IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 21 February 2013 (all times are UTC)
| 01:09 | dead_inside has joined IRC (dead_inside!~dead_insi@host-98-127-28-44.grf-mt.client.bresnan.net) | |
| 01:13 | khildin has left IRC (khildin!~khildin@ip-80-236-212-199.dsl.scarlet.be, Quit: I'm gone, bye bye) | |
| 01:14 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 252 seconds) | |
| 01:29 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 01:44 | vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Quit: leaving) | |
| 02:08 | hays_ is now known as hays | |
| 02:18 | dead_inside has left IRC (dead_inside!~dead_insi@host-98-127-28-44.grf-mt.client.bresnan.net, Quit: Leaving...) | |
| 03:30 | adrianorg__ has left IRC (adrianorg__!~adrianorg@177.134.58.112, Ping timeout: 255 seconds) | |
| 03:36 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 03:36 | Parker955_Away is now known as Parker955 | |
| 04:20 | Parker955 is now known as Parker955_Away | |
| 04:33 | vnc786 has joined IRC (vnc786!~chatzilla@49.248.129.178) | |
| 04:55 | sha has joined IRC (sha!~sha@e177116104.adsl.alicedsl.de) | |
| 04:58 | sha_ has left IRC (sha_!~sha@e177117035.adsl.alicedsl.de, Ping timeout: 240 seconds) | |
| 05:05 | staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Ping timeout: 240 seconds) | |
| 05:24 | ogra_ has left IRC (ogra_!~ogra_@p5098ed03.dip0.t-ipconnect.de, Ping timeout: 248 seconds) | |
| 05:26 | seamuz has joined IRC (seamuz!~seamus@89.237.49.94) | |
| 05:36 | vmlintu has joined IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi) | |
| 05:43 | work_alkisg is now known as alkisg | |
| 06:27 | Enslaver has left IRC (Enslaver!~Enslaver@c-98-196-42-169.hsd1.tx.comcast.net, Ping timeout: 248 seconds) | |
| 06:35 | <seamuz> Hello all!
| |
| 06:35 | Why would my thin-client on boot have a screen that says "This thin client hasn't been configured yet."?
| |
| 06:36 | When I click ok it works like a charm
| |
| 06:36 | I forgot to do something?
| |
| 06:40 | <alkisg> !release
| |
| 06:40 | <ltsp> release: please mention the linux distro and release you're using :)
| |
| 06:44 | highvoltage has joined IRC (highvoltage!~highvolta@ubuntu/member/highvoltage) | |
| 06:51 | <alkisg> sbalneav: around for a quick ldm/xorg question?
| |
| 06:58 | Or, stgraber... or anyone else that knows xorg: :)
| |
| 06:58 | We run `xinit ldm(.c)`, and support X crashes. Can we have an on_exit() handler in ldm that gets a chance to cleanup stuff, or ldm will unconditionally die (and the ssh connection along with it...)
| |
| 07:00 | komunista has joined IRC (komunista!~slavko@adsl-195-168-244-224.dynamic.nextra.sk) | |
| 07:01 | <alkisg> Sounds like stdlib atext() might work... trying...
| |
| 07:07 | Nope, doesn't get called with ctrl+c... we probably need a trap() there...
| |
| 07:17 | * alkisg tries to do it from shell instead... ;) | |
| 07:42 | <alkisg> Heh, fortunately crashing xorg is very easy, so it's easy to get a test case... :P
| |
| 08:13 | meamy has joined IRC (meamy!~hannes@pd95cdee4.dip0.t-ipconnect.de) | |
| 08:13 | ogra_ has joined IRC (ogra_!~ogra_@p5098ed03.dip0.t-ipconnect.de) | |
| 08:27 | mikkel has joined IRC (mikkel!~mikkel@80-71-132-15.u.parknet.dk) | |
| 08:44 | ajith has joined IRC (ajith!b7522f02@gateway/web/freenode/ip.183.82.47.2) | |
| 08:44 | <ajith> HI
| |
| 08:46 | ajith has left IRC (ajith!b7522f02@gateway/web/freenode/ip.183.82.47.2, Client Quit) | |
| 08:50 | <highvoltage> heh
| |
| 08:56 | dievel has joined IRC (dievel!~dievel@2.229.104.66) | |
| 09:09 | dievel has left IRC (dievel!~dievel@2.229.104.66, Remote host closed the connection) | |
| 09:10 | baptiste has left IRC (baptiste!baptiste@paul-sud.asso.ups-tlse.fr, Ping timeout: 255 seconds) | |
| 09:23 | Gremble has joined IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginmedia.com) | |
| 09:25 | highvoltage has left IRC (highvoltage!~highvolta@ubuntu/member/highvoltage, Remote host closed the connection) | |
| 09:36 | dobber_ has joined IRC (dobber_!~dobber@89.190.199.210) | |
| 09:48 | baptiste has joined IRC (baptiste!baptiste@paul-sud.asso.ups-tlse.fr) | |
| 09:58 | <vmlintu> lately everything has been unstable, it seems like..
| |
| 09:58 | alkisg: have you had stability problems with nbd?
| |
| 09:58 | nbd-server that is
| |
| 09:59 | <alkisg> vmlintu: no
| |
| 09:59 | (without hardware networking problems involved, right?)
| |
| 09:59 | <vmlintu> nbd-server stopping to respond for new connection
| |
| 09:59 | connections
| |
| 10:00 | We got hit by this on ubuntu 12.04: http://sourceforge.net/mailarchive/forum.php?thread_name=20130208073853.GB15504%40grep.be&forum_name=nbd-general
| |
| 10:02 | Basically the main process of nbd-server dies and new connections fail after that until you restart whole nbd-server
| |
| 10:02 | <alkisg> vmlintu: at 3-4 occations I had to stop nbd-server, *kill* it as well, and then restart it for the clients to be able to boot again. Is that what you're talking about? It was rare though...
| |
| 10:03 | <vmlintu> Yes, sometimes it would require that
| |
| 10:03 | <alkisg> I've seen segfaults of nbd-server in the system logs as well
| |
| 10:03 | <vmlintu> That happened on some of our test systems 5-10 times a day..
| |
| 10:03 | <alkisg> Ouch
| |
| 10:03 | No wonder schools here are reverting to 10.04 :P
| |
| 10:03 | <vmlintu> are they?
| |
| 10:04 | We patched now nbd with this to temporarily get rid of the problem: https://github.com/opinsys/opinsys-debs/blob/master/packages/nbd/debian/patches/opinsys-ignore-sigpipe
| |
| 10:04 | dievel has joined IRC (dievel!~dievel@2-229-104-66.ip196.fastwebnet.it) | |
| 10:05 | <alkisg> vmlintu: I think we should cooperate wrt supporting our installations...
| |
| 10:05 | E.g. we have the ts.sch.gr PPA where we send such fixed packages until they are SRU'ed
| |
| 10:06 | So my idea is that we can "exchage" fixed packages via PPAs...
| |
| 10:06 | <vmlintu> we are currently patching these packages in 12.04/12.10: https://github.com/opinsys/opinsys-debs/tree/master/packages
| |
| 10:07 | Yes, it'd be great to get the fixes out faster
| |
| 10:07 | <alkisg> We have these ones to our PPA: https://launchpad.net/~ts.sch.gr/+archive/ppa/?field.series_filter=precise
| |
| 10:07 | Do you have them in a PPA? Or are you manually compiling + installing your fixes?
| |
| 10:08 | <vmlintu> We have our own builder that builds them so that we don't need to wait for PPA compiling
| |
| 10:08 | <alkisg> That's nice for upload times but it hinders cooperation
| |
| 10:09 | I.e. you can't copy other people's work easily
| |
| 10:09 | While with PPAs you can e.g. find a fixed nbd-server somewhere (even in raring) and just copy it
| |
| 10:09 | We copy packages from newer Ubuntu released quite often, when those solve some bug
| |
| 10:09 | *releases
| |
| 10:10 | <vmlintu> The packages are in a repository, but it contains now also proprietary packages that we are not allowed to distribute
| |
| 10:10 | Maybe I should separate those out and make the repository public
| |
| 10:10 | <alkisg> We have a separate repository for proprietary software
| |
| 10:10 | Yup, that's what we did :)
| |
| 10:11 | So schools here have 2 additional repositories in their sources, our open one and our closed one
| |
| 10:11 | <vmlintu> We could of course upload the same debian source packages to a ppa also
| |
| 10:12 | <alkisg> Sure, you could even automate that
| |
| 10:12 | With some script
| |
| 10:12 | Nowadays we're not using dput at all anymore though, only launchpad "recipes"
| |
| 10:12 | They even support getting the source from external branches, even git
| |
| 10:13 | <vmlintu> That's something I haven't heard of
| |
| 10:13 | <alkisg> https://code.launchpad.net/~ts.sch.gr/+recipe/ltsp
| |
| 10:13 | This one gets the ltsp source from upstream, the vagrantc's packaging from debian, and it produces a .deb
| |
| 10:13 | All in 2 lines :)
| |
| 10:15 | garymc has joined IRC (garymc!~chatzilla@host81-148-18-165.in-addr.btopenworld.com) | |
| 10:15 | <alkisg> Btw in reality we have a 3rd PPA, "ts.sch.gr proposed" where we build the packages, test them, and when we see that they work OK, we then copy them to "ts.sch.gr ppa"
| |
| 10:16 | <vmlintu> I'll have a look at the recipe stuff with others to see if we could use that to automate things..
| |
| 10:16 | Right now we take releases from github automatically with watch files: https://github.com/opinsys/opinsys-debs/blob/master/packages/puavo-ds/debian/watch
| |
| 10:17 | But yes, collaborating on this could be helpful
| |
| 10:18 | <alkisg> vmlintu: which DE are you using?
| |
| 10:19 | <vmlintu> DE=desktop environment?
| |
| 10:19 | <alkisg> Yes, gnome-fallback, lxde etc
| |
| 10:19 | <vmlintu> gnome-fallback with webmenu as the menu
| |
| 10:20 | This is with quantal
| |
| 10:20 | <alkisg> Gnome-fallback here too.. so you're not using gnome-panel?
| |
| 10:20 | <vmlintu> Yes, there's gnome-panel, but we have replaced the menu
| |
| 10:20 | http://labs.opinsys.com/blog/2012/11/01/bringing-html5-to-the-desktop-with-appjs/
| |
| 10:21 | It's an html5 based application launcher. You can define the menu structure in a separate configuration file so that one can mix normal applications and web links.
| |
| 10:22 | <alkisg> Are you having empty gnome-panel at times too? https://bugs.launchpad.net/ubuntu/+source/gnome-panel/+bug/1078679
| |
| 10:23 | <vmlintu> I'd have to check with others to know for sure
| |
| 10:24 | We are just starting to roll out 12.10 and bulk of the systems are on 10.04
| |
| 10:25 | If everything goes right, we'll pxe boot also the ltsp servers
| |
| 10:31 | <alkisg> Why 12.10 instead of 12.04, which is an LTS release?
| |
| 10:32 | * alkisg avoids non-lts releases, devs stop caring about them when the next non-lts release is out | |
| 10:33 | <vmlintu> 12.04 wasn't really stable either, so we'll be updating to every release from now on
| |
| 10:34 | We are using 12.04 on boot servers, but everything using the ltsp images will be based on 12.10: ltsp servers, thin/fat clients and laptops
| |
| 10:34 | <alkisg> That's true, 10.04 was way more stable than 12.04, but I'm afraid e.g. 13.04 won't be more stable than 12.04
| |
| 10:34 | It won't even support gnome-fallback anymore, will it?
| |
| 10:35 | And neither unity-2d, forcing LTSP'ers to use lxde, mate...
| |
| 10:36 | <vmlintu> I'll have to start digging into 13.04 soon
| |
| 10:36 | Most probably we'll be running multiple ubuntu versions in parallel on the same servers in the future
| |
| 10:37 | <alkisg> How many sites/clients are you supporting/
| |
| 10:37 | ?
| |
| 10:38 | <vmlintu> around 170 schools now, some 6500 computers if I remember correctly..
| |
| 10:39 | <alkisg> Yeah we should definetely join forces, at least notifying each other about problems we see
| |
| 10:39 | Around 400 schools here for now...
| |
| 10:41 | <vmlintu> yes.. I just checked with others and uploading our packages also to a ppa should be just a dput configuration change
| |
| 10:43 | ychEpalMoirwn has joined IRC (ychEpalMoirwn!c23fd8aa@gateway/web/freenode/ip.194.63.216.170) | |
| 10:44 | <ychEpalMoirwn> Hello! Is there anyone here ?
| |
| 10:46 | I am trying to install teamviewer on an ltsp server.
| |
| 10:46 | <alkisg> That doesn't sound LTSP related though... :)
| |
| 10:47 | <ychEpalMoirwn> one second alkisg
| |
| 10:49 | <vmlintu> I need to get some lunch..
| |
| 10:51 | <vnc786> hello everyone
| |
| 10:51 | <vmlintu> alkisg: I'll have a chat with others to see how we could get the patched packages in an open repository
| |
| 10:53 | <ychEpalMoirwn> hi vnc786
| |
| 10:54 | <vnc786> i have fat clients running ubuntu 12.04 64 bit and they are connected to LCD thru VGA cable now i want to add addition display thru HDMI cable
| |
| 10:55 | i hot plugged HDMI cable to existing fat clients but nothing work
| |
| 10:55 | any help on this ??
| |
| 10:55 | <ychEpalMoirwn> too difficult for me.
| |
| 10:56 | do you know how to get the hdmi working on a normal ubuntu installation ?
| |
| 10:57 | <vnc786> i first tried on ubuntu but it didnt work and now i m trying on fat clients
| |
| 10:58 | my requirements is : i need to share one fat clients system screen with a big display ...?
| |
| 10:59 | <ychEpalMoirwn> ok. wait for any more expert than me...
| |
| 10:59 | If I were you I would try xrandr
| |
| 11:00 | <vnc786> i did xrandr it showed me HDMI connection...
| |
| 11:00 | <ychEpalMoirwn> that's good.
| |
| 11:00 | Can you enable it for a single fat client using xrandr ?
| |
| 11:02 | <vnc786> w8 i m checking
| |
| 11:02 | <ychEpalMoirwn> sorry, I need to go...
| |
| 11:03 | <vnc786> np
| |
| 11:03 | <ychEpalMoirwn> hope someone will come in.
| |
| 11:03 | it's my first time also here!
| |
| 11:03 | anyway bye
| |
| 11:03 | ychEpalMoirwn has left IRC (ychEpalMoirwn!c23fd8aa@gateway/web/freenode/ip.194.63.216.170, Quit: Page closed) | |
| 11:19 | <Hyperbyte> vnc786, your solution is in xrandr. You can set xrandr options per-client via lts.conf, XRANR_* settings.
| |
| 11:19 | meamy has left IRC (meamy!~hannes@pd95cdee4.dip0.t-ipconnect.de, Quit: meamy) | |
| 11:22 | <vnc786> Hyperbyte: ok so 1st i have to put entry in lts.conf
| |
| 11:23 | adrianorg__ has joined IRC (adrianorg__!~adrianorg@177.134.59.247) | |
| 11:24 | <Hyperbyte> No, 1st you have to experiment with the xrandr command, see what options/settings you have to make so the displays work for you
| |
| 11:25 | Run 'xrandr' and 'xrandr --help' on the affected client, read online guides about xrandr, or ask in #ubuntu
| |
| 11:25 | And when you have the correct xrandr modes/commands, insert those in lts.conf for the affected client
| |
| 11:28 | anunnaki is now known as anunnaki_afk | |
| 11:31 | <vnc786> ok in the middle: i was just messing with fat clients + hdmi and and it has stopped booting i have removed HDMI cable the screen stop at "trying to load : pxelinux.cfg...def..:ok" and it stucks there in the syslog there is one entry coming like this
| |
| 11:31 | Feb 21 16:58:12 EV-TC-SRV01 in.tftpd[23524]: tftp: client does not accept options
| |
| 11:33 | it was working properly from morning just i was testing with HDMI which i hot plugged and rebooted which didnt work so i reverted back to default
| |
| 11:45 | <alkisg> alkisg: http://paste.ubuntu.com/1698469/
| |
| 11:46 | alkisg is now known as work_alkisg | |
| 11:56 | rickogden has joined IRC (rickogden!~Rick@host.hifirevolution.com) | |
| 11:58 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 12:10 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 12:11 | Gremble has left IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginmedia.com, Quit: I Leave) | |
| 12:11 | Gremble has joined IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginmedia.com) | |
| 12:16 | Gremble has joined IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginmedia.com) | |
| 12:18 | Gremble has left IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginmedia.com, Client Quit) | |
| 12:22 | <vnc786> i have connected hdmi cable to fat clients but display is not coming what step i have to do ?? any help
| |
| 12:31 | <baptiste> what's the graphics card
| |
| 12:32 | and do you have several outputs to choose from
| |
| 12:34 | yes, as Hyperbyte says, it should be RandR's job though on my desktop (proprietary nvidia) that should be workable with driver options instead.
| |
| 12:37 | to anwser myself, you said vga + hdmi. so you want dual monitor ?
| |
| 12:51 | komunista has left IRC (komunista!~slavko@adsl-195-168-244-224.dynamic.nextra.sk, Quit: Leaving.) | |
| 12:51 | komunista has joined IRC (komunista!~slavko@adsl-195-168-244-224.dynamic.nextra.sk) | |
| 12:54 | <vnc786> baptiste: fat clients are running well on acer lcd thru vga cable no issue in that
| |
| 12:55 | now i have one more LCD which I have connected to a fat clients thru HDMI cable but i am not able to see any thing
| |
| 12:59 | engblom has left IRC (engblom!~user@unaffiliated/engblom) | |
| 13:07 | F-GT has left IRC (F-GT!~phantom@ppp59-167-136-109.static.internode.on.net, Ping timeout: 252 seconds) | |
| 13:10 | F-GT has joined IRC (F-GT!~phantom@ppp59-167-136-109.static.internode.on.net) | |
| 13:13 | vnc786 has left IRC (vnc786!~chatzilla@49.248.129.178, Quit: ChatZilla 0.9.90 [Firefox 11.0/20120410121533]) | |
| 13:29 | mgariepy has left IRC (mgariepy!mgariepy@ubuntu/member/mgariepy, Ping timeout: 245 seconds) | |
| 13:42 | mgariepy has joined IRC (mgariepy!mgariepy@ubuntu/member/mgariepy) | |
| 13:51 | komunista has left IRC (komunista!~slavko@adsl-195-168-244-224.dynamic.nextra.sk, Quit: Leaving.) | |
| 13:59 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 14:15 | monteslu_ has joined IRC (monteslu_!~monteslu@ip68-109-174-213.ph.ph.cox.net) | |
| 14:19 | monteslu has left IRC (monteslu!~monteslu@ip68-109-174-213.ph.ph.cox.net, Ping timeout: 252 seconds) | |
| 14:20 | vnc786 has joined IRC (vnc786!~chatzilla@49.248.129.178) | |
| 14:21 | <vnc786> thanks everyone issue solved..
| |
| 15:13 | jammcq has joined IRC (jammcq!~jam@c-69-245-75-255.hsd1.mi.comcast.net) | |
| 15:14 | <jammcq> Helloooo friends
| |
| 15:14 | <alkisg> Hey jammcq
| |
| 15:15 | <jammcq> hey alkisg
| |
| 15:15 | alkisg: what kind of weather do you have in greece these days?
| |
| 15:15 | <alkisg> Rainy :-/
| |
| 15:15 | jammcq: could I have your thoughts wrt the sshfs issue and a possible solution I think I found?
| |
| 15:15 | <jammcq> cold? warm?
| |
| 15:16 | sshfs issue?
| |
| 15:16 | <alkisg> ~6 ° C here, not too bad
| |
| 15:16 | <jammcq> hmm, it's 21F here right now
| |
| 15:16 | <alkisg> Yeah that one with ldm and the possible data loss
| |
| 15:16 | <jammcq> much colder where sbalneav is
| |
| 15:16 | <alkisg> So here's my idea... xinit runs X and LDM. LDM runs `ldm-script xsession`, which then calls some ldm/rc.d/X* scripts,
| |
| 15:17 | ...and X95-run-xsession
| |
| 15:17 | * jammcq clears the cobwebs from his brain to remember how all that stuff works | |
| 15:18 | <alkisg> The problem is that if xorg dies, ldm dies too, and the ssh connection closes,
| |
| 15:18 | so scripts that run _afterwards_ and need the ssh connection to clean up stuff, cannot do that
| |
| 15:18 | Can't unmount sshfs, can't run cleanup on the server etc etc
| |
| 15:18 | <jammcq> hmmm
| |
| 15:18 | <alkisg> So... if in some X94* shell script I put a `trap 0 INT QUIT etc etc`... and xorg crashes... will that trap run AND keep ldm alive for a bit, so that it has a chance to cleanup before LDM dies?
| |
| 15:20 | <jammcq> what signal does the kernel send the child processes when the parent dies?
| |
| 15:20 | is it QUIT or KILL ?
| |
| 15:20 | as long as you are catching the right signal, it should work
| |
| 15:21 | but what if ldm caught those signals?
| |
| 15:21 | isn't LDM going away soon?
| |
| 15:21 | <alkisg> It could run xsetiohandler or whatever else the xorg crash function is called, sure, but we'd then need to save/restore the environment, while with a shell trap it's cleaner
| |
| 15:22 | LDM is going away, but e.g. 12.04 will be around for 4+ years, and people will have data loss with that bug
| |
| 15:22 | It also occurs even if xorg doesn't crash
| |
| 15:22 | <jammcq> what data are they losing?
| |
| 15:22 | <alkisg> It goes like this:
| |
| 15:22 | User with localapps or fat client logs in, sshfs connection is set up etc etc
| |
| 15:23 | At some point he logs out, xorg dies, and ldm dies too
| |
| 15:23 | The bug is that our scripts though the ssh connection would be alive at that point
| |
| 15:23 | But it's not; it died with LDM
| |
| 15:23 | So the fusermount -u calls fail, but that's not the worse thing
| |
| 15:23 | Some user processes are still running, and try to write to $HOME
| |
| 15:24 | So since the sshfs connection is broken, they end up writing locally, in the client tmpfs
| |
| 15:24 | Some data loss is there, but that's not the worse either
| |
| 15:24 | <jammcq> heh
| |
| 15:24 | <alkisg> The worst problem is that on the next time the user logs in, there's a /home/username local folder already there because of the previous processes writing locally,
| |
| 15:24 | staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu) | |
| 15:25 | <alkisg> so our code thinks that /home/username was mounted by NFS or some user-defined script, so it doesn't use SSHFS to mount $HOME from the server,
| |
| 15:25 | and the user ends up seeing a completely clean environment without his data
| |
| 15:25 | And, any files he creates on that "clean" environment, are actually on the tmpfs, lost after client reboot
| |
| 15:26 | So he loses everything he does on that session
| |
| 15:26 | And that happens on every localapp/fat client login after the first one... which is a pretty bad bug
| |
| 15:26 | <jammcq> umm... sounds pretty bad
| |
| 15:26 | so when X dies, does it just restart X or do they need to reboot the client?
| |
| 15:26 | <alkisg> So, lightdm or not, that one needs to get fixed + backported :)
| |
| 15:27 | The problem is that we want to have LDM stay up for a bit after xorg dies, so that the ssh connection stays up
| |
| 15:27 | So that we properly clean up things
| |
| 15:28 | <jammcq> well... putting a trap in X94* isn't going to keep LDM up
| |
| 15:28 | <alkisg> Even if it's his child hooking the signal?
| |
| 15:28 | <jammcq> seems like LDM needs to trap the signals, do it's cleanup and then exit
| |
| 15:29 | x dies, sends a signal to all its children
| |
| 15:29 | that causes LDM to die, which then sends a signal to all it's children
| |
| 15:29 | <alkisg> *then*? Or wait for its childs to die first?
| |
| 15:30 | <jammcq> ldm is a zombie at that poing, waiting for it's children to die
| |
| 15:30 | but ldm has many children, right?
| |
| 15:30 | <alkisg> And does "zombie" mean that the ssh socket will still be up?
| |
| 15:30 | <jammcq> I don't think so
| |
| 15:30 | <alkisg> Or that only involves the stdio descriptors?
| |
| 15:30 | Hmmm...
| |
| 15:31 | <jammcq> I think ldm is pretty much dead at that point, just waiting for it's children to die before the process is completely removed from the process table
| |
| 15:31 | <alkisg> The ssh process is one of LDM children, the same as the ldm-script process
| |
| 15:31 | <jammcq> but those children are independent of each other, right?
| |
| 15:31 | <alkisg> Right
| |
| 15:32 | <jammcq> I think LDM should detect the KILL/QUIT signal
| |
| 15:32 | act like the user is doing a normal log off
| |
| 15:32 | and once all the children have done their cleanup, then LDM should finish dying
| |
| 15:32 | <alkisg> There's the problem then, we need sbalneav or stgraber for that
| |
| 15:32 | <jammcq> ok
| |
| 15:33 | <alkisg> No other devs know glib/xorg/whatever else is involved to install xsetiohandlers
| |
| 15:33 | <jammcq> let's knock on sbalneave's door, see what he thinks
| |
| 15:34 | <alkisg> Also, that's why I say having LDM in shell would be cleaner... no need for hacks to save/restore the environment then
| |
| 15:34 | Because now we'll need to save the environment (SSH_HOME etc) in X94*, and restore it on the first script called by the LDM handler
| |
| 15:34 | <jammcq> I don't know enough about LDM to know if a shell script can do everything
| |
| 15:35 | <sbalneav> Morning all
| |
| 15:35 | Bo
| |
| 15:35 | <alkisg> Hey Scotty
| |
| 15:35 | <jammcq> hey buddy
| |
| 15:35 | <sbalneav> What can I do for you all?
| |
| 15:35 | <alkisg> Grab a coffee mug and join us :)
| |
| 15:35 | Can you read the backlog?
| |
| 15:35 | <sbalneav> Let me run and grab a coffee two secs.
| |
| 15:35 | yeah, gimme 5...
| |
| 15:38 | So, to summarize: ldm should keep running for a bit even after X dies, to allow for cleanup. Did I get that right?
| |
| 15:39 | <alkisg> Yup
| |
| 15:39 | <sbalneav> ok, one sec...
| |
| 15:39 | <jammcq> thats what i'm thinking
| |
| 15:39 | <sbalneav> let me look at the code...
| |
| 15:39 | <alkisg> sbalneav: ideally, if a shell script trap can keep ldm running for a bit, we prefer that instead
| |
| 15:40 | <jammcq> does a script run LDM ?
| |
| 15:40 | xinit -> X -> some script -> LDM ?
| |
| 15:40 | <alkisg> screen.d => xinit => ldm => ssh.c connection => ldm-script xsession => X94-shell-trap => X95-run-xsession => presume a crash here => question, will X94 trap keep LDM running for a bit?
| |
| 15:41 | <jammcq> even if tehre was a script in there, LDM needs to know that it's Xserver has died, so it can clean up
| |
| 15:41 | <alkisg> A shell script can do a better cleanup than LDM
| |
| 15:41 | If LDM has to be the one to call the shell scripts that do the cleanup (as it does now, `ldm-script stop`),
| |
| 15:42 | ...then we need an X94-save-environment (i.e. export -p > somefile) and a K00-restore-environment (i.e. source somefile)
| |
| 15:42 | ...which is kinda hackish
| |
| 15:42 | <jammcq> what happens now if a user logs out?
| |
| 15:44 | <alkisg> X95-run-xsession continues with its execution, X95+ scripts are called, they finish, so `ldm-script xsession` returns control to LDM
| |
| 15:44 | (LDM used a waiting exec() call there)
| |
| 15:44 | Ah, moment...
| |
| 15:44 | Since it's a waiting call... can LDM actually run anything until that returns?
| |
| 15:44 | I.e. I really think the shell trap will work...
| |
| 15:45 | *blocking call
| |
| 15:46 | <jammcq> how many direct children does ldm have?
| |
| 15:46 | jsut the X95 script?
| |
| 15:46 | * jammcq can't tpye today | |
| 15:46 | <alkisg> The ssh connection too, not sure about others, let me check...
| |
| 15:47 | <jammcq> when ldm gets the signal from it's parent, it will send a signal to all its children
| |
| 15:47 | so ssh will die pretty quickly
| |
| 15:47 | before X95 has a chance to do anything
| |
| 15:48 | <alkisg> Even if LDM at that point is in a blocking exec() call?
| |
| 15:48 | Won't glib process the signals after the call returns?
| |
| 15:49 | <jammcq> I dont' think so, but a more experienced hacker (like scotty) would know better than me
| |
| 15:49 | blocking doesn't mean it's ignoring signals
| |
| 15:49 | * alkisg will know soon enough... trying a shell trap with a big sleeeeeeep in it... | |
| 15:49 | <jammcq> it just means it'll wait there for that process to complete before moving along in the normal flow
| |
| 15:52 | <alkisg> So, if I put a `trap "sleep 1000"`, and a few minutes later the SSH process is still up, we're OK, right?
| |
| 15:52 | <sbalneav> Huh. Just looking through the source code. I particularily like how my name's been removed from all the copyright headers.
| |
| 15:53 | * jammcq wonders why that would happen | |
| 15:54 | <sbalneav> ldm doesn't fork/exec. We'll need to make it do that in order to have it survive X dying.
| |
| 15:54 | <alkisg> Probably with the LDM "plugin-style" rewrite
| |
| 15:54 | <jammcq> alkisg: how are you seeing that the SSH connection is still up? is it able to pass traffic or is it just showing up in netstat ?
| |
| 15:55 | <alkisg> jammcq: I haven't yet; I'm just now trying to test that
| |
| 15:55 | <jammcq> sbalneav: how does ldm call the scripts? system() ?
| |
| 15:55 | alkisg: the connection may appear to still be there, but it might be in the process of shutting down
| |
| 15:55 | if it's still able to pass data, then it should be ok
| |
| 15:57 | <alkisg> jammcq: 2 minutes after selecting logoff: `ssh -S /var/run/ldm_socket_1826_server server ls` ==> works
| |
| 15:57 | So it appears fine
| |
| 15:58 | sbalneav: so I think it's much better to implement this as a shell trap
| |
| 15:58 | <jammcq> but LDM is still running, right?
| |
| 15:58 | <alkisg> Ah let me try with "kill" instead of "logoff" as well...
| |
| 15:58 | <sbalneav> OK, well, if you can do it with a shell trap, then fine.
| |
| 15:59 | * jammcq is gonna go rewrite the Linux kernel in bash | |
| 15:59 | <jammcq> :)
| |
| 16:00 | <alkisg> The problem with mixing C and shell, is that shell scripts don't get to keep their variables between different invocations
| |
| 16:04 | Results:
| |
| 16:05 | Logout: a shell trap can keep LDM up and running for as long as it likes
| |
| 16:05 | Kill Xorg: it only has a couple of secs to act
| |
| 16:05 | sbalneav: ok, let's do it your way :)
| |
| 16:06 | <jammcq> alkisg: at that point, it's not even acting. it just takes a couple secs to completely die
| |
| 16:06 | <sbalneav> OK, well, I'll have a look at it today. We
| |
| 16:06 | <alkisg> jammcq: it can execute stuff, e.g. I tried `env > /tmp/file`... didn't try using the ssh socket though
| |
| 16:06 | <sbalneav> 'll need to:
| |
| 16:06 | 1) Daemonize ldm itself
| |
| 16:07 | <jammcq> yeah, i'm almost certain you won't be sending any data through the ssh tunnel at that point
| |
| 16:07 | <sbalneav> 2) trap the X error.
| |
| 16:07 | <jammcq> sbalneav: is it an X error or a signal from the kernel?
| |
| 16:07 | or maybe it's too soon to tell what you'll get
| |
| 16:07 | <sbalneav> yeah, I'll have to test.
| |
| 16:08 | anyway, trap, and make sure we run the exit scripts before actually dying
| |
| 16:08 | <alkisg> XSetIOErrorHandler()
| |
| 16:08 | I think that's the one to call for trapping the X error...
| |
| 16:10 | We want to run `rcfiles("stop")` on the trap handler, but also upon normal LDM exit, and we want the ssh connection to still be up at that point
| |
| 16:10 | <sbalneav> ok, I'll dig into it.
| |
| 16:11 | * alkisg can handle the shell scripts part :) | |
| 16:12 | <sbalneav> Did you see the other day that I have lightdm authenticating with the libpam-sshauth? I haven't gotten it to go onto the session yet, but I'll get that going.
| |
| 16:12 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 16:13 | <alkisg> Cool! I've noted down a few things more that we can work on at the hackathon: http://wiki.ltsp.org/wiki/Dev:Hackathon_2013.03#Ideas
| |
| 16:19 | Gremble has joined IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginmedia.com) | |
| 16:25 | bobby_C has joined IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at) | |
| 16:37 | <sbalneav> alkisg: Looks like a good list.
| |
| 16:37 | I'll add a few notes, if that's ok.
| |
| 16:37 | <alkisg> Of course it is
| |
| 16:49 | dobber_ has left IRC (dobber_!~dobber@89.190.199.210, Remote host closed the connection) | |
| 17:02 | bobby_C has left IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at, Ping timeout: 276 seconds) | |
| 17:08 | <sbalneav> Hmmm. I'm not a dev in ltspedia :)
| |
| 17:11 | <jammcq> umm
| |
| 17:11 | I wonder if I can make you one
| |
| 17:13 | sbalneav: what's your username on the wiki?
| |
| 17:14 | you are a administrator, bureaucrat and developer
| |
| 17:14 | <sbalneav> ScottBalneaves
| |
| 17:15 | <jammcq> there's also a 'Sbalneav'
| |
| 17:20 | <sbalneav> Ah, do it to ScottBalneaves.
| |
| 17:23 | komunista has joined IRC (komunista!~slavko@adsl-195-168-244-224.dynamic.nextra.sk) | |
| 17:24 | <jammcq> sbalneav: done
| |
| 17:32 | EDinNY has joined IRC (EDinNY!~ed@207.207.28.129) | |
| 17:39 | <EDinNY> Is there a bot in here?
| |
| 17:40 | <alkisg> Yup, ltsp, why?
| |
| 17:41 | <EDinNY> Want to demonstrate a bot to a class in internet security
| |
| 17:42 | <alkisg> !ltsp
| |
| 17:42 | <ltsp> ltsp: LTSP is the Linux Terminal Server Project, the open source thin client solution. You can find it at http://www.ltsp.org
| |
| 17:42 | <alkisg> But better do that in #test
| |
| 17:42 | Not in a "real" channel
| |
| 17:43 | <EDinNY> thanks
| |
| 18:01 | vagrantc has joined IRC (vagrantc!~vagrant@75-150-46-245-Oregon.hfc.comcastbusiness.net) | |
| 18:01 | vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc) | |
| 18:17 | <baptiste> !ask
| |
| 18:17 | <ltsp> ask: Don't ask to ask a question, simply ask it, and if someone knows the answer, they'll respond. Please hang around for at least a full hour after asking a question, as not everybody constantly monitors the channel.
| |
| 18:18 | <baptiste> that's a nice feature
| |
| 18:20 | then, even if you're not a bot the /exec -o command (at least in irssi) can be useful
| |
| 18:22 | alexqwesa__ has left IRC (alexqwesa__!~alex@alexo-veto.broker.freenet6.net, Ping timeout: 276 seconds) | |
| 18:27 | EDinNY has left IRC (EDinNY!~ed@207.207.28.129, Ping timeout: 264 seconds) | |
| 18:32 | hays_ has joined IRC (hays_!~quassel@unaffiliated/hays) | |
| 18:32 | hays has left IRC (hays!~quassel@unaffiliated/hays, Quit: No Ping reply in 180 seconds.) | |
| 18:43 | anunnaki_afk is now known as anunnaki | |
| 18:59 | Gremble has left IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginmedia.com, Quit: I Leave) | |
| 20:15 | Gremble has joined IRC (Gremble!~Ben@cpc35-aztw23-2-0-cust207.18-1.cable.virginmedia.com) | |
| 20:19 | staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Read error: Connection reset by peer) | |
| 20:29 | enslaver has joined IRC (enslaver!~enslaver@c-98-196-42-169.hsd1.tx.comcast.net) | |
| 20:32 | <enslaver> Ok so i'm aiming for a March 17th release date (Code name: Clover) for the Redhat release, that should give us the hackathon week to incorporate everything, give it any last minute changes and prepare the packages for release.
| |
| 20:34 | <vagrantc> cool!
| |
| 20:35 | staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu) | |
| 20:35 | <enslaver> I don't think i'll be putting any of the 'cool changes' i've made for the first release, its more of a basic working start. then i'll come back and add iscsi and all that neat stuff
| |
| 20:36 | i've gotten mock working to build the chroot, iscsi target booting, 2 other copy-on-write file system choices, no lightdm yet tho
| |
| 20:37 | I'm going to focus on fat clients and lightdm at the same time
| |
| 20:38 | That should deliver me into the Fedora world quite a bit
| |
| 20:40 | <warren> enslaver: which distros?
| |
| 20:40 | <enslaver> of fedora?
| |
| 20:41 | Starting with Fedora 18
| |
| 20:41 | <warren> enslaver: I mean the march 17th release
| |
| 20:42 | <vagrantc> enslaver: so not even bothering with fatclients with LDM ?
| |
| 20:42 | <enslaver> warren: My distro's I'm running for devel are Centos, Scientific and RH official
| |
| 20:42 | vagrantc: They are part of the mock chroots i'll be incorporating
| |
| 20:43 | I even wrote a wrapper install script that i'll be publishing as well for easy install, so someone can just wget http://url | bash
| |
| 20:44 | all it really does is change some sysctl stuff and rpm install the ltsp-release rpm (EPEL soon hopefully)
| |
| 20:45 | <warren> why wget | bash ? that's incredibly insecure
| |
| 20:46 | <enslaver> i see that a lot on sites for installs
| |
| 20:46 | later on i'll write a 0install version
| |
| 20:46 | that way all the distros can use
| |
| 20:48 | maybe even a ccboot version of ltsp? :)
| |
| 20:58 | j/k btw, don't start thinking I'm some windows fan
| |
| 21:02 | <alkisg> LTSP for windows would be a best seller though :D
| |
| 21:03 | <enslaver> We may have to redesign the shopping cart to support that many purchases
| |
| 21:04 | i'd do it but my active directory server is down and i can't even login
| |
| 21:04 | <alkisg> No worries we can reimplement AD in shell in a few days
| |
| 21:05 | <enslaver> Posix compliant AD?
| |
| 21:05 | :P~
| |
| 21:05 | <alkisg> LSB, even :D
| |
| 21:06 | <enslaver> gasp
| |
| 21:06 | Uh you mean WSB right?
| |
| 21:07 | <vagrantc> What?! Standards? base
| |
| 21:08 | vmlintu has left IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi, Ping timeout: 240 seconds) | |
| 21:08 | <enslaver> or would that be MSB
| |
| 21:08 | jammcq has left IRC (jammcq!~jam@c-69-245-75-255.hsd1.mi.comcast.net, Quit: leaving) | |
| 21:10 | <Phantomas> (btw, I can redesign the shopping cart and write it in ASP, running in an IIS, if you give me 50% of LTSP's income)
| |
| 21:11 | <alkisg> sbalneav: if you finish the LDM exit hook today, please notify me somehow (e.g. commit it :)), I have some free time tomorrow to work on rewritting the cleanup shell scripts that the hook will call
| |
| 21:12 | <Phantomas> alkisg: I'm rewritting epoptes in Visual Basic btw
| |
| 21:13 | <alkisg> Phantomas: I hope you're using the cool vb6 and not the new lame .net stuff, right?
| |
| 21:13 | So that we're still compatible with windows 95...
| |
| 21:14 | <Phantomas> of course! And an addon system in C#
| |
| 21:14 | * vagrantc fails to validate the date as 2013-04-01 | |
| 21:14 | <Phantomas> Here in Greece it's 01-04-2013
| |
| 21:15 | <enslaver> is there an rpm for epoptes?
| |
| 21:16 | <vagrantc> Phantomas: and here in the use it'd be 04/01/2013
| |
| 21:16 | US
| |
| 21:16 | that's why i used an internationalized ISO standard :P
| |
| 21:16 | <Phantomas> vagrantc: yep, could you convert it in inches please? :P :P
| |
| 21:17 | <vagrantc> Phantomas: i only have codeto convert to furlongs
| |
| 21:17 | Phantomas: but you could write a module to convert from furlongs to parsecs to inches, i'm sure.
| |
| 21:19 | <Phantomas> nah, I prefer miles as a unit for everything, or gallons!
| |
| 21:20 | enslaver: I think there is
| |
| 21:23 | <enslaver> I got an idea off red hat's website, they use a java based NX viewer applet, incorporating that would allow a management interface in php/java
| |
| 21:23 | Parker955_Away is now known as Parker955 | |
| 21:23 | <Phantomas> http://software.opensuse.org/download.html?project=server:ltsp&package=epoptes
| |
| 21:26 | garymc has left IRC (garymc!~chatzilla@host81-148-18-165.in-addr.btopenworld.com, Ping timeout: 240 seconds) | |
| 21:28 | <enslaver> ah nice, thats a good src to start with
| |
| 21:29 | epoptes the same as italc?
| |
| 21:30 | <Phantomas> with a better name
| |
| 21:31 | <vagrantc> and arguably a better implementation...
| |
| 21:31 | <Phantomas> ah, forgot about that :p
| |
| 21:31 | <enslaver> italic requires hast?
| |
| 21:31 | yast*
| |
| 21:34 | <Phantomas> I don't think yast has anything to do with italc, but use epoptes! :D
| |
| 21:35 | <enslaver> I was just reading the opens use description, i read the actual italc site and see now what it is
| |
| 21:35 | i could easily port epoptes to el6 and fc18
| |
| 21:36 | lemme make some rpms
| |
| 21:37 | Phantomas: did you code epoptes?
| |
| 21:38 | <Phantomas> enslaver: alkisg did, I am in the marketing department of Epoptes Inc.
| |
| 21:38 | report all the bugs to him
| |
| 21:39 | okay, just kidding, I wrote a couple of lines too :)
| |
| 21:41 | (but not those with the bugs)
| |
| 21:42 | <enslaver> lol
| |
| 21:43 | ok good since you know the code a bit i might have a Q for you while porting this
| |
| 21:45 | update-desktop-files ? that a ubuntu thing?
| |
| 21:45 | <Phantomas> porting it to what? Are you trying to make rpm packages?
| |
| 21:45 | <enslaver> el6
| |
| 21:47 | python 2.6 supported?
| |
| 21:52 | <Phantomas> yep
| |
| 22:00 | <enslaver> last tag was at 328, latest revision is 378, is 378 revision stable?
| |
| 22:01 | alexqwesa__ has joined IRC (alexqwesa__!~alex@109.172.12.47) | |
| 22:01 | <Phantomas> No, it hasn't been tested.
| |
| 22:24 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.) | |
| 22:31 | alexqwesa__ has left IRC (alexqwesa__!~alex@109.172.12.47, Quit: Хана X'ам !!!) | |
| 22:31 | alexqwesa has joined IRC (alexqwesa!~alex@109.172.12.47) | |
| 22:54 | komunista has left IRC (komunista!~slavko@adsl-195-168-244-224.dynamic.nextra.sk, Quit: Leaving.) | |
| 23:15 | Parker955 is now known as Parker955_Away | |
| 23:45 | <enslaver> response = [('handles', amp.ListOf(amp.Unicode()))]
| |
| 23:45 | ./daemon/commands.py line: 41
| |
| 23:46 | What python module is ListOf part of?
| |
| 23:50 | mikkel has left IRC (mikkel!~mikkel@80-71-132-15.u.parknet.dk, Quit: Leaving) | |
| 23:51 | <Phantomas> enslaver: twisted.protocols.amp
| |
| 23:52 | <enslaver> that must be what I'm missing, theres no protocols included with twisted
| |
| 23:53 | <Phantomas> enslaver: do you have python-twisted-core?
| |
| 23:53 | <enslaver> yes
| |
| 23:53 | /usr/lib64/python2.6/site-packages/twisted/protocols/amp.py
| |
| 23:55 | <Phantomas> so there is a protocols module, right?
| |
| 23:55 | but when you do import twisted.protocols.amp it doesn't work?
| |
| 23:55 | <enslaver> the import works, yet no ListOf in there
| |
| 23:56 | <Phantomas> which version of python-twisted-core do you have?
| |
| 23:56 | <enslaver> python-twisted-core-8.2.0-4.el6.x86_64
| |
| 23:59 | <Phantomas> wow, pretty old
| |
| 23:59 | <vagrantc> RHEL makes debian look young again :)
| |