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


Channel log from 9 January 2013   (all times are UTC)

00:05Parker955_Away is now known as Parker955
00:52
<Tommy100>
anybody up yet ?
01:39vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Quit: leaving)
01:57Tommy100 has left IRC (Tommy100!83cb4632@gateway/web/freenode/ip.131.203.70.50, Ping timeout: 245 seconds)
01:59adrianorg has left IRC (adrianorg!~adrianorg@177.134.56.253, Ping timeout: 252 seconds)
04:18vagrantc has joined IRC (vagrantc!~vagrant@c-98-232-129-196.hsd1.or.comcast.net)
04:18vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc)
04:19alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
04:20
<vagrantc>
alkisg: any recent fixes you think we really should have in wheezy?
04:20
<alkisg>
vagrantc: yes, but I don't know how to minimize them...
04:20
<vagrantc>
yeah, minimal is key.
04:20
<alkisg>
vagrantc: a couple of more patches will be needed to address the sshfs issue,
04:21
<vagrantc>
alkisg: the patch to support older initramfs without /run was incomplete...
04:21
<alkisg>
and an "allow-nonempty" parameter for /home/username mounting as well, to prevent lost user files
04:21
<vagrantc>
alkisg: lost user files?
04:21
<alkisg>
vagrantc: ah, is it in a runnable state? I haven't started testing yet
04:22
I hope I'll have time to finish this today or tomorrow
04:22
vagrantc: it goes like this:
04:22
User logs in on a client either using localapps or fat clients,
04:22
<vagrantc>
alkisg: wheezy is in decent shape, sid's LTSP is better, but i need to re-upload.
04:22
alkisg: in order to get the fixes in sid to make it to wheezy, i'll need to add a versioned dependency on newer initramfs-tools
04:22
<alkisg>
when he logs out, /home/username is forced-unmounted before the user processes end,
04:23
so when those end, they write back their data to /home/username *locally*
04:23
<vagrantc>
doh.
04:23
<alkisg>
Upon the next login, our code sees that /home/username is not empty, and sshfs mounting doesn't work
04:24
<vagrantc>
i got the impression there were problems, just didn't realize it actually resulted in data loss.
04:24
<alkisg>
So at _least_ an "allow-non-empty" sshfs parameter needs to go to wheezy (and backported to precise)
04:24
The result is that the user files are saved in the tmpfs locally
04:24
<vagrantc>
which is effectively data loss...
04:24
<alkisg>
They don't see any of their real files/settings, and anything they do doesn't get saved on the server
04:24
Yup
04:25
"versioned dependency" ==> if you're talking about backporting, I only care about precise+wheezy, I don't mind otherwise...
04:25
<vagrantc>
yeah, the versioned dependency is no big deal, and for the backport i can just remove the patch.
04:26
alkisg: the key is if precise's initramfs has /run mounted.
04:26
er, if it exists and is used
04:26
<alkisg>
I think it does
04:27
<vagrantc>
so that should be fine.
04:27
<alkisg>
initramfs-tools 0.99ubuntu13, booting a client to verify...
04:27
<vagrantc>
and if it's not, just leave those patches out
04:27
<alkisg>
This on my "server": tmpfs on /run type tmpfs (rw,noexec,nosuid,size=10%,mode=0755)
04:28
<vagrantc>
lsinitramfs /boot/initrd.img-$version | egrep /run
04:29
<alkisg>
...wow, lsinitramfs.... and I got to all this trouble to learn the gunzip | cpio commands... :D
04:29
var/run var/run/watershed bin/run-init
04:29
<vagrantc>
cd $(mktemp -d) ; zcat ... | cpio -i ?
04:30
alkisg: sounds like it'll need the patch
04:30
alkisg: er, shouldn't apply the patch
04:30
but i dunno, right now i just want to get those fixes into wheezy.
04:30
<alkisg>
vagrantc: can't the patch be written in a way that it works on all versions?
04:30
<vagrantc>
alkisg: it can, but that's not going to happen for wheezy at this point.
04:31
<alkisg>
vagrantc: in the initramfs (break=bottom), /run is mounted as tmpfs and does contain udev and 2 other dirs
04:31
udev, initramfs, and net-eth0.conf, specifically
04:31
<vagrantc>
alkisg: oh, good. that should be fine then.
04:31
<alkisg>
Cool
04:33
<vagrantc>
alkisg: so that sshfs related bug is prompting all the thoughts about rearranging the order of some scripts?
04:34
<alkisg>
vagrantc: that's the most pressing one, yes, but there were some other issues in the past wrt script ordering
04:35
<vagrantc>
i got one of them fixed already
04:35
<alkisg>
I ended up just saving/restoring the environment though, I didn't get to make ldm run outside the X session
04:35
Which one?
04:35
<vagrantc>
the one in ltsp_config.d
04:36* alkisg also has a "design a new configuration system" in his ltsp todo list... :-/
04:36
<vagrantc>
yeah...
04:37avijit has joined IRC (avijit!73739243@gateway/web/freenode/ip.115.115.146.67)
04:37Parker955 is now known as Parker955_Away
04:37
<alkisg>
And hopefully then the ltsp-cluster bits can be written as external plugins (if at all needed)
04:38
<vagrantc>
yeah, i've always wondered why those couldn't just be written as plugins ... i'd rather see ltsp/ldm/ltspfs's plugin system work than requiring ltsp-cluster-spcific code in ltsp itself.
04:40
alkisg: regarding what's missing in wheezy, http://anonscm.debian.org/loggerhead/pkg-ltsp/ltsp-debian-packaging/annotate/head:/changelog
04:41
alkisg: 5.4.2-2 is what's in wheezy, so the last two changelog entries are waiting to go in... and with a versioned dep on initramfs-tools, i think the release-team will let it in
04:42
maybe i should remove the incomplete backwards compatibility code from the patch...
04:42jammcq has left IRC (jammcq!~jam@c-98-209-67-190.hsd1.mi.comcast.net, Read error: Operation timed out)
04:44
<alkisg>
I thought we included a patch for both initramfs versions a few months ago...
04:45
<vagrantc>
alkisg: yes, an i included it, but it turns out to be incomplete
04:45
alkisg: it relies on /run/ existing
04:46
<alkisg>
Can't we `mkdir -p /run` if it doesn't?
04:46
<vagrantc>
simple "mkdir -p /run/" would fix it.
04:46
<alkisg>
:)
04:47cyberorg has joined IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg)
04:48
<vagrantc>
2382 Alkis Georgopoulos 2012-08-21
04:48
Background jetpipe (LP: #996533).
04:48
?
04:48
didn't that result in a failure to boot or something?
04:51
hmmm. 2419 Vagrant Cascadian 2012-12-07
04:51
Debian: Fix --debootstrap-keyring option for ltsp-build-client.
04:51jammcq has joined IRC (jammcq!~jam@c-98-209-67-190.hsd1.mi.comcast.net)
04:52
<vagrantc>
meh. i should just drop support for multistrap entirely.
04:52
it's so crude and already lead to some bugs...
04:53
alkisg: by the way, did you see my crazy hack to load the NBD image directly into ram and run without an NBD server?
04:53
<jammcq>
oooh, that sounds very interesting
04:53
<alkisg>
vagrantc: re: r2382, the failure to boot was before that fix, yeah
04:53
<vagrantc>
it's a model used by PXES and TCOS and such, and i finally got around to implementing it for LTSP, though it needs some polish to work right
04:54
<alkisg>
About the nbd-to-ram, indeed, you could probably even commit that part upstream inside some condition
04:54
Clever ;0
04:54
:)
04:55sha has joined IRC (sha!~sha@e177118020.adsl.alicedsl.de)
04:55
<vagrantc>
it would also benefit from further excludes to ltsp-update-image ... i.e. exclude /usr/share/doc, /usr/share/man and maybe some other things
04:56
<alkisg>
vagrantc: what's the main benefit/usage for that, though?
04:57
survive after server reboots?
04:57
<vagrantc>
that's one, yes.
04:57
<alkisg>
Why not just have a client reboot script for that?
04:57
<vagrantc>
alkisg: also many more thin clients per NBD server.
04:57
<jammcq>
I'm trying to help someone that wants to run hundreds of thin clients on a single boot server
04:57staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Ping timeout: 260 seconds)
04:57
<alkisg>
But clients do keep nbd "sectors" in their cache indefinately, no?
04:58
So I'd imagine that copying all the nbd image actually strains the server...
04:58
<vagrantc>
alkisg: don't know that it's handled on a per-sector level, maybe at the filesystem level.
04:58
<alkisg>
...because it makes the clients read all the image, even parts of it that will never be needed
04:58
<vagrantc>
that's a potential risk if the caching is good
04:59
<jammcq>
I'd be more worried about the entire image using up ram
04:59
<vagrantc>
alkisg: it also means no network traffic once the image is loaded.
04:59
it's really snappy once loaded :)
04:59sha_ has left IRC (sha_!~sha@d151070.adsl.hansenet.de, Ping timeout: 264 seconds)
05:00
<vagrantc>
but yes, the additional ram costs, for a standard thin client image was just below 300MB of ram just for the squashfs.
05:00
<alkisg>
Hmmm it'd be cool for kiosks... but I don't think I'd use it when a server needs to be around (and booted)...
05:00
E.g. for rdesktop cases, I'd put the nbd image in the windows server
05:00
<vagrantc>
alkisg: anyways, how important do you think r2382 is?
05:01
i'm thinking i'll just push the bare minimal of what's there now so we get something in wheezy
05:01
<alkisg>
vagrantc: I just committed what other people here reported that works, I haven't tested it,
05:01
...but if what they say is true, without it, specifying PRINTER_0_DEVICE makes the clients unable to boot
05:01
Although ogra was wondering how could that be true (jetpipe is supposed to daemonize itself)
05:02work_alkisg has left IRC (work_alkisg!~alkisg@plinet.ioa.sch.gr, Quit: Leaving.)
05:02
<vagrantc>
alkisg: i think i was able to reproduce it hanginng by putting anything in PRINTER_0_DEVICE
05:02
even totally invalid information
05:02
<alkisg>
Cool. So yeah better try to include that in wheezy.
05:03
<vagrantc>
i'll probably get the current stuff in first, at risk of it not mmaking it.
05:03
at risk of the other fixes not making it
05:04
<alkisg>
vagrantc: btw did you have any news from wouter? Did he stop working on nbd? Previously he was quick to reply on bug reports, and now it's been months since our last one..
05:07
<vagrantc>
alkisg: dunno. wouter's active in #debian-devel now and then
05:08
<alkisg>
vagrantc: ah, and sometime, maybe next week, could we look at the launchpad translations/project page/branches things?
05:08
<vagrantc>
alkisg: friday?
05:09
alkisg: or monday?
05:09
<alkisg>
Sounds good, both of them
05:10
<vagrantc>
so the fix in ubuntu regarding jetpipe was significantly rewrite it...
05:12
so the backgrounding of jetpipe wasn't good enough?
05:13
stgraber: ??
05:13
<alkisg>
bb in 30'
05:14alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
05:21telex has left IRC (telex!~telex@freeshell.de, Read error: Connection reset by peer)
05:21avijit has left IRC (avijit!73739243@gateway/web/freenode/ip.115.115.146.67, Quit: Page closed)
05:22telex has joined IRC (telex!~telex@freeshell.de)
05:25
<stgraber>
vagrantc: background worked but was a terribly ugly workaround as the jetpipe code was supposed to do that itself (as it depended on python-daemon)
05:26
vagrantc: the fix I did during the hackfest is dropping python-daemon and instead do a standard double-fork in pure python which then drops the need for the workaround
05:27
vagrantc: I don't think the code change was actually that big, though it may have caused a full re-indentation and I did use the opportunity to fix all issues found by pyflakes and pep8 (code style)
05:27
I didn't change anything in the way jetpipe works though, so from a user point of view, it should be identical
05:28* stgraber -> out
05:35telex has left IRC (telex!~telex@freeshell.de, Ping timeout: 248 seconds)
05:36telex has joined IRC (telex!~telex@freeshell.de)
05:40
<vagrantc>
stgraber: thanks for the update.
05:40work_alkisg has joined IRC (work_alkisg!~alkisg@plinet.ioa.sch.gr)
05:41* vagrantc will try something like that in phase 2.
05:42work_alkisg is now known as alkisg
06:04
<alkisg>
sbalneav, stgraber: endsession() in ssh.c gets called from ldm, so it doesn't get called in case of a crash
06:05
I can remove it from there and call it from a shell script instead
06:05
Or would one of you prefer to put some onexit() handler?
06:10
vagrantc: btw about the "ldm was implemented in C because of security issues" argument... the password value isn't cleared in ldm.c either, so... :D
06:11
<vagrantc>
alkisg: i'm merely parrotting arguments people made to me when i suggested we just write it in shell :)
06:12
alkisg: before ldm existed, i had already been using sdm :)
06:12
alkisg: s,had already been using,had already written,
06:12
<alkisg>
vagrantc: well, for 14.04, if lightdm isn't yet implemented, I'm willing to reimplement ldm in shell... :)
06:12
*lightdm/pamssh
06:13
<vagrantc>
the "LTSP is just the wind blowing through the trees, we don't really do anything ourselves" release.
06:13
<alkisg>
Oooh nice metaphor :)
06:15
OK I'm going to ignore X crashes for now and leave them up to our C guys...
06:25bauerski has joined IRC (bauerski!~witekb@frodo.psp.opole.pl)
06:28mah454 has joined IRC (mah454!~quassel@2.191.25.125)
06:36mah454 has left IRC (mah454!~quassel@2.191.25.125, Read error: Connection reset by peer)
06:52
<alkisg>
From localapps cleanup: rm $LOCALAPPSD_PIDFILE ==> what's that? I don't see any other references in the sources...
07:25MastaaK has joined IRC (MastaaK!c3dd3a29@gateway/web/freenode/ip.195.221.58.41)
07:27mah454 has joined IRC (mah454!~quassel@2.191.25.125)
07:27
<MastaaK>
Hi everyone, I tried to make an LTSP server (Ubuntu 12.04 on VirtualBox)with DHCP. The DHCP work fine, but when I tried to boot on network (internal network of VBox) my VM says she have nothing to boot on... But the DHCP tell her an right IP...
07:30
I tried to modify the dhcpd.conf to call the pxelinux.0 and the nbi.img which aren't in the default directory (on my server they are in /i386/boot/ and in the dhcp.cond they don't define the /boot/.
07:30
So if someone can tell me how to fix that or just a link that will be helpful
07:31
<vagrantc>
so, you're not using the standard location for tftp files?
07:32
and you're passing the correct "filename" parameter in dhcpd.conf?
07:33
<MastaaK>
When I make the client that was the default location... and I change the location in /etc/ltsp/dhcpd.conf too
07:34
Sorry, I forget to say it's a x86 server and x86 client
07:35
Did you know if I need to upgrade the pxe of Vbox or something?
07:36
<vagrantc>
i'm just having a hard time understanding what the situation is...
07:37
<MastaaK>
Sorry that I don't speack well, maybe you're french?
07:37
<vagrantc>
at least on debian and ubuntu systems, they'll appear in /srv/tftp/ltsp/i386/ or /var/lib/tftpboot/ltsp/i386/, and there will also be files in /opt/ltsp/i386/boot/
07:38
no parley francais?
07:38
solo un poco espanol
07:38
<MastaaK>
Pardon?
07:38
<vagrantc>
MastaaK: i don't speak french, only a little spanish
07:41
<MastaaK>
Okay, I'm trying what you say and I tell you if it works... or not
07:42mah454 has left IRC (mah454!~quassel@2.191.25.125, Read error: Connection reset by peer)
07:43
<vagrantc>
MastaaK: so in dhcpd.conf, it should specify /ltsp/i386/pxelinux.0
07:43
doesn't matter if it's in /srv/tftp or /var/lib/tftpboot, that's where tftpd usually serves up files.
07:44
MastaaK: the /opt/ltsp/i386/boot is only typically used for NFS, or the source for building an NBD image.
07:45
<MastaaK>
Okay, but pxelinux.0 are on /ltsp/i386/boot/pxelinux.0, is that noraml?
07:45
normal*
07:45
<vagrantc>
MastaaK: in your dhcpd.conf?
07:45
MastaaK: normally, it's just /ltsp/i386/pxelinux.0
07:45* alkisg notes that virtualbox already has a dhcp server, so MastaaK needs to be careful with his VM networking
07:46
<alkisg>
It should be: server: 2 nics, one in nat, the other in internal networking,
07:46
client: internal networking
07:46
<MastaaK>
no, in my dhcpd.conf it was marked what you said, but in my folder pxelinux.0 aren't in the directory that dhcpd.conf said
07:46
<vagrantc>
alkisg: when using dnsmasq, we could actually just have symlinks from /var/lib/tftpboot/i386/ -> /opt/ltsp/i386/boot/
07:47
<alkisg>
vagrantc: when that does exist, yeah, but it doesn't, in ltsp-pnp...
07:47
vagrantc: in ltsp-pnp we'd symlink to /boot, if not for vmlinuz access rights
07:47
<vagrantc>
alkisg: in ltsp-pnp it could just point to /boot ?
07:47
<alkisg>
rw------- 1 root root 4863712 Δεκ 5 20:52 vmlinuz-3.2.0-35-generic
07:47
Not world-readable
07:47
<vagrantc>
seems silly.
07:48
all this read-only junk in /boot seems silly.
07:48
<alkisg>
I agree... since anyone can just download the kernel from the repositories
07:48
<vagrantc>
or /var/cache/apt/archives if you haven't cleaned it
07:49
MastaaK: did you move the files in your tftp dirs?
07:49
MastaaK: what are they different?
07:49
<MastaaK>
i doesn't move anything
07:49
<vagrantc>
why are they different... ?
07:50
<MastaaK>
They are the same file
07:50
<vagrantc>
the defaults would not create such files.
07:50
MastaaK: so what is different about your setup?
07:51
<MastaaK>
How that?
07:52
<vagrantc>
MastaaK: you have a setup that is different from the usual setup, so how did it get that way?
07:52
<MastaaK>
I only install ubuntu, put open-ssh, ltsp standalone server and dhcp3-server and i made the client
07:53
<vagrantc>
MastaaK: so you don't *want* it that way?
07:53
MastaaK: you could remove the tftp files and run ltsp-update-kernels and see if they get recreated in the same way.
07:54
<MastaaK>
okay I try that but which way did you speack?
07:54
<vagrantc>
MastaaK: are they in /var/lib/tftpboot or /srv/tftp ?
07:55
MastaaK: sudo rm -rf /var/lib/tftpboot/ltsp && sudo ltsp-update-kernels
07:55
if it's in /var/lib/tftpboot
07:56
<MastaaK>
They are in /var/lib
07:56
<alkisg>
vagrantc: the reason we're not using an "ltsp-session" server-side script is that we want to be able to login to any ssh-capable server, even if ltsp isn't installed there, right?
07:57
vagrantc: so, if we generated+scp'ed an "ltsp-session" script to the server, and run that instead of the echo LTSPROCKS; ltspfsmounter cleanup etc, would it be ok?
07:57
<vagrantc>
alkisg: where would you put it?
07:58
alkisg: it needs to be a place that's writeable by the user, and executable as well.
07:58
<MastaaK>
The terminal says: Updating /var/lib/tftpboot directories for chroot: /opt/ltsp/i386
07:58
<alkisg>
vagrantc: somewhere in /tmp/
07:58
<MastaaK>
All right isn't it?
07:58
<vagrantc>
MastaaK: that's weird.
07:58
<MastaaK>
Arg!
07:58
<vagrantc>
alkisg: sometimes /tmp is mounted noexec, i think.
07:58
MastaaK: do you have files in /etc/ltsp/ ?
07:58
<alkisg>
vagrantc: sh file, it doesn't need to be executable
07:59
<MastaaK>
vagrantc: Yes, debconf.seeds dhcpd.conf ltsp-build-client.conf ltsp-update-image.conf
07:59
<vagrantc>
alkisg: now you're talking :)
08:00
MastaaK: what's in ltsp-build-client.conf and ltsp-update-image.conf ?
08:00
MastaaK: oh wait a minute! i misread what you said earlier.
08:01
<MastaaK>
ltsp-build-client: # needs some special care, so let's use it as an example. LATE_PACKAGES=" ubuntu-restricted-extras gimp nfs-client " # This is needed to answer "yes" to the Java EULA. # We'll create that file in the next step. DEBCONF_SEEDS="/etc/ltsp/debconf.seeds" # This uses the server apt cache to speed up downloading. # This locks the servers dpkg, so you can't use apt on # the server while building the chroot. M
08:01
<vagrantc>
MastaaK: how does /var/lib/tftpboot/ltsp look now?
08:02cyberorg has left IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg, Quit: cyberorg)
08:02
<MastaaK>
A folder named i386 come in with pxelinux.0, nbi.img and many othe
08:03cyberorg has joined IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg)
08:03
<vagrantc>
MastaaK: that sounds more like it should be.
08:03
MastaaK: you'll want to change your dhcpd.conf back.
08:04
<MastaaK>
So now in my dhcpd.conf I say to find the boot sequence in /var/lib/tftpboot/ltsp right?
08:04
<vagrantc>
MastaaK: just /ltsp/i386/pxelinux.0
08:04
MastaaK: leave out the /var/lib/tftpboot part.
08:04
MastaaK: hopefully that will help.
08:05
MastaaK: although what alkisg said earlier, virtualbox may have it's own dhcp server that will interfere
08:05
<MastaaK>
And in root path /var/lib/tftpboot/ltsp/i386/ areright?
08:06
<vagrantc>
no, root-path "/opt/ltsp/i386";
08:06
option root-path "/opt/ltsp/i386";
08:06
<MastaaK>
Okay, i restart the networking service and i tried
08:06
<vagrantc>
although with 12.04, if you're using NBD, i don't think root-path matters.
08:07
alkisg: ltsp-config dnsmasq continues to amaze me.
08:08
alkisg: i had an amd64 server setup, and it configured it for amd64 chroots. i deleted the amd64 chroot, created an i386 chroot, ran "ltsp-config dnsmasq --overwrite" and it figured out that i386 was the only available chroot :)
08:08
<alkisg>
;)
08:10
<vagrantc>
MastaaK: hope that helps! i've got to go now.
08:11
<MastaaK>
vagrantc: That doesn't help very much but I innderstand much easier how it works and thanks for your time!
08:11
<vagrantc>
other folks can hopefully help any last remaining parts
08:13vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Quit: leaving)
08:27autoditac has joined IRC (autoditac!~rouven@pd95cdee4.dip0.t-ipconnect.de)
09:27dobber has joined IRC (dobber!~dobber@213.169.45.222)
09:36shawnp0w1rs has joined IRC (shawnp0w1rs!~spowers@151.236.4.166)
09:42andygraybeal has left IRC (andygraybeal!~andy.gray@obsidian.casanueva.com, *.net *.split)
09:42shawnp0wers has left IRC (shawnp0wers!~spowers@linuxjournal/staff/shawnp0wers, *.net *.split)
09:50andygraybeal has joined IRC (andygraybeal!~andy.gray@obsidian.casanueva.com)
09:53yalu_ has joined IRC (yalu_!~yalu@91.180.85.133)
09:54yalu has left IRC (yalu!~yalu@91.180.83.86, Ping timeout: 248 seconds)
10:21Gremble has joined IRC (Gremble!~Ben@cpc29-aztw23-2-0-cust144.18-1.cable.virginmedia.com)
10:42Gremble has left IRC (Gremble!~Ben@cpc29-aztw23-2-0-cust144.18-1.cable.virginmedia.com, Quit: I Leave)
11:07alkisg is now known as work_alkisg
11:27adrianorg has joined IRC (adrianorg!~adrianorg@187.113.216.197)
11:37bobby_C has joined IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at)
12:35[GuS] has joined IRC ([GuS]!~MysT@213-117-16-190.fibertel.com.ar)
12:35[GuS] has joined IRC ([GuS]!~MysT@unaffiliated/gus/x-663402)
12:36PhoenixSTF has joined IRC (PhoenixSTF!~rudi@78.29.134.164)
12:48andygraybeal has left IRC (andygraybeal!~andy.gray@obsidian.casanueva.com, Remote host closed the connection)
12:58lotharn has left IRC (lotharn!~nick@24.154.55.32, Ping timeout: 272 seconds)
12:59lotharn has joined IRC (lotharn!~nick@24.154.55.32)
13:03MastaaK has left IRC (MastaaK!c3dd3a29@gateway/web/freenode/ip.195.221.58.41, Quit: Page closed)
13:03andygraybeal has joined IRC (andygraybeal!~andy.gray@obsidian.casanueva.com)
13:05vmlintu has left IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi, Ping timeout: 260 seconds)
13:07jammcq has left IRC (jammcq!~jam@c-98-209-67-190.hsd1.mi.comcast.net, Quit: leaving)
13:16highvoltage has left IRC (highvoltage!~highvolta@ubuntu/member/highvoltage, Ping timeout: 264 seconds)
13:16highvoltage has joined IRC (highvoltage!~highvolta@ubuntu/member/highvoltage)
13:18vmlintu has joined IRC (vmlintu!~vmlintu@37-136-136-150.nat.bb.dnainternet.fi)
13:35gvy has joined IRC (gvy!~mike@altlinux/developer/mike)
13:38Gremble has joined IRC (Gremble!~Ben@cpc29-aztw23-2-0-cust144.18-1.cable.virginmedia.com)
13:47alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
13:50Gremble has left IRC (Gremble!~Ben@cpc29-aztw23-2-0-cust144.18-1.cable.virginmedia.com, Quit: I Leave)
14:37bauerski has left IRC (bauerski!~witekb@frodo.psp.opole.pl, Quit: Leaving.)
14:51
<sbalneav>
alkisg: So, this weekend, I'm going to reconfigure both my work machine, and home machine to Debian, and get my devel environments set up for Deb. I'll dig into that list of 4 bugs this weekend.
14:51
Once I get them put to bed, I want to got nose-down on getting the pam stuff going.
14:52
<alkisg>
\o/
14:53
Veeeeery nice sbalneav!!!
14:53dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174)
14:54
<alkisg>
sbalneav: if you get the initial lightdm/pamssh work up and running, I'll probably be able to help with the sessions stuff
14:54
I.e. creating /usr/share/xsessions/* desktop files from init-ltsp.d
14:54
Installing appropriate lightdm hooks, cleaning up the ldm scripts etc
14:57* alkisg wonders if this one: https://launchpad.net/libpam-freerdp would work out of the box with lightdm...
15:34vmlintu has left IRC (vmlintu!~vmlintu@37-136-136-150.nat.bb.dnainternet.fi, Ping timeout: 265 seconds)
15:52alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
15:52alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
15:57alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Ping timeout: 240 seconds)
16:13alexqwesa has joined IRC (alexqwesa!~alex@109.172.12.47)
16:31Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas)
16:33dobber has left IRC (dobber!~dobber@213.169.45.222, Remote host closed the connection)
16:47alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
16:48highvoltage has left IRC (highvoltage!~highvolta@ubuntu/member/highvoltage, Remote host closed the connection)
16:55komunista has joined IRC (komunista!~slavko@adsl-195-098-005-235.dynamic.nextra.sk)
17:12mikkel has joined IRC (mikkel!~mikkel@80-71-132-15.u.parknet.dk)
17:32staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu)
17:34autoditac has left IRC (autoditac!~rouven@pd95cdee4.dip0.t-ipconnect.de, Quit: Ex-Chat)
17:39dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Read error: Operation timed out)
17:42highvoltage has joined IRC (highvoltage!~highvolta@ubuntu/member/highvoltage)
17:48komunista has left IRC (komunista!~slavko@adsl-195-098-005-235.dynamic.nextra.sk, Ping timeout: 240 seconds)
17:53dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174)
18:03komunista has joined IRC (komunista!~slavko@adsl-195-168-235-217.dynamic.nextra.sk)
18:16vagrantc has joined IRC (vagrantc!~vagrant@75-150-46-245-Oregon.hfc.comcastbusiness.net)
18:16vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc)
18:41
<alkisg>
Hrm... it looks like LDM r1454 will be needed after all, because when the ssh connection is (cleanly) ended, the sshfs $LDM_HOME is unmounted, and pulse + pulse-gconf-helper which stay for a bit after the session, end up writing their data locally...
18:42komunista has left IRC (komunista!~slavko@adsl-195-168-235-217.dynamic.nextra.sk, Quit: Leaving.)
19:24alexqwesa_ has joined IRC (alexqwesa_!~alex@109.172.12.47)
19:24yalu_ is now known as yalu
19:24alexqwesa has left IRC (alexqwesa!~alex@109.172.12.47, Quit: Хана X'ам !!!)
19:26garymc has joined IRC (garymc!~chatzilla@host86-140-178-9.range86-140.btcentralplus.com)
19:26
<alkisg>
vagrantc: I'm trying `debuild -b -tc` locally to get a new ldm executable, and it's working, but then it tries to launch wwm from /usr/local/libexec/ldm/wwm... What am I doing wrong, should I be running debuild as root in order foir it to look at /usr/lib/ldm/wwm?
19:34dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Quit: Computer has gone to sleep.)
19:35
<knipwim>
the /usr/local -> /usr would be the --prefix setting i guess
19:35
looking for the libexec...
19:36
<alkisg>
...sudo debuild did work... I hope there aren't any bad sideeffects there :-/
19:41mikkel has left IRC (mikkel!~mikkel@80-71-132-15.u.parknet.dk, Quit: Leaving)
19:43
<alkisg>
Meh... the ssh process is a child of ldm, so it's killed with ldm even if the ssh socket isn't closed
19:45
<vagrantc>
alkisg: you have fakeroot installed?
19:45
<alkisg>
vagrantc: yes
19:45
<vagrantc>
hmmm... debuild should use that automatically.
19:45
<alkisg>
It does
19:46
<vagrantc>
alkisg: i've never encountered that problem...
19:46
<alkisg>
When I tried running it with fakeroot, it told me that nested fakeroots aren't supported
19:46
With sudo debuild, the generated ldm.deb worked fine
19:47
<vagrantc>
i always build with "debuild -us -uc -S" and then sbuild on the resulting .dsc
19:47
once in a while i'll build it manually
19:47
<alkisg>
vagrantc: "sbuild on the resulting .dsc" ==> exact command for that?
19:48
Just sbuild and the .dsc ?
19:49alexqwesa_ has left IRC (alexqwesa_!~alex@109.172.12.47, Quit: Хана X'ам !!!)
19:49
<vagrantc>
depends on how you have sbuild configured, but i usually do: sbuild -d (UNRELEASED|unstable|testing) -c (sid|wheezy) foo.dsc
19:49
<alkisg>
Ty
19:50
<vagrantc>
it'll make some assumptions based on the debian/changelog, and so you'd need your schroot configurations setup to default to the right thing.
19:50
i.e. it's feasible to have "sbuild foo.dsc" do the righht thing
19:50
but i often enough need to do the wrong thing that i prefer to be explicit.
19:51
i.e. development/testing builds i never intend to upload
19:55ltspuser_37 has joined IRC (ltspuser_37!59680fc2@gateway/web/freenode/ip.89.104.15.194)
19:56[GuS] has left IRC ([GuS]!~MysT@unaffiliated/gus/x-663402, Quit: Konversation terminated!)
19:57
<vagrantc>
gak!
19:57
the powerpc64 vmlinux for precise is 25MB?
19:57
yaboot can't boot it anymore :(
20:01vmlintu has joined IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi)
20:13jammcq has joined IRC (jammcq!~jam@c-98-209-67-190.hsd1.mi.comcast.net)
20:17
<vmlintu>
alkisg: I remember you mentioning this once and now we ended up doing our own tftp server that supports dynamic scripts to generate the results (e.g. pxelinux.cfg/01-xx-xx-xx-xx-xx-xx and lts.conf): https://github.com/opinsys/puavo-tftp
20:20
<alkisg>
vmlintu: there's a python tftp-server module
20:20
<vagrantc>
nice.
20:20
<alkisg>
With that, it would be possible to have a single source of information for everything
20:20
TFTP (pxelinux, menus, 01-mac-address entries),
20:20
lts.conf, ldm-server,...
20:20
ltsp-cluster, ...
20:21
...without having to generate files for everything, the URL would be enough
20:22
<vagrantc>
sounds pretty nice.
20:22
<alkisg>
Although I'm not sure how stable the implementation is.. if it isn't, we'll have to resort to some "ltsp-apply-config" script that would generate the tftp config out of some /etc/ltsp configuration dir...
20:23
...and a python simplesocketserver for lts.conf + ldm-server + cluster
20:24
<vmlintu>
our implementation is done with ruby
20:25
<alkisg>
vmlintu: but you're reusing some existing tftp server, right?
20:25
Like tftpd-hpa or dnsmasq...
20:25
<vmlintu>
alkisg: no, it's full implementation
20:25
we've been testing it now to serve the pxelinux.cfg configs and lts.conf information from LDAP
20:26
<alkisg>
Ah, nice! Do the files need to be in the file system, or can they be served dynamically/generated from the code?
20:26
<vmlintu>
alkisg: it serves also the kernels and initrd images
20:26
<alkisg>
vmlintu: e.g. suppose we do something like "tftp get /ltsp/i386/lts.conf/mac=xxx/ip=yyy", can you return something based on the URL, or would you have to create the whole path?
20:26
<vmlintu>
alkisg: it supports both - you can give regular expressions and a script is called if they match
20:27dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174)
20:27
<alkisg>
vmlintu: and the script has to create the file, or can it return data in its stdout?
20:27
<vmlintu>
stdout
20:27
https://github.com/opinsys/puavo-tftp/blob/master/puavo-tftp.yml
20:27
that's an example configuration file
20:28
<alkisg>
That's very cool - I think the python implementation had some restrictions there, in order to implement the TSIZE tftp extension
20:28
<vmlintu>
puavo-tftpd supports also TSIZE
20:30
We have now lts.conf script that generates the contents from LDAP and the client is detected by searching the arp table for the client's ip address
20:30
<alkisg>
So it waits for the whole script stdout first, in order to determine the size of the output? Very good...
20:30
<vmlintu>
So the client does not have to send it's mac address either when fetching lts.conf
20:31
<alkisg>
vmlintu: is puavo-tftpd your own implementation, or someone else's ?
20:31
<vmlintu>
Yes, it waits for the script to finish
20:31
It's ours
20:31
<alkisg>
Cool! :)
20:31
<vmlintu>
And I just realised that the license file is missing from the repository. It's GPLv2+
20:32
<alkisg>
It makes me want to learn ruby! :D
20:32
<vmlintu>
All the dependencies are in Ubuntu repos, so no external gems are needed
20:32
<alkisg>
vmlintu: the idea about the /path/mac=xxx/ip=yyy is that the client can also send ram=rrr and cpu=ccc etc
20:33
Then the server can decide better about which configuration the client should get
20:33
E.g. if it would be a fat client, on which server it should connect, which image it needs (maybe based on its lspci output...)
20:33
...based on the boot phase (starting, login, session, logout...)
20:37
<vmlintu>
alkisg: Yes, that could also work.. We already have that information in LDAP
20:40
It's up to the script to parse the path, puavo-tftpd just need a regex that matches the path
20:41
<vagrantc>
yay. ltsp unblocked and on it's way to wheezy.
20:41
ugh, i should have set medium priority for such a trivial change... 10 days later...
20:41
<jammcq>
yay!
20:42
<vmlintu>
But we'll write some more documentation about puavo-tftp shortly so that also others can test it
20:42
I have to get going now
20:43dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Ping timeout: 260 seconds)
20:48dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174)
20:49vmlintu has left IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi, Ping timeout: 264 seconds)
20:53ltspuser_37 has left IRC (ltspuser_37!59680fc2@gateway/web/freenode/ip.89.104.15.194, Ping timeout: 245 seconds)
20:53alexqwesa_ has joined IRC (alexqwesa_!~alex@109.172.12.47)
21:13garymc has left IRC (garymc!~chatzilla@host86-140-178-9.range86-140.btcentralplus.com, Quit: ChatZilla 0.9.89 [Firefox 17.0.1/20121128204232])
21:20dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Ping timeout: 265 seconds)
21:46kev_j has joined IRC (kev_j!~kevin@63.225.43.68)
21:52gvy has left IRC (gvy!~mike@altlinux/developer/mike, Quit: Leaving)
21:59alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
22:04vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Quit: leaving)
22:15dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174)
22:19Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 276 seconds)
22:19dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Client Quit)
22:34dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174)
22:41dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Quit: Computer has gone to sleep.)
22:46dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174)
22:49dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Client Quit)
22:58staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Ping timeout: 255 seconds)
23:02kev_j has left IRC (kev_j!~kevin@63.225.43.68, Quit: Leaving)
23:18markit has joined IRC (markit!~marco@88-149-177-66.v4.ngi.it)
23:53markit has left IRC (markit!~marco@88-149-177-66.v4.ngi.it, Quit: Konversation terminated!)
23:57alexqwesa_ has left IRC (alexqwesa_!~alex@109.172.12.47, Quit: Хана X'ам !!!)