19:01:31 <h01ger> #startmeeting
19:01:31 <MeetBot> Meeting started Mon Jan 11 19:01:31 2010 UTC.  The chair is h01ger. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:01:31 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
19:01:32 <h01ger> huhu, MeetBot
19:01:34 <MeetBot> h01ger: Error: Can't start another meeting, one is in progress.
19:01:43 <h01ger> cheers
19:01:54 * pere = Petter Reinholdtsen
19:02:00 <h01ger> #topic 1. please indicate your presence - who will write the summary?
19:02:02 * vagrantc = Vagrant Cascadian
19:02:03 * debalance = Philipp Huebner
19:02:06 * h01ger = Holger Levsen
19:02:11 * Werner = Morten Werner Forsbring
19:02:20 * and1bm Andreas Mundt
19:02:40 <FBI> debian-edu: 3 de-build-guest committed revision 61208 to debian-edu: automatic commit after build from lenny-test-dvd
19:03:10 <h01ger> any volunteers to write+post the summary? meetbot.debian.net will help you :)
19:03:46 <h01ger> hi OleA, the meeting hasnt really started yet. still looking for a volunteer to write the summary
19:04:03 <h01ger> http://wiki.debian.org/DebianEdu/Meeting
19:04:06 <h01ger> has the agenda
19:04:11 <OleA> hi :)
19:06:09 * h01ger thinks fixing bugs and testing is more important than summary writing atm and proposes to do the meeting anyway. meetbot should give some summary too...
19:06:46 <pere> a meeting without a summary did not happen, so I see no point in calling random chatter a meeting when there is no summary.
19:07:02 <h01ger> meetbot provides one
19:07:02 <MeetBot> h01ger: Error: "provides" is not a valid command.
19:07:09 <Werner> :)
19:07:27 <h01ger> but ok, if we dont get started in the next 8min i'll watch some tv instead
19:07:46 <vagrantc> i can't say my attempts have ever felt like adding much value to meetbot's summaries
19:07:47 <h01ger> s/watch tv/read a book/ ;-)
19:08:24 <vagrantc> though i wouldn't say i've done it well, either
19:08:35 <pere> vagrantc: your summaries have been fine. :)
19:08:57 <vagrantc> pere: it's pretty much just taking meetbot's results and pasting them into the wiki...
19:09:02 <vagrantc> little formatting
19:09:10 <h01ger> and mailing to the list :)
19:10:20 <OleA> I'll write the summary
19:10:39 <h01ger> \o/
19:10:47 <h01ger> #info OleA will write the summary
19:10:59 <h01ger> #topic 2. lenny status
19:11:13 <h01ger> #topic 2.1 lenny status: #1409: Iceweasel does not save exception rule for https://www/lwat
19:11:33 * jever = Jürgen Leibner
19:11:50 <h01ger> hi jever
19:11:53 * OleA = Ole-Anders Andreassen
19:11:58 <jever> hi to all
19:12:22 * klausade = klaus ade johnstad
19:12:35 <h01ger> according to the ticket, there is still work to do for #1409
19:14:00 <pere> I suspect it is working now, but an error is logged during installation anyway.
19:14:09 * h01ger needs to update (read: reinstall) his test setup to be able to comment on this. and the last days there were too many changes to do that sensible :)
19:14:17 * klausade is insttalling lenny-test netinst from today.
19:14:18 <FBI> debian-edu: 3 de-build-guest committed revision 61211 to debian-edu: automatic commit after build from lenny-test-amd64-i386-powerpc-netinst
19:14:51 <h01ger> next topic?
19:14:59 <pere> the error is from snakeoil-on-ice executed within ltsp-make-client, and it fail because http://www is unavailable.
19:15:21 <pere> suspect the check in snakeoil-on-ice simply is checking things in the wrong order or something.
19:15:51 <pere> next, yeah.
19:15:54 * h01ger nods. added to the ticket
19:15:55 <h01ger> #topic 2.2 status lenny: 1408: Diskless Workstation not working yet
19:16:15 <pere> this one seem to be the dumping ground for everything diskless related.
19:16:35 <pere> the original problem was ldap connection related, and I never figured out what was wrong.
19:16:54 <pere> I did however find a lot of bugs with dvd installs leading to broken diskless chroots.
19:17:09 <pere> I have fixed all of these, and believe dvd installs should be working.
19:17:19 * debalance will test 2morrow
19:17:27 <h01ger> should we close it or is it still useful? i dont think we should keep unreproducible bugs at p2
19:17:34 <pere> I suspect the original problem was with dvd installs, or with virtualbox installs, but do not know.
19:17:42 <pere> was never able to reproduce the original problem.
19:18:02 <debalance> I was. always.
19:18:05 <OleA> I hve not had any problem logging in to diskless workstations for the last 3 weeks
19:18:15 <debalance> I'd suggest I test it 2morrow, with real hardware.
19:18:21 <klausade> i'm with OleA.
19:18:30 <h01ger> debalance, even if you are. if you cannot make other people reproduce the problem, its not p2
19:18:42 <debalance> I could
19:18:45 <debalance> but only with vbox
19:18:56 * h01ger uses vbox and couldnt reproduce it
19:19:09 <pere> did anyone reproduce it with the netinst cd?
19:19:28 * debalance never tried
19:19:55 <h01ger> #info debalance will check if #1408 is still there tomorrow
19:19:56 <klausade> pere: could not reproduce it with netinst.
19:19:56 <vagrantc> i'll find out hopefully by the end of the meeting
19:19:58 <pere> my dvd installed diskless workstations now boots, so I believe that part is fixed, if it was the problem.
19:20:03 <h01ger> vagrantc, please update the bug then
19:20:04 <vagrantc> with netinst
19:20:23 <h01ger> #topic 2.3 status lenny: 1138: We need to check and act on the non-free packages on the DVD
19:21:05 <pere> #info h01ger said he was going to look at this
19:21:34 <pere> #info I created a wiki page to keep information on what we figured out for each non-free package.
19:22:10 <h01ger> next? ;)
19:22:29 <pere> #link http://wiki.debian.org/DebianEdu/NonFreePackages
19:22:34 * h01ger opens said wiki page to look at it later/tomorrow
19:22:46 <pere> I am sure h01ger would like help with the evaluation.
19:22:52 <h01ger> oh, yes!
19:22:59 <pere> I suspect we should keep those we can distribute and drop the rest.
19:23:18 <pere> having stupid packages asking license questions during installation is out of the question (ref sun java).
19:23:56 * h01ger nods
19:24:18 <h01ger> #topic 2.4 status lenny: 1415: Cannot unlock KDE lock session
19:25:36 <OleA> totdays install didn't end normal, so I have to test again tomorrow
19:25:40 <h01ger> #info needs investigation, afaics
19:25:55 <pere> I suspect that was caused by me breaking diskless workstations several times the last few days.
19:26:19 <pere> some times cfengine would not run, the last bug gave failing boot sequence.
19:26:31 <pere> all these should be fixed, so check and see if the problem still exist.
19:26:36 <debalance> there was a bug about missing /etc/pam.d/kde in Debian which caused the same behaviour
19:26:39 <pere> I was able to lock and unlock the root account, at least.
19:26:52 <pere> today after fixing the boot sequence.
19:27:07 <OleA> root account was no problem on the same install
19:27:14 <pere> did not have time to create a non-privileged user, so I do not know if the problem only affected non-roots.
19:27:28 <h01ger> debalance, do you happen to remember the bug#?
19:28:16 <pere> if cfengine had not be running, pam would lack ldap setup.
19:28:36 <debalance> not exactly, but I think I can find it again
19:28:37 <pere> but that would make it impossible to log in too, so I do not know what could cause this.
19:28:50 <h01ger> debalance, please add to the skole-bug...
19:28:57 <debalance> the scenario of log-in working, but unlocking a session not, is when /etc/pam.d/kdm is there, but /etc/pam.d/kde is not
19:29:03 <debalance> h01ger: as soon as I've found it agin
19:29:06 <h01ger> :)
19:29:09 <h01ger> thanks already
19:29:57 <h01ger> #topic 2.5 status lenny: 1412: Errors when building the LTSP chroot are not reported to the user
19:30:41 <pere> seemed to me that ltsp-build-client failed to return an error, and thus the udeb failed to pass the error on. :(
19:30:52 <debalance> h01ger: #534714, it's filed against kdm4, but exactly same behaviour
19:31:21 <h01ger> pere, so still open? (the error reporting..)
19:31:59 <h01ger> debalance, hm
19:32:06 <pere> h01ger: I believe so.  I had a download failure which failed to be reported, causing a broken ltsp chroot.
19:32:08 <vagrantc> the problem is most likely due to ltsp-build-client running as a backgrounded process, and the hairbrained approach necessary to get a meaningful progress bar.
19:32:31 <h01ger> which *meaningful* progress bar? ;)
19:32:44 <FBI> debian-edu: 3 de-build-guest committed revision 61214 to debian-edu: automatic commit after build from lenny-test-dvd
19:33:01 <pere> h01ger: the one causing users to not believe the installer is hanging when installing the ltsp chroot.
19:33:05 <vagrantc> the ltsp-client-builder progress bar ... use to have two steps: 50% and 100% ... which would stall long enough that people thought it was broken.
19:33:32 <vagrantc> so i managed to get it a little more fine-grained ... but it's tricky to do well.
19:33:40 <h01ger> pere, what vagrant says. it stays so long at 50% that *i* believe its broken...
19:33:46 <OleA> alt+F4 gives hint about things that are happening ;)
19:33:50 <h01ger> aeh, missread..
19:34:16 <debalance> the chroot really takes a lot of time
19:34:53 <pere> we are talking about two different things, I believe.
19:34:56 <h01ger> yes
19:34:59 <vagrantc> so i *fixed* that bug, but it seems it doesn't properly report errors now
19:35:07 <OleA> today my install stalled at 54% ! Not 50% but 54%
19:35:10 <vagrantc> (as in, for the last 6 months or so)
19:35:12 <debalance> pere: when doing an ltspserver only install, will the chroot also be set up for diskless during installation?
19:35:14 <pere> the ltsp progress bar is fairly quick, while the diskless workstation step is horribly slow and stops at 54% for a long time.
19:35:29 <pere> debalance: yes, now all ltsp servers also support diskless workstations.
19:35:40 <debalance> okay, good to know.
19:35:42 <pere> debalance: by default they boot thin clients, thought.
19:35:55 <debalance> *documentation*?
19:35:56 <pere> we should document that and tell admins how to switch to diskless workstations.
19:36:10 <vagrantc> so, the *bug* that we're talking about is a consequence of an improved progress bar
19:36:25 <h01ger> debalance, http://wiki.debian.org/DebianEdu/Documentation/Lenny/HowTo/NetworkClients has some documentation
19:36:26 <pere> which is adding '3' to the kernel argument list to tell the client to use runlevel 3.
19:36:26 <vagrantc> the bug being that it doesn't detect ltsp-build-client failures
19:36:55 <pere> vagrantc: perhaps.  do not know the cause of this bug.
19:37:03 <h01ger> ok, /me concludes #1412 still needs some work
19:37:09 <h01ger> next?
19:37:44 <vagrantc> or possibly that the debconf priority needs to be temporarily adjusted?
19:37:59 <vagrantc> maybe it reports the error, but debconf just ignores it?
19:38:21 <OleA> doesn't detect ltsp-build-client failures "every time"
19:38:58 <vagrantc> well, most of the time, it works, yes?
19:39:08 <OleA> that is my impression
19:39:22 <vagrantc> so then this bug is hidden...
19:40:03 <vagrantc> i'll try and do some more testing with the current udeb and see if i can reproduce it.
19:40:51 <h01ger> #topic 3. alpha3
19:41:10 <h01ger> lets copy an image tonite, test it, and call it alpha3 tomorrow?
19:42:08 <h01ger> its just needs a time when noone has changes pending, else its a bit pointless / less useful
19:43:13 <vagrantc> it seems like there's been a lot of changes lately ... is that hopefully slowing?
19:43:59 <debalance> h01ger: you called me 'debacle' in #1408
19:44:00 <vagrantc> i think there's been at least 2 builds since i started testing this install, for example
19:44:08 * vagrantc chuckles
19:44:09 <h01ger> debalance, sorry :)
19:44:14 <debalance> no problem ;)
19:44:37 * h01ger will try to find a good moment to make an alpha3
19:44:50 <h01ger> having an image is the first step, the 2nd is writing the announcemnet :)
19:45:30 <h01ger> #topic rc1
19:45:50 <h01ger> if alpha3 looks promising, i'd like to see rc1 RSN
19:46:32 <h01ger> where looks promising = not absolutly unreleasable.
19:47:05 <OleA> :)
19:48:04 <h01ger> #topic 4. rc1
19:48:05 <h01ger> even
19:48:20 <h01ger> #topic 5. any other business
19:48:36 <h01ger> #topic 5.1 any other business: should we modify lwat to allow editing passwords per default
19:49:06 <OleA> update the wiki about extending the range of statics?
19:49:26 <h01ger> its there, i think
19:49:39 <debalance> 5.1: if so, the password should not be displayed in cleartext IMO
19:49:56 <h01ger> debalance, i disagree
19:50:03 <OleA> and a lot of other /!\ FIXME's
19:50:03 <h01ger> or asked for twice
19:50:35 <debalance> people who want to set the password directly usually want to keep it, and thus not having others standing by see it
19:50:44 <h01ger> OleA, considering that the documentation for etch was (almost: a lot) worse, i dont think the documentation is a blocker per se. rather the opposite :)
19:51:49 <vagrantc> another point of business: should we test stable-proposed-updates by default in the CD image?
19:51:58 <debalance> h01ger, the documentation is in the local repo anyway, right? so it can be updated after the release?
19:52:01 <h01ger> #info no agreement how we should deal with lwat + changing passwords. needs more discussion, probably on the list.
19:52:03 <h01ger> debalance, yes
19:52:22 <debalance> then it's definitely not blocking
19:52:24 <h01ger> #topic 5.2 AOB:  should we test stable-proposed-updates by default in the CD image?
19:52:28 <OleA> FIXME: add a pointer to DebianEdu/Documentation/Lenny/GettingStarted#DNSManagementwithlwat and describe it there.
19:52:53 <vagrantc> i think i posted to the list about stable-proposed-updates ...
19:53:15 <h01ger> vagrantc, i thought its not worth bothering
19:53:33 <h01ger> vagrantc, the only interesting thing to test, imo, would be d-i upgrades
19:53:36 <vagrantc> h01ger: my only fear is if there's an update in stable-proposed-updates, we release, and then it breaks something
19:53:37 <OleA> FIXME: a chapter "DNS Management with lwat" needs to be written, or better, be incorporated in the above "machine management with lwat" chapter
19:54:12 <h01ger> but whats in proposed-updates is usually not important for us
19:54:29 <h01ger> vagrantc, i'm watching s-p-u via d-release@...
19:55:30 <vagrantc> well, i brought it up, i'm not dead-set on it :)
19:55:34 <h01ger> :)
19:55:35 <vagrantc> if it's not worth the trouble, so be it
19:55:42 <h01ger> #topic 6. next meeting
19:56:04 * janr thinks it woukd be wise to extend the max no. of statics from 50 to 200.  this way slx will support larger schools out of the box running diskless
19:56:33 <h01ger> vagrantc, also more testing "maybe-stuff" might make us delay the release further. i'd rather release now and fix breackage tomorrow, then keep testing and not release while etch still has security support...
19:56:43 <vagrantc> h01ger: i hear you
19:57:15 <pere> sorry, got a bit occupied here.
19:57:15 <h01ger> janr, please file a bug. but i wouldnt want to change that now, before the release. but squeeze is really near :)
19:57:24 <h01ger> (squeeze freeze scheduled in 2 month...)
19:57:34 <h01ger> pere, have fun with backlog :)
19:57:46 * h01ger suggests next monday, same time, for the next meeting
19:58:30 * debalance likes janr's idea, but thinks it belongs into the squeeze release
19:58:54 * h01ger likes the idea too...
19:59:08 <h01ger> #agreed next meeting, next monday, 19 utc
19:59:26 <h01ger> for certain values of "agreed" ;-)
19:59:43 <h01ger> thank you all for attending and fixing bugs and testing and working on the docs!
19:59:46 <h01ger> cheers!
19:59:47 <vagrantc> the no massive dissent in a rapid fashion value
19:59:48 <h01ger> #endmeeting