upstart stuff is calling the scripts in a wrong order (like killing the network before unmounting the root with nbd-disconnect) but changeing the init scripts had no effect
08:16
<alkisg>
Hyperbyte: have you noticed that with ltsp-pnp?
08:16
<JustSomeGuy>
Hyperbyte, ltsp
08:18markit has joined IRC (markit!~marco@88-149-177-66.v4.ngi.it)
08:20
<JustSomeGuy>
we tested some hardware http://pastebin.com/aVHUcDse but there is no pattern
08:23
one more thing if i shutdown the computer with poweroff -f or reboot -f the system shuts down fine (it no option because of nfs shares that should be disconnected properly but maybe its a hint)
08:24
<Hyperbyte>
alkisg, yes, there's one client, a slower HP one with very few ram - if I shut it down, it spits out a whole bunch of squashfs errors. Usually it shuts down regardless, so I haven't paid much attention otherwise.
08:24
This particular client is running the rdesktop screen script, not ldm.
08:24
<alkisg>
Hyperbyte: right, it's ok to have the squashfs error if it shuts down anyway
08:24
So what I wanted to see if the fixes in ltsp-pnp resolved a problem in the nbd-disconnect script shipped with 12.04
08:25
<Hyperbyte>
Well I think it's also frozen once or twice while shutting down, but I'm not entirely sure since this is an 'always on' machine. :)
08:25
<alkisg>
Hmm ok then we might need some more troubleshooting on where to put that script
08:26
JustSomeGuy: it's not something to be solved quickly, so better file a bug about it...
08:26
!ubuntu-bug
08:26
<ltsp>
alkisg: ubuntu-bug: To file a bug report for Ubuntu LTSP, go to https://bugs.launchpad.net/ubuntu/+source/ltsp
08:26
<JustSomeGuy>
alkisg, ok thx
08:26
<Hyperbyte>
alkisg, why is it important to disconnect nbd at all?
08:26
<alkisg>
If we don't disconnect it, it leaves the swap files etc on the server
08:27
And an nbd-server process for each client, to die 2 h later
08:27
<Hyperbyte>
Only for a certain timeout period I presume?
08:27
Two hours, mmmm
08:27
<alkisg>
No, the swap files stay there afaik
08:27
Until server reboot.
08:27
<Hyperbyte>
Oh, that'd suck
08:28
<alkisg>
In theory the workaround works, but we might have missed something, or upstart might trigger it when we don't want to
08:29
The correct fix would be in the nbd package, not in ltsp, but it's hard to do
08:29
<Hyperbyte>
What workaround are you using now?
08:30
<alkisg>
Wait for all services to die, and have the nbd-client service run last, which disconnects the nbd mount and right then uses the REISUB trick to tell the kernel to shutdown
08:30
(echo b > /sys/somewhere)
08:31
The next better one, in nbd, would be not to disconnect the nbd mount, but have the client signal the server that it will shutdown it a while, and then have the server close the connection after e.g. 1 minute
08:32
<Hyperbyte>
Oooo, I like that.
08:32
<alkisg>
(and clean up the swaps etc)
08:32
<JustSomeGuy>
what can i do about the hangups?
08:32
<alkisg>
File a bug report, not much else unless you can send patches...
08:32
<JustSomeGuy>
ok
08:32
thx
08:34
<alkisg>
Ah, or you can revert to using `poweroff -fp` and have the swap files and nbd-server processes linger on..
08:35
<JustSomeGuy>
where can i change that? i read about that somewhere and changed it in the acpi config but that doesnt affect the shutdown button on the unity ui
08:36
<alkisg>
No, you need to copy /etc/init/ltsp-shutdown.conf from an older ltsp package
08:36sep has left IRC (sep!~sep@40.211.jostedal.no, Ping timeout: 260 seconds)
08:36
<alkisg>
And ltsp-reboot.conf
08:36
<JustSomeGuy>
ah ok
08:40sep has joined IRC (sep!~sep@40.211.jostedal.no)
08:45
<mealstrom>
JustSomeGuy: if i understand youve got problem of deleting swap files on server after client poweroffs?
08:48
<markit>
alkisg: just reading... any new ltsp bug I'd better be aware of when update my schools?
08:49
<JustSomeGuy>
mealstrom, no the problem is that under some hardware configurations the system just hangs up when i try to shutdown
08:50
<mealstrom>
JustSomeGuy: oh. so i've misunderstood question. ok
09:00khildin has joined IRC (khildin!~khildin@ip-83-134-230-135.dsl.scarlet.be)
09:21
<JustSomeGuy>
alkisg, the workaround with the conf files works finde thx (and i can unmount my nfs fileshares before shutdown yay)
09:21
alkisg, should i still file a bugreport?
09:27vmlintu has left IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi, Ping timeout: 244 seconds)
09:50stgraber has left IRC (stgraber!~stgraber@ubuntu/member/stgraber, Remote host closed the connection)
09:50stgraber has joined IRC (stgraber!~stgraber@ubuntu/member/stgraber)
09:57
<JustSomeGuy>
alkisg, https://bugs.launchpad.net/ubuntu/+source/ltsp/+bug/1039442
09:57
<alkisg>
JustSomeGuy: (just back) yup, nice
09:57
<JustSomeGuy>
hope thats enougth information :/
09:58
<alkisg>
JustSomeGuy: the pastebin pastes expire after a while, so it'd be better to attach any info into the bug report itself
09:58
<JustSomeGuy>
will do that
10:00
<alkisg>
I don't know how fast will someone, or me, look into this, but it's nice to have a person that can reproduce it, cause it seems rare, I haven't been able to reproduce the problem myself
10:02pod2 has joined IRC (pod2!836fb95f@gateway/web/freenode/ip.131.111.185.95)
10:03
<JustSomeGuy>
yeah its not high priority now we will just buy hardware that we know works and use the workaround with poweroff -fp
10:03
thx for your help!
10:06
<cyberorg>
alkisg, hi, i had sent this http://susepaste.org/75236235 for ldm long time back, could you commit it please so i can remove from our package
10:07
<alkisg>
Hi cyberorg, you might want to ping stgraber or commit it yourself, I don't work on ldm much (waiting for libpam-ssh + lightdm), what is it for?
10:07
(I think I only did 1 ldm commit, ever...)
10:07
<cyberorg>
alkisg, this was sent last year, about build warnings http://susepaste.org/27061420
10:08
does no dev read ltsp-developer list?
10:08
<alkisg>
Gotcha... yeah best contact one of the ldm maintainers either in ubuntu or debian or gentoo etc
10:08
<cyberorg>
stgraber, ^^ please commit
10:10
<alkisg>
(or file it in a bug report against upstream ldm, if you don't want to commit it yourself...)
10:11
<cyberorg>
the patch still applies cleanly
10:12
i assume sending on ltsp-developer list should be enough, do we have to be so formal? :)
10:15vmlintu has joined IRC (vmlintu!~vmlintu@dsl-hkibrasgw4-fe5cdc00-2.dhcp.inet.fi)
10:16
<alexqwesa>
hello everybody, i want suggest patch for ltsp ( /usr/share/ltsp/screen.d/start_local_wm )
10:16
http://susepaste.org/50019954
10:17
<cyberorg>
alexqwesa, i think something like that is already done if FATCLIENT is set
10:19
*LTSP_FATCLIENT
10:22
<alexqwesa>
cyberorg: ??, there i can find docs about it ?
10:23jimjimovich2 has joined IRC (jimjimovich2!~administr@77.239.224.238)
10:24
<alexqwesa>
the most alike to start_local_wm its kiosk - but it's not exactly that i needed
10:24
<jimjimovich2>
alkisg: I did everything per the ltsp-pnp wiki instructions. How do I get clients to boot as Fat clients?
10:27
<cyberorg>
alexqwesa, can't find docs, grep LTSP_FATCLIENT /srv/kiwi-ltsp-nfs-i386/usr/share/ltsp/* -R
10:31
it currently runs twm on suse image as that is the only wm available locally
10:33
<alexqwesa>
cyberorg: i want run ldm and local session both (SCREEN_08 = ldm and SCREEN_05 = start_local_wm )
10:34
so, user can switch between them
10:34
and how to setup which DE will use in FAT_CLIENT ??
10:34
<cyberorg>
ah, with FAT you get default session running localwm
10:35
you can select from preference menu at ldm
10:36
but you can also run remoteapps from local fat session, not sure if it is implemented yet, alkisg do you know?
10:36
yes it is, ltsp-remoteapps
10:40qwebirc1575 has joined IRC (qwebirc1575!5195e8e1@gateway/web/freenode/ip.81.149.232.225)
10:41
<qwebirc1575>
hello
10:41
<alkisg>
jimjimovich2: they're supposed to work out of the box, > 500 mb ram => fats, < 500 => thins
10:41
Remoteapps works, yeah
10:42
<qwebirc1575>
i am trying to find some information in regards to implementing LTSP into my network
10:42
<alkisg>
Hi qwebirc1575
10:43
<qwebirc1575>
i have a virtually null budget, and want to reduce costs as much as possible
10:43
<pod2>
I'm using debian packages 5.4.2-2 with a fresh i386 client build. Some clients work fine, but one type of hardware fails very early in the boot process
10:43
<alexqwesa>
cyberorg: i add LTSP_FATCLIENT = True and SCREEN_06 = menu into lts.conf - and got this error: http://www.23hq.com/7828204/8102667_bd9807a8dd7c2f51da0f1ea5807eedd9_large.jpg
10:43
<mealstrom>
qwebirc1575: nice start. continue please
10:44
<jimjimovich2>
anyone come up with a good solution for the atom based motherboards that report a non-existant laptop screen?
10:44
<pod2>
it loads the kernel and initrd image and then just stops. It's not hung - it responds to ping and ctrl-alt-del. It just sits there doing nothing.
10:44
<alkisg>
jimjimovich2: I think Hyperbyte has a fix handy, but it was fixed upstream in later versions
10:44
The idea is to disable the other output
10:45
<cyberorg>
alexqwesa, remove SCREEN_06 = menu
10:45
<alkisg>
pod2: try removing "quiet" from pxelinux.cfg/default
10:45
<qwebirc1575>
looking to deploy a debian based LTSP server to handle 30 clients, with the possibility of expanding to around 50
10:45
<alkisg>
!quiet-splash | echo pod2
10:45
<ltsp>
pod2 quiet-splash: to disable the splash screen in Ubuntu, in order to see any boot error messages, run `sudo gedit /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default` and remove quiet splash .
10:45
<jimjimovich2>
alkisg: yeah, I remember I tried back in June and couldn't get it to work
10:45
<alkisg>
jimjimovich2: output of xrandr?
10:45
<pod2>
it doesn't have a splash screen, but I'll try that anyway
10:46
this is debian rather than ubuntu
10:46
<alkisg>
pod2: I know, it's just that we don't have a !quiet factoid for debian... :)
10:46
<qwebirc1575>
currently most of the machines are mac mini's (PPC G4) with the rest being dell optiplex 170L's
10:46
<alkisg>
!learn quiet as to enable the Debian kernel and initramfs to show up more messages, in order to better troubleshoot the boot process, remove "quiet" from /var/lib/tftpboot/ltsp/i386/lts.conf
10:46
<ltsp>
alkisg: The operation succeeded.
10:46
<alkisg>
pod2: ^ :)
10:46
<qwebirc1575>
basically looking for some advice on best pracs for setup
10:47
<alkisg>
qwebirc1575: client RAM, and server specs?
10:47
<Hyperbyte>
Haha, "Wishing this patch a happy anniversary" :-D
10:48
<qwebirc1575>
clients have 512 ram min, server is going to be a dell dual core xenon 3ghz with 2xgbe nics
10:49
<alkisg>
Hyperbyte, cyberorg: That's nothing, my patch from 2008 waited 2 years in the ltsp-developers list until I became a developer and committed it myself :P
10:49
<cyberorg>
alkisg, :)
10:49
<alkisg>
qwebirc1575: if you have a switch with 1 gigabit port, you're fine, you'd just need 2 chroots if I understood correctly, that you have 2 client ARCHes
10:49
<mealstrom>
alkisg: /lib/init/upstart-job this will help to debug init.d scripts on ubuntu/debian
10:49
<Hyperbyte>
cyberorg, I think you get the underlying message. ;-)
10:49
<mealstrom>
alkisg: http://pastebin.com/cGm1AcNT
10:50
<qwebirc1575>
yeah
10:50
<alkisg>
mealstrom: what for? the nbd-client process? It's more complicated unfortunately than just upstart debugging...
10:50
<cyberorg>
Hyperbyte, yeah, that is why we do all kiwi-ltsp development on https://sourceforge.net/projects/kiwi-ltsp/
10:50
<pod2>
same thing without 'quiet'
10:50
<qwebirc1575>
mixed domain, PPC G4/G5 and Intel
10:50andygraybeal has joined IRC (andygraybeal!~andy.gray@obsidian.casanueva.com)
10:50
<alkisg>
cyberorg: why don't you commit it yourself? Distro maintainers will just test if it's ok for them...
10:51
<Hyperbyte>
cyberorg, I -think- alkisg meant you should just become an LTSP developer. :P
10:51
<jimjimovich2>
alkisg: http://pastebin.com/wx5qdtYt
10:51
<pod2>
it downloads the kernel and initrd, errors about modules unix and af_packet not found. The kernel does it's dhcp request and then an error about module overlayfs not found and it stops there
10:52
<alkisg>
pod2: overlayfs?!
10:52
Debian uses aufs...
10:52
<cyberorg>
Hyperbyte, may be someday, for now we don't change anything in upstream ltsp
10:52
<alkisg>
pod2: do you see the client getting an ip address? maybe it's missing the nic module
10:52
<Hyperbyte>
cyberorg, why not? I understand LTSP is made in a such a way it's very easy to add-in specific distro code
10:52
*distro-specific
10:52
<pod2>
definitely gets the correct ip
10:53
the working clients give the same module errors as well
10:53
<cyberorg>
Hyperbyte, we don't need any distro specific code to ltsp, all the default and common bits are good for us
10:53
<Hyperbyte>
cyberorg, so kiwi-ltsp is just unmodified ltsp?
10:53
<alkisg>
pod2: ah, what I mean is if it gets an ip address in the *initramfs*, not in the pxe stage
10:53
<qwebirc1575>
do you have any good sources for LTSP implimentation, i have had a good look around and there seems to be so many different options, all a bit confusing really
10:53
<alkisg>
qwebirc1575: there's a good .pdf in the docs, see /topic
10:54
<cyberorg>
yes, ltsp part is unmodified
10:54
<pod2>
the working clients have the next line: negotiation size = 512mb
10:54
<alkisg>
cyberorg: I think that LTSP welcomes distro-specific commits, so why avoid that?
10:54
<alexqwesa>
cyberorg: works, but now i can't load remote sessions(via ssh) with ldm
10:54
<pod2>
sorry, yes both stages get the same correct ip
10:54
<qwebirc1575>
what do you mean by /topic?
10:54
<alexqwesa>
maybe use LTSP_FATCLIENT_FOR_ttyN = true and LTSP_FATCLIENT_FOR_ttyM = false
10:54
<alkisg>
pod2: ...is it possible that you send an Ubuntu initramfs to these clients? It sounds like you're using NBD, while debian uses NFS by default
10:55
(and Ubuntu also prefers overlayfs while debian only has aufs)
10:55
<cyberorg>
alkisg, there is no question of avoiding it, upstream is doing great job, everything we need works as it is on suse :)
10:56
<pod2>
alkisg: it's possible - how would I tell?
10:56
<Hyperbyte>
cyberorg, I'd still just work on the "LTSP" project if I were you rather than do everything in a seperate project. LTSP has good support for that, and it'll greatly improve support for OpenSUSE LTSP. :)0000
10:56
<alkisg>
cyberorg: sure, but in your place, I'd go ahead and commit the changes, since they even seem distro-independed, and I'd expect distro maintainers to check them before their releases
10:56
<Hyperbyte>
Anyway - that's my opinion.
10:56
<alkisg>
pod2: first of all, are you deliberately using NBD, or by mistake?
10:57
<qwebirc1575>
alkisgL do you have the full url that you are meaning for /topic>
10:57
<Hyperbyte>
And I'm no developer and not using OpenSUSE so it really doesn't matter to me. I've just never understood why there's need for k12linux, kiwi-ltsp, etc....
10:57
<pod2>
whatever the debian default is - I have no preference
10:57
<alkisg>
qwebirc1575: I meant the irc topic (you type /topic on your client to see it), but anyway here's the link: http://sourceforge.net/apps/mediawiki/ltsp/index.php?title=Ltsp_LtspDocumentationUpstream
10:58
<jimjimovich2>
alkisg: I actually found an old chat log where you helped someone else with this problem ... and it works :)
10:58
<alkisg>
jimjimovich2: cool
10:58
pod2: then you did something wrong, maybe ran ltsp-update-image at some time, and switched to nbd
10:58
<pod2>
i found some instructions to remove /etc/ltsp/update-kernels.conf and then run ltsp-update-kernels. I did that, and the boot process definitely changed. But still didn't work
10:58
the changelog for 5.4.2 also says that file should be there now
10:59
<alkisg>
True, it needs to be there
10:59
<cyberorg>
Hyperbyte, kiwi bits takes care of suse specific configurations, image building, image management etc, ltsp part is 100% upstream so we get all the goodness of LTSP project
10:59
<alkisg>
pod2: pastebin your /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default
11:00
<jimjimovich2>
hmmm ... unity 3d doesn't work when i boot as a fat client. strange.
11:01
<Hyperbyte>
cyberorg, you either didn't read my previous statement or are choosing to ignore it: with LTSP's great upstream possibilities for distro-specific code, I don't understand the -need- for k12linux and kiwi-ltsp.
11:01
<alkisg>
jimjimovich2: if 2d works and 3d doesn't, maybe it's a problem in your graphics card not supporting all the bits needed by compiz
11:01
<jimjimovich2>
alkisg: I'll check on a different machine. These Intel cards on the atom motherboards are really flaky
11:02
<pod2>
ah, I think I found the problem. I have a separate dhcp server which has different config. I'll fix that and try again :)
11:02
<jimjimovich2>
compiz seems to be working in the "classic" desktop.
11:02
<cyberorg>
Hyperbyte, because there are lot of features specific to kiwi, not sure how they would fit in upstream ltsp
11:04
Hyperbyte, http://susepaste.org/view/raw/23931057 for example
11:06
Hyperbyte, there is nothing distro specific we require in ltsp bits
11:07
<alkisg>
ltsp-update-image --type=clickfs, --revert etc, all those would fit upstream too
11:08
We're already supporting loopback images, nbd images, an in the future btrfs images, clickfs wouldn't add much overhead...
11:08
ltsp-config --dhcp is there
11:08
ltsp-config nfs for exports... etc etc
11:08
For the boot menu, vagrantc has committed extensive support
11:09
Anyway, if all that is already provided by kiwi, dunno
11:11
<cyberorg>
alkisg, the link ^^ lists current features supported, of course there will be some that will fit with ltsp-* scripts, some wont
11:11
<markit>
cyberorg: btw, what is kiwi?
11:12
<cyberorg>
markit, in short it is system imaging tool
11:12
markit, http://opensuse.github.com/kiwi/
11:13
<markit>
cyberorg: thanks
11:13
googling I've found a "easy-ltsp" project
11:13
maybe a way to easely install and configure ltsp... I did scripts for that, sigh
11:13
<cyberorg>
markit, that was gsoc project, GUI for editing lts.conf
11:14
<markit>
ah, I see
11:16|Paradox| has left IRC (|Paradox|!~iamparado@c-71-62-18-35.hsd1.va.comcast.net, Read error: No route to host)
11:16szhwx has joined IRC (szhwx!~iamparado@c-71-62-18-35.hsd1.va.comcast.net)
11:16szhwx is now known as |Paradox|
11:26
<cyberorg>
alexqwesa, may be send in your patch for LTSP_FATCLIENT_SCREEN
11:27
<pod2>
ok, dhcp server is fine. It's pointing back to my ltsp server and the server logs definitely show it downloading pxelinux.cfg/default from there
11:27
<alkisg>
pod2: pastebin that file
11:27
<pod2>
here is the pxelinux.cfg/default file: http://pastebin.com/1BEzCYAx
11:28
<alkisg>
pod2: something is definately wrong, if the client says "negotiation: 500+ mb" etc, which shows that it's using nbd
11:28
<cyberorg>
alexqwesa, not sure how that would work though, it would require multiple ldm?
11:28
<alkisg>
pod2: which image does the client load? Can you verify from the tftp logs that it loads initrd.img-3.2.0-3-486?
11:29
<pod2>
http://pastebin.com/ytN9L2vJ
11:30
172.29.224.32 is the client
11:31
<alkisg>
pod2: put break=bottom or break=init in pxelinux.cfg (in line 15), you'll get a busybox shell
11:31
Aaah
11:31
Never mind
11:31
The nbd mount was the swap
11:32
How much RAM does the client have?
11:32
<pod2>
not a lot - I'll check
11:34
48mb plus a bit shared with the graphics
11:34
<alkisg>
Yeah it's too little
11:34
<alexqwesa>
cyberorg: i suggest: in addition to SCREEN_$ttyN = ldm you can specifity LTSP_FATCLIENT_$ttyN (and all another LDM_* vars with suffix _$ttyN like - LDM_AUTOLOGIN_$ttyN)
11:34
<pod2>
they've worked fine up until the last few days
11:34
<alkisg>
pod2: I think the minimum for debian is 64mb, while for recent ubuntu 128 (or somewhere 90 to be precise)
11:35
Hmmm
11:35
pod2: do other clients work fine?
11:35
<pod2>
yes, other ones are fine
11:36
it's all the clients of this hardware type
11:36
neoware ca5
11:36
<alkisg>
pod2: try adding this to pxelinux.cfg/default, line 15 (append): ltsp.break=50-fstab
11:36
Reboot the client and see if you get a shell
11:36
<cyberorg>
alexqwesa, before you go hacking post on ltsp-discuss and see if people are interested and devs accept the patches
11:36
<alkisg>
pod2: also, remove the "quiet" part of this line...
11:40jimjimovich2 has left IRC (jimjimovich2!~administr@77.239.224.238)
11:40zevlag has left IRC (zevlag!~zevlag@intranetgateway.etv.net, Read error: Connection reset by peer)
11:41zevlag has joined IRC (zevlag!~zevlag@intranetgateway.etv.net)
11:41
<cyberorg>
alexqwesa, but why have 2 sessions when you can run ltsp-localapps from normal session or ltsp-remoteapps from fatclient session?
11:42
<pod2>
alkisg: that made no difference, which makes me think I'm doing something wrong...
11:43
<cyberorg>
also not sure what would happen if you run local and remote session as same user simultaneously, /home/username is same
11:46
<pod2>
ok, sorted that problem
11:47
lots more message detail and it stopped at a shell
11:47
last line was loading aufs
11:49
<alexqwesa>
cyberorg: i want run one lightwight session(fluxbox e17 or e16) local and full session(gnome kde) via ssh - both at the same time, local session - is like guest login, and it should load at boot
11:50
<cyberorg>
alexqwesa, as different user?
11:52
<alkisg>
pod2: nice, in that place you're in the real file system, past initrams
11:52
pod2: I'd advice booting another client with more ram, run `ps aux` there, and then use RM_SYSTEM_SERVICES to disable from lts.conf the services you don't need, in order to free up some RAM for the older clients
11:52
<cyberorg>
there is SCREEN_=startx , if you have DM installed you can log in locally
11:53
<alkisg>
pod2: with "exit" from there, you can have the client boot process continued, and I guess it'll halt a bit later when it runs out of ram
11:54
<alexqwesa>
cyberorg: yes, (but in my cause it really don't matter since session type is different)
11:55
<pod2>
yeah, exit just says logout and then stops
11:56
<cyberorg>
alexqwesa, so use startx, on one screen and ldm on the other
11:56
<pod2>
what would have changed recently to make them need more ram?
11:56
<alexqwesa>
cyberorg: startx - don't allow chose WM
11:57
<alkisg>
pod2: installing some service in the chroot? some update needing more ram? dunno
12:04
<pod2>
I've found one of these same clients that has 300MB ram and that fails the same way
12:06
thanks for the help so far by the way
12:07DIoX|DaZ has left IRC (DIoX|DaZ!~KaKa@server.civicclub.lt, Quit: changing servers)
12:08
<alexqwesa>
cyberorg: hmm, maybe i should patch startx to read LOCAL_DE_FOR_DISPLAY_N and use it instead xterm ?
12:09
<cyberorg>
alexqwesa, startx -> xdmcp so not local
12:10
<pod2>
back in an hour or so
12:10pod2 is now known as pod2_afk
12:15brunolambert has joined IRC (brunolambert!blambert@nat/revolutionlinux/x-ayecbsqehqvmcloq)
12:17artista_frustrad has joined IRC (artista_frustrad!~fernando@189.125.205.147)
12:36piet has joined IRC (piet!551063c2@gateway/web/freenode/ip.85.16.99.194)
12:37
<piet>
@ alkisg: Hi Alkis. I'm still trying to install the target server (after image copy). But there still must be something missing . . .
12:38
A user can login on the server, but trying to log in via ltsp restarts the login in ltsp . . .
12:38
<muppis>
piet, updated ssh-keys?
12:39
<piet>
@ alkisg: On the source server with the same users : no problem.
12:39
<alkisg>
piet: right, did you update the ssh keys?
12:39
Or, copy you 1st server keys to the second
12:43piet_ has joined IRC (piet_!551063c2@gateway/web/freenode/ip.85.16.99.194)
12:43DIoX|DaZ has joined IRC (DIoX|DaZ!~KaKa@server.civicclub.lt)
12:43piet has left IRC (piet!551063c2@gateway/web/freenode/ip.85.16.99.194, Ping timeout: 245 seconds)
12:44
<alkisg>
(03:39:46 μμ) alkisg: piet: right, did you update the ssh keys?
12:44
(03:39:52 μμ) alkisg: Or, copy you 1st server keys to the second
12:44
<piet_>
@alkisg: can You pls be more precise about updating ssh?
12:44
<alkisg>
!ltsp-update-sshkeys
12:44
<ltsp>
alkisg: ltsp-update-sshkeys: If you changed your server IP, you need to run ltsp-update-sshkeys, and if you're using NBD (Ubuntu) you also need ltsp-update-image afterwards
12:45komunista has joined IRC (komunista!~slavko@adsl-195-168-247-068.dynamic.nextra.sk)
12:47
<piet_>
@ akisg: Do You remember my constallation? Two servers copying the image to the target. So afterwards I have to call ltsp-update-sshkeys on the target?
12:48
@ alkisg: Do You remember my constallation? Two servers copying the image to the target. So afterwards I have to call ltsp-update-sshkeys on the target?
12:48
<alkisg>
piet_: yes, because it has another ip (and server ssh keys)
12:49
Unless if you're using LDM_SERVER=server in lts.conf
12:49
<ogra_>
iirc there was a way to provider a known_hosts file that has the entries in it
12:49
<alkisg>
Then it only checks the server name, which is always "server" for ltsp clients, avoiding the need for ip validation
12:49
<ogra_>
if you provide that at client chroot build time and have all servers in it, that should work too
12:50
<alkisg>
ogra_: true, but he's interested in only transferring the chroot (nbd)
12:50
...after updates
12:50
<ogra_>
alkisg, yeah, but since the IPs of both servers wont change ...
12:50
<alkisg>
So ltsp-update-sshkeys --import-whatever + ltsp-update-image won't work, since he won't have the chroot in the second server, only the image
12:50
ogra_: the 2nd server keys are different
12:51
<ogra_>
and ?
12:51
<alkisg>
He'd need to copy them from server1 to server 2 to make them the same
12:51
in order for the clients to trust them
12:51
<ogra_>
just means you have two stanzas in the file
12:51
<alkisg>
Ah, you mean in the 1st server, before the image copying
12:51
Right, that's easily doable
12:51* ogra_ thought that was vagrants usecase for which he implemented that function
12:52
<alkisg>
Copy the 2nd server keys to the 1st one, before running ltsp-update-image and before transferring the squashfs image
12:52
<ogra_>
right, have a known_hosts file with entries for all servers in the image and never have to update the keys again
12:52
<alkisg>
Yeah I was thinking that it wouldn't help to put that file on the second server, didn't think of putting it on the 1st one
12:52
<ogra_>
... and be able to copy around the img
12:53
well, i wasnt thinking about servers, just about image content ...
12:53
but yeah, indeed it needs to be available on the server that was used to build the image
12:54qwebirc1575 has left IRC (qwebirc1575!5195e8e1@gateway/web/freenode/ip.81.149.232.225, Ping timeout: 245 seconds)
12:55
<piet_>
@ alkisg: I'm not an expert so what will be the easiest/best way to do it?
12:55
<alkisg>
piet_: did you transfer the chroot (/opt/ltsp/i386) to the 2nd server, or just the image (/opt/ltsp/images/i386.img)?
12:56
<piet_>
@ alkisg: like You told me just the image
12:56
<alkisg>
piet_: there's an lts.conf option to bypass the ssh host checking, if you're not afraid of someone faking your server, you can just use that,
12:57
...or, the "formal" way, is to add the 2nd server keys to the 1st one, regenerate the image, and then transfer it again
12:57
Which one you prefer?
12:58
<piet_>
@alkisg lts.conf sounds easier. Do You know the order by heard?
12:58
<alkisg>
Let me search for it..
12:58[GuS] has joined IRC ([GuS]!~MysT@213-117-16-190.fibertel.com.ar)
12:58[GuS] has joined IRC ([GuS]!~MysT@unaffiliated/gus/x-663402)
12:58
<alkisg>
piet_: LDM_SSHOPTIONS="-o StrictHostKeyChecking=no -o CheckHostIP=no -o LogLevel=silent"
12:58
You put that in the second server and you're done, no other changes necessary
12:59F-GT has left IRC (F-GT!~phantom@ppp121-44-213-42.lns20.syd7.internode.on.net, Ping timeout: 256 seconds)
12:59brunolambert has left IRC (brunolambert!blambert@nat/revolutionlinux/x-ayecbsqehqvmcloq, Quit: brunolambert)
13:00brunolambert has joined IRC (brunolambert!blambert@nat/revolutionlinux/x-hgiyynewjgofjhuw)
13:01
<piet_>
@alkisg and leave the SERVER=ipaddress off out of the 2. server? Or does it no matter?
13:02
<alkisg>
piet_: I meant "LDM_SERVER=server", literally, there, but no, leave it out, as that won't work due to the different server keys (it would require copying the server keys once)
13:02
Ah, you mean SERVER=xxx
13:03
How do you configure the client server? From lts.conf or from dhcp? Or both?
13:03
<piet_>
@alkisg both
13:03
<alkisg>
piet_: well, leave it there for now, doesn't hurt
13:04
Normally you can configure it with one of those 2 methods, no need for both of them
13:05
<piet_>
@ alkisg: well, but I think there are configurtion options that I can't do in lts.conf and vice versa . . .
13:05
<alkisg>
Sure, I only meant the server part, not the other details
13:06
<piet_>
@ alkisg: ok, misunderstanding . . .
13:06pod2_afk is now known as pod2
13:08piet_ has left IRC (piet_!551063c2@gateway/web/freenode/ip.85.16.99.194, Quit: Page closed)
13:09
<pod2>
ok, recap time. Server running ltsp 5.4.2 on debian. Fresh i386 client build. Most clients work. One hardware type stops booting after loading the aufs module. Client has plenty (300mb) ram
13:18piet has joined IRC (piet!551063c2@gateway/web/freenode/ip.85.16.99.194)
13:20
<piet>
@alkisg: After using Your parameter in ltsp a client trying to boot from 2. server says somthing like <link traling errors occur> a short moment long.
13:21bauerski has left IRC (bauerski!~witekb@frodo.psp.opole.pl, Quit: Leaving.)
13:21
<alkisg>
piet: I didn't get that, you mean that it doesn't reach the login screen? (LDM)?
13:22
<piet>
@ alkisg No it reaches the screen, but shows a short moment long that message and then behaves unchanged not accepting the login.
13:23
<alkisg>
piet: to solve it faster, let me get a vnc screen to the client
13:23
Put these in server 2 lts.conf:
13:23
SCREEN_02=shell
13:23
SCREEN_07=ldm
13:24
Reboot client, press alt+ctrl+f2 to get a root shell, and ping me here to continue
13:26komunista has left IRC (komunista!~slavko@adsl-195-168-247-068.dynamic.nextra.sk, Read error: Operation timed out)
13:29
<piet>
@ alkisg: gimme Your address again pls
13:30
<alkisg>
apt-get --yes install socat screen
13:30
socat SYSTEM:"sleep 1; exec screen -xRR ra",pty,stderr tcp:alkisg.dyndns.org:25547 & screen -l -S ra
13:35piet has left IRC (piet!551063c2@gateway/web/freenode/ip.85.16.99.194, Quit: Page closed)
13:40dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174)
13:46piet has joined IRC (piet!551063c2@gateway/web/freenode/ip.85.16.99.194)
13:50qwebirc11671 has joined IRC (qwebirc11671!5195e8e1@gateway/web/freenode/ip.81.149.232.225)
13:50
<qwebirc11671>
Hi
13:51
<piet>
@ alkisg: You have to be patient. It seems as if I'm not able to get that line correct into the clients terminal, cause the printing cuts it off. On top of it its thowing errors . . .
13:51
<qwebirc11671>
quick question to anyone that knows, is it possible to have a different distro as the client os? i.e Debian as the LTSP server, and Ubuntu as the Thin Client
13:53
<alkisg>
piet: np
13:53
qwebirc11671: yes
13:53
<qwebirc11671>
can you point me in a direction of docs or info regarding this? i have hit a brick wall :(
13:54
<alkisg>
The easiest way is to create the chroot in an ubuntu server, and then transfer it
13:55
Note that after login, the clients would work on the Debian server
13:55
(unless they're fats)
13:55
<qwebirc11671>
yes i understand that all of the processing for the clients would be on the debian server
13:56
<alkisg>
Why an Ubuntu chroot then?
13:56
I would understand the opposite, for RAM reasons, but not this...
13:56
<qwebirc11671>
but i think the migration from windows/OS X to linux would be easier with ubuntu, as a few of the staff seem to think that an os without a 'flashy' interface is old and rubbish :(
13:56
<piet>
@ alkisg: I hope that was it. Does it work?
13:57
<alkisg>
piet: nope
13:57
Did apt-get install work, or did it mention problems?
13:57
qwebirc11671: but you wouldn't have a flashy OS, as they'd show the debian server GUI
13:57
(not that I think there's a big difference between their UIs...)
13:57
<qwebirc11671>
ahh right i see
13:58
<alkisg>
!fatclients
13:58
<ltsp>
alkisg: fatclients: You may find some info about the Ubuntu/LTSP implementation of fat clients at https://help.ubuntu.com/community/UbuntuLTSP/FatClients
13:58
<alkisg>
(unless you do that)
13:58
<qwebirc11671>
i see
13:59
so essentially, it would be alot easier to just setup the ltsp server on an ubuntu box
13:59
<alkisg>
Yeah
13:59
<qwebirc11671>
ok no problem, clears a few things up
13:59
<alkisg>
piet: need me to tell you some easier method?
13:59
<qwebirc11671>
in regards to windows thin clients with LTSP?
14:00
is it possible to run windows as LTSP thin clients?
14:01piet has left IRC (piet!551063c2@gateway/web/freenode/ip.85.16.99.194, Ping timeout: 245 seconds)
14:01
<qwebirc11671>
i have a requirement to run sage on some clients, and would rather not dive into the pool of wine etc to achieve this
14:02
so would like to run windows instances, maybe in a seperate terminal window to cater for sage
14:10
<dead_inside>
you can run remote desktop on another window
14:10
but like running a windows vm on a thin client…. that would have to be one powerful thin client
14:23
<pod2>
remote desktop works great, but of course you still need a windows server to connect to
14:23ricotz has joined IRC (ricotz!~rico@unaffiliated/ricotz)
14:23Parker955_Away is now known as Parker955
14:24staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu)
14:25
<dead_inside>
i guess i am spoiled with forgetting many years ago just how many servers i control
14:35
<stgraber>
alkisg, cyberorg: I commited that patch and also fixed all the compiler warnings for the rest of the code, so the change is really quite big (deprecated glib and x11 symbols and a bunch of other internal API changes), all I know is that it happily builds without warnings now, but I haven't tried it
14:36Parker955 is now known as Parker955_Away
14:38
<cyberorg>
stgraber, thanks :)
14:41markit has left IRC (markit!~marco@88-149-177-66.v4.ngi.it, )
14:44dtrask has joined IRC (dtrask!a9f46b02@gateway/web/freenode/ip.169.244.107.2)
14:45
<dtrask>
alkisg: you here?
14:45
alkisg: still having issues....clients at login screen not showing up
14:49
anyone seen alkisg?
14:50
<alkisg>
dtrask: back
14:50
<dtrask>
cool
14:50
<alkisg>
dtrask: vnc?
14:50
To have a quick look on what's wrong with the keys
14:50
E.g. did you run ltsp-update-image?
14:50
<dtrask>
wanna do a remote? It's still not working even after copying the cert and so forh
14:50
yes
14:50
<alkisg>
OK, run: x11vnc -connect alkisg.dyndns.org
14:51
<dtrask>
want me to use the remote help feature in epoptes?
14:51
<alkisg>
Ah, yeah, easier this way
14:51
Just put alkisg.dyndns.org there
14:52
<mgariepy>
hey dtrask, if you want i can take 5 minutes to help you out with the menueditor.
14:52
<dtrask>
hang on....on a VM....going to move to a client
14:52* jocarter finally had a chance to look at the new ltsp site, looks nice!
14:52
<dtrask>
um...ok...gotta do the epoptes thing first hang on
14:53
<alkisg>
dtrask: why, it doesn't run from a vm?
14:53
<dtrask>
dunno....entered your domain and it won't connect
14:53Parker955_Away is now known as Parker955
14:53
<alkisg>
Let me check if it's a dyndns problem...
14:53
<dtrask>
i;m in bridged mode...running as a thin-client in the VM
14:53
<mgariepy>
dtrask, no problem, just ping me when you have time
14:53
<dtrask>
k
14:54
mgariepy: thx will do
14:54
<cliebow>
wooha!
14:54
<alkisg>
dtrask: can you run this from an xterm? x11vnc -noshm -connect 89.210.202.102
14:55
<dtrask>
what do you mean
14:55
<alkisg>
dtrask: can you boot/login a client so that it shows up in epoptes?
14:56
<dtrask>
that's the problem....several are on, but they don't show up....oh...I can log one in...hang on
14:58
<alkisg>
dtrask: did you login?
14:59
If you restarted the epoptes service, the clients also need to be rebooted
14:59
dtrask: faster method:
14:59
run this command from an xterm on a logged in client:
14:59
x11vnc -noshm -connect 89.210.202.102
15:02
dtrask: reboot a client
15:03
(I just restarted the epoptes service, maybe e.g. it still had the older keys loaded, if you generated new ones without restarting it)
15:04
OK, all seem fine now
15:04
That was it
15:04
Let me also try the chroot thing...
15:04
<dtrask>
ok...also...I lost all my "groups" I had in epoptes (classroom groups)
15:05
<alkisg>
OK if you see that problem again, file a bug report about it
15:06
...why does the problem NEVER happen when I try to troubleshoot it? :(
15:06
<dtrask>
dunno....kinda' like when the car makes a noise...it doesn't when you take it to the mechanic
15:06
<alkisg>
OK, I think that was all, right?
15:07
<dtrask>
let's see if the "off" machines show up
15:07
<alkisg>
They'll do once you add them to groups
15:08
<dtrask>
Yay!...thx
15:08Bootless has left IRC (Bootless!~AnDyLap@yath1870.cronon.net, Quit: Leaving.)
15:08
<alkisg>
np, if you want also close the question
15:08* alkisg closes vnc...
15:08
<dtrask>
feature request....it would be nice to be able to drag a whole bunch of machines to a group rather than one by one
15:08
<alkisg>
Yeah Phantomas has that in mind, but GTK gave him some problems and we postponed it
15:08
<dtrask>
no rush....happy now :-)
15:09
thx for all your help
15:09
<alkisg>
You're welcome
15:10
<pod2>
still stuck with my non-booting clients :(
15:10
what happens in the boot process after the aufs module loads? where can I look to find this out?
15:11
<alkisg>
pod2: I don't have much more time today to help, but here's a quick tip:
15:11
Put again the ltsp.break=50-fstab
15:11
There, run: openvt bash -l
15:11
Type exit to continue with the booting
15:12
When it hangs, press alt+ctrl+f2 to switch to the root shell that openvt opened
15:12
And there try `ps aux` and other stuff
15:12
<pod2>
thanks, I'll try that
15:13JustSomeGuy has left IRC (JustSomeGuy!~JustSomeG@port-87-193-233-146.static.qsc.de, Quit: Verlassend)
15:15vmlintu has left IRC (vmlintu!~vmlintu@dsl-hkibrasgw4-fe5cdc00-2.dhcp.inet.fi, Ping timeout: 272 seconds)
15:26dobber has left IRC (dobber!~dobber@213.169.45.222, Remote host closed the connection)
15:32
<pod2>
fixed it! \o/
15:36
the problem is jetpipe of all things
15:37dtrask has left IRC (dtrask!a9f46b02@gateway/web/freenode/ip.169.244.107.2, Quit: Page closed)
15:37
<pod2>
http://pastebin.com/dkfmqave <- how to fix it
15:38quiliro has joined IRC (quiliro!~quiliro@186.4.149.245)
15:39
<alkisg>
pod2: https://bugs.launchpad.net/ltsp/+bug/996533
15:40
<pod2>
yes, that's the one
15:40
I had a similar problem a while back with rdesktop sessions not working. Took me *ages* to find that page
15:41
<alkisg>
And you had PRINTER set for all clients, or just the ones that don't boot?
15:41* alkisg is still amazed that recent debian versions still boot with 48 mb ram... :D
15:42
<pod2>
hehe, yes this is now back to the original 48mb box :)
15:42
<alkisg>
So the PRINTER thing only affects some of the clients?
15:42
<pod2>
printer was only set for some clients
15:43
<alkisg>
Got it
15:43
<pod2>
coincidentally, all the same type of hardware
15:43
<alkisg>
I thought about it, but thought was too rare to mention it
15:44
<pod2>
it seems to be a recent change as it was working for ldm sessions before
15:44
i only had the problem with rdesktop previously
15:45
thanks again for all your help
15:45
<alkisg>
You're welcome
15:47
<pod2>
next problem is that rdesktop appears to have been replaced by xfreerdp - I should be able to deal with that though
15:47Lumiere has joined IRC (Lumiere!~jstraw@unaffiliated/jstraw)
15:48
<ogra_>
alkisg, erm, why isnt the jetpipe call backgrounded in the 50-jetpipe script ?
15:49
iirc jetpipe is blocking
15:49
so a simple & should fix it
15:51
<alkisg>
ogra_: no idea, didn't touch that part :)
15:51
pod2: does it fix it for you?
15:51
I.e. putting a & right after jetpipe...
15:52
If so, comment it on the bug report so that we commit it
15:52
<pod2>
is that at the end of this line: /usr/sbin/jetpipe ${JETPIPE_ARGS} ${PRINTER_DEVICE} ${PORT}
15:52
<alkisg>
Yeah: /usr/sbin/jetpipe ${JETPIPE_ARGS} ${PRINTER_DEVICE} ${PORT} &
15:53
<ogra_>
yes, with a space behind ${PORT}
15:53
<pod2>
I'll have a go....
15:56
yes, that seems to do it
15:58
<alkisg>
OK, let me commit that then
16:03Parker955 is now known as Parker955_Away
16:03
<pod2>
I've added a comment to the bug page
16:05
<alkisg>
And I've pushed the fix :)
16:05
<ogra_>
\o/
16:06
<pod2>
very small :)
16:06
I feel like I've made a small contribution back now
16:08
<alkisg>
Sure, very nice! (and good catch ogra_!)
16:08
<ogra_>
:)
16:10
<alkisg>
(although I should have left YOU to do the commit :P :D)
16:11
<ogra_>
lol
16:14
<pod2>
aha, the xfreerdp message just appeared because rdesktop wasn't installed in the chroot
16:14
I've installed it and everything is working and back to normal \o/
16:23cliebow has left IRC (cliebow!~cliebow@WatchGuard.ellsworth-hs.ellsworth.k12.me.us, Quit: Ex-Chat)
16:23cliebow has joined IRC (cliebow!~cliebow@WatchGuard.ellsworth-hs.ellsworth.k12.me.us)
16:30alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
16:31alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
16:31Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas)
16:42vagrantc has joined IRC (vagrantc!~vagrant@75-150-46-245-Oregon.hfc.comcastbusiness.net)
16:42vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc)
16:47qwebirc46723 has joined IRC (qwebirc46723!c7daf20f@gateway/web/freenode/ip.199.218.242.15)
16:47qwebirc46723 is now known as sean-w
16:49
<sean-w>
Hi all. Is there a way to provide an LTSP over RDP for remote clients?
16:49
<elias_a>
sean-w: Why RDP?
16:50
sean-w: Or in other words - why use LTSP if you plan to use RDP.
16:51
<sean-w>
alkisg: To allow multiple remote users to access the same graphical Linux system from their windows PC. Any windows system has a RDP client
16:52
opps. wrong auto complete elias_a
16:55
<vagrantc>
sean-w: short answer is yes, you can use the RDP session type and there's some RDP server for linux.
16:55
opensuse pioneered a lot of that stuff
16:56
xrdp i think?
16:56
there's an rdp plugin for ldm...
16:58
<sean-w>
vagrantc: Thanks.
16:58
<elias_a>
sean-w: I see your point.
16:59
If you want to boot LTSP remotely over Internet that is also possible.
16:59
gPXE is the magic word.
17:00
I know guys are fussing around with it in Bruxelles.
17:01komunista has joined IRC (komunista!~slavko@adsl-195-168-247-068.dynamic.nextra.sk)
17:03pod2 has left IRC (pod2!836fb95f@gateway/web/freenode/ip.131.111.185.95, Quit: Page closed)
17:16komunista has left IRC (komunista!~slavko@adsl-195-168-247-068.dynamic.nextra.sk, Ping timeout: 256 seconds)
17:18alkisg_ has joined IRC (alkisg_!551063c2@gateway/web/freenode/ip.85.16.99.194)
17:23
<alkisg_>
@alkisg Sorry, alkis, I'm giving up. Totally desperate, but I don't find a way to let You into the running server with internet. Whatever I do I always have the canopener in the can(too complicated to describe in here). Couldn't You try to have a look into the server without a running client ?
17:31
<Hyperbyte>
alkisg_, I'm sure you're not doing it on purpose, but I don't think it's very appropriate to use "alkisg_" as your own nickname. Type: /nick yourownnickname
17:33alkisg_ is now known as piet
17:34
<Hyperbyte>
Better. :)
17:34
<piet>
@Hyperbyte sorry, don't know what happened . . . (I really hate this way of communication)
17:35
@alkisg Sorry, alkis, I'm giving up. Totally desperate, but I don't find a way to let You into the running server with internet. Whatever I do I always have the canopener in the can(too complicated to describe in here). Couldn't You try to have a look into the server without a running client ?
17:35
<Hyperbyte>
piet, this way of communication becomes much easier if you install an IRC client on a server somewhere that's always on, and then run it inside a screen session or something and just attach whenever you feel like asking questions. :)
17:36
I'm running this IRC client on my virtual server and whenever I'm around I connect to the session and see if there's something going on that interests me, or if someone left messages. :)
17:38
<piet>
@Hyperbyte I'm new to this so absolutely unexpirienced. Took me the whole afternoon to try to make alkis get into my system. Will have a look at these things when I'm ready soving my problems . . .
17:38epoptes_user7 has joined IRC (epoptes_user7!be971cba@gateway/web/freenode/ip.190.151.28.186)
17:40asmok has joined IRC (asmok!~asmok@adsl-85-157-204-13.regionline.fi)
17:45sean-w has left IRC (sean-w!c7daf20f@gateway/web/freenode/ip.199.218.242.15, Quit: Page closed)
17:50khildin has left IRC (khildin!~khildin@ip-83-134-230-135.dsl.scarlet.be, Quit: I'm gone, bye bye)
17:50
<alkisg>
piet: here?
17:51
<piet>
@Hyperbyte @alkisg yes
17:51
<alkisg>
OK let me catch up with the irc logs...
17:52
piet: ok sure
17:53
piet: x11vnc -connect alkisg.dyndns.org
17:53
Do that with an open client, even if you can't connect it to me, I'll find a way to get access to it ;)
18:03quiliro has left IRC (quiliro!~quiliro@186.4.149.245, Ping timeout: 252 seconds)
18:04
<alkisg>
piet: write here in IRC if you need anything else, I'm not watching the vnc window...
18:04brunolambert has left IRC (brunolambert!blambert@nat/revolutionlinux/x-hgiyynewjgofjhuw, Quit: brunolambert)
18:05brunolambert has joined IRC (brunolambert!blambert@nat/revolutionlinux/x-mtytnopuczdvnvsr)
18:11
<piet>
@ alkisg Thanks for Your help. Found the error: Had a hard link on lts.conf for easier handling but for some reason the link was broken. So the change didn't make it to the right place. Sorry for Your extrawork!!
18:11
<alkisg>
piet: np, the reason the client couldn't connect to me is that it's probably behind the 2nd server nic and you don't have forwarding enabled in your server
18:12
When I thought about it, you weren't here so I couldn't give you the hint
18:15
<piet>
@ alkisg. yes Hyperbyte allready told me that there are easier ways to communicate. You gyes
18:15
@ alkis You folkes are far away from what I can do here. Remember I'm still a newby
18:16
@ alkisg You folkes are far away from what I can do here. Remember I'm still a newby
18:16
<alkisg>
piet: no worries you're doing fine
18:18
<piet>
@alkisg When I'm thru with setting these servers up in our suboffice I come back to You to get me hints how to set up a good communication. Promised :)
18:18
<alkisg>
Hehe, cool, ok
18:20komunista has joined IRC (komunista!~slavko@adsl-195-098-012-021.dynamic.nextra.sk)
18:28
<knipwim>
alkisg: you know why we use ntpdate instead sntp in init-ltsp.d/50-set-time ?
18:29
erm, ntpdate instead of sntp
18:30
<alkisg>
knipwim: no idea, debian and ubuntu override that file
18:30
...they just modify the (preinstalled) ntpdate configuration file instead of running anything
18:31
...and let the ntpdate service do its job
18:31
<knipwim>
i see
18:32
<alkisg>
knipwim: so if the "main" 50-set-time doesn't apply to any distro at all, you could probably modify it to suit whatever's in gentoo
18:33
Or, add a check whether ntpdate or sntp is available
18:33
if [ -x ...
18:33vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Read error: Operation timed out)
18:33
<knipwim>
i'll look into it
18:34
i could use the same method debian uses, only by default, gentoo uses ntpdate for syncing
18:34
i found sntp is much faster
18:35
<alkisg>
Well you can check/modify any of the configuration files that are present
18:35
And let the services do their job, assuming they run on normal installations
18:36
Normally ntpdate runs in the background, so it doesn't take any real user time, does it?
18:36
<knipwim>
yeah, just wondering if i should do that, or, adapt the default one to use sntp
18:36
hmm
18:37
<alkisg>
The defaults are slow (I think it tries 3 times) in order to be more accurate, afaik
18:37
<knipwim>
you're probably right ;)
18:37
<alkisg>
While if you tell it to run just once, it's very faster
18:43asmok has left IRC (asmok!~asmok@adsl-85-157-204-13.regionline.fi, Quit: Lähdössä)
18:46piet has left IRC (piet!551063c2@gateway/web/freenode/ip.85.16.99.194, Quit: Page closed)
18:51Parker955_Away is now known as Parker955
19:07
<knipwim>
hmm, my $SYSLOG_HOST is not preserved from my lts.conf
19:07
50-syslog fails and it's not in /var/cache/ltsp/ltsp_config_env
19:08
don't know if it should be there though
19:14quiliro has joined IRC (quiliro!~quiliro@186.4.149.245)
19:23
<alkisg>
I think that part about saving/restoring env vars sucks a bit...
19:23
It's supposed to speed up things, and it ends up reading/writing files for each var access
19:31vagrantc has joined IRC (vagrantc!~vagrant@75-150-46-245-Oregon.hfc.comcastbusiness.net)
19:31vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc)
19:42cliebow has left IRC (cliebow!~cliebow@WatchGuard.ellsworth-hs.ellsworth.k12.me.us, Quit: Ex-Chat)
19:58[GuS] has left IRC ([GuS]!~MysT@unaffiliated/gus/x-663402, Quit: Konversation terminated!)
20:02dberkholz has left IRC (dberkholz!~dberkholz@gentoo/developer/dberkholz, Quit: Lost terminal)
20:25quiliro has left IRC (quiliro!~quiliro@186.4.149.245, Ping timeout: 240 seconds)
20:32irule has joined IRC (irule!~chatzilla@187.139.65.58)
20:32
<irule>
hi
20:32
ltsp rocks
20:33
<vagrantc>
yay
20:36
<irule>
how may I setup 1 debian 6 as server (new) and a buch of old pcs with debian 6 also installed on the hard drive without removing the HD?
20:37
I was not given money to get nics with EPROM slots, told to wear out the hard drives first he
20:38
<alkisg>
!ipxe
20:38
<ltsp>
alkisg: ipxe: iPXE is the successor to the etherboot/gPXE project, and can be used to netboot clients that don't have a NIC ROM with a PXE stack. To add it to grub, see !grub-ipxe. To add it to the Windows boot loader, see !win32-loader. To download floppy, CD or USB images, visit http://ipxe.org or install the ipxe package.
20:38
<alkisg>
I.e. you don't need eprom slots
20:38
<irule>
I have full blown linux installed on the clients
20:38
<alkisg>
Then you just need a file and a grub entry if you want to netboot them with ltsp
20:38
(ipxe.lkrn)
20:39
<irule>
what if I just X -query the server?
20:39
<alkisg>
You lose sound, usb sticks etc
20:39
<irule>
yikes!
20:39
<alkisg>
Central configuration for clients, e.g. resolution, and lots more
20:39
<irule>
what file do I need?
20:39
<alkisg>
Also also updates, it's much easier to update a chroot than several HDs
20:40
ipxe.lkrn is provided by the ipxe package in debian, or you can download it from ipxe.org
20:40
<elias_a>
alkisg: Does ipxe also read from DNS servers in case you want to boot over Internet?
20:41
<alkisg>
Yup
20:41
<elias_a>
Ok. Have to read the docs.
20:42
alkisg: Have you tried booting with ipxe over usual crappy xdsl lines but using a local proxy between the local clients and xdsl line?
20:42
Would a local proxy make it any faster?
20:44
I have been thinking this as a possible solution for national tests.
20:45
"Just boot your machines to this environment" :)
20:45
<alkisg>
brb
20:46alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
20:48alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
20:51
<elias_a>
I have to go to bed now. Let's talk in the morning.
20:51
<alkisg>
OK, bb, sorry had some debugging to do
20:53
<elias_a>
alkisg: Well, I'm still here :)
20:53
tired but still kind of awake.
20:54
alkisg: So what do you say: would a local proxy do any good?
20:54
<alkisg>
elias_a: 2 mins please to finish the bug report
20:54
<elias_a>
pretty much everything is the same when booting up the clients.
20:54
Ok.
20:58
<alkisg>
OK, done, let me see the question again...
20:59
elias_a: are you talking about thin clients (small chroot, lots of bandwidth after login), fat clients (big chroot, little bandwidth after login) or in general?
21:00
<elias_a>
In general.
21:00
<alkisg>
For thin clients, there shouldn't be any need for caching, as the bandwidth requirements after login are greater than the boot requirements
21:00
<elias_a>
I guess I am talking about a bit fatter clients, then.
21:00
<alkisg>
For fat clients, the chroot is a bit big to cache on the fly, it might be better to have it synced on a local nbd server beforehand
21:00
<elias_a>
Ok.
21:01
<alkisg>
The local nbd server may be a very simple device, it might even run windows...
21:01
xnbd supports the "proxy" scheme that you mentioned, but I haven't tried it
21:01
<elias_a>
What I have in mind is: what about have one national platform for national computerized tests.
21:02
Seems like I am going to do some testing :)
21:02
<alkisg>
I'd do that web-based :D Or, if you need a real environment, with something like NX
21:02
It's way faster over *dsl than X
21:02Parker955 is now known as Parker955_Away
21:02
<elias_a>
Yep - i know.
21:03
The other reason for using a LTSP-like system is that I want to keep it FLOSS.
21:04
<alkisg>
x2go is FLOSS, isn't it?
21:04
<elias_a>
IMHO it is a very good argument for supporting Linux platform to say "look - this company had a flight simulator hidden in a spreadsheet calculator. Do you really want to guarantee their sw does only what it is supposed to do?"
21:05
I don't have to mention any names of companies, do I?
21:05
<alkisg>
:)
21:06
<elias_a>
I'll dig a bit into this. Now I am so tired that I will get evil soon (after mentioning a certain party) :P
21:06
Gonna hit the bed ->
21:06
Good night!
21:07
<alkisg>
Καληνύχτα :)
21:09warren has left IRC (warren!~warren@fedora/wombat/warren, Read error: Operation timed out)
21:10Parker955_Away is now known as Parker955
21:21artista_frustrad has left IRC (artista_frustrad!~fernando@189.125.205.147, Quit: Leaving)
21:32brunolambert has left IRC (brunolambert!blambert@nat/revolutionlinux/x-mtytnopuczdvnvsr, Quit: brunolambert)
21:39dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Quit: Leaving...)
21:44ricotz has left IRC (ricotz!~rico@unaffiliated/ricotz, Quit: Ex-Chat)
21:51Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 260 seconds)
21:54warren has joined IRC (warren!~warren@fedora/wombat/warren)
21:57irule has left IRC (irule!~chatzilla@187.139.65.58, Quit: ChatZilla 0.9.88.2 [Firefox 14.0.1/20120713134347])
21:58komunista has left IRC (komunista!~slavko@adsl-195-098-012-021.dynamic.nextra.sk, Quit: Leaving.)
22:01alexqwesa has left IRC (alexqwesa!~alex@alexo-veto.broker.freenet6.net, Ping timeout: 245 seconds)
22:01alexqwesa has joined IRC (alexqwesa!~alex@2001:5c0:1000:b::5221)
22:02Parker955 is now known as Parker955_Away
22:06Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas)
22:08alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
22:15andygraybeal_ has joined IRC (andygraybeal_!~andy@h206.212.213.151.dynamic.ip.windstream.net)
22:15F-GT has joined IRC (F-GT!~phantom@ppp121-44-213-42.lns20.syd7.internode.on.net)
22:29F-GT has left IRC (F-GT!~phantom@ppp121-44-213-42.lns20.syd7.internode.on.net, Ping timeout: 268 seconds)
22:38Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 256 seconds)
22:46hays has left IRC (hays!~quassel@unaffiliated/hays, Ping timeout: 268 seconds)
22:49hays has joined IRC (hays!~quassel@unaffiliated/hays)
22:52Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas)
23:05andygraybeal_ has left IRC (andygraybeal_!~andy@h206.212.213.151.dynamic.ip.windstream.net, Quit: Ex-Chat)
23:06andygraybeal_ has joined IRC (andygraybeal_!~andy@h206.212.213.151.dynamic.ip.windstream.net)
23:08andygraybeal has left IRC (andygraybeal!~andy.gray@obsidian.casanueva.com, Remote host closed the connection)
23:44Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 245 seconds)