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


Channel log from 21 January 2014   (all times are UTC)

01:29Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas)
03:22andygraybeal_ has left IRC (andygraybeal_!~andy@h23.205.213.151.dynamic.ip.windstream.net, Ping timeout: 272 seconds)
03:23cyberorg has left IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg, Ping timeout: 246 seconds)
03:25Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 265 seconds)
03:36cyberorg has joined IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg)
03:55alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
04:36alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
04:55lifeboy has left IRC (lifeboy!~roland@105-236-162-164.access.mtnbusiness.co.za, Ping timeout: 246 seconds)
05:18work_alkisg has left IRC (work_alkisg!~alkisg@plinet.ioa.sch.gr, Quit: Leaving.)
05:31work_alkisg has joined IRC (work_alkisg!~alkisg@plinet.ioa.sch.gr)
06:12Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas)
06:18alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
06:20Ark74 has joined IRC (Ark74!~Ark74@189.214.42.128.cable.dyn.cableonline.com.mx)
07:11alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Remote host closed the connection)
07:43Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Quit: Leaving.)
08:15gbaman has left IRC (gbaman!~gbaman@host81-134-114-61.in-addr.btopenworld.com, Remote host closed the connection)
09:03bennabiy has left IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com, Ping timeout: 252 seconds)
09:04bennabiy has joined IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com)
09:28gbaman has joined IRC (gbaman!~gbaman@host86-181-59-248.range86-181.btcentralplus.com)
09:40bobby_C has joined IRC (bobby_C!~bobby@212-186-178-75.dynamic.surfer.at)
09:56monteslu has left IRC (monteslu!~monteslu@ip68-109-175-67.ph.ph.cox.net, Ping timeout: 260 seconds)
09:59work_alkisg is now known as alkisg
10:13alexxtasi has joined IRC (alexxtasi!~alex@unaffiliated/alexxtasi)
10:21cyberorg has left IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg, Ping timeout: 264 seconds)
10:31andygraybeal_ has joined IRC (andygraybeal_!~andy@h23.205.213.151.dynamic.ip.windstream.net)
10:34gbaman has left IRC (gbaman!~gbaman@host86-181-59-248.range86-181.btcentralplus.com, Remote host closed the connection)
10:36bobby_C has left IRC (bobby_C!~bobby@212-186-178-75.dynamic.surfer.at, Ping timeout: 252 seconds)
10:38monteslu has joined IRC (monteslu!~monteslu@ip68-109-175-67.ph.ph.cox.net)
10:41bobby_C has joined IRC (bobby_C!~bobby@212-186-178-75.dynamic.surfer.at)
10:52alkisg has left IRC (alkisg!~alkisg@plinet.ioa.sch.gr, Remote host closed the connection)
10:53ogra_ has left IRC (ogra_!~ogra_@p5098ed03.dip0.t-ipconnect.de, Ping timeout: 272 seconds)
10:54work_alkisg has joined IRC (work_alkisg!~alkisg@plinet.ioa.sch.gr)
10:55work_alkisg is now known as alkisg
11:03Gremble has joined IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginm.net)
11:03Gremble is now known as Guest76697
11:16willianmazzardo has joined IRC (willianmazzardo!~textual@187.4.15.116)
11:19ogra_ has joined IRC (ogra_!~ogra_@p5098ed03.dip0.t-ipconnect.de)
11:23adrianorg has left IRC (adrianorg!~adrianorg@177.156.57.190, Quit: leaving)
11:27adrianorg has joined IRC (adrianorg!~adrianorg@177.156.57.190)
11:53khildin has joined IRC (khildin!~khildin@ip-80-236-242-89.dsl.scarlet.be)
12:10mikkel has joined IRC (mikkel!~mikkel@93.176.85.50)
12:12alexqwesa has left IRC (alexqwesa!~alex@109.172.15.11, Ping timeout: 272 seconds)
12:17gbaman has joined IRC (gbaman!~gbaman@host86-181-59-248.range86-181.btcentralplus.com)
12:23bobby_C has left IRC (bobby_C!~bobby@212-186-178-75.dynamic.surfer.at, Ping timeout: 248 seconds)
12:41freedomrun has joined IRC (freedomrun!~freedomru@unaffiliated/freedomrun)
12:43alkisg is now known as work_alkisg
12:54gbaman has left IRC (gbaman!~gbaman@host86-181-59-248.range86-181.btcentralplus.com, Remote host closed the connection)
12:59telex has left IRC (telex!~telex@freeshell.de, Remote host closed the connection)
13:16andygraybeal_ has left IRC (andygraybeal_!~andy@h23.205.213.151.dynamic.ip.windstream.net, Ping timeout: 272 seconds)
13:26gbaman has joined IRC (gbaman!~gbaman@host86-181-59-248.range86-181.btcentralplus.com)
13:30alexqwesa has joined IRC (alexqwesa!~alex@109.172.15.11)
13:42gbaman has left IRC (gbaman!~gbaman@host86-181-59-248.range86-181.btcentralplus.com, Remote host closed the connection)
13:55bobby_C has joined IRC (bobby_C!~bobby@178-191-185-199.adsl.highway.telekom.at)
14:09bobby_C has left IRC (bobby_C!~bobby@178-191-185-199.adsl.highway.telekom.at, Ping timeout: 264 seconds)
14:24Guest76697 has left IRC (Guest76697!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginm.net, Quit: I Leave)
14:39Ark74 has left IRC (Ark74!~Ark74@189.214.42.128.cable.dyn.cableonline.com.mx, Quit: Saliendo)
15:02orygynz has joined IRC (orygynz!d433a2a5@gateway/web/freenode/ip.212.51.162.165)
15:02
<orygynz>
hi
15:02
anybody ?
15:08
<elias_a>
vmlintu: Nice blog post! "Managing Thousands of Linux Desktops With Puavo" http://labs.opinsys.com/blog/2014/01/21/managing-thousands-of-linux-destops-with-puavo/
15:08
orygynz: Yes?
15:21orygynz_ has joined IRC (orygynz_!d433a2a5@gateway/web/freenode/ip.212.51.162.165)
15:21
<orygynz_>
sorry disconnected
15:21
still here ?
15:21
<||cw>
orygynz: you are here
15:22
elias_a: that sounds quite cool
15:22
<orygynz_>
:)
15:22
i have some problem with my ltsp server on Xubuntu
15:22orygynz has left IRC (orygynz!d433a2a5@gateway/web/freenode/ip.212.51.162.165, Ping timeout: 272 seconds)
15:23
<orygynz_>
its to connect my client to ad authentification
15:24alexxtasi has left IRC (alexxtasi!~alex@unaffiliated/alexxtasi)
15:25orygynz_ has left IRC (orygynz_!d433a2a5@gateway/web/freenode/ip.212.51.162.165, Client Quit)
15:25orygynz has joined IRC (orygynz!d433a2a5@gateway/web/freenode/ip.212.51.162.165)
15:26
<orygynz>
:-|
15:26willianmazzardo has left IRC (willianmazzardo!~textual@187.4.15.116, Quit: Textual IRC Client: http://www.textualapp.com/)
15:27
<||cw>
ad auth is a pain. I've Centrify's free version and that seems to work reasonably well, but I haven't used it for think clients yet, just for samba and ssh, so it should work
15:27
think/thin/
15:27
<orygynz>
the integration is ok for the ltsp server, i can connect my ad users
15:27
but not with thin client
15:28
and i always found configuration with ldap
15:28
or i used winbind in pam and its works
15:28vmlintu has joined IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi)
15:42
<||cw>
you mean for localapps or something?
15:49
<orygynz>
i have a ltsp server
15:50
and i can connect with my users local or from my windows'users server
15:50
but if i try to connect with a thin client
15:51orygynz_ has joined IRC (orygynz_!d433a2a5@gateway/web/freenode/ip.212.51.162.165)
15:51
<orygynz_>
you don't speak french ?
15:55orygynz has left IRC (orygynz!d433a2a5@gateway/web/freenode/ip.212.51.162.165, Ping timeout: 272 seconds)
15:58
<orygynz_>
cw ?
15:58
i have some trouble with web irc
15:58
i don't know if you can read this
16:01gbaman has joined IRC (gbaman!~gbaman@host81-134-114-61.in-addr.btopenworld.com)
16:03
<||cw>
I do not
16:03mattcen has left IRC (mattcen!~mattcen@c110-22-201-130.sunsh4.vic.optusnet.com.au, Ping timeout: 252 seconds)
16:04
<||cw>
are you using https://kiwiirc.com/client/irc.freenode.net/?channel=#ltsp ?
16:05
hm, I though freenode offically moved to kiwi from qw, seems not. kiwi works a lot better
16:06
<orygynz_>
no im in irc from ltsp.org, but im not at home
16:06
<||cw>
anyway, I'd think if the ldap is working for ssh with domain accounts, it would work for ltsp too
16:07
<orygynz_>
i know... but i don't find the problem
16:07
in auth.log
16:07
i juste have
16:07
just
16:07
client ltsp sshd[2415] connection closed by 0.0.0.0 [preauth]
16:09
and i don't understand why the local users don't work on thin client...
16:11alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
16:11
<||cw>
so local and domain users both fail?
16:12
<orygynz_>
yes on thin client
16:12
<||cw>
that could be as simple as a ssh host key mismatch in the thin client image
16:12
<orygynz_>
yes its possible
16:12
ssh service status is running
16:12
<||cw>
that's the a common cause of " sshd[2415] connection closed by 0.0.0.0 [preauth]"
16:13
<orygynz_>
i try with ltsp-update-sshkeys
16:13
<||cw>
right, was just going to suggest that
16:13
<orygynz_>
where i can see if its this command do is job ?
16:14
<||cw>
check the /etc/ssh/ file in the chroot before and after
16:15
<orygynz_>
ok
16:15
i try
16:15
and the file ssh_know_hosts seems to be new
16:15
the hour changed
16:16
in
16:16
opt ltsp i386 etc ssh
16:17
its crazy !
16:17
:/
16:19
i don't understand...
16:22orygynz has joined IRC (orygynz!d433a2a5@gateway/web/freenode/ip.212.51.162.165)
16:22
<orygynz>
i missed something ?
16:22
no other ideai ? :/
16:22
idea
16:23
<alkisg>
Go to some other pc and type `ssh user@server`, can you login that way?
16:24orygynz_ has left IRC (orygynz_!d433a2a5@gateway/web/freenode/ip.212.51.162.165, Ping timeout: 272 seconds)
16:25
<||cw>
and switch to the kiwi client, it's for more stable
16:25
https://kiwiirc.com/client/irc.freenode.net/?channel=#ltsp ?
16:26* alkisg never had issues with http://webchat.freenode.net/?channels=ltsp
16:32
<orygynz>
ssh users@server works
16:32
sorry, i have to go
16:32
<alkisg>
Which distribution and version are you using?
16:32
OK, bb
16:32
<orygynz>
thanks for your time, i will come back later
16:32
Xubuntu 12.04.3
16:32
<alkisg>
!screen_02
16:32
<ltsp>
screen_02: To get a root shell on an Ubuntu thin client: https://help.ubuntu.com/community/UbuntuLTSP/ClientTroubleshooting#Using_a_shell_SCREEN
16:33
<alkisg>
Try ssh from there when you have time, then switch to vt7 and try again via ldm
16:33
This bypasses the keys problem
16:33
<orygynz>
ok thanks
16:33
!quit
16:33
<ltsp>
Error: You don't have the owner capability. If you think that you should have this capability, be sure that you are identified before trying again. The 'whoami' command can tell you if you're identified.
16:33orygynz has left IRC (orygynz!d433a2a5@gateway/web/freenode/ip.212.51.162.165, Quit: Page closed)
16:34
<alkisg>
vmlintu: nice article! :)
16:34
Did you say you have nvidia-current etc installed? Do you also have nvidia-173 and fglrx?
16:35
<vmlintu>
alkisg: thanks
16:35mikkel has left IRC (mikkel!~mikkel@93.176.85.50, Quit: Leaving)
16:36
<vmlintu>
alkisg: was the nvidia/fglrx question to me?
16:36
<alkisg>
yes
16:36
<vmlintu>
alkisg: yes, we have nvidia-current installed and some images have also fglrx, but we have pretty much no use for it
16:37
We are using kernel.org kernels nowadays, so fglrx requires a bit more work
16:38
<alkisg>
Wow, why so?
16:38
(kernel.org)
16:38
<vmlintu>
alkisg: device drivers mostly.. too many bugs in the versions Ubuntu is providing
16:38
And also NFS
16:39
<alkisg>
vmlintu: why not install a debian kernel instead?
16:39
I too saw issues with the ubuntu kernels
16:39
Ubuntu is trying to do too much for its manpower, and many things have been falling apart...
16:39
<vmlintu>
Haven't thought about using debian kernels, actually
16:40
<alkisg>
They also support cmov, non-pae systems etc etc
16:40
aufs, not overlayfs, but it even seems more stable
16:40
<vmlintu>
kernel.org kernels have been working nicely actually with overlayfs patched on top of it
16:41
This is what we are using: https://github.com/opinsys/opinsys-linux
16:41
<alkisg>
vmlintu: actually , do you have a few minutes to chat in general about your solution? As you know, we're in very similar situations, so I think it'd be good to exchange knowledge etc...
16:42
You're not using ltsp-cluster at all, right?
16:42
<vmlintu>
alkisg: sure
16:42
No, there's no ltsp-cluster at all
16:42
<alkisg>
Do you have a public repository?
16:42
ppa or something?
16:42
<vmlintu>
http://archive.opinsys.fi/
16:43
<alkisg>
And, using only LTS versions?
16:43
Or, did you ever thing about migrating to Debian?
16:43
<vmlintu>
There's Jenkins following the Github repos and it compiles and pushes to the apt repo
16:44
<alkisg>
Hmm I can't find any packages in http://archive.opinsys.fi/pool/ ...
16:44
<vmlintu>
Currently we are using 12.04 on the boot servers and 12.10 on everything else. We are planning to migrate to 14.04 next.
16:45
alkisg: it has git branches as sub directories. http://archive.opinsys.fi/git-master/ has the master branches
16:45
<alkisg>
Sorry, I meant repository of packages, .deb files...
16:45
Haha
16:45
OK I just saw what you meant
16:46
About nvidia-current, does it cause any problems on boot servers or non-nvidia clients?
16:47
<vmlintu>
alkisg: About Debian.. We've been thinking about it also, but haven't decided anything on that yet
16:47
<alkisg>
If not, maybe why should push the update-alternatives code upstream in ltsp for everyone?
16:47
We were thinking about moving to debian, if you guys also are going to do it at some point, it's an additional reason, as there will be more users with the same use case :)
16:48
<vmlintu>
alkisg: our boot servers do not have desktop or anything else installed.. basically just network services
16:48
<alkisg>
It's getting harder and harder to undo all the ubuntu modiciations that break things..
16:48
What about non-nvidia clients? Everything working fine there?
16:50
Wow you guys have a *lot* of stuff under https://github.com/opinsys/ - maybe it's time to upstream some of those? :)
16:50
<vmlintu>
alkisg: we have patched the nvidia package itself to get it working.. the packaging is here: https://github.com/opinsys/opinsys-debs/tree/master/packages/default/nvidia-graphics-drivers/debian
16:51
<alkisg>
vmlintu: cool, that's what I was thinking. Patching, or repackaging, the nvidia-* and fglrx drivers to a separate package that doesn't update alternatives etc, and leaves ltsp to do it
16:52
But I was afraid to invest too much time there...
16:52
<vmlintu>
Basically it works so that if nvidia is not defined, none of the libraries get used
16:52
<alkisg>
Does it need initramfs changes?
16:53
Haha, you've packaged cmaptools too?
16:53
<vmlintu>
Actually, I don't remember if it the current version changes initramfs. At some point it did, but we did rework it at one point.
16:54
<alkisg>
We got permission to package that, and uploaded it to our repository...
16:54
We really need to join forces on many things
16:54
What changes would you like to upstream to LTSP?
16:55
<vmlintu>
There are some packages that we do not put in archive.opinsys.fi because they have licensing issues. So we use those only internally.
16:57
<alkisg>
Ah, and another thing... We're thinking of switching to samba for authentication instead of libpam-sshauth or ldap... have you tried that?
16:58
<vmlintu>
We really need the laptop support in the images, so that would be the first on my list..
16:58
We are now using pam_krb5 and pam_sss depending on the device type. sssd for laptops
16:59
<alkisg>
You mention something about .diff images sent to the laptops by the servers, how is that implemented?
17:00
What I was thinking: download new.img or rsync its contents, if it's compressed btrfs instead of squashfs, while the laptop is up and running with old.img
17:00
<vmlintu>
It's rdiff between old and new image
17:01
<alkisg>
So you have 3 aufs/overlayfs mounts in the client?
17:01
i386.img, diff.img and tmpfs?
17:01
<vmlintu>
No, we download the rdiff and create the new.img from old.img using the rdiff
17:02
<alkisg>
It's squashfs?
17:02mattcen has joined IRC (mattcen!~mattcen@c110-22-201-130.sunsh4.vic.optusnet.com.au)
17:02
<vmlintu>
yes, squashfs currently
17:02
<alkisg>
I find it weird that you can diff it
17:02
I thought since it's compressed, _all_ content would be different...
17:03
<vmlintu>
Well, it's a compromise between download speed, required cpu power, disk usage, etc..
17:03
<alkisg>
Do the laptops have local ext4 partitions?
17:04
and/or swap partitions/
17:05
<vmlintu>
alkisg: /images, /home, /tmp, /state and swap
17:05
<alkisg>
Cool
17:05
<vmlintu>
images has old and new images
17:05
home has normal user directories
17:05
<alkisg>
Can they also boot windows?
17:06
And login happens with pam caching?
17:06
<vmlintu>
state has things like puavo device credentials, ssl client certificates, logs, etc
17:06
we do not really support dualbooting windows on laptops
17:07
sssd caches the kerberos logins and its cache is put under /state
17:07
<alkisg>
Cool. So, users that haven't logged in before, cannot do it while the laptop is offline, ok, no big deal there.
17:08
<vmlintu>
So offline logins work after user has logged in once. sssd can even fetch kerberos ticket for the user when the network comes back up.
17:09
E.g. samba mounts from the boot server work with kerberos ticket so the user doesn't need to type password there
17:09pppingme has left IRC (pppingme!~pppingme@unaffiliated/pppingme, Ping timeout: 260 seconds)
17:10
<alkisg>
Does kerberos need ldap? I mean, if we want to keep it as simple as possible, can we only use samba + kerberos? Or, samba + sssd? What would be the simplest way there?
17:10pppingme has joined IRC (pppingme!~pppingme@unaffiliated/pppingme)
17:11
<alkisg>
There's also the possibility to just mount the samba home dir, and copy the ssh keys of the client in order for passwordless ssh to any server, but it sounds a bit hackish...
17:11
<vmlintu>
alkisg: laptops are actually not using ldap now, but Puavo sessions that include user information + some configurations
17:11
alkisg: kerberos doesn't need ldap
17:12
<alkisg>
Cool
17:12
How come you haven't yet "applied" to be an upstream ltsp developer? :)
17:13
Or at least, open launchpad bug reports for all the code that you'd like to see merged upstream?
17:14
<vmlintu>
that's a good question
17:15
We have actually changed most of the LTSP code so much that I'm not sure if it's LTSP anymore..
17:16
<alkisg>
We've wondered many times if it's worth it to develop in upstream LTSP anymore, having rewritten so much of its code
17:17
It would be much simpler to start from scratch
17:17
We stayed because of the community, not only the devs but the users too
17:17
Your changes solve actual problems
17:17
<vmlintu>
My plan was to write more blog posts about our solution so that the changes would be understandable to others too
17:17
<alkisg>
So I think LTSP should support them in one form or another...
17:18
<vmlintu>
But our code is also integrated to Puavo, which is not good from LTSP's point-of-view
17:18
<alkisg>
You can use hooks
17:19
For example, sch-scripts installs a greek version of lts.conf, when the user runs ltsp-config lts.conf
17:19
There's no mention of sch-scripts anywhere in ltsp-config, it just has appropriate hooks
17:20
<vmlintu>
yes, our code should be made more like that..
17:21
<alkisg>
For example I also want to remove all the ltsp-cluster code from ltsp, but first I'll need to create appropriate hooks for all the configuration phases it needs
17:21
<vmlintu>
I never got ltsp-cluster to work myself
17:22
We have replaced it with puavo's load balancer now
17:22
<alkisg>
All those can be integrated to ltsp
17:22
It can just have hooks in the appropriate places... configuration, load balancers, pam, whatever
17:23
<vmlintu>
we are now configuring quite a bit of stuff here: https://github.com/opinsys/puavo-ltsp/blob/master/client/puavo-configure-client
17:23willianmazzardo has joined IRC (willianmazzardo!~textual@187.4.15.116)
17:23
<vmlintu>
Some of the configuration is done from upstart jobs also
17:24* alkisg doesn't like upstart, it's distro-specific
17:24
<alkisg>
When new ltsp versions are released, do you need a lot of time to adapt them to your needs?
17:24
Or your code works fine even as ltsp is changing?
17:25
E.g. does puavo client depend on ltsp-client, or it replaces it completely?
17:25
<vmlintu>
puavo-ltsp-client replaces most of ltsp-client
17:26
<alkisg>
So ltsp-client is installed, or it isn't?
17:26
<vmlintu>
ltsp-client is installed, but e.g. we are not using ldm, but lightdm
17:27
But we are not using init=/sbin/init-ltsp, but init=/sbin/init-puavo
17:27
<alkisg>
Why? Couldn't you put puavo in /usr/share/ltsp/init-ltsp.d/00-puavo?
17:27
<vmlintu>
To answer your question - I haven't been really updating to the new ltsp versions
17:27freedomrun has left IRC (freedomrun!~freedomru@unaffiliated/freedomrun, Quit: So long and thanks for all the fish)
17:28
<alkisg>
That's where I'm getting at... you've basically forked ltsp
17:28
Which makes both of them have less manpower
17:29
<vmlintu>
The dhcp client code killed laptops
17:30
<alkisg>
I don't know what part you're speaking of, but a bug report should be filed there...
17:30
<vmlintu>
The laptop support required extensive changes to so many places, that I really cannot say that it's only a bug report
17:31Gremble has joined IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginm.net)
17:31
<alkisg>
It doesn't matter if it's a "master bug report with many little ones" or some other form
17:31Gremble is now known as Guest13852
17:32
<vmlintu>
My plan was to write about these in a series of blog posts so that different bits could be understood
17:32
<alkisg>
The main idea is if people want to join forces with upstream, or fork... forking is fine to quickly draft/solve things , or when upstream is not responsive, but then it's harder to maintain and gives less manpower to both projects
17:33
<vmlintu>
yes.. forking was not really our plan
17:34* alkisg was tempted to fork, or rather, rewrite, ltsp, many times in the past...
17:36
<alkisg>
Thanks for the chat vmlintu, I hope you'll ask to join ltsp-upstream at some point in the future :)
17:36
<vmlintu>
Looking back now, we started going to different direction when we started using kerberos and ltsp couldn't be changed to use it
17:37
<alkisg>
I think it should be changed though
17:37
<vmlintu>
I know that LTSP is about being simple, but kerberos really enabled us to do everything else
17:37
<alkisg>
LTSP may ship with e.g. libpam-sshauth as a base, but it should support anything pam-*
17:39
<vmlintu>
If LTSP were to go towards kerberos, I'd be really interested in participating
17:39
And laptops too
17:39
<alkisg>
vmlintu: if LTSP supported pam, wouldn't that be enough for kerberos?
17:40
LTSP is about having a simple base *that has hooks about anything* :)
17:42
<vmlintu>
PAM support does most of it, yes
17:42
<alkisg>
So yup ltsp is already going that way
17:43
It only lacks manpower, not direction
17:44
We've dropped xatomwait, cdpinger, will drop ldm etc... and at that point with all that code dropped we'll support more than we ever did before
17:45
LTSP can be made small and simple, but able to support a very wide range of use cases
17:46
<vmlintu>
what do you think of netbooting the ltsp servers?
17:46
<alkisg>
I've adviced dgross to do that 2-3 years ago :)
17:46
<vmlintu>
are you doing it?
17:46
<alkisg>
I've no need for it
17:46
Schools here have 1 server
17:47
And their internet connections are usually 2 mbps, to little to netboot stuff
17:47
*too
17:48rmav has joined IRC (rmav!89cfe03e@gateway/web/freenode/ip.137.207.224.62)
17:49
<alkisg>
Greece has lots of small schools, 50 to 400 students, so 1 server is usually enough. If more are needed, they're usually completely separated even if they're on the same school
17:49
<vmlintu>
What we do is that the boot server has minimal network services + ldap + kerberos kdc + kvm. Most boot servers have enough power to run a virtual kvm-ltspserver and some places have also physical hardware netbooting
17:49
<rmav>
I'm looking for some help getting Fluxbox working on an LTSP client.
17:49
<vmlintu>
In this setup we can change Ubuntu version of the LTSP server without doing actual installation
17:49
<alkisg>
rmav: write the problems you're experiencing, and wait to see if anyone answers.. :)
17:50
<vmlintu>
alkisg: or we can run multiple ubuntu versions side-by-side when testing updates
17:50
<alkisg>
Schools here also don't have a lot of money, so in a good scenario they only give 500€ for a new modest server and a switch
17:51
And they switch their existing workstations from e.g. windows 2000 to ltsp thin clients...
17:51
So putting a VM layer would require more money, which we don't have
17:51
We advice them to use 2 partitions, one for the previous installation and one for the new
17:52
<vmlintu>
here we own most of the servers and the school owns the client machines
17:52
<alkisg>
So if e.g. 12.04 fails for them, they reboot the server and work in 10.04 in a couple of minutes, and we fix the problem later on when we have time
17:53laprag_ has left IRC (laprag_!~ragnar@ti0071a400-4884.bb.online.no, Remote host closed the connection)
17:53
<alkisg>
vmlintu: which parts are public, funded by the government there? Are most schools private? Your company?
17:53
<vmlintu>
Cities fund their schools. There are no private schools here.
17:54
<alkisg>
Ah, same here, except for very few ones
17:54
With the crisis cities don't have money for schools, though...
17:54
<vmlintu>
I didn't understand all parts of your question, though
17:55
<alkisg>
My team is part of the ministry of education here, and we mostly don't use private companies to support schools...
17:55
Is that the case there too?
17:55
<rmav>
Our application servers are clustered Ubuntu 12.04. The main issue is ldm isn't showing any available sessions, only Default and Failseafe xterm. I used to be able to just start fluxbox from the user profile when I was using Ubuntu 10.10, but this version seems to be tied more tightly to the Unity desktop. I need to be able to run fluxbox for a controlled online testing environment, so only some user accounts need it.
17:55
<alkisg>
rmav: check if ldm-server is installed on your server
17:56
<vmlintu>
alkisg: it depends - some cities outsource all IT for their schools, some do it internally, some leave it to the "volunteer" math teacher
17:56
<alkisg>
And try `ldminfod localhost` on your server
17:57
<rmav>
Of course it is. ldminfod returns a number of available sessions, but the client doesn't list them. I've been running an LTSP cluster for the last 4 years.
17:57Gadi has joined IRC (Gadi!~gadi@pool-108-21-187-8.nycmny.fios.verizon.net)
17:57
<alkisg>
Wowwww Gadi!!!
17:57
!g
17:57
<ltsp>
g: Gadi!!!!!!!!!!!!!!!!!!!!!!!!
17:57
<Gadi>
alkisg!!
17:58
<alkisg>
Loooooooooong time no see!!!
17:58
<Gadi>
too long
17:58
way too long
17:58
<alkisg>
How are you buddy?
17:58
<Gadi>
thanks God, how are you?
17:58
<alkisg>
rmav: and if you run it from the client SCREEN_02?
17:58
Gadi: a bit overworked, but fine otherwise
17:59
<Gadi>
so, basically, nothing's changed ;)
17:59
<alkisg>
Vagrant has mentioned that you've changed jobs?
17:59
Haha
17:59
True
18:01kwmiebach has left IRC (kwmiebach!sid16855@gateway/web/irccloud.com/x-qxvjtzatrecfjsnh, Ping timeout: 245 seconds)
18:03kwmiebach has joined IRC (kwmiebach!sid16855@gateway/web/irccloud.com/x-ouumzbrlflctnngw)
18:03
<rmav>
alkisg: SCREEN_02 isn't configured, only SCREEN_01 and SCREEN_07. lmdinfod isn't available on the chroot client in SCREEN_01.
18:04
<alkisg>
Let me find you the netcat line...
18:05
nc server 9571
18:05
It should return the sessions
18:06
<rmav>
Well, that doesn't return anything. Obviously 'server' isn't resolving.
18:07
<alkisg>
From screen_01?
18:07
Try with the IP
18:07
<rmav>
alkisg: Yes, from scree_01
18:07
I'll try the IP
18:08
alkisg: Still nothing.
18:08
<alkisg>
Check firewall etc
18:09
$ grep 9571 /etc/inetd.conf
18:09
9571 stream tcp nowait nobody /usr/sbin/tcpd /usr/sbin/ldminfod
18:09
That's on the server, it should reply to clients
18:11
<vmlintu>
alkisg: how seriously are you thinking of debian?
18:12
<alkisg>
vmlintu: the main thing holding me back is the lack of ubiquity there
18:12telex has joined IRC (telex!~telex@freeshell.de)
18:13
<rmav>
alkisg: xinetd is installed. There is no entry for ldminfod.
18:13
<vmlintu>
alkisg: haven't really used it myself..
18:13
<alkisg>
vmlintu: But the ubuntu live cd has various other issues that I'm starting to get tempted to create a simple installer and a grub-based debian live cd...
18:14
We already have boot code in ltsp, we wouldn't even need casper
18:14
And the live cd could function as a live ltsp server too
18:15
<vmlintu>
alkisg: we are using preseeded usb sticks to do boot server installs now
18:15
<alkisg>
We've been remastering our own live CDs, unpacking the ubuntu .iso, doing changes, repacking it
18:15
Because it was too broken, it didn't even support typing greek from 10.10 and on
18:16
And the Ubuntu developers ignore all the filed bug reports...
18:16
<vmlintu>
But I've been thinking of creating a read-only squashfs/btrfs image that could be used on boot servers the same way as we use on laptops..
18:16
<alkisg>
We've been pretty disappointed by many of the Ubuntu-developed software, unity, indicators, lightdm, ubiquity etc, they do minimal testing and they don't fix issues even years after they appear
18:17
<vmlintu>
We built our own application menu for the desktop
18:17
lightdm works now after fixing some fork issues in it
18:18
But yes, the quality hasn't always been great
18:18
<alkisg>
Lightdm, accountsservice, ubiquity... we can't type greek there. With gdm everything worked fine.
18:19
<vmlintu>
hmm.. I should check if the russian support works everywhere.. I've been testing it only after login..
18:19
<alkisg>
If Debian decides to adopt systemd instead of upstart, I think it'll be the final straw at least for me, I'll move there :)
18:20
Lightdm breaks the keyboard layout after login too. So, with autologin it's fine, but when actually using lightdm, the keyboard layout switching doesn't work inside the user session
18:20
<vmlintu>
weird
18:21
<alkisg>
The developers think that all countries use a single layout, e.g. "ru". We're using two, "us,gr"
18:21
So they only keep the first one, "us"
18:21
And now in 14.04 it's even worse, the Ubuntu keyboard indicator doesn't work either
18:21
So that's the 4th program they break wrt international keyboard support
18:22
And all of them work fine e.g. in Debian...
18:22
<vmlintu>
How is the greek layout setup? Are there different variations of it so that you need two?
18:23
<alkisg>
The characters are not latin, so we need a keyboard shortcut (alt+shift) which changes from typing english, to typing greek
18:23
english <alt+shift> ελληνικά <alt+shift> english again
18:23
<vmlintu>
so us,gr defines that alt+shift switches between those two?
18:24
<alkisg>
cat /etc/default/keyboard
18:24
XKBLAYOUT="us,gr"
18:24
XKBVARIANT=","
18:24
XKBOPTIONS="grp:alt_shift_toggle,grp_led:scroll"
18:24
It's the XKBOPTIONS that defines the alt+shift part
18:25
<vmlintu>
ok.. never had to deal with those myself
18:25
<alkisg>
It's been working in xorg for ages
18:25
Ubuntu broke it 4 years ago, and they're not working on fixing it
18:25vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc)
18:26* alkisg waves to vagrantc
18:26
<vmlintu>
xorg does quite a few things nicely there.. I was really surprised when icelandic letters just work..
18:27
<alkisg>
vagrantc: so, vmlintu says that if LTSP is OK with supporting kerberos, he'll start contributing more upstream :D
18:27
<vmlintu>
I better start learning shell again :)
18:28
<alkisg>
xorg and gdm 2 were fine. gdm 3 broke things, but they fixed it in just a few months
18:28
Lightdm yet didn't, years later...
18:28
vmlintu: I think python is fine, at least for server-side parts
18:28
Ah, ruby... dunno
18:29
<vmlintu>
Puavo is now all ruby
18:29
<alkisg>
Is an "apt-get install ruby" enough? I.e. it only adds 2-3 MB?
18:30
<vmlintu>
alkisg: https://github.com/opinsys/puavo-ltsp/blob/master/debian.default/control#L30
18:31
<alkisg>
Any particular reason to select ruby over e.g. python, which seems to be a bit more standard in the linux world?
18:32
<vmlintu>
When we started coding Puavo, python's ldap support was not even comparable..
18:32
<alkisg>
Strange... but a good reason, of course
18:33
<vmlintu>
It's built on top of ruby on rails and activeldap
18:35
So because the server side is ruby, we are building also the client side mostly on ruby
18:35vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Ping timeout: 276 seconds)
18:35
<vmlintu>
I should try to create a demo organisation in Puavo for you so that you can try it out
18:37
<rmav>
alkisg: I tried reinstalling ldm-server and got this waring:
18:37
alkisg: Setting up ldm-server (2:2.2.9-1ubuntu0.1) ... Note: xinetd currently is not fully supported by update-inetd. Please consult /usr/share/doc/xinetd/README.Debian and itox(8).
18:38
alkisg: So I've installed openbsd-inetd and now have ldminfod in /etc/inetd.conf
18:38
<alkisg>
rmav: yes inetd is usually preferred and more easy with ltsp than xinetd
18:38
So if it's the same to use, completely purge xinetd...
18:38
*to you
18:39
<rmav>
alkisg: Purged. Which services should I restart to see results?
18:40
<alkisg>
openbsd-inetd
18:42vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc)
18:52
<rmav>
alkisg: Thanks so much for the insight. The Ubuntu installer should just replace xinetd with openbsd-inetd. I should be able to get everything working in the next half hour.
18:53rmav has left IRC (rmav!89cfe03e@gateway/web/freenode/ip.137.207.224.62, )
19:04Guest13852 has left IRC (Guest13852!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginm.net, Quit: I Leave)
19:08vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Ping timeout: 245 seconds)
19:25vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc)
20:00willianmazzardo has left IRC (willianmazzardo!~textual@187.4.15.116, Quit: Computer has gone to sleep.)
20:30khildin has left IRC (khildin!~khildin@ip-80-236-242-89.dsl.scarlet.be, Remote host closed the connection)
20:40vmlintu has left IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi, Read error: Operation timed out)
20:40willianmazzardo has joined IRC (willianmazzardo!~textual@187.4.15.116)
21:02staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Ping timeout: 248 seconds)
21:08staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu)
21:40vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Ping timeout: 276 seconds)
21:59alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Remote host closed the connection)
22:02gbaman has left IRC (gbaman!~gbaman@host81-134-114-61.in-addr.btopenworld.com, Remote host closed the connection)
22:12gbaman has joined IRC (gbaman!~gbaman@host81-134-114-61.in-addr.btopenworld.com)
22:22PhoenixSTF has joined IRC (PhoenixSTF!~rudi@78.29.154.124)
22:40alexqwesa has left IRC (alexqwesa!~alex@109.172.15.11, Ping timeout: 248 seconds)
22:54vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc)
23:19willianmazzardo has left IRC (willianmazzardo!~textual@187.4.15.116, Quit: Textual IRC Client: http://www.textualapp.com/)
23:26andygraybeal_ has joined IRC (andygraybeal_!~andy@h23.205.213.151.dynamic.ip.windstream.net)
23:33staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Ping timeout: 246 seconds)
23:42staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu)
23:52telex has left IRC (telex!~telex@freeshell.de, Quit: leaving)
23:54telex has joined IRC (telex!~telex@freeshell.de)
23:56telex has left IRC (telex!~telex@freeshell.de, Client Quit)
23:58Gremble has joined IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginm.net)
23:58Gremble is now known as Guest93977