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


Channel log from 19 January 2014   (all times are UTC)

00:08freedomrun has joined IRC (freedomrun!~freedomru@unaffiliated/freedomrun)
01:03bennabiy has joined IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com)
01:32Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 252 seconds)
01:34Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas)
01:38rgttx has joined IRC (rgttx!iamparadox@c-71-206-130-134.hsd1.va.comcast.net)
01:39|Paradox| has left IRC (|Paradox|!iamparadox@c-71-206-130-134.hsd1.va.comcast.net, Read error: Connection reset by peer)
01:39rgttx is now known as |Paradox|
02:17alexqwesa has left IRC (alexqwesa!~alex@109.172.12.47, Ping timeout: 248 seconds)
02:22andygraybeal has left IRC (andygraybeal!~andy@h23.205.213.151.dynamic.ip.windstream.net, Quit: Ex-Chat)
02:55freedomrun has left IRC (freedomrun!~freedomru@unaffiliated/freedomrun, Quit: So long and thanks for all the fish)
03:01Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Quit: Leaving.)
03:06Parker955 is now known as Parker955_Away
03:07Parker955_Away is now known as Parker955
05:08kwaltman has joined IRC (kwaltman!~kwaltman@108.198.96.229)
05:17cyberorg has left IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg, Ping timeout: 260 seconds)
05:43alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
06:00kwaltman has left IRC (kwaltman!~kwaltman@108.198.96.229)
06:16Ark74 has joined IRC (Ark74!~Ark74@189.214.42.128.cable.dyn.cableonline.com.mx)
06:23alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Remote host closed the connection)
06:24alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
06:31Parker955 is now known as Parker955_Away
06:31Parker955_Away is now known as Parker955
06:51
<alkisg>
Eeewww /usr/share/upstart/sessions ???!!! Putting things in /etc/xdg/autostart is not enough anymore, we need distro-specific code?
06:52
<vagrantc>
alkisg: well, i must say, i find that lenny doesn't work anymore as almost good news :P
06:52
<alkisg>
Haha
06:52
<vagrantc>
alkisg: though it won't really be feasible to backport 5.5 to squeeze ...
06:53
<alkisg>
Ah... why so?
06:53
It'd be nice to have it in archives in the future
06:53
<vagrantc>
alkisg: well, using the official backports infrastructure
06:53
alkisg: typically only backports from one release to the next ... i.e. backports for squeeze come from wheezy
06:53
alkisg: backports for wheezy come from jessie
06:53
<alkisg>
Aaaah ok
06:54
So you can't have wheezy backports go to squeeze :P
06:54
<vagrantc>
though sometimes there are "sloppy" archives that allow other stuff
06:54
<alkisg>
OK ok forget it it's not that significant
06:54
<vagrantc>
dunno if those are enabled or not
06:54
<alkisg>
The most difficult thing was compiling ldm
06:54
<vagrantc>
of course, you can always publish an unofficial repository
06:54
<alkisg>
I bet ltsp 6 will be super-easy to backport anywhere...
06:55
<vagrantc>
that would be once nice benefit
06:55
alkisg: didn't get to the ltsp 5.5+ update yet
06:56
alkisg: not sure when i'll have a chance... kind busy this week
06:56
<alkisg>
vagrantc: no problem we have days left till 14.04 stops importing from debian
06:56
And I want to solve a couple of things more
06:56
nbd-disconnect being one of them...
06:57
<vagrantc>
alkisg: might not be able to get to it till february, honestly
06:57
<alkisg>
The code was wrong, it was looking for nbdroot, and we were using root=/dev/nbd
06:57
I fixed it, and now that it's running, it prevents people's systems from shutting down?!!!
06:57
<vagrantc>
heh
06:57* vagrantc always found nbd-disconnect to be suspicious
06:58
<alkisg>
vagrantc: 16 February 6th Debian Import Freeze
06:58
<vagrantc>
is it importing from jessie or sid?
06:58
<alkisg>
If you and stgraber can have a newer ltsp in trusty till then, it's fine, if not, I don't mind, I have my ppa - I'll just stop supporting people here that are using ubuntu and don't have the greek schools ppa in their sources :p
06:59
I think LTS releases import from testing, and non lts from sid
06:59
So, jessie
06:59
<vagrantc>
although i suspect stgraber can manually request an import from anywhere
06:59
but obviously, the sooner the better
06:59
<alkisg>
It's just bureaucracy... :)
07:00
Btw if debian ends up preferring systemd over upstart, I'll decide that I want to switch to debian for most of my systems :D
07:00
<vagrantc>
i should be able to dedicate some time the first few days of february, or maybe even get in an upload the week before
07:01* vagrantc finds the whole discussion silly
07:01
<alkisg>
vagrantc: ok if you can, ping me a couple of days before, to do some last minute uploads + testing
07:01
upstart breaks a lot of things in my user sessions
07:01
So while I don't care about the discussion, I do care about what they end up choosing...
07:02
<vagrantc>
people talk about sysvinit like the sky is falling or something
07:03* vagrantc is concerned all the dozens of init systems will just be passing phases
07:03
<alkisg>
I think systemd got too spread into various other projects to get removed from them any time soon...
07:03
<vagrantc>
basically, i'll adapt to whatever's decided for Debian
07:04
<alkisg>
And ubuntu just doesn't have the manpower to maintain half-systemd and half-upstart solutions
07:04
<vagrantc>
alkisg: which is the sort of anti-social coding I don't want to support
07:04
it's an insidious sort of vendor lock-in
07:05
<alkisg>
True. But they're at least active part of upstream for most of those projects.
07:05
<vagrantc>
*anyways
07:05
*
07:05
alkisg: get in your fixes in ltsp-trunk as soon asn possible!
07:06
<alkisg>
I'll try... but those left need much testing
07:06
<vagrantc>
just in case i find a spare few hours to upload
07:06
<alkisg>
So do allow me until the next weekend if you can
07:06
<vagrantc>
sure
07:07
i think my last tests at least covered your ltsp-update-kernels changes
07:07
<alkisg>
Everything ok there?
07:07
<vagrantc>
seemed to be, if my memory serves well :)
07:08
<alkisg>
I didn't update /usr/share/ltsp/cleanup.d/50-vmlinuz though, it's on my todo list
07:08* vagrantc wants to generalize the image-to-ram feature
07:08
<vagrantc>
i.e. AOE/NBD whatever
07:09
<alkisg>
Does it currently check the mount method?
07:10
<vagrantc>
it's all hard-coded to root=/dev/nbd0
07:10
<alkisg>
Ah, we don't include the script functions there so 'panic' is unknown
07:10
<vagrantc>
could probably generalize it by checking root=/dev/nbd|root=/dev/ether/ with a case statement or something
07:11
<alkisg>
Aaaah right.... stat can show the device
07:11
<vagrantc>
nbd-client -d exists
07:12
can check /proc/cmdline for root= and parse that, unless it's available as a variable...
07:12
<alkisg>
It's silly that nbd doesn't have a proper mount wrapper like nfs
07:12
So that plain unmount can disconnect it
07:13
<vagrantc>
sounds like a wishlist bug
07:14
of course nbd is currently failing to build on *all* 32-bit architectures :(
07:15
fix apparently in the works...
07:15
<alkisg>
Haha... and I was waiting for the authfile fixes to upload them to the ppa...
07:16
<vagrantc>
yeah, wait a little longer :)
07:18
alkisg: so basically 5.5.0 breaks grub because it's not able to extract the version from the symlinks?
07:18
alkisg: seems like that'd be good to fix in grub
07:18
<alkisg>
Yes, grub fails
07:19
True, but it'd be hard to backport the fix everywhere
07:19
And I've already sent ltsp 5.5 to all 12.04 school installations
07:20
<vagrantc>
alkisg: you've patched it to work? or you've just broken it everywhere?
07:20
i could just upload a new 5.5.0 with that patched, at least...
07:20
<alkisg>
http://bazaar.launchpad.net/~ltsp-upstream/ltsp/ltsp-trunk/revision/2539
07:21
You mean now?
07:21
<vagrantc>
maybe tomorrow
07:22
<alkisg>
OK, good enough
07:22
I'll try to upload the most pressing fixes today
07:22
<vagrantc>
it's an "important" if not release-critical bug, given that it can make a system unbootable.
07:22
<alkisg>
Well only ltsp-pnp systems, but yeah
07:23
<vagrantc>
alkisg: i'm talking about just updating with 2539 patched in the packaging, and leaving alll the other stuff for later
07:23
<alkisg>
(and they just select another proper entry from grub)
07:23
<vagrantc>
alkisg: yeah, that's why it'd be important
07:23
alkisg: rather than definitely release-critical
07:24
alkisg: i.e. leaving the other stuff for 5.5.1
07:24
<alkisg>
vagrantc: why not upload 5.5.1 instead?
07:24
<vagrantc>
alkisg: more changes == more testing
07:24
<alkisg>
And iiiiif there's time and enough commits, you could upload 5.5.2 later...
07:24
OK, but they're not that many
07:24
<vagrantc>
it's true
07:25
i'll need to test all those pointless translations!
07:25
<alkisg>
Hahaha
07:25
Basically I'd like to fix cleanup.d/50-vmlinuz and nbd-disconnect today, those are the pressing ones
07:25
So, 8 commits to test
07:26Patina has left IRC (Patina!~tomas@dhcp-5-103-105-144.seas-nve.net, Ping timeout: 264 seconds)
07:26
<vagrantc>
well, if you got them done, i *might* have time to upload tomorrow or monday...
07:26
<alkisg>
I'll try!
07:27Patina has joined IRC (Patina!~tomas@dhcp-5-103-105-144.seas-nve.net)
07:27
<alkisg>
vagrantc: why did you say that the nbd-disconnect logic is suspicious?
07:27
Do you think I should just delete it?
07:28
The idea is, disconnect nbd-client in the last moment, and use sysrq to shutdown after that
07:28
It works fine everywhere that I tested, but I get reports that in some systems it doesn't shut down...
07:30
Although, I liked the "lazy disconnect" idea much more, where nbd-client sends a message to nbd server to automatically disconnect after e.g. 1 minute
07:30
<vagrantc>
alkisg: it seems very prone to race conditions, no?
07:30
<alkisg>
The problem is the temporary swap files...
07:31
I don't see why though... processes are supposed to get a TERM signal
07:31
<vagrantc>
alkisg: i.e. it's relying on a binary loaded from the NBD root filesystem to shutdown the nbd connection...
07:31
<alkisg>
I hook that, disconnect nbd, and do the sysrq
07:31
nbd-client is surely accesible *before* nbd-client -d
07:32
The system needs to read it from nbd, before it runs it, and disconnects nbd
07:32
So it can't have a race condition there
07:32
<vagrantc>
i guess so
07:32
<alkisg>
There might be a race condition on the term signal,
07:32
<vagrantc>
as long as it works ...
07:32
and kills things fast enough
07:33
<alkisg>
where nbd-client get the TERM signal from upstart etc before nbd-disconnect
07:33
<vagrantc>
i.e. if the "nbd-client -d" failed for some reason...
07:33
<alkisg>
Then, sure, nbd-disconnect won't be able to read nbd-client...
07:33
<vagrantc>
alkisg: so, the newer code started introducing problems after you fixed it to not hard-code root=/dev/nbd0?
07:34
<alkisg>
Yup. Basically before fixing it, it wasn't running at all for anyone
07:34
So it was the same as if it wasn't there at all
07:34
Except for the swap devices...
07:35
I'd tested that before the switch to root=/dev/nbd0, I didn't test it later on
07:35
<vagrantc>
i do seem to recall Debian working fine with NBD before nbd-disconnect was even written
07:36
shutdown/reboot was handled a bit different with sysvinit vs. upstart
07:36
or maybe it was leaving open connections that i just hadn't noticed ... hard to remember
07:36
<alkisg>
Yup
07:36
That was the only issue
07:37
nbd-client -d was never called, so the swap files weren't deleted
07:39
When nbd-server was running from inetd, it wasn't a big problem because it supported tcp keep-alive
07:39
I think for a quick fix I'll disable it _unless_ some lts.conf directive is used
07:40
E.g. DISCONNECT_NBD_ON_SHUTDOWN=True
07:40
<vagrantc>
disable nbd-disconnect?
07:40
<alkisg>
It was disabled until a week ago
07:40
So yup...
08:06
vagrantc: we have INITRD_NAME, but not NBI_NAME
08:06
Can we assume that we can replace "initrd" with "nbi" and get it? Since it's being deprecated etc?
08:23
!kvm
08:23
<ltsp>
kvm: Virtual thin client: kvm -vga vmware -ctrl-grab -no-shutdown -kernel /var/lib/tftpboot/ltsp/i386/vmlinuz -initrd /var/lib/tftpboot/ltsp/i386/initrd.img -append "ro init=/sbin/init-ltsp root=/dev/nbd0 ip=:10.160.67.10::"
08:31
<alkisg>
OK one done, one to go... :)
08:32
<vagrantc>
alkisg: i wouldn't bother with nbi support, honestly...
08:32
<alkisg>
vagrantc: see my last commit, I think it's good enough for at least debian + ubuntu
08:32
It doesn't even mention "pae" anymore... :)
08:32
And, it fixed the vmlinuz symlink here in an ltsp-update-image -c / test
08:34
On to nbd-disconnect... fortunately kvm allows me to see the last image of the VM after it shuts down
08:36* vagrantc waves
08:36
<alkisg>
bb!
08:36vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Quit: leaving)
08:53alkisg1 has joined IRC (alkisg1!~alkisg@ppp089210180048.access.hol.gr)
08:53alkisg1 has joined IRC (alkisg1!~alkisg@ubuntu/member/alkisg)
08:54alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Ping timeout: 272 seconds)
08:57alkisg1 has left IRC (alkisg1!~alkisg@ubuntu/member/alkisg, Client Quit)
08:59alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
09:00bennabiy has left IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com, Read error: Connection reset by peer)
09:05bennabiy has joined IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com)
09:20alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Remote host closed the connection)
09:23bobby_C has left IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at, Quit: Goin' down hard)
10:47ogra_ has left IRC (ogra_!~ogra_@p5098ed03.dip0.t-ipconnect.de, Ping timeout: 248 seconds)
10:47ogra_ has joined IRC (ogra_!~ogra_@p5098ed03.dip0.t-ipconnect.de)
10:50alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
10:51alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Remote host closed the connection)
10:53alkisg has joined IRC (alkisg!~alkisg@ppp089210180048.access.hol.gr)
10:53alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
10:59alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
11:03cyberorg has joined IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg)
11:04alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
11:43alexqwesa has joined IRC (alexqwesa!~alex@109.172.12.47)
13:31PhoenixSTF has joined IRC (PhoenixSTF!~rudi@78.29.154.124)
14:06andygraybeal has joined IRC (andygraybeal!~andy@h23.205.213.151.dynamic.ip.windstream.net)
15:06awilliam1 has joined IRC (awilliam1!~awilliams@unaffiliated/mistik1)
15:07awilliams has left IRC (awilliams!~awilliams@unaffiliated/mistik1, Ping timeout: 265 seconds)
15:07awilliam1 is now known as awilliams
16:18khildin has joined IRC (khildin!~khildin@ip-213-49-85-26.dsl.scarlet.be)
16:28alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Remote host closed the connection)
16:30awilliam1 has joined IRC (awilliam1!~awilliams@unaffiliated/mistik1)
16:31awilliams has left IRC (awilliams!~awilliams@unaffiliated/mistik1, Ping timeout: 260 seconds)
16:31awilliam1 is now known as awilliams
16:42awilliam1 has joined IRC (awilliam1!~awilliams@unaffiliated/mistik1)
16:43awilliams has left IRC (awilliams!~awilliams@unaffiliated/mistik1, Ping timeout: 272 seconds)
16:43awilliam1 is now known as awilliams
16:54
<stgraber>
alkisg: we are in sync with Debian for ltsp and ldm. We should be for nbd too if the package was actually building: https://launchpad.net/ubuntu/+source/nbd/1:3.6-1
16:55freedomrun has joined IRC (freedomrun!~freedomru@unaffiliated/freedomrun)
17:19alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
17:21
<alkisg>
stgraber: the current ltsp in debian, 5.5.0, is buggy, so I was talking with vagrantc about if he could release 5.5.1 before trusty import freeze
17:29
<stgraber>
alkisg: ah, ok. The Debian Import Freeze only affects auto-sync, not manual syncs so there's no rush. As long as any new feature is in before Feature Freeze
17:29
<alkisg>
Ah, cool, let me check when feature freeze is...
17:30
Although, there aren't really new features now, mostly bug fixes
17:30
18 February 20th FeatureFreeze
17:30
Very good
17:34Selveste1 has left IRC (Selveste1!~Selveste1@static-5-103-136-165.seas-nve.net, Ping timeout: 272 seconds)
17:34
<stgraber>
not sure if you're in touch with nbd upstream, but it may be worth having them look at: https://launchpad.net/ubuntu/+source/nbd/1:3.6-1
17:35
as long as those 3 builds don't succeed, the new nbd won't land in trusty
17:43
<alkisg>
I think wouter is already working on those
18:23Selveste1 has joined IRC (Selveste1!~Selveste1@static-5-103-136-165.seas-nve.net)
18:28alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Remote host closed the connection)
18:32gothaggis|work has left IRC (gothaggis|work!~gothaggis@corp.heavyhammer.com, Ping timeout: 252 seconds)
18:35khildin has left IRC (khildin!~khildin@ip-213-49-85-26.dsl.scarlet.be, Quit: I'm gone, bye bye)
18:43
<bennabiy>
alkisg: what currently is buggy about 5.5... I am about to do a test chroot to see if we still have compatibility with mint 15 +
19:15
!fat-clients
19:15
<ltsp>
I do not know about 'fat-clients', but I do know about these similar topics: 'epoptes-fat-clients'
19:15
<bennabiy>
!fatclient
19:15
<ltsp>
I do not know about 'fatclient', but I do know about these similar topics: 'fatclients', 'fatclient-printers'
19:15
<bennabiy>
!fatclients
19:15
<ltsp>
fatclients: You may find some info about the Ubuntu/LTSP implementation of fat clients at https://help.ubuntu.com/community/UbuntuLTSP/FatClients
19:29
<bennabiy>
!ltsp-pnp
19:29
<ltsp>
ltsp-pnp: ltsp-pnp is an alternative (upstream) method to maintain LTSP installations for thin and fat clients that doesn't involve chroots: https://help.ubuntu.com/community/UbuntuLTSP/ltsp-pnp
19:58boospy has joined IRC (boospy!~kvirc@77.235.70.194)
20:10vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc)
20:13bobby_C has joined IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at)
20:51alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
20:51andygraybeal has left IRC (andygraybeal!~andy@h23.205.213.151.dynamic.ip.windstream.net, Quit: Ex-Chat)
20:51andygraybeal_ has joined IRC (andygraybeal_!~andy@h23.205.213.151.dynamic.ip.windstream.net)
20:53
<alkisg>
vagrantc: so, I committed the first fix, and I have this ready: http://pastie.org/8648794
20:53
...but I can't test it because shutdown never failed for me
20:53
I asked some teachers that did experience the issue to test it, I hope they'll reply tomorrow
20:53
So, if you want to upload, I can push it untested...
20:54
(untested, meaning, I see no difference while testing here locally)
20:54
bennabiy: ltsp 5.5.0 breaks grub booting with ltsp-pnp, but the current ltsp-trunk has fixes for that
20:56
Btw, nbd-disconnect did work for the "disconnect swap" part till now, which is the most significant part, to not leave 512 MB swap files on the server
20:57
It's just the "disconnect root" part which didn't work; which isn't very important,
20:57* vagrantc nods
20:57
<vagrantc>
alkisg: that patch seems to assume upstart, no?
20:57
<alkisg>
and, for a few users, the "don't use poweroff -fp" part caused a regression, while for others the sysrq part caused a regression
20:58
vagrantc: if it does, it's not intentional
20:58
<vagrantc>
so upstart reads stdout ?
20:58
<alkisg>
I did remove the "-p" parameter to poweroff, so that it's compatible with debian...
20:58
<vagrantc>
or stderr?
20:58
<alkisg>
Ah you mean the comment?
20:58
It's there in case stgraber ever wants to "correct" the upstart job
20:58
upstart grabs stdout/stderr etc, yes
20:59
But it doesn't matter the message isn't significant...
20:59
<vagrantc>
how could that even break the nbd-swap?
20:59
<alkisg>
nbd-swap was and is ok
20:59
The only issues are if the root nbd is cleanly disconnected or not, and if the clients shutdown properly or not
21:00
Now I tried to replace the sysrq with poweroff -f, assuming it does some extra magic, because those users that have the "not shutting down" issue, do report that "poweroff -fp" works for them
21:00
poweroff -fp was upstart specific, so I changed it to -f...
21:02
Now I'm using the "cache the command and hope that it works" hack... let's see if the users report that it fixes their shutdown issue
21:19alkisg1 has joined IRC (alkisg1!~alkisg@ubuntu/member/alkisg)
21:19alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Disconnected by services)
21:19alkisg1 is now known as alkisg
21:24
<stgraber>
c
21:28alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Ping timeout: 248 seconds)
21:34alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
21:38
<alkisg>
vagrantc: stgraber says that import freeze only affects automatic syncs, so manual syncs can be done until featurefreeze on February 20th
21:54Parker955 is now known as Parker955_Away
21:55Parker955_Away is now known as Parker955
22:18alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
22:30bobby_C has left IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at, Quit: Goin' down hard)
22:39Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas)
23:18Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Quit: Leaving.)
23:29boospy has left IRC (boospy!~kvirc@77.235.70.194, Quit: KVIrc 4.3.1 Aria http://www.kvirc.net/)
23:59freedomrun has left IRC (freedomrun!~freedomru@unaffiliated/freedomrun, Ping timeout: 272 seconds)