[12:54] <KatteKrab> howdy peoples
[12:55] <KatteKrab> RichEd: ping?
[02:39] <bddebian> Heya
[02:54] <freet15> Hi
[04:16] <BHSPitLappy> hi, all
[08:55] <cbx33> mornin all
[08:55] <cbx33> hey willvdl
[08:55] <cbx33> anything intersting in the meting after I left?
[09:03] <BHSPitLappy> yo
[10:20] <cbx33> hey peeps
[10:22] <highvoltage> hey peep
[10:23] <cbx33> hi highvoltage
[10:23] <cbx33> howz it going
[10:24] <highvoltage> going well thanks
[10:25] <highvoltage> I have a lot to do but I'm having trouble getting out of bed :)
[10:25] <highvoltage> need to get my passport sorted out, hair cut, among other semi-pointless things. but I think I'll do that tomorrow.
[10:25] <highvoltage> been bugsquashing this morning, much more fun :)
[10:30] <cbx33> heheeh
[11:01] <kgoetz> can i ask ubuntu ltsp questions here? i'm not getting much luck with #ltsp
[11:01] <cbx33> sure
[11:01] <cbx33> fire away
[11:02] <kgoetz>  my LTSP client (dapper based) is getting to "* Loading hardware drivers..." then just says "nfs:
[11:02] <kgoetz> server 192.168.0.2 not responding, still trying" after a few minutes (and keeps repeating the nfs error).
[11:02] <kgoetz> any thoughts? :(
[11:02] <cbx33> sounds as if it's telling the truth ;)
[11:02] <cbx33> have you confirm the nfs is exported ok?
[11:04] <kgoetz> how can i confirm it? it used to work ok  :\
[11:04] <cbx33> hmmm
[11:04] <cbx33> well do you know how to mount an nfs share manually
[11:04] <kgoetz> no :\
[11:04] <cbx33> ok....do you have another machine on the network?
[11:05] <cbx33> hang on two secs
[11:05] <kgoetz> i can grab another laptop to plug in. brb.
[11:06] <cbx33> hanve you tried restarting the nfs server?
[11:06] <kgoetz> yes, both the service and the system its on
[11:07] <cbx33> ok
[11:07] <cbx33> i can't remember the syntax exactly and I don;t have an nfs share anywhere handy to test
[11:07] <cbx33> it's something like
[11:08] <cbx33> mount -t nfs 182.23.45.23:/path/to/nfs /mnt/directory
[11:08] <cbx33> try that
[11:10] <kgoetz> sudo mount -t nfs 192.168.0.2:/opt/ltsp /mnt works (ls /mnt gives i386). thats on the local host untill i get the lappy booted
[11:10] <cbx33> ok good
[11:10] <cbx33> so the nfs works
[11:10] <cbx33> what about grepping the logs to see what happens?
[11:12] <kgoetz> i dont see anything special in syslog, is ther another place?
[11:13] <kgoetz> hm. new error
[11:13] <kgoetz> nfs: rpc call returned error 101 <- on laptop pxe booting
[11:14] <cbx33> hmmm
[11:14] <cbx33> are you sure the dns is working on? - I've had reverse DNS errors before
[11:15] <kgoetz> its an ltsp-server-standalone setup. i hope that stuff is dealt with
[11:15] <kgoetz> it was in my other setups :/
[11:15] <cbx33> ok
[11:16] <kgoetz> only difference (apart from one being a laptop,a nd one a desktop) is that the desktop is getting a static dhcp address
[11:16] <cbx33> hmmm
[11:16] <kgoetz> actually, i'm wrong
[11:17] <cbx33> is the address being assigned by another dhcp server?
[11:17] <kgoetz> they are both in dhcp configs
[11:17] <kgoetz> no, its on a crossover cable
[11:17] <cbx33> ahhh
[11:18] <cbx33> is it communicating ok over that cable?
[11:18] <kgoetz> well, its pxe booting, and going fine to the nfs bmounting bit
[11:19] <kgoetz> * seems to be communicating fine
[11:19] <cbx33> hmm....
[11:19] <cbx33> so tftp is working
[11:19] <cbx33> you sure that the nfs path is being servered right from the dhcp server
[11:20] <kgoetz> its serving /opt/ltsp/i386, iirc correct?
[11:21] <cbx33> yeh i think so
[11:22] <kgoetz> would the host bein  outside the dhcp range cuase issues?
[11:23] <cbx33> I'm not too hot on the set
[11:23] <cbx33> look in
[11:23] <cbx33> see what path / host range is accepted
[11:23] <cbx33>  /etc/exports on the server
[11:24] <kgoetz>  /opt/ltsp *(ro,no_root_squash,async)
[11:25] <cbx33> ahh....
[11:25] <cbx33> does it need to be /opt/ltsp/i386 now?
[11:25] <cbx33> I'm not sure tbh
[11:26] <kgoetz> that matches with what i have on a working production server, so i think its right :|
[11:27] <cbx33> hmmm
[11:27] <kgoetz> brb. heard thunder, making sure dogs are ok
[11:27] <cbx33> have you tried googleing for the exact error message?
[11:30] <kgoetz> cbx33: got some tabs open atm. hadnt looked at them yet
[11:31] <cbx33> ok
[11:32] <cbx33> sorry about that
[11:32] <kgoetz> wb
[11:47] <kgoetz> i'm going to take > 1 hour to dist-upgrade :/ i might not be albe to work on this again until tomorrow :(. thanks for the help cbx33
[11:47] <cbx33> np
[11:48] <cbx33> sorry i couldn't do more
[11:48] <kgoetz> thanks for yoru effort :)
[11:49] <cbx33> let me know what happens
[11:52] <kgoetz> will do :)
[12:41] <willvdl> highvoltage, passport? where you of to?
[03:23] <cbx33> hey all
[03:26] <utk> Hey, can you help me out ?
[03:41] <bddebian> Heya
[03:41] <highvoltage> heya bddebian
[03:41] <bddebian> Hi highvoltage
[03:42] <highvoltage> edubuntugirl: pipedream is Jan Groenewald
[03:42] <edubuntugirl> highvoltage: sure thing
[03:42] <highvoltage> edubuntugirl: wizzy is Andy Rabagliati
[03:42] <edubuntugirl> highvoltage: righto
[03:58] <sbalneav> ogra: Hey, vagrant tells me you've come up with a python alternative to lp_server?
[04:02] <cbx33> ahhh python
[04:02] <pipedream> heh
[04:05] <cbx33> hey sbalneav
[04:05] <sbalneav> ICMP ECHO REPLY
[04:06] <cbx33> howz it going?
[04:06] <cbx33> have you ever done any work on the ldm glade file?
[04:07] <sbalneav> No!  Learning glade + python is VERY high on my list of things to do!  I REALLY need to learn how to do gui devel, and glad + python seems the perfect way to get into it.  Got any good tutorials?
[04:07] <cbx33> sbalneav, hold there 2 seconds
[04:07] <cbx33> http://www.progbox.co.uk/wordpress/?p=183
[04:09] <sbalneav> Cool!  Thankx
[04:09] <cbx33> hope it helps
[04:09] <cbx33> had some great feeback from it
[04:09] <sbalneav> s/x$/s/
[04:09] <cbx33> s/x/s/ would have been sufficient no?
[04:09] <cbx33> :p
[04:12] <sbalneav> Sure, but precision always counts :)
[04:12] <cbx33> hehe
[04:13] <ogra> sbalneav, https://launchpad.net/people/ogra/+branch/ltsp/feisty-ltsp-jetpipe
[04:13] <cbx33> ogra, I'm here now if you're available for meeting?
[04:13] <cbx33> or 7 if not
[04:13] <sbalneav> Had I said: "Cool!  xylophones are sexy, thankx", yours would have not worked :)
[04:13] <cbx33> sbalneav, true
[04:15] <ogra> sbalneav, for now it does only usb and parallel printers, is missing all kinds of error catching etc, the code does only show the principle but needs a lot fleshing out ...
[04:15] <cbx33> however "Thankx that's great I'll look later thankx again would have failed too :p
[04:15] <ogra> cbx33, 7 is better, i'm fighting with a weird screensaver bug atm
[04:15] <cbx33> heheh np
[04:15] <sbalneav> ogra: I need serial printers, so guess who's going to help you with this? :)
[04:15] <cbx33> ogra, did you get my message about BETT?
[04:16] <sbalneav> fetch phase 1/4 :)
[04:17] <ogra> sbalneav, just look at the diff from the last revision ... that should be everything ...
[04:17] <ogra> cbx33, no ?
[04:18] <sbalneav> ogra: I think this is a great idea.  solves the whole licensing issue nicely.
[04:19] <cbx33> I've convinced my boss to let us take some kids up there
[04:19] <ogra> sbalneav, yep
[04:22] <sbalneav> ogra: heh, nice and simple.  OK, all we need to do is pull in python.serial, and parse command line options to set up the serial port properly.
[04:22] <ogra> yeah
[04:22] <cbx33> it's gonna be a great BETT show for edubuntu
[04:22] <cbx33> ;)
[04:23] <ogra> i didnt know about python.serial :D ... i thought about a pipe to setserial, but found that to ugly :)
[04:28] <sbalneav> python.serial looks easy.  It opens a file like object just like the open does, so all we'll have is just some code to select wether or not it's a "simple" device like /dev/lp... or /dev/usb/lp..., or a /dev/ttyS....  Then just do the proper open.  We'll probably only add about 15 lines of code to do that.
[04:28] <sbalneav> I'll look at that today.
[04:29] <sbalneav> One comment: why don't we pull jetpipe out of the ltsp tree, and just make it a standalone package.  It would be useful generally, and not just in the context of ltsp. :)
[04:33] <sbalneav> heh, copyright notice is longer than the code.
[04:35] <ogra> sure, lets make it standalone ... for now i just wanted a quick fix for feisty users so they can print ... the merge with debian wiped lp_server from our code as well and i didnt want to just re-add it
[04:36] <sbalneav> yeah, understood.
[04:36] <sbalneav> Which one?
[04:36] <ogra> https://bugs.launchpad.net/distros/ubuntu/+source/gnome-screensaver/+bug/76623
[04:37] <sbalneav> Oh, that's easy.
[04:37] <ogra> i cant reproduce it, but half the world seems to have it
[04:37] <sbalneav> gconftool-2 --direct \
[04:37] <sbalneav>             --config-source xml:readwrite:/etc/gconf/gconf.xml.mandatory \
[04:37] <sbalneav>             --type string \
[04:37] <sbalneav>             --set /apps/gnome-screensaver/mode blank-only
[04:38] <sbalneav> That solves all problems with screensavers, I find :)
[04:38] <ogra> oh, crap, wrong bug No.
[04:38] <cbx33> heheh
[04:38] <ogra> https://launchpad.net/distros/ubuntu/+source/gnome-screensaver/+bug/76632
[04:38] <ogra> 76623 has a patch and all, i just need to merge it ...
[04:39] <cbx33> not a problem here
[04:39] <cbx33> can you get soemone to do a trace on it?
[04:40] <ogra> there is a trace attached ...
[04:40] <cbx33> oh sorry
[04:40] <cbx33> bbiab
[04:40] <ogra> (oh, its attached to one of the duplicates, sorry)
[04:41] <cbx33> ahh
[04:41] <cbx33> i thought i didn't see one there
[04:41] <cbx33> any more thoughts on the gnome canvas thing?
[04:43] <bddebian> cbx33: Have an error message?
[04:43] <cbx33> basicalyl a required catalog gnomecanvas could not be found
[04:43] <cbx33> blah
[04:45] <bddebian> cbx33: Well that's not very helpful :)
[04:46] <cbx33> heheh
[04:46] <cbx33> nope
[04:46] <cbx33> i got to dash
[04:46] <cbx33> will look closer later
[04:46] <bddebian> Laterz
[05:47] <highvoltage> !seen pips1
[05:47] <ubotu> I last saw pips1 (n=AZ2304@actz87.neoplus.adsl.tpnet.pl) 3d 20h 53m 19s ago, quiting: "Azureus 2.3.0.4"
[08:05] <cbx33> ogra, meeting?
[08:06] <ogra> gimme a minute
[08:06] <cbx33> ok np
[08:06] <ogra> but yes
[08:07] <cbx33> crimsun_, ever had it where a small mic sounds like it's got a ring mod on it?
[08:17] <crimsun_> cbx33: check your capture settings
[08:18] <cbx33> ok
[08:18] <cbx33> thanks crimsun_
[08:20] <ogra> cbx33, oook
[08:20] <cbx33> ogra,  oggggg
[08:21] <ogra> i assume you read the SCP spec
[08:21] <cbx33> yes
[08:21] <cbx33> meeting here or in pm?
[08:21] <ogra> here is fine
[08:21] <ogra> so, the first and most important part is the split into front and backend
[08:22] <cbx33> agreed
[08:22] <ogra> i'm not concerned if any of the other features need to wait until feisty+1, but we need the infrastructure for multiple server support now
[08:22] <ogra> so that should be the focus
[08:22] <cbx33> ok
[08:22] <cbx33> right ok
[08:22] <cbx33> well I can split that up fairly easily
[08:23] <cbx33> not a huge task really
[08:23] <ogra> yep
[08:23] <cbx33> do you have any idea about how you want classes/functions arrangeD?
[08:23] <cbx33> or shall I do what I think there?
[08:23] <ogra> do what you think, but i guess you need to end up with frond and backe3nd classes
[08:24] <ogra> *front
[08:24] <cbx33> yeh
[08:24] <cbx33> are we splitting into two pacakges?
[08:24] <ogra> hmm
[08:24] <ogra> good question
[08:24] <cbx33> which would eventually become 3
[08:24] <cbx33> k,g,base
[08:24] <ogra> will it make sense to have a commandline gui, soemthing with curses ?
[08:24] <cbx33> hmmm
[08:24] <ogra> hmm, KDE, right
[08:25] <cbx33> I'm not sure
[08:25] <ogra> yeah, then lets have a split package
[08:25] <cbx33> i think it's the most sensible way to go
[08:25] <cbx33> I think command line would be great
[08:25] <ogra> at least SCP-backend SCP-gnome
[08:25] <ogra> the kde part is up to others
[08:25] <cbx33> I'm not saying I'm gonna write the kde frontend
[08:25] <cbx33> though I think I could
[08:26] <ogra> python-qt shouldnt be to hard, but thats really something the KDE community can do
[08:26] <cbx33> command line would be good....as I would have...if it was fully implemented here....need to have ssh access to it
[08:26] <cbx33> that would be a good idea
[08:26] <ogra> well, commandline is defined in the spec
[08:26] <ogra> but not with curses ui ...
[08:27] <cbx33> yeh
[08:27] <ogra> but i think it wouldnt make sense anyway
[08:27] <cbx33> is curses tieable with python?
[08:27] <ogra> raw commandline is fine ... for sending messages etcv
[08:27] <cbx33> ok
[08:28] <cbx33> ok so the split + package split is main concern
[08:28] <cbx33> now we need to look at Multiple server control
[08:28] <ogra> leright
[08:28] <ogra> -le
[08:29] <cbx33> so
[08:29] <cbx33> when you say multiple servers
[08:29] <cbx33> from the spec.....i see it meaning the following
[08:29] <cbx33> one front end can talk to several servers
[08:29] <cbx33> but only one at a time
[08:29] <cbx33> am I right here?
[08:29] <ogra> right
[08:30] <ogra> in later releases we can spoil that on the UI side and make it talk to multiple servers transparently
[08:30] <cbx33> do you think at some stage it would be useful to do true multiple servers?
[08:30] <cbx33> yes
[08:30] <cbx33> I really think SCP could be a big selling point for Edubuntu
[08:30] <ogra> for now one server<->one client is fine
[08:30] <cbx33> ok
[08:30] <ogra> lets grow slowly, but with good quality ;)
[08:30] <cbx33> what else on the list do you see being priority for feisty
[08:31] <cbx33> if I had the time I'd work on it all day everyday
[08:31] <cbx33> ;)
[08:31] <cbx33> get it to be a fantastic app
[08:31] <cbx33> sadly I have to goto work
[08:31] <ogra> well, blank/lock seems like a low hanging fruit
[08:31] <cbx33> will need tie in to x11vnc?
[08:31] <ogra> i.e. like implementable in less than 30min
[08:32] <ogra> yeah, we'll need the vnc integration
[08:32] <cbx33> how do you see that being implemented
[08:32] <cbx33> right ok
[08:32] <ogra> well it was described in the former spec
[08:32] <cbx33> yes it was
[08:32] <cbx33> now I don;t think I'm qualified to touch x11vnc
[08:32] <ogra> ah, and copied over :)
[08:32] <cbx33> but I would like to know what's goign on there
[08:33] <ogra> thats fine, i'll care for that part
[08:33] <cbx33> ok...
[08:33] <cbx33> if you could lemme see a diff when you're done
[08:33] <cbx33> that'd be great
[08:33] <cbx33> I'm keen to learn
[08:33] <ogra> indeed
[08:33] <ogra> i'll keep the work in bzr, you can diff every small change ;)
[08:34] <cbx33> ok
[08:34] <Burgwork> breaking it out as a backend means somebody could right a frontend, or an ebox plugin, etc.
[08:34] <cbx33> yes
[08:34] <Burgwork> a web frontend, rather
[08:34] <ogra> Burgwork, yep
[08:34] <Burgwork> ogra: have you looked at ebox?
[08:34] <ogra> even though i'd love to see vnc through html working first :P
[08:34] <cbx33> haha
[08:34] <Burgwork> everybody I have seen do it uses java
[08:35] <cbx33> that would be awesome.....
[08:35] <ogra> Burgwork, only at the website, i didnt install it
[08:35] <cbx33> but..
[08:35] <Burgwork> ogra: imbrandon has a setup and I am about to get one going
[08:35] <ogra> well, anyway, web is beyond our focus atm, but the API will be there
[08:35] <Burgwork> looks really interesting, is actively developed, have a number of DDs on staff
[08:35] <cbx33> Burgwork, care to blog/howto about it once you're done?
[08:35] <Burgwork> cbx33: will do
[08:35] <Burgwork> it requires some hacking to run on Feisty, due to upstart
[08:35] <cbx33> is something I'm keen at looking at too
[08:35] <cbx33> ok
[08:36] <cbx33> ogra, are we going for this pop up install x11vnc pacakge thing?
[08:36] <ogra> yep
[08:36] <cbx33> explain why it can't be a dep?
[08:36] <Burgwork> can't you use vino?
[08:36] <ogra> because you cant install a dep inside a chroot
[08:36] <cbx33> oh i see
[08:36] <cbx33> yeh sorry
[08:36] <ogra> ;)
[08:36] <cbx33> can't it be a dep for the chrrot?
[08:37] <cbx33> i mean when building the chroot?
[08:37] <ogra> no
[08:37] <ogra> that would make *all* ltsp installs vnc servers
[08:38] <ogra> thats not what we want
[08:38] <cbx33> ok
[08:38] <ogra> it needs to be installed dependant on SCP being installed
[08:38] <cbx33> ok
[08:38] <ogra> what do you think about userlist filtering ?
[08:38] <Burgwork> ogra:  have you seen http://fedoraproject.org/wiki/Desktop/FastUserSwitching
[08:38] <Burgwork> ?
[08:39] <ogra> seems easy enough to me
[08:39] <cbx33> yeh i think so
[08:39] <ogra> Burgwork, nope, i havent
[08:40] <Burgwork> Davidz is working on it
[08:40] <ogra> aha
[08:40] <Burgwork> it is an FC7 goal
[08:40] <Burgwork> it looks interesting for the "who is where" use case
[08:40] <ogra> i'm usually not a big fan of davidz's GUI works ... lets see
[08:41] <ogra> cbx33, ok, so i see the prio as follows:
[08:41] <Burgwork> is not a gui
[08:41] <ogra> * Multiple server control (split front/backend)
[08:41] <cbx33> yup
[08:41] <ogra> * Blank / Lock Workstations
[08:42] <ogra> (because its so easy)
[08:42] <ogra> * Remote desktop access
[08:42] <ogra> * Filter functionallity of the userlist
[08:42] <ogra> * Tiled class overview via VNC
[08:42] <sbalneav> ogra: I'm working through jetpipe.  I'm going to make it command line equivalent to the old lp_server.  I.e. a -w, -d <device> -t <stty opts>, etc.  That ok with you?
[08:42] <cbx33> is that bottom one doable in feisty you think?
[08:42] <ogra> i'll care about the remote desktop access stuff and about renaming etc
[08:43] <ogra> sbalneav, totally, i was just to lazy :)
[08:43] <sbalneav> No problem.
[08:43] <cbx33> ogra, you lazy...?
[08:43] <ogra> cbx33, the bottom one is a nice to have
[08:43] <sbalneav> Ogra's the *last* person in the world I;d call lazy.
[08:43] <cbx33> me too
[08:43] <cbx33> ogra, I think it would be awesome
[08:44] <ogra> and sharing teachers screen is likely for feisty+1
[08:44] <cbx33> posibility to make those thumbnails clickable?
[08:44] <cbx33> for a full screen preview?
[08:44] <ogra> yep
[08:44] <cbx33> possible to take control of screen?
[08:44] <ogra> but thats a huge task
[08:44] <ogra> and i'm fearing we end up with it half implemented
[08:44] <cbx33> which ?
[08:44] <ogra> the tiled VNC thing
[08:44] <cbx33> just making it bigger
[08:44] <cbx33> oh
[08:44] <cbx33> what do you see as the problems?
[08:45] <ogra> so if you want to work on it, please do it in a separate bzr branch
[08:45] <cbx33> ok
[08:45] <ogra> vnc is a beast
[08:45] <cbx33> heheh
[08:45] <cbx33> a beast we shall tame
[08:45] <ogra> well, but it will take its time ...
[08:45] <cbx33> of course
[08:45] <ogra> i dont want to rape it ;)
[08:45] <cbx33> I'm not saying it'll be done over night ;)
[08:46] <ogra> just tame :)
[08:46] <cbx33> ok
[08:46] <cbx33> so I now have something to work on after christmas
[08:46] <ogra> yeah
[08:46] <cbx33> doubt I'll get anything done before
[08:46] <cbx33> if I'm being perfectly honest
[08:47] <ogra> nobody forces you to ;)
[08:47] <cbx33> but i like to
[08:47] <cbx33> and I hate letting you all down if I dont
[08:47] <cbx33> oh ogra one last thing....any word on how to get this glade working
[08:47] <cbx33> for the ldm screen
[08:47] <cbx33> ?
[08:48] <Burgwork> cbx33: always better to under promise and over deliver
[08:48] <ogra> to be honest, i have no idea whats wrong there
[08:48] <cbx33> grr
[08:48] <cbx33> sounds like a bug to me
[08:48] <cbx33> do you know what the catalogs it's talking about are?
[08:48] <cbx33> and where they are located?
[08:48] <ogra> it may be that gnomecanvas isnt supported anymore by recent glade implementations
[08:48] <cbx33> that would suck
[08:48] <ogra> well, gnomecanvas is dying in favor of cairo
[08:48] <cbx33> leaves us with an out of date ldm glade file
[08:48] <Burgwork> I wonder if I could get Userful to merge our stuff into SCP...
[08:49] <ogra> but cairo isnt there yet
[08:49] <cbx33> Burgwork, explain?
[08:49] <ogra> so there is an unfilled gap
[08:49] <cbx33> yeh
[08:49] <Burgwork> ogra: we have a web-based thingy like SCP, called DiscoverAssisst
[08:49] <ogra> you can try to redo the ldm engine in cairo if you have to much spare time ;)
[08:50] <cbx33> dued I've never even used cairo
[08:50] <cbx33> not about to start screwing with that yet
[08:50] <cbx33> too much else to do
[08:50] <cbx33> though I'd like to
[08:50] <ogra> Burgwork, aha, well, i guess you could merge stuff, yes
[08:50] <cbx33> I wouldn't really know where to start
[08:51] <ogra> a cairo rewrite is on my list for ldm ... but its far down at the bottom atm ...
[08:51] <cbx33> heh
[08:51] <cbx33> maybe when you get to that stage I could help out
[08:53] <ogra> it will take some releases to get that far down the list :)
[08:53] <ogra> but i'll think of you if i get there
[08:53] <cbx33> ok
[08:54] <Burgwork> ogra: if you split the backend, life gets more interesting
[08:54] <ogra> ok, i'm afk until distro meeting ...
[08:54] <Burgwork> http://userful.com/products/discoverassist
[08:54] <ogra> Burgwork, split ?
[08:54] <Burgwork> split SCP, sorry
[08:54] <cbx33> Burgwork, SCP is goign to be split
[08:54] <ogra> ah
[08:54] <ogra> yes
[08:54] <ogra> ok ... bbl
[08:56] <cbx33> thanks ogra
[08:56] <cbx33> it took us a while but has been really good
[08:59] <Burgwork> cbx33: once you split the backend, are there plans to make it work on standalone machines?
[09:00] <cbx33> explain?
[09:02] <Burgwork> make it so, if you have a lab of 30 machines, you can install the backend on each and the front end on one and have it work?
[09:03] <Burgwork> assuming each of those 30 machines is a fat client
[09:04] <Burgwork> ogra: have you guys talked with the stateless Linux people? Fedora is doing some cools things with systems management
[09:16] <cbx33> yes i was interested in stateless at one point
[09:18] <Burgwork> cbx33: basically, with ltsp fat clients, that is basically the same as stateless
[09:18] <cbx33> ah ok
[09:23] <willvdl> folks, I'm out. catch you tomorrow afternoon
[09:49] <cberlo> Hi folks.  Got a (hopefully) quick question:  how would I kill all processes from users that aren't logged on (execept daemons and system users)?
[09:50] <cbx33> write a script
[09:55] <Burgwork> cberlo: upon the user logging out, kill all their running processes
[10:09] <cberlo> Burgwork: Okay, and with ldm, how do I detect when they logout?
[10:09] <cberlo> (sorry, got tied up for a bit there)
[10:09] <Burgwork> I have no idea
[10:09] <cberlo> cbx33: a script would be a good idea.  I'll try that.  What would you put in it?  :)
[10:11] <sbalneav> Do all your users run the same kind of desktop?
[10:12] <sbalneav> i.e. gnome-session?
[10:12] <sbalneav> I guess, more to the point is: what's being left around that you feel you need to kill off?
[10:14] <Burgwork> sbalneav: in a public environment, you want to kill anything that people leave running, for security reasons
[10:15] <sbalneav> Burgwork: I'm trying to find out WHAT's being left around.  If it's only the users' gconfd, there's no harm in that staying there: it doesn't need to restart next time then.  If it's something else, then that may point to a bug, or a misconfiguration.
[10:16] <Burgwork> right
[10:18] <cberlo> sbalneav: bonobo is leaving a process behind that seems to affect nautilus upon re-logging in.  Basically, I'm not pulling a proper desktop, and plugging in USB storage devices isn't visually registering (but they are accessible if you know where to look, which I'm assuming my students won't).
[10:20] <sbalneav> If you kill off the bonobo process, do things work correctly, then?
[10:20] <cberlo> Don't have access to the server at the moment (just spent the afternoon testing a new lab with it, and noticed as I logged in and out of the account I used for testing that it occasionally left this one bonobo process running; killing it seemed to fix the Desktop glitches I was experiencing).
[10:20] <cberlo> sbalneav: In short, yes.  :)
[10:21] <cberlo> sbalneav: I had thought to use ps aux to narrow down the running processes, but I wind up with an assortment of daemons and such in the mix, so grepping to remove what I know should not be killed is a lengthy process.
[10:22] <cberlo> sbalneav: Basically what I was trying to do is grab a list of who's running processes vs. who's logged in (except for the root process) and kill any processes that were running but didn't belong to anyone logged in.
[10:22] <sbalneav> OK, well, there's two courses of action here.  Most importantly is to file a bug on Bonobo, as it should exit correctly upon session termination.  The second would be running a script to clean up.  Something along the following lines.
[10:22] <sbalneav> #!/bin/bash
[10:23] <sbalneav> for each user in $(getent passwd | awk '$3 > 999 {print $1}'); do
[10:24] <BHSPitLappy> got a little question
[10:25] <BHSPitLappy> we have an ubuntu desktop running at home, and I was curious if I could install edubuntu-desktop and have its environment run for just a particular user.
[10:25] <sbalneav>    pgrep -u ${user} gnome-session || pkill -u ${user} bonobo
[10:25] <sbalneav> done
[10:26] <cberlo> As soon as I have Internet access on that server again I will file an appropriate bug.  About that script...  what's the awk supposed to do?  I'm not fully versed in awk...
[10:26] <sbalneav> I typed that off the top of my head, so you'd need to play around with it a bit.
[10:27] <sbalneav> the awk looks at your passwd file, and any user that has a userid greater than 999 (i.e. 1000 and above) it will return the username
[10:27] <sbalneav> you can try it on the command line by itself
[10:27] <cberlo> right, just figured that part out.  :)
[10:27] <cberlo> Running it on my own laptop here, and watching the results.
[10:27] <ogra> that might kill yourself :)
[10:27] <sbalneav> that means you wont look for any root, etc processes.
[10:27] <sbalneav> yeah, you'd want ot be running that as root
[10:27] <sbalneav> not as a user :)
[10:28] <cberlo> Yes, so fairly safe to play with as me...
[10:28] <sbalneav> well, except for the pkill part.  You'll return your name, and kill your own processes
[10:28] <sbalneav> which will log you out.
[10:28] <cberlo> So the pgrep greps the processes for gnome-session then kills bonobo if it doesn't find it?
[10:29] <sbalneav> yes
[10:30] <sbalneav> that's a very simplistic way of doing it, but...
[10:30] <cberlo> Okay, here's my revision:  for user in $(getent passwd | awk '$3 > 999 {print $1}'|cut -f1 -d:); do echo $user; pgrep -u ${user} gnome-session || pkill -u ${user} bonobo; done
[10:30] <cberlo> Now, where can I put that so that it runs when someone logs off?  Or should I just stick it in a cron job?
[10:30] <sbalneav> ah, yeah, forgot the separator in awk.
[10:30] <sbalneav> awk -F: '$3......
[10:31] <sbalneav> then you can ditch the cut.
[10:32] <cberlo> good.  I'll learn awk yet!  :)
[10:33] <cberlo> So, ogra, what runs at logout that I can stick that little script in?
[10:33] <ogra> nothing, you need to stick it into ldm
[10:33] <cberlo> ogra: You'd be proud of me:  I used Edubuntu LTSP 6.10 without hacking the ldm out of it this time!  30 user setup.
[10:33] <cberlo> ogra: Okay, so how do I modify ldm, then?
[10:34] <ogra> nice :)
[10:35] <cberlo> Oh, and a side question:  what kind of memory would you suggest for the following configs:  10 user, 30 user, 60 user?  My estimate was 1GB, 2GB and 4GB.  Fair?
[10:38] <sbalneav> That seems reasonable.
[10:38] <cberlo> I gotta run.  I'll have to pick your brains again tomorrow.  Thanks for the assistance, sbalneav!  A lot cleaner than what I was working with.  And I'll get that bug report in ASAP!
[10:44] <ogra> ah, crap, he's gone
[10:44] <ogra> he could have used LDM_REMOTECOMMAND for what he wanted
[10:44] <ogra> err
[10:45] <ogra> LDM_REMOTECMD
[11:08] <blue-frog> hi aquickie if I may.. where the login window GUI saves the security information (automatic login...)?
[12:04] <Marco``> hi ppl
[12:04] <Marco``> is edubuntu available in multilanguage?
[12:05] <Marco``> I need it in german
[12:07] <Marco``> #ubuntu was faster ;)
[12:07] <Marco``> ty anyway