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


Channel log from 18 January 2013   (all times are UTC)

00:05mark___ has joined IRC (mark___!7b75ba88@gateway/web/freenode/ip.123.117.186.136)
00:07mark___ has left IRC (mark___!7b75ba88@gateway/web/freenode/ip.123.117.186.136, Client Quit)
00:27Parker955_Away is now known as Parker955
00:40PhoenixSTF has left IRC (PhoenixSTF!~rudi@78.29.134.164, Remote host closed the connection)
00:49andygraybeal_ has joined IRC (andygraybeal_!~andy@h117.195.213.151.dynamic.ip.windstream.net)
00:55andygraybeal_ has left IRC (andygraybeal_!~andy@h117.195.213.151.dynamic.ip.windstream.net, Ping timeout: 276 seconds)
01:03
<otwieracz>
Currently:
01:03
Pseudo-terminal will not be allocated because stdin is not a terminal
01:03
Thats how SSH forwarding dies.
01:29ltspuser_32 has joined IRC (ltspuser_32!62d38da6@gateway/web/freenode/ip.98.211.141.166)
01:29
<ltspuser_32>
hello
01:29
anyone here?
01:30ltspuser_32 has left IRC (ltspuser_32!62d38da6@gateway/web/freenode/ip.98.211.141.166, Client Quit)
01:33
<vagrantc>
patience is so difficult...
01:33
or rather, patience for the impatient is difficult.
01:35shawnp0wers has left IRC (shawnp0wers!~spowers@151.236.4.166, Quit: leaving)
01:45shawnp0wers has joined IRC (shawnp0wers!~spowers@linuxjournal/staff/shawnp0wers)
02:14vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Quit: leaving)
02:23shawnp0wers has left IRC (shawnp0wers!~spowers@linuxjournal/staff/shawnp0wers, Ping timeout: 272 seconds)
03:09adrianorg__ has left IRC (adrianorg__!~adrianorg@177.156.57.29, Ping timeout: 276 seconds)
03:16shawnp0wers has joined IRC (shawnp0wers!~spowers@linuxjournal/staff/shawnp0wers)
03:30shawnp0wers has left IRC (shawnp0wers!~spowers@linuxjournal/staff/shawnp0wers, Remote host closed the connection)
03:36shawnp0wers has joined IRC (shawnp0wers!~spowers@151.236.4.166)
03:36shawnp0wers has joined IRC (shawnp0wers!~spowers@linuxjournal/staff/shawnp0wers)
04:20Parker955 is now known as Parker955_Away
05:04staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Ping timeout: 260 seconds)
05:16Guest59642 has joined IRC (Guest59642!~chatzilla@49.248.129.178)
05:17telex has left IRC (telex!~telex@freeshell.de, Ping timeout: 260 seconds)
05:24telex has joined IRC (telex!~telex@freeshell.de)
05:36telex has left IRC (telex!~telex@freeshell.de, Ping timeout: 256 seconds)
05:39telex has joined IRC (telex!~telex@freeshell.de)
05:40telex has joined IRC (telex!~telex@freeshell.de)
05:42work_alkisg is now known as alkisg
05:54alkisg has left IRC (alkisg!~alkisg@plinet.ioa.sch.gr, Quit: Leaving.)
06:07work_alkisg has joined IRC (work_alkisg!~alkisg@plinet.ioa.sch.gr)
06:24bauerski has joined IRC (bauerski!~witekb@frodo.psp.opole.pl)
07:24ltspuser_66 has joined IRC (ltspuser_66!de6b6df3@gateway/web/freenode/ip.222.107.109.243)
07:26
<ltspuser_66>
Hi there, I want to configure my local bluetooth dongle to use bluetooth headset at ltsp-client. is this possible?
07:27work_alkisg is now known as alkisg
07:50vmlintu has joined IRC (vmlintu!~vmlintu@cs180030.pp.htv.fi)
07:58dobber has joined IRC (dobber!~dobber@213.169.45.222)
08:10ltspuser_66 has left IRC (ltspuser_66!de6b6df3@gateway/web/freenode/ip.222.107.109.243, Quit: Page closed)
08:15
<otwieracz>
Hey.
08:19F-GT has joined IRC (F-GT!~phantom@ppp121-44-167-233.lns20.syd7.internode.on.net)
08:23meamy has joined IRC (meamy!~hannes@pd95cdee4.dip0.t-ipconnect.de)
08:31sep has joined IRC (sep!~sep@40.211.jostedal.no)
08:49
<otwieracz>
knipwim: ping
08:50alkisg has left IRC (alkisg!~alkisg@plinet.ioa.sch.gr, Changing host)
08:50alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
09:08
<alkisg>
stgraber, jammcq, or anyone else with C knowledge: ldm.c spawns the ssh process with g_spawn_async_with_pipes()
09:08
That's probably why ssh dies with ldm
09:08
So we can't wait until X has died to call rc_files("stop")
09:09
The proper solution is for ldm to get redesigned so that it runs outside of X
09:09
I can do that in shell, but not in C. But stgraber prefers that ldm is not rewritten in shell.
09:10
So currently I've no other option but to revert rc_files("stop") to be called while ldm is still running,
09:10
and at that point to forcibly kill all user processes, wait for them to finish, unmount gvfs/fuse mounts etc
09:10
That's very unclean, but it's still better than the lost data problem we have now
09:11
So... if ldm is to be kept in .C, we need a coder to maintain it...
09:25
<Hyperbyte>
alkisg, how many lines is the ldm code? Is it a lot of code?
09:26
<alkisg>
Hyperbyte: no, it's small
09:26
250 lines, with 200 of them being just c-maintainance stuff
09:27
error checking, malloc, free, reading environment... won't be needed at all in shell
09:27
The problem is that stgraber and some of his colleagues splitted ldm into plugins
09:28
Those would need to be rewritten too
09:28
(ssh and rdesktop plugins)
09:28
Those can be rewritten in shell as well, no problem there,
09:28
and the greeter could be reimplemented in C, so that it's easier to maintain/theme etc
09:28
*in pygtk, sorry
09:29
Previously, they *were* implemented in shell and python
09:29
I don't know why they switched to C, it doesn't make any sense to me
09:36
<Hyperbyte>
Can you explain this to me a bit further?
09:36
You said ssh dies with ldm. What do you mean with this, what happens exactly? And why is running outside of X the solution?
09:38
<alkisg>
Hyperbyte: ldm runs inside X. And it spawns the ssh socket. So, if ldm dies, ssh dies. If X dies, then ldm AND ssh die.
09:38
After xorg dies, we want to do some cleanup. To cleanly unmount stuff like the sshfs mount
09:39
So... we would like to have the ssh connection still alive there, after xorg and the ldm greeter have died
09:39
At that point we would have access to the server to cleanup the remoteapps stuff, then we would cleanly unmount gvfs and the fuse mount etc
09:40
While currently, those mounts get killed with ldm, and our cleanup scripts don't work,
09:41
so the result is what markit reported, that one can see /home/username/.stuff (like .pulse, .gconf) on the client *locally* on the tmpfs
09:41
That's caused because when the user processes terminated, the sshfs connection has already died, so they can't write their data to the server when they close their files
09:41
So, that's the first time that data gets lost. There's a second time though,
09:42
When the same user tries to re-login, our code sees that /home/username/* already exists on the client, so it decides that it doesn't need sshfs,
09:42
so the user gets a desktop without his settings/files, and whatever he creates from then on, are lost on logout/reboot, as they're stored on the tmpfs
09:43
The solution would be to have ssh spawned in such manner that it would keep running after xorg dies
09:43
And for that to happen, ldm would need to run outside of X
09:43
I could do that in shell, but I don't know glib to do it in C
09:44
So the problem comes down to "we don't have an ldm coder, because it's written in C"
09:44
So we have noone to properly fix it, and all we can do currently is to implement hackish workarounds in shell
09:46blueonyx has joined IRC (blueonyx!~blueonyx@141.43.149.125)
09:47
<blueonyx>
hi, i have a fat client that spams my servers syslog with "nbdinfod connect"s. on the client i see xinit popping up again and again
09:48
but the client doesnt even have a physical screen
09:52
so what process runs /bin/sh /usr/share/ltsp/screen.d/ldm?
09:53
<alkisg>
screen-session
09:53
/usr/share/ltsp/screen_session
09:57
<blueonyx>
why does it get started over and over again?
09:57
<alkisg>
Do other fat clients work?
09:57
<blueonyx>
my users use x only through ssh
09:57
yes
09:57
<alkisg>
Maybe that one has a xorg problem
09:57
<blueonyx>
i have 3 and only 1 does this
09:57
<alkisg>
Check /var/log/Xorg.0.log.old
09:57
Or .7.
09:58
<blueonyx>
oh yea seg faults
10:01
<knipwim>
otwieracz: at work now
10:02
<otwieracz>
uhm.
10:02
<knipwim>
but this afternoon i'll probably be free to help
10:03
otwieracz: but i see blueonyx also having segfaults, maybe it's related
10:06
<otwieracz>
knipwim: No segfaults now, fixed it.
10:06
But now I'm getting:
10:06
<blueonyx>
xorg segfaults? how? :D
10:06
<otwieracz>
Pseudo-terminal will not be allocated because stdin is not a terminal.
10:07
blueonyx: Disabled driver acceleration.
10:07
X_OPTION_1="NoAccel"
10:07
Or something like this.
10:07
But I still can not log in.
10:07
X forward failed.
10:07
With this message about tty.
10:09
<blueonyx>
stackoverflow suggests "ssh -t -t"
10:19
<otwieracz>
Yep.
10:19
But *where*
10:23
<blueonyx>
where do get the error from?
10:23
<otwieracz>
ldm.
10:24
ldm std{out,err}
10:28
<knipwim>
otwieracz: i'm not entirely sure, but you might need the tunnel thing from the kernel
10:29
<otwieracz>
you mean tun/tap?
10:29
<knipwim>
CONFIG_INET_TUNNEL=y
10:29
<otwieracz>
But, for what?
10:29
<knipwim>
bacause X is forwarded over ssh
10:30
<otwieracz>
But I've got that tty errors.
10:32
I've modprobed them
10:32
(tunnel4 and tunnel6)
10:32
The same.
10:35
http://wklej.orgid/929110/
10:35
http://wklej.org/id/929110/
10:35
Below line 10 LDM is restarting.
10:43
<alkisg>
otwieracz: also try with LDM_DIRECTX=True in lts.conf
10:43
<otwieracz>
ok
10:44
<blueonyx>
is /etc/lts.conf really parsed on the fat client? my XSERVER and other X variables dont do a thing, and i cant reboot the client :/
10:45
<otwieracz>
file /home/otwieracz/.Xauthority does not exist
10:47alkisg is now known as work_alkisg
10:56adrianorg__ has joined IRC (adrianorg__!~adrianorg@177.204.77.9.dynamic.adsl.gvt.net.br)
11:10
<Hyperbyte>
I set X_BLANKING=3600 in lts.conf
11:10
I thought this would blank the screen after an hour of inactivity, but people are complaining that it goes blank much, much quicker than that... 10 minutes or so
11:11
Does anyone have any ideas why this doesn't work with a 1 hour timeout?
11:34Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas)
11:35
<meamy>
hi team, i got problems to install the the current nvidia driver for my fat client, could that be related to LTSP? The problems seems do be that the nouveau driver is not completely deactivate, all HOWTOs for that i find online didn’t have an affect on this problem, could it be that the nouveau driver is some how fix in the initrd.img for example?I know that makes not much sense but I am running out of option :D
11:35meamy has left IRC (meamy!~hannes@pd95cdee4.dip0.t-ipconnect.de, Quit: meamy)
11:36bobby_C has joined IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at)
11:47ltspuser_25 has joined IRC (ltspuser_25!c31772ca@gateway/web/freenode/ip.195.23.114.202)
11:47
<ltspuser_25>
Hi
11:47
I have a question
11:48
I know that I can change the ldm theme and put other images in background and logo...
11:48
but can I change the login and password fileds??
11:48
now I need to put the user first and then appears the password field
11:49
can I change the theme to appear the two fields (user and password)?
11:50
so when I have a ldm session I can see, on login, the login and password fileds
11:56
anyone?
12:02Guest59642 has left IRC (Guest59642!~chatzilla@49.248.129.178)
12:08JuJuBee has left IRC (JuJuBee!~mike_knic@24-148-112-102.ip.mhcable.com)
12:08
<Hyperbyte>
ltspuser_25, you'd have to rewrite parts of ldm
12:09
I think if you just leave it as is, you'll find your users will adapt.
12:09Billfeld has joined IRC (Billfeld!~Ozym@krlh-5f71da48.pool.mediaWays.net)
12:10
<ltspuser_25>
but I need to rewrite many things?
12:10
if you could tell me I can try...
12:10Billfeld has left IRC (Billfeld!~Ozym@krlh-5f71da48.pool.mediaWays.net)
12:12
<Hyperbyte>
ltspuser_25, sure, download the upstream LTSP source, edit it and recompile.
12:12
It's written in C.
12:13
<ltspuser_25>
so I can see it's not easy :P
12:14
<Hyperbyte>
I didn't try to discourage you earlier without a good reason. ;-)
12:14
TBH, more and more operating systems use one field for username/password.
12:14
Or interfaces in general. I wouldn't bother with it, people will get used to it.
12:16vnc786 has joined IRC (vnc786!~chatzilla@49.248.129.178)
12:17
<vnc786>
i have 10 machines with ubuntu 12.04 running through LTSP now i want to configure gnome settings of one user and apply that settings to others. how do i do that ?
12:21
<tarzeau>
hahaha
12:22
believe me. gnome. sucks.
12:22
we have 3000 users...
12:22
and we just don't configure ANYTHING
12:22
since all settings are getting incompatible with any future version of gnome, it just doesn't make sense. the efforts are for nothing
12:22
better go out, date girls, whatever, just not gnome configuration
12:23
<ogra_>
date girls that can do gnome configuration !
12:23
<tarzeau>
good idea, but really give up configuring gnome
12:24
better date one that's good in bed, or at least in the kitchen
12:58Parker955_Away is now known as Parker955
12:59bobby_C has left IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at, Ping timeout: 252 seconds)
13:07Parker955 is now known as Parker955_Away
13:08alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
13:23jammcq has left IRC (jammcq!~jam@c-69-245-75-255.hsd1.mi.comcast.net, Quit: Lost terminal)
13:28Da-Geek has joined IRC (Da-Geek!~Da-Geek@exch.imparando.com)
13:44gothaggis|limbo has left IRC (gothaggis|limbo!~gothaggis@corp.heavyhammer.com, Ping timeout: 256 seconds)
13:48gothaggis has joined IRC (gothaggis!~gothaggis@corp.heavyhammer.com)
14:03[conrad] has joined IRC ([conrad]!~mrice@rrcs-70-62-71-144.midsouth.biz.rr.com)
14:05
<[conrad]>
Is there a utility available that I could use from a thin client to allow me to find the value of a parameter available in lts.conf? ltsp-info is confirming the lts.conf file is found, but I'm not certain values I'm supplying in the file are being respected/applied.
14:10
<alkisg>
getltscfg -a
14:19dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174)
14:27bauerski has left IRC (bauerski!~witekb@frodo.psp.opole.pl, Quit: Leaving.)
14:37ltspuser_25 has left IRC (ltspuser_25!c31772ca@gateway/web/freenode/ip.195.23.114.202, Quit: Page closed)
14:51bobby_C has joined IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at)
14:54
<[conrad]>
alkisg: Thanks, much appreciated.
15:00blueonyx has left IRC (blueonyx!~blueonyx@141.43.149.125)
15:06vmlintu has left IRC (vmlintu!~vmlintu@cs180030.pp.htv.fi, Ping timeout: 248 seconds)
15:28staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu)
15:57jammcq has joined IRC (jammcq!~jam@c-69-245-75-255.hsd1.mi.comcast.net)
15:57
<jammcq>
hey all
15:57
highvoltage: congrats!
16:05komunista has joined IRC (komunista!~slavko@adsl-195-168-235-217.dynamic.nextra.sk)
16:05adrianorg__ has left IRC (adrianorg__!~adrianorg@177.204.77.9.dynamic.adsl.gvt.net.br, Ping timeout: 252 seconds)
16:06
<knipwim>
jammcq: his birthday?
16:07bobby_C has left IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at, Ping timeout: 240 seconds)
16:07
<jammcq>
knipwim: no, he's back on the App review board for Ubuntu
16:09
<Hyperbyte>
:o
16:09* Hyperbyte takes a bow
16:15simpleirc has joined IRC (simpleirc!~simpleirc@mobile-166-137-150-134.mycingular.net)
16:22Enslaver has joined IRC (Enslaver!~Enslaver@mobile-166-137-150-134.mycingular.net)
16:25[conrad] has left IRC ([conrad]!~mrice@rrcs-70-62-71-144.midsouth.biz.rr.com)
16:29Enslaver has left IRC (Enslaver!~Enslaver@mobile-166-137-150-134.mycingular.net, Read error: Connection reset by peer)
16:29simpleirc has left IRC (simpleirc!~simpleirc@mobile-166-137-150-134.mycingular.net, Read error: Connection reset by peer)
16:41dobber has left IRC (dobber!~dobber@213.169.45.222, Remote host closed the connection)
16:48mikkel has joined IRC (mikkel!~mikkel@80-71-132-15.u.parknet.dk)
17:07mikkel has left IRC (mikkel!~mikkel@80-71-132-15.u.parknet.dk, Quit: Leaving)
17:24kostas_ has joined IRC (kostas_!4f671c1e@gateway/web/freenode/ip.79.103.28.30)
17:31vagrantc has joined IRC (vagrantc!~vagrant@c-98-232-129-196.hsd1.or.comcast.net)
17:31vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc)
17:31
<highvoltage>
jammcq: oh, thanks I guess :)
17:32
jammcq: it's like, congrats! here's some more work for you! hehe :)
17:32Enslaver has joined IRC (Enslaver!~Enslaver@adsl-70-241-84-25.dsl.hstntx.swbell.net)
17:36
<alkisg>
vagrantc: got time to chat about the translation functions? _() etc ?
17:36
<vagrantc>
alkisg: sure!
17:36
<alkisg>
knipwim: ^ ?
17:37
vagrantc: thoughts on that? https://bugs.launchpad.net/ltsp/+bug/1008237
17:37
You're in favor of _() and of $(), right?
17:37
And, any ideas on how to align multiple-line usage?
17:37
Like the tftp-boot-dir in the example, or see ltsp-chroot --help
17:38
<Enslaver>
Hey alkisg do you know who the fat client work is done by?
17:38
<alkisg>
And finally, are we going to use "text %s" or "text $VAR"?
17:38
<Enslaver>
I think I will take over for Warren, in a fashion.
17:39
<alkisg>
Enslaver: from what I remember, stgraber did a proof of concept implementation, highvoltage implemented the first plugin, and then I helped from then on
17:39
So if you have any questions there, feel free to ping me
17:39
<Enslaver>
Yeah I could use a minute of your time when you get one
17:40
Thats one of the major things that hasn't been ported over in fedora/RHEL that I was able to get semi-working.
17:40
<alkisg>
Shoot, since I'm here in IRC anyways talking about the translations... np with mixing subjects
17:41
<Enslaver>
ok one of the things i noticed was the implementation was obviously designed for ubuntu, the paths and commands are different in RHEL which was not an issue
17:41
<alkisg>
Enslaver: we're using distro-overridable functions now
17:42
<Enslaver>
The main thing I am having problems with is getting the concept down of how it does its mime associations
17:42
<alkisg>
Did you get some recent ltsp to boot in RHEL?
17:42
One with init-ltsp.d ?
17:42
<Enslaver>
distro-overridable? free desktop stuff?
17:42
yes
17:43
<alkisg>
Enslaver: and you had trouble with the code in /client/share/ltsp/init-ltsp.d ?
17:43
We shouldn't have any distro-depended code there
17:43
If there is, we need to extract it to distro-specific dirs
17:43
<Enslaver>
I have LTSP 5.2 running on our production server and the latest trunk on my development
17:43staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Quit: Leaving)
17:44
<alkisg>
Like e.g. client/Ubuntu/share/ltsp/init-ltsp.d
17:44
Enslaver: what mime associations? For localapps?
17:44
<Enslaver>
remoteapps
17:44
and localapps
17:45
<alkisg>
I think those were done by Gadi and mgariepy
17:45
<Enslaver>
I basically had to re-write remote apps for RHEL
17:45
<alkisg>
From what I remember I suggested implementing a generic ltsp-open command
17:45
But I haven't looked at the implementation, we're using plain fat clients here
17:45
No localapps/remoteapps
17:46
<Enslaver>
using nbd?
17:46
<vagrantc>
alkisg: oops, got sidetracked making breakfast
17:46* vagrantc reads backscroll
17:47
<alkisg>
Enslaver: yup
17:47
<Enslaver>
we were using thin clients over NFS, but the traffic is unmanageable and slowed us down, switching to NBD improved it but web browsing was terrible using flash
17:47
now we run a mix of local apps that call remote apps
17:48
but I can't seem to get the file type associations working across them
17:48
I saw where someone started to implement a tmpdir that would make its own .desktop files
17:49
in the file ~/ltsp-trunk/client/localapps/ldm-rc.d/X01-remoteapps
17:49
that's where a lot of my changes took place
17:49
for instance in RHEL there is no /usr/lib/mime
17:49
<vagrantc>
alkisg: definitely like _() and using $()
17:50
<Enslaver>
There is no update-mime command, although that may be a function that is called, yet not includede
17:51
<vagrantc>
alkisg: i'm leaning towards the printf %s style translations, just because it seems more consistant across languages
17:51
alkisg: across programming languages, that is...
17:51
<Enslaver>
Let me grab the latest nightly, things might be different since i last looked
17:51
<alkisg>
vagrantc: what I like about `` is that it's _not_ nestable (and that's what we want for translations), and that we can get used to seeing `` == translated string
17:51
<vagrantc>
alkisg: but that means some languages will have word-ordering issues?
17:51
<alkisg>
While $() is also reused for other commands
17:52
vagrantc: with our biiig texts... no, I wouldn't think so :D
17:52
I don't think we even have 1 string with 2 vars
17:52
<vagrantc>
alkisg: i'll never get used to seeing ', or is it `, or is it...
17:52
<alkisg>
vagrantc: and if we really have to, we can ship our own ltsp_printf function
17:52
<vagrantc>
alkisg: there are half-a-dozen characters that can be confused with `
17:53
<alkisg>
Which is a couple of lines in C code, the standard C library supports reordering vars
17:53
So it's just a .c file with a printf in it
17:54
vagrantc: ok, so, _(), $(), "%s", and, any ideas about wrapping usage functions?
17:54
<vagrantc>
alkisg: the example in LP#1008237 seems unreadable to me. i would much prefer $()
17:54
<alkisg>
vagrantc: we could use fmt + a bit of shell code to wrap translated usage functions
17:54
<vagrantc>
alkisg: i'm really not sure how to handle wrapped lines...
17:55
alkisg: in general, you want the smallest string that makes sense for translations
17:55
<alkisg>
$ ltsp-update-image --help
17:55
-f, --config-nbd Generate appropriate nbd-server configuration files.
17:55
It's automatically set if NFS isn't used or if other LTSP
17:55
generated nbd-server configuration files already exist.
17:55
-h, --help This message.
17:55
If we put $(_ 'It's automatically set if NFS isn't used or if other LTSP') in the translated strings,
17:56
it might not be aligned in all languages
17:56
So I'd go for ourfmt("It's automatically set if NFS isn't used or if other LTSP generated nbd-server configuration files already exist.") instead
17:56
<Enslaver>
did ltsp and ldm break off into 2 separate projects?
17:56komunista has left IRC (komunista!~slavko@adsl-195-168-235-217.dynamic.nextra.sk, Ping timeout: 252 seconds)
17:57
<alkisg>
vagrantc: It would make the code a bit more unreadable, but it would work with all languages + terminal widths
17:57
<vagrantc>
Enslaver: long, long ago, yes.
17:58
alkisg: so we don't bother to wrap lines in the strings, we format them on output?
17:58
<alkisg>
vagrantc: yes
17:58
<Enslaver>
vagrantc: and you're the ldm guru from what i see, correct? were you the one that came up with the concept?
17:59
<alkisg>
Enslaver: there are 3 branches, ltsp, ldm, ltspfs
17:59
<vagrantc>
Enslaver: i wrote something similar years before ldm existed
17:59
<alkisg>
Ah, and ltsp-docs
17:59
<vagrantc>
alkisg: ltspfs was always it's own project, and ldm split off from ltsp
17:59
<Enslaver>
Bravo on ldm, i gotta say, moving away from xdmcp was the best thing for ltsp
18:00
<vagrantc>
it's had it's difficulties
18:00* vagrantc is not really any more an ldm guru than any of the other committers
18:00
<Enslaver>
I can immagine
18:00
<vagrantc>
encryption takes a pretty heavy toll
18:00
many distros default to disabling encryption, i think
18:00
<Enslaver>
No one has pinned a remote desktop protocol down in a while, people have tried (RDP, NX, etc..)
18:01
<vagrantc>
but LDM_DIRECTX still is somewhat slower than using XDMCP for some reason
18:01
Enslaver: and now spice?
18:01
<Enslaver>
How many people you think will be using directs calls over XDMCP or even LDM?
18:01
<vagrantc>
Enslaver: "directs calls" ?
18:01
<Enslaver>
I can't say anything bad about spice
18:02
never used it
18:02
directX, i mean
18:02
<vagrantc>
ah,
18:02
LDM_
18:02
<alkisg>
(07:49:40 μμ) Enslaver: that's where a lot of my changes took place ==> if you want put them to pastebin to see how we can merge the 2 variants
18:02
<vagrantc>
er, LDM_DIRECTX has nothing to do with directX
18:02
<Enslaver>
they should be on my trunk on launchpad
18:03
<alkisg>
Link?
18:03
<vagrantc>
it is... simple, LDM with a DISPLAY variable set to go unencrypted
18:03
<Enslaver>
lemme see if warren posted them
18:03adrianorg has joined IRC (adrianorg!~adrianorg@187.113.244.29)
18:04* vagrantc is happy to see renewed involvement from the fedora family
18:04JuJuBee has joined IRC (JuJuBee!~mike_knic@24-148-112-102.ip.mhcable.com)
18:04
<vagrantc>
was kinda like the long-lost colony
18:04
<Enslaver>
you referring to access_control rights or carrying the display manager over an encrypted protocol or tunnel?
18:04
<alkisg>
Btw, is k12linux different than fedora?
18:04
Do we need 2 dirs there?
18:04
<Enslaver>
k12 is a spinoff of a fedora implementation
18:04
<vagrantc>
Enslaver: essentially, LDM is a GUI frontend for "ssh -X"
18:05
Enslaver: LDM logs in, but sets environment variables (and auth credentials) so that X goes "directly" over TCP
18:05
Enslaver: er, LDM_DIRECTX logs in...
18:05
<Enslaver>
Ah yes, i caught that, basically it makes all the gnome-desktop calls to setup the environment using ssh, which has been a hassle for us due to the ssh key issue
18:06
i noticed it much faster using blowfish encryption btw in the ssh
18:06* vagrantc seems to recall we originally preferred blowfish with *old* versions of ldm
18:06
<vagrantc>
using ECDSA would also likely be faster
18:07
<alkisg>
vagrantc: once me, knipwim and you change all the `eval_gettext` to $(_ "text %s" "$var") etc etc, can you also handle .pot generation code? It shouldn't be hard, but I'm completely unfamiliar with that part of the code...
18:07
<Enslaver>
But I gotta ask, why encrypt? I know its default in ssh, but unless you're compressing the packets its gonna be much much slower
18:08
<alkisg>
ssh doesn't support the null cipher
18:08
That's where LDM_DIRECTX comes in
18:08
So that X traffic doesn't go through the ssh tunnel
18:08
<vagrantc>
alkisg: i can try
18:08
<alkisg>
But directly to the client
18:08
<jammcq>
hey guys
18:08
<alkisg>
Hi jammcq
18:08* vagrantc waves to jammcq
18:09
<jammcq>
vagrantc: I loaded up a wheezy system
18:09
seems cool so far
18:09
<vagrantc>
jammcq: glad to hear it!
18:09
<jammcq>
I haven't installed ltsp yet
18:09
<vagrantc>
jammcq: wait at least a day- a new improved version should hit wheezy tomorrow
18:09
<jammcq>
you said something about there are some bugs that will take a few days for the fixes to make it into the repositories
18:09
<Enslaver>
jammcq: get it a cough drop and put on CentOS 6.3 :)
18:09
<jammcq>
any bug I should worry about?
18:09
anything that'll stop me from booting a thin client?
18:10
<vagrantc>
jammcq: basic thin clients should work fine
18:10
<jammcq>
cool
18:10
<vagrantc>
jammcq: some of the helper scripts, ltsp-config, have some pretty messed up bugs
18:10
<jammcq>
Enslaver: you pm'd me. sorry I didn't respond. crazy busy here
18:10
<Enslaver>
no worries, i know you're busy
18:11
been chatting with alkisg and vagrant.
18:11
<jammcq>
ah, 2 of the best
18:11
<Enslaver>
I've made a decision, I think i will take over for Warren, in a fashion
18:11
<jammcq>
fantastic!
18:12
the big trick will be getting packages accepted upstream
18:12
<vagrantc>
upstream?
18:12
<jammcq>
fedora might not be too tough
18:12
<Enslaver>
Heres the 1 thing
18:12
<jammcq>
well... into the distro
18:12
<Enslaver>
I really don't want to develop for fedora, I want everything I do to be on an enterprise class OS, locking versions of LTSP to only major releases
18:13
<jammcq>
but I wouldn't expect RH to accept them into RHEL so I don't know how Centos handles that
18:13
<Enslaver>
Doesn't warren work for RH?
18:13
<jammcq>
not anymore
18:13
he hasn't been at RH for quite a while
18:13
<Enslaver>
Not a big deal, I have my RH connections
18:13
<jammcq>
I think
18:13
I know he's in grad school in hawaii
18:14
<Enslaver>
But i probably wouldn't even submit it to be part of the OS
18:14
<jammcq>
when using RHEL, can you specify other repositories for pulling packages?
18:14
<Enslaver>
I can let you guys handle that, but we would need a rpm repo setup
18:14
yes
18:14
<jammcq>
that shouldn't be too bad then
18:14
<Enslaver>
just like dpkg/apt we use yum
18:14
well, we have apt also
18:15
/etc/yum.repos.d holds all .repo files
18:15
<JuJuBee>
Is it possible to have thin and fat clients from same server?
18:15
<Enslaver>
yes
18:16
<JuJuBee>
Enslaver: thanks, any good docs on how to do this?
18:16
<Enslaver>
JuJu: Look at the doc in relation to running 4.2 and 5 simultaneously, you can run 5 and 5 together the same way
18:16
<alkisg>
vagrantc: btw about the "sshfs gets disconnected before processes die, and that might cause data loss" issue, the only thing I can do with shell code is to forcibly kill all user processes at rc_files("stop")... I'm going to test+commit that in the weekend, unless someone has a better suggestion... For more explanations, see today's logs
18:16
<JuJuBee>
Enslaver: ok thanks
18:16
<alkisg>
JuJuBee: you don't need 2 chroots
18:17
Fat chroots can handle thin clients as well
18:17
<Enslaver>
JuJu: The ltsp-build-client can also do this for you
18:17
<alkisg>
You just specify LTSP_FATCLIENT=False for them
18:17
JuJuBee: as a matter of fact, there's also a FAT_RAM_THRESHOLD variable, set to 400 by default, which separates thins/fats based on client RAM
18:18
<Enslaver>
I gotta remember, not everyone is using rhel/fedora :)
18:20
question, why isn't ltsp-update-sshkeys called from ltsp-update-image?
18:20
<alkisg>
Good question :)
18:20
It should be
18:20
Enslaver: did you follow http://area51.stackexchange.com/proposals/46547/ltsp-linux-terminal-server-project ?
18:20
It needs 1 follower and 29 questions to become accepted
18:20
<Enslaver>
whats that?
18:20
<jammcq>
Enslaver: do you know about StackExchange?
18:20
<alkisg>
It's a questions/answers system for ltsp
18:21
<jammcq>
please please please, when voting for questions, vote for those with LESS THAN 10 votes
18:21
we need to get more of the questions up to the 10-vote level
18:22
<vagrantc>
Enslaver: you're proposing to maintain LTSP for RHEL outside of RHEL? that kind of goes against the principles of LTSP 5+
18:22
<Enslaver>
not familiar with stackexchange
18:22
i'm on launchpad
18:22
<vagrantc>
Enslaver: the whole point of LTSP 5 was to integrate into the distro.
18:22
<Enslaver>
vagrantc: tell something that isn't maintained outside of RHEL?
18:23
<jammcq>
vagrantc: that really won't happen with RHEL unless someone inside RH steps up to help
18:23
but if we could show RH that the work has been done, then we might get them to do something
18:23
<Enslaver>
I'm pretty sure, red hat, just like other distro's just people that take a collection of packages and put them together on top of a linux kernel
18:23
<vagrantc>
i can understand the proof of concept with outside repositories, but long-term it should get merged in
18:24
<jammcq>
yes, I agree with vagrantc
18:24
<Enslaver>
long term I'm sure it will
18:24
<jammcq>
the ultimate goal is to actually include it into RH
18:24
<vagrantc>
ok, just wanted to clarify the goal
18:24
<Enslaver>
they won't accept my package until they are sure it is stable and integrates well
18:25
<vagrantc>
i wouldn't expect any less :)
18:25
i think warren primarily worked through fedora because it was easier to get changes in
18:26
is CentOS essentially a rebuild of RHEL, or do they do significant CentOS-specific changes?
18:27
<Enslaver>
Nothing centos specific other than sed 's/redhat/centos/g' :)
18:27
they do leave the credits there and leave out any intellectual property
18:28
<vagrantc>
alkisg: i've heard a lot of the discussion regarding the ssh connection killing stuff, and suspect i don't have anything to contribute there that hasn't already been said
18:28
<Enslaver>
The problem with working with fedora, or was in the past was the changes are happening soo fast when you published an rpm for fedora 8 it would be broken in 3 months
18:29
<vagrantc>
Enslaver: yeah, i seem to recall warren uploading untested changes to ltsp to a "stable" fedora release, saying it should be in the repositories within 24 hours...
18:29
<Enslaver>
If we can implement a fully working snapshot of LTSP and make it stable and enterprise ready, i'm sure it will be adopted quickly
18:29
<vagrantc>
which boggled my mind.
18:30
<Enslaver>
Yes we are aware :( We are more careful than that here. Everything is tested and then retested before anything is implemented
18:30
I work for a multi billion dollar steel company doing millions of dollars a day in transactions, if we as so much even went down for 3 hours were *cked
18:31
<vagrantc>
Enslaver: well, looking forward to see what comes of it. happy to answer questions that help!
18:32
<Enslaver>
So i'm setting up a virtual machine with the latest ltsp-trunk on centos 6.3, i'll start implementing it today but for anything I have issues with could one of you guys possibly log on and see whats going on there? I know it will be related to that remote/local apps
18:33
I also will setup a virtual machine that boots the LTSP using dhcp that will have vnc access
18:33
Thats how i test most of my development work
18:33* alkisg isn't familiar with anything non-debian based, and hasn't been using localapps/remoteapps, but for anything else, I'd be glad to help
18:34
<vagrantc>
yeah, i'd be a bit lost, but will do what i can
18:36highvoltage has left IRC (highvoltage!~highvolta@ubuntu/member/highvoltage, Ping timeout: 260 seconds)
18:36highvoltage has joined IRC (highvoltage!~highvolta@ubuntu/member/highvoltage)
18:40JuJuBee has left IRC (JuJuBee!~mike_knic@24-148-112-102.ip.mhcable.com)
18:41
<Enslaver>
excellent, i can fill in any gaps os related
18:42
i've been using linux for almost 20 years now, god i feel old
18:42
<vagrantc>
heh
18:42
<effenberg>
you feel old, i'm old
18:43
<Enslaver>
what is old in the computer days these days? 40?
18:43
Because i feel 70
18:43
<vagrantc>
unix is even over 40
18:44
<Enslaver>
yeah i remember working on xenix and univac systems back in the day
18:45
xenix, aka MSUX
18:46
<effenberg>
# uname -rs
18:46
SCO_SV 3.2
18:46
YAY
18:46
<Enslaver>
# uname -n
18:46
18:46
I win
18:47
lol
18:47
<effenberg>
your host should get a name
18:48
<Enslaver>
names are so 2012
18:49komunista has joined IRC (komunista!~slavko@adsl-195-168-235-217.dynamic.nextra.sk)
18:49
<Enslaver>
Jim: how are you related to diskless workstations, do you work for them?
18:51
<jammcq>
Enslaver: I was one of the founders, but I left 5 yrs ago
18:52
<Enslaver>
I take it the Jammin 125 was named after you? :)
18:53
<jammcq>
umm, loosly I guess
18:53
<Enslaver>
We still have a few of those around, we run all diskless workstations here
18:54
We almost moved away from ltsp last year due to 5 only being able to run in ubuntu
19:04redbaritone has joined IRC (redbaritone!~dhall@mu163.aidt.edu)
19:09
<Enslaver>
Ok i'm starting my LTSP5 re-implementation for RHEL6, does anyone have a map, even if its for ubuntu of where each file is supposed to be located?
19:10
<alkisg>
Something like that? $ dpkg -L ltsp-client-core| pastebinit
19:10
http://paste.ubuntu.com/1546397/
19:10
<Enslaver>
that'll work, ty
19:10
<alkisg>
$ dpkg -L ltsp-server | pastebinit
19:10
http://paste.ubuntu.com/1546398/
19:11alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
19:14
<knipwim>
Enslaver: we have a wiki document for that
19:15
http://wiki.ltsp.org/wiki/Source_Hierarchy
19:15
and http://wiki.ltsp.org/wiki/Packaging and underlying docs
19:16
<vagrantc>
look at that!
19:16
<Enslaver>
I'll check there, The first link looks familiar
19:17
I think the issue i ran into was in the ltsp-build-client last time
19:18adrianorg has left IRC (adrianorg!~adrianorg@187.113.244.29, Ping timeout: 255 seconds)
19:20
<vagrantc>
the short of ltsp-build-client is ... it should bootstrap a minimal chroot, and then install the packages needed for ltsp
19:20
<whyzzyrd>
first unix-like I used was linux, but I did use RSX-11 and RT-11 after that
19:21
i 'acquired' a PDP-11/34
19:22adrianorg has joined IRC (adrianorg!~adrianorg@177.132.216.99)
19:22
<Enslaver>
couldn't the 'client' just be ubuntu? while keeping the server rhel?
19:23
<vagrantc>
knipwim: good work on the packaging docs
19:23
Enslaver: it's possible, yes...
19:23
Enslaver: somewhat outside the spirit of LTSP5...
19:23
<Enslaver>
yeah i'd never do it that way but i was curious
19:24
<knipwim>
vagrantc: thx :)
19:24
<vagrantc>
i've done a Debian chroot on an Ubuntu server (due to old thin-clients where old x86 hardware didn't work with ubuntu)
19:24
<knipwim>
although it's not quite finished
19:24
<vagrantc>
knipwim: sure, but it's a great start... people walking through it can hopefully contribute improvements
19:25
<Enslaver>
also why the ltspbr0 bridge? also why not just have servers broadcast dhcp using the OS's own dhcp configuration?
19:26
<otwieracz>
Hello.
19:26
knipwim: Have some time now?
19:26
<vagrantc>
Enslaver: i think that was one of warren's ideas ... i don't much use that sort of thing myself
19:26
<knipwim>
some, although i doubt i can be of assistance :)
19:26
<Enslaver>
We had to disable all that to get it working
19:27
<otwieracz>
knipwim: Hello.
19:27
knipwim: No ideas for my problem?
19:27
<vagrantc>
Enslaver: it's by no means a "standard" LTSP thing
19:27
<Enslaver>
ok good
19:27
<otwieracz>
knipwim: (that about pseudo-terminal)
19:27
<vagrantc>
i think the intention was to easily set up a bridge for testing with virtual machines, without risking mangling the existing network
19:28
<knipwim>
otwieracz: could you post the page again?
19:28
<vagrantc>
which i do sometimes, but i just configure it myself
19:28
<otwieracz>
knipwim:
19:28
http://wklej.org/id/929110/
19:28
After line 10 X (and LDM) is restarting.
19:28
<Enslaver>
well the good news for me then is i run both on virtual machines, client and server so it shouldn't be an issue
19:30
<knipwim>
otwieracz: where and when do you see this output?
19:31
<otwieracz>
knipwim: in /opt/ltsp/i686/usr/share/ltsp/screen.d/ldm:
19:31
#xinit $xinitrc /usr/sbin/ldm -- ${DISPLAY} vt${TTY} -auth ${XAUTHORITY} ${X_ARGS} -br >/dev/null 2>&1
19:32
xinit $xinitrc /usr/sbin/ldm -- ${DISPLAY} vt${TTY} -auth ${XAUTHORITY} ${X_ARGS} -br >>/var/log/ldm_debug.log 2>&1
19:32
So, I've found this in ldm_debug.
19:33* vagrantc forgets that vagrantc did one of the first LTSP5 "proof" of concept implementations
19:34
<vagrantc>
mostly just to stir the muck, figuring if someone from the Debian community could pull off a half-working LTSP implementation, surely someone from the fedora communitiy could produce a working onne
19:35
er, first protean ltsp5 implementations for fedora
19:38
<knipwim>
otwieracz: drawing at straws here, but, line 32 "Parse error on line 4 of section Device in file /var/run/ltsp-xorg.conf" ?
19:39
<otwieracz>
hm.
19:39
what?
19:39
<knipwim>
erhm, line 31 of the logfile you posted
19:39
<otwieracz>
Oh, thats right.
19:39
Lemme check.
19:41
Yeah, I'll probably fix it.
19:41
Wait a second.
19:42
http://wklej.org/id/929695/
19:44
Pseudo-terminal will not be allocated because stdin is not a terminal.
19:44
But this is on client or on server?
19:44
On server root or in server chroot?
19:44
<knipwim>
there has to be something else wrong, i'm seeing all kinds of error i've never seen
19:45
meanwhile, i found this, https://bugs.gentoo.org/show_bug.cgi?id=444442
19:46
<otwieracz>
I've already workaround this by disabling accel.
19:50adrianorg has left IRC (adrianorg!~adrianorg@177.132.216.99, Read error: Connection reset by peer)
19:50
<knipwim>
ok, the last logfile, that's output on the terminal right, after you rebooted the client? no manual actions?
19:51adrianorg has joined IRC (adrianorg!~adrianorg@177.134.62.190)
19:51
<otwieracz>
Yep, reboot, log-in, Ctrl+Alt+F1, root, password, and contents of /var/log/ldm_debug.log pasted.
19:52
<knipwim>
and on screen7 you see xorg restarting?
19:52
<otwieracz>
yep
19:52
Logging in, waiting for a while, and restart.
19:52
(while is about one second)
19:53
<knipwim>
can you post the contents of Xorg.7.log?
19:53
<vagrantc>
weird, i cannot find my original posting to ltsp-developer in the sourceforge archives, but it's referenced by other posts in the archive :(
19:53
<knipwim>
or whatever Xorg.log is there
19:54
<otwieracz>
http://wklej.org/id/929713/
20:02
<knipwim>
otwieracz: and your lts.conf?
20:03
<otwieracz>
euclid ~ # wklej /tftpboot/ltsp/i686/lts.conf
20:03
Link: http://wklej.org/id/929730/
20:06shogunx has left IRC (shogunx!~shogunx@2001:4978:106:1:b431:66eb:4a59:56af, Remote host closed the connection)
20:12
<knipwim>
i'm wondering if it has something to do with the LDM_LANGUAGE
20:13
gentoo hasn't really been tested for multilingual support
20:13
to be honest, this is really beyond me
20:14* vagrantc always forgets the language settings for LDM if they affect the greeter or the logged in session
20:14
<vagrantc>
i think there are variables for both
20:15
<knipwim>
according to the manual, LDM_LANGUAGE sets the server env, LANG is for the LDM ui
20:21
<vagrantc>
good work, docs!
20:21
<knipwim>
otwieracz: also, if didn't set a locale, the en_GB's are probably removed
20:22
and afk ...
20:32
<Enslaver>
ok first issue, ltsp-build-client is looking for the detect_vendor function but ltsp-build-client only includes ltsp-server-functions not ltsp-common-functions
20:33
<vagrantc>
Enslaver: i *think* ltsp-server-functions should source the apppropriate common-functions...
20:33* vagrantc doublecheks
20:35
<Enslaver>
Common functions shared by LTSP scripts
20:35
if [ -f /usr/share/ltsp/ltsp-server-common-functions ]; then
20:35
. /usr/share/ltsp/ltsp-server-common-functions
20:35
elif [ -f ../../common/ltsp-common-functions ]; then
20:35
# Sourcing build env common functions, so help2man can be called for ltsp-build-client
20:35
# the help2man script is called from /server/man
20:35
. ../../common/ltsp-common-functions
20:35
fi
20:36
<vagrantc>
Enslaver: ok, so i have this in my package build rules: install -m 644 common/ltsp-common-functions $(CURDIR)/debian/ltsp-client-core/usr/share/ltsp/ltsp-client-common-functions
20:36
Enslaver: essentially, common/ltsp-common-functions should get copied to ltsp-client-common-functions (ditto for server)
20:36
<Enslaver>
gotcha, i'll include it in my build rules as well
20:37
whats the server common functions file?
20:37
<vagrantc>
we did that to avoid having a tiny common package shared between the two
20:37
<Enslaver>
ltsp-common-server-functions?
20:37
<vagrantc>
install -m 644 common/ltsp-common-functions $(CURDIR)/debian/ltsp-server/usr/share/ltsp/ltsp-server-common-functions
20:37
ltsp-server-common-functions
20:40
and ltsp-server-vendor-functions are any functions you need to override what's in any of the previous *-functions files
20:41
<otwieracz>
knipwim: So, try to remove LDM_LANGUAGE?
20:52komunista has left IRC (komunista!~slavko@adsl-195-168-235-217.dynamic.nextra.sk, Quit: Leaving.)
20:55mgariepy has left IRC (mgariepy!mgariepy@ubuntu/member/mgariepy, Quit: Leaving)
20:59
<Enslaver>
where exactly is TFTPDIRS supposed to be set? warren is setting it in ltsp-update-kernels.conf but it needs to be set before 000-verify-tftpdir is called in the ltsp-build-client/plugins/distro
21:02
<vagrantc>
Enslaver: ltsp-build-client plugins, you're talking about?
21:02
<Enslaver>
yes
21:03
I think I'm gonna have to completely get rid of all of those vendor specific ones
21:03
none of them apply any more really
21:03
<vagrantc>
Enslaver: it should have a default of /srv/tftp, /var/lib/tftpboot and /tftpboot ...
21:03ltspuser_95 has joined IRC (ltspuser_95!5ce9fbc5@gateway/web/freenode/ip.92.233.251.197)
21:03
<vagrantc>
Enslaver: so whichever of those dirs are present, it populates
21:04Parker955_Away is now known as Parker955
21:04
<vagrantc>
Enslaver: so the only reason i would see for needing to set it is to make sure only one of them is used.
21:04
Enslaver: also, there are a lot of things that Fedora did in-chroot, that should now be handled in init-ltsp.d hooks at boot time.
21:08
<Enslaver>
Here we go, traced it back to ltsp-common-functions, it sources /etc/sysconfig/ltspdist
21:09
<vagrantc>
didn't want to use lsb_release for some reason...
21:10jammcq has left IRC (jammcq!~jam@c-69-245-75-255.hsd1.mi.comcast.net, Ping timeout: 252 seconds)
21:12jammcq has joined IRC (jammcq!~jam@c-69-245-75-255.hsd1.mi.comcast.net)
21:20
<Enslaver>
y?
21:21
<vagrantc>
Enslaver: i think warren said it had too big a dependency tree
21:22
<Enslaver>
:/
21:22
i'd prefer using LSB
21:22
That would solve all the issues with fedora/rhel
21:26
<ltspuser_95>
hello, please pardon my sudden entrance. Does anyone have experience booting igel 3210 clients? "Could not get nbi.img " This link preview.tinyurl.com/a6fuwnv shows a success story with same client hw
21:37
<Enslaver>
For whatever reason, ltsp-common-functions is not sourcing /etc/ltsp/ltspdist properly
21:37
err /etc/sysconfig i mean
21:40
ltsp-build-client sources ltsp-server-functions which sources ltsp-server-common-functions which has the detect_vendor function, which looks for /etc/sysconfig/ltspdist and is supposed to source it, but thats not getting sourced properly or something
21:44
<ltspuser_95>
is there a log for ltsp I can view? syslog does show the dhcp lease but nothing else
21:56
<vagrantc>
ltspuser_95: it's pretty unusual to fall back to the nbi.img, most clients should be using pxelinux.0
21:57
ltspuser_95: i don't know much about that particular hardware
22:00
<ltspuser_95>
thanks vagrantc for the pointer. I'll try and simplify and try another system to narrow down the issue. "BOOTP:connect to 192.168.1.1 and get /ltsp/i386/nbi.img" is what the client attempts.
22:02
<vagrantc>
ltspuser_95: i also added the following line to my dhcpd.conf:
22:02
# this should add vendor-class-identifier to dhcpd.leases files.
22:02
set vendor-string = option vendor-class-identifier;
22:02
then i can see the vendor strings in the /var/lib/dhcp/dhcpd.leases files
22:05
ltspuser_95: that's useful for setting specific options for specific clients.
22:10Parker955 is now known as Parker955_Away
22:11
<ltspuser_95>
vargrantc: thank you added that, useful
22:12adrianorg has left IRC (adrianorg!~adrianorg@177.134.62.190, Ping timeout: 246 seconds)
22:13
<knipwim>
Enslaver: the ltsp-server-common-functions sources /usr/share/ltsp/ltsp-server-vendor-functions
22:13
in Fedora, it consists of an overwrite for detect_arch()
22:15
http://bazaar.launchpad.net/~ltsp-upstream/ltsp/ltsp-trunk/view/head:/server/Fedora/share/ltsp/ltsp-server-vendor-functions
22:17xet7 has left IRC (xet7!~xet7@a91-156-233-189.elisa-laajakaista.fi, Quit: Lähdössä)
22:22
<Enslaver>
I found the problem, it only sources /etc/sysconfig/ltspdist in the detect_vendor function which isn't called until later on in the script. The fix was to remove it from detect_arch and added a check for it at the bottom of ltsp-common-functions then source it
22:22
I also found another issue in ltsp-update-image, when ionice is called, its called without the -c3 option which errors out in fedora/rhel
22:25adrianorg has joined IRC (adrianorg!~adrianorg@177.156.58.63)
22:29jammcq has left IRC (jammcq!~jam@c-69-245-75-255.hsd1.mi.comcast.net, Quit: leaving)
22:30
<Enslaver>
So i branch'd off a trunk of my own called ltsp-enslaver, now how do i submit my changes for approval?
22:32adrianorg has left IRC (adrianorg!~adrianorg@177.156.58.63, Read error: Connection reset by peer)
22:32adrianorg_ has joined IRC (adrianorg_!~adrianorg@177.134.58.12)
22:33dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Quit: Leaving...)
23:24Enslaver has left IRC (Enslaver!~Enslaver@adsl-70-241-84-25.dsl.hstntx.swbell.net, Quit: Leaving)
23:32Enslaver has joined IRC (Enslaver!~Enslaver@mobile-166-137-150-110.mycingular.net)
23:40
<Hyperbyte>
mhmm
23:41
Hi Enslaver :-)
23:48whyzzyrd has left IRC (whyzzyrd!~grimreape@chon2-adsl.demon.co.uk, Quit: sleep)
23:54Enslaver has left IRC (Enslaver!~Enslaver@mobile-166-137-150-110.mycingular.net, Ping timeout: 252 seconds)
23:56staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu)