Log from the IRC meeting 2006-01-19
19:00 < pere> So, lets start. Please, can everyone reading this message
respond with 'I am here' so the know who is present and
who is not.
19:00 < finnarne> I'm here
19:01 < pere> The agenda is available from
<URL:http://wiki.debian.org/DebianEduMeeting20060119>.
Feel free to update it during the meeting with information
and links.
19:01 < pere> We should limit this meeting to 60 minutes. Because of
this I will move forward on the agenda when I believe we
have covered the topic at hand.
19:01 < pere> finnarne: so, only the two of us present. this will be
a quick meeting. :)
19:01 < finnarne> well - not a lot of people here yet, jemtland will
join us shortly i think
19:02 < ThierryForb> present - french team
19:02 < finnarne> white - you are writing a summary, dont you ?
19:02 < white> right
19:02 < pere> white promised to write a summary, and finnarne promised
to make the IRC logs available after the meeting, yes.
19:02 < pere> First topic, what is the current status of debian-edu.
We should limit this to 15 minutes, I believe.
19:02 < white> I am here
19:03 < [-oskar-]> hi
19:03 < pere> who wants to give us an idea of the current status?
19:03 < white> i think there are 16 bugs left
19:03 < finnarne> Current status ? I've not tested a new image latelty
19:03 < Werner_> I am here
19:03 < white> i think we agreed in calling all pre06 bugs release stopping bugs
19:04 < finnarne> not all?
19:04 < white> hopefully not all
19:04 < pere> did anyone test the dailiy built CD lately?
19:04 < Werner_> I tested it on saturday
19:05 < white> so who made the last test
19:05 < white> Werner on saturday
19:05 < pere> what is working, what is not?
19:05 < Werner_> the installation of server+thin-client-server worked,
two thin client booted simultaniously
19:05 < finnarne> I know there is some changes committed into svn that
will break the installation of thin clients
19:05 < white> we should skip the barebone stuff and do it after the release
19:05 < finnarne> yes
19:05 < Werner_> white: agree
19:05 < white> we should also talk about the ntp configuration
19:06 < finnarne> I think we will have to do it the ugly way - with
cfengine hacks
19:06 < Werner_> white: bug # ?
19:06 < pere> finnarne: what was commited breaking the installation?
19:06 < white> Werner_: 1024
19:07 < finnarne> jemtland: found one of my script and commited them as
a replacement.
19:07 < pere> Werner_: did you have any problems with NFS timeouts when
you tested?
19:07 < white> pere: shall we look at all pre06 bugs now and summarize
which of them are important for the next release?
19:07 < pere> finnarne: are your script broken?
19:07 < finnarne> We could revert it, but it should be splitt (like
you started)
19:07 < Werner_> pere: no
19:07 < finnarne> pere: when taken out of context and soem additional
files are missing, yes
19:07 < pere> white: I believe we should focus on this meeting now,
and go looking for info.
19:08 < pere> s/for info/for info later/
19:08 < white> ok
19:08 < Werner_> shall we talk about the changes that will break
thin-client-setup first, NTP afterwords?
19:09 < white> yes please
19:09 < finnarne> the quick fix is to revert the checkin from jemtland.
19:09 < pere> so, if I get this right, the CD is working at least for
main-server+ltsp-server, and there are dubioius changes
in svn for the ltsp build at the moment?
19:10 < pere> any other status information you are aware of?
19:10 < white> i think we need a working solution now and we can hack
on it after a release
19:10 < finnarne> yes, and we have broken ntp setup (calls pool.ntp.org)
19:10 < finnarne> it worksm but could cause some problems
19:10 < pere> it has done so for quite some time. how seriuos is it?
19:11 < finnarne> serious enough if you're on a isdn line
19:11 < white> Werner_: does the debian-edu package run and depend on
kdepim-plugins now so that we can close this bug?
19:11 < finnarne> and anoying if your hw is not syncronized in the first place
19:11 < Werner_> white: one thing at the time..
19:11 < white> finnarne: but it is not breaking anything
19:11 < white> Werner_: sure sorry
19:12 < pere> we seem to be moving away from the general status issue
and on to individual packages. lets to that then.
which packages do we need to discuss? I've seen ntp,
kdepim and ltsp mentioned. any others?
19:12 < finnarne> udev
19:12 < jemtland> I'm here
19:12 < finnarne> and the shared config in /usr/share/debian-edu from
debian-edu-config
19:13 < pere> so, for problematic packages. lets do them one after
another. ntp first. what is the problem and how do we
fix it?
19:13 < finnarne> lets use cfengine(2) to point all ntp(date|-server) to ntp
19:13 < Werner_> the problem is afaik all hosts does ntp requests to
pool.ntp.org ..
19:14 < finnarne> and I'm not sure if all profiles gets ntp-server installed
19:14 < finnarne> and on many installations ntp is blocked in some firewall
19:14 < pere> Werner_: right. this happened when the new maintainer
ripped out the debconf preseeding support in ntp.
19:14 < Werner_> so shall we make sure all profiles get ntp-server
installed and hack this correct (ntp.intern) with
cfengine(2) then?
19:15 < finnarne> yes
19:15 < pere> only the networked profiles should install ntp, I believe.
19:15 < pere> not standalone.
19:15 < Werner_> right
19:15 < white> agree
19:15 < finnarne> pere: right
19:16 < Werner_> I'll paste this into bugzilla bug #1024 .. who has time
to do this?
19:16 < pere> there should be some old cfengine code in svn to do this.
We used to edit the config using cfengine, and then we
got the maintainer of ntp to support preseeding, and
then the maintainer changed and the new one ripped out
preseeding support.
19:17 < pere> any volnuteer to fix and test this?
19:18 < pere> right, no one. lets move on to udev. what is the problem
and how do we fix it?
19:18 < white> and which bug number?
19:18 < finnarne> #974
19:18 < finnarne> some cfengine again
19:19 < pere> what is the problem?
19:19 < finnarne> I think I've documented what needs to be done on the bug
19:19 < finnarne> sound is not working
19:19 < finnarne> cdrom is only read/writable by root and cdrom group
19:19 < finnarne> stuff lik that
19:20 < white> sounds like you attached a patch
19:20 < finnarne> anyone knows whay we need read and write permission
on /dev/rtc ? (same bug)
19:20 < finnarne> white: no - not a patch - just a memo on what to do.
19:20 < pere> why is this an udev problem?
19:21 < finnarne> because we fixed this with devfsd, and now we uses
udev instead
19:21 < finnarne> for 2.6-kernel
19:21 < white> can't we add the users into the audio group? or am i
missing something?
19:21 < finnarne> and we want 2.6 kernel for usb devices ++
19:21 < pere> how did we fix access problems using devfsd?
19:21 < finnarne> cfengine I think
19:21 < finnarne> white: I have 600 user and growing- do you want them
added to audio and cdrom group ?
19:22 < white> if they need sound access ... yes
19:22 < finnarne> by hand for each user ....
19:22 < white> no
19:22 < [-oskar-]> :)
19:22 < white> wether automatically or with a script
19:22 < pere> I believe redhat changes the permission/owner on the
devices when the user log into the console. another
solution is to add the user to a set of groups when he
log in. a third is to make the devices accessable to all.
not sure if the latter is a good one.
19:23 < pere> I saw some discussion on some debian list about race
conditions in the redhat solution. not sure about the
details.
19:23 < white> i don't agree with the last one although it is the easiest way
19:23 < finnarne> with devfs we allow access to everyone
19:23 < white> but i am not sure
19:24 < finnarne> It's the quick fix for a quick release
19:24 < white> what can happen to the system if the user has access to
/dev/dsp ?
19:24 < pere> someone _else_ can destroy the hearing of people with
earphones plugged in, I guess.
19:25 < white> on the other hand normally the admin can add the pupils to
the group but i think this is too much for our admins
because it should happen automatically
19:25 < pere> so, anyone volunteering to look into this device access
issue and come up with a fix?
19:25 < white> well i would add every user to the group audio by default
or provide a small script so that the admin has to right
the user names into a file and then they are added
19:26 < white> s/right/write/
19:26 < finnarne> white: what would that help over using mode 0666 ?
19:26 < pere> any ideas for solutions not mentioned so far?
19:26 < jemtland> If everyone in the system is added to the sound group,
why cant we just leave the device open for everyone ?
19:27 < Werner_> no .. I say we go for 666.
19:27 < finnarne> I vote for extending cf/cf.dsp, to support udev as well.
19:27 < white> jemtland finnarne Werner_: yes i think you are right
19:27 < pere> so, enough of udev/access problems. we need to move on.
no-one took on the task, but we have some ideas for
solutions.
19:27 < white> but then the admin has to change all if he doesn' want it
19:27 < pere> next on the agenda is 'ltsp'. what is wrong and how do
we fix it?
19:27 < finnarne> ldm is taking to long to log in
19:28 < pere> finnarne: how long is too long?
19:28 < finnarne> #1010
19:29 < pere> any other problems with ltsp? finnarne you mentioned a
broken script in svn. what is broken with it?
19:30 < finnarne> it's taking 60 seconds longer to log in using ldm than
using X -query ltspserver
19:30 < pere> finnarne: so it is 1 seconds and 61 seconds, or 3 minutes
and 4 minutes?
19:30 < finnarne> pere: jemtland checked in a version of debian-edu-ltsp
which sets up diskless workstations.
19:30 < pere> finnarne: what kind of hardware have this problem?
19:31 < finnarne> the script are missing some additionla config files,
and I think it will break (although I have not
tested it)
19:31 < pere> finnarne: lets wait with that issue and focus on the slow
login for a while.
19:31 < finnarne> pere: normal second hand workstations used as thin
clients
19:32 < laprag> Have anyone spoken to ogra about this?
19:32 < pere> finnarne: 486, pentium, pentium II or what?
19:32 < finnarne> I think it's a PIII
19:32 < laprag> He is apparently working on it for the new Ubuntu version
due in April
19:32 < finnarne> with enough memory
19:32 < pere> laprag: ogra=Oliver Grawert?
19:33 < laprag> Yes
19:33 < laprag> The problems are well known and acknowledged
19:33 < pere> finnarne: right. ldm encrypts the connectoin, and that
might be the problem.
19:33 < finnarne> the easy fix is to set up an lts.conf
19:33 < pere> but we do want encrypted connections to avoid sending
passwords in clear text over the net.
19:34 < finnarne> pere: not on the cost of people chosing something else.
19:34 < pere> ok, anyone willing to look into the slow ldm issue, and
make contact with ogra to look at solutions?
19:34 < white> but security comes first
19:35 * pere want to move on to the broken install script issue unless
someone have new info on the slow ldm issue.
19:35 < laprag> I can speak to ogra, but do we want to make our own
solution or wait for upstream?
19:35 < finnarne> quick fix for #1010 is on comment#3 on the bug I think
19:36 < pere> laprag: I do not know what we want, as I do not know the
options. great if you can find out how long we have to
wait for upstream.
19:36 < laprag> ok, I'll speak to him and find out
19:36 < finnarne> thx
19:36 < pere> ok, next issue on ltsp. finnarne claim a fix commited by
jemtland will break the ltsp build. did anyone test this?
jemtland?
19:37 < Werner_> I don't think that fix is on the CD yet.
19:37 < jemtland> I have not tested the new package. Have taken the file
from a working installation on one of the shools here
in Hamar.
19:37 < finnarne> no, it's not
19:37 < finnarne> (on the cd)
19:38 < laprag> which bug is this?
19:38 < pere> finnarne: you know this script a bit, right? how much
work is it to get it working?
19:38 < jemtland> no bug-
19:38 < finnarne> I need ok from someone to work on it ...
19:39 < white> finnarne: i can give you every ok i have ;)
19:39 < pere> finnarne: do you estimate 30 minutes, a day, a week?
19:39 < finnarne> when that's ok - I guess I need 30 minutes - 1 hour
19:39 < laprag> What fix is it? Fix what?
19:39 < jemtland> laprag: commit r5359
19:40 < pere> anyone got time to test that the script is broken or not,
and have a look at fixing it? finnarne is obviously tied
down with this issue at the moment.
19:40 < jemtland> I have talked to Halvor today. Asked if Skolelinux
Drift can give faj a couple of hours to get this on
the cd
19:40 < jemtland> He is going to talk to faj about this tomorrow.
19:40 < pere> are there other fixes in svn being held back because of
this commit?
19:41 < Werner_> finnarne: can you give a short explanation of what
it does?
19:41 < jemtland> thick clients, with the / system from tjener.
19:41 < finnarne> installs a full workstation in the ltsp chroot
19:41 < jemtland> root system
19:42 < finnarne> should have working usb support etc.
19:42 < white> that is nice
19:42 < pere> so this is the diskless workstation solution, right?
19:42 < finnarne> yes
19:43 < finnarne> based on ltsp5 (or whatever it's called - muekow)
19:43 < pere> are we in a hurry to get the svn version fixed, or can we
wait for a solution?
19:43 < pere> or to put it another way, are there other fixes in svn
being held back because of this commit?
19:43 < pere> are there other issues with ltsp we need to be aware of now?
19:43 < finnarne> well - to get other fixes into debian-edu-config we
either need to move the script aside, pull back an
old one or fix it
19:44 < jemtland> Not that I know of. There has not been any commits
against debian-edu-config after my commit.
19:44 < pere> finnarne: I know, but are there such fixes waiting in svn
at the moment?
19:44 < jemtland> Ss far as I have been able to see.
19:44 < pere> ok.
19:44 < finnarne> ntp, udev, ?
19:45 < pere> jemtland, you triggered this. do you have time to look
into the issue to either get a fixed version into svn
or revert the changes until someone have time to fix
the script?
19:45 < pere> I assume this means testing the current version and perhaps
get finnarne hands untied.
19:46 < pere> or find out what is wrong and fix it yourself.
19:46 < jemtland> I'll look into getting finnarne's hands untied.
19:46 < finnarne> :)
19:47 < pere> jemtland: and please add a comment on top of
sbin/ltsp-make-client about the commit if you end up
reverting it for the time being.
19:47 < pere> ok we move on. kdepim. what is wrong and how do we fix it?
19:47 < white> kdepim needs to be on the cd
19:47 < finnarne> and get installed
19:47 < white> in sid there kmail has a dependency against it
19:47 < white> but not in sarge
19:47 < pere> white: why do we need it?
19:48 < finnarne> imap support for kmail
19:48 < white> bug 995
19:48 < Werner_> I've added it to the desktop-kde, but not built the
debian-edu source package yet..
19:48 < pere> how big is it, what will fall out if it is added?
19:49 < finnarne> it's alreadyu added I think
19:49 < Werner_> it was too many packages missing when running 'rm
debian/control; make debian/control' so I didn't commit
19:49 < Werner_> and build the package(s).
19:49 < Werner_> pere: if I can get 15-30 minutes from you tomorrow,
I guess it is fixed?
19:49 < pere> I had planned to remove a few packages based on the counts
from <URL:http://developer.skolelinux.no/popcon/>, but
never gat that far.
19:50 < pere> celestia is rarely used, so removing it was high on
my agenda.
19:50 < Werner_> pere: so do you have some time to look at it with
me tomorrow?
19:50 < pere> sad, but as long as no-one is using it, it does not make
sense to spend 15MiB of CD space on it.
19:50 < pere> Werner_: yes, we can have a look at it tomorrow. I have
a concert to attend to, but we can do it earlier in the day.
19:51 < Werner_> sounds good .. I'm leaving at 1630 or so, so before
that would be great.
19:51 < pere> any other problematic packages? is wlus in a good shape?
19:51 < finnarne> it's mostly working - no changes actually
19:51 < finnarne> I've seen some strange things with samba
19:52 < finnarne> and some strange things with group creation
19:52 < white> is the jradmin stuff working?
19:52 < finnarne> not within wlus
19:52 < white> :(
19:52 < finnarne> there are a couple of scripts in debian-edu-config/tools
that might help
19:52 < white> that is an high requested feature
19:53 < white> to give others than root usercreation access
19:53 < pere> ok. we are running out of time. lets do a round around
the table, please give two sentences explaining what you
are currently working on related to debian-edu.
19:54 < finnarne> I've spent some time testing slbackup and rdiff-backup lately
19:54 < laprag> I'll talk to ogra about the login issue
19:54 < k4x> preparing Erkelenz
19:54 < white> still working on some other annoying bugs, i will leave
most of them unspecified because they shouldn't blog
the release
19:54 < pere> I've been working on the norwegian spell checking system
lately, trying to get the volunteer project going to
maintain it. I've also worked on the boot process, trying
to find ways to speed up the boot and make it better for
thin clients and diskless workstations (aka read-only
root systems).
19:55 < jemtland> I'm mostly trying to follow up on the bugs into
bugzilla. Mostly testing installations. Trying to
locate bugs, against later builds.
19:55 < Werner_> I've been working with finnarne on rdiff-backup/slbackup
(and hope we have a solution now), also trying to fix
the kdepim package, and general bugs from bugzilla.
19:55 < white> i recommend all to create a status for releasing and not
bringing more features things into the sarge release ...
19:55 < jemtland> Werner_: is that against #1025 ?
19:55 < finnarne> is all profiles tested in both norwegian and german
after we switched to use cfengine2 ?
19:55 < white> let's do it after the release
19:56 < ThierryForb> working (with our team) on the french addon cd :
we'll show you this work in Erkelenz :)
19:56 < jemtland> should we plan for a new meeting ?
19:56 < pere> I've spent some time sponsoring edu-packages from white
into debian. ThierryForb, are the french team working on
getting the special packages into debian?
19:57 < ThierryForb> yes
19:57 < white> i can make reviews if necessary
19:57 < ThierryForb> more or less 10
19:57 < pere> jemtland: yes, we should schedule a new meeting, and we
should discuss what we need to focus on for the short
term future.
19:57 < Werner_> jemtland: #1025 yes..
19:57 < finnarne> should we plan the next meating after erkelenz ?
19:57 < pere> ThierryForb: please get in touch with white and me if
you need sponsoring.
19:57 < ThierryForb> of course
19:58 < white> finnarne: i think after erkelenz
19:58 < finnarne> jemtland: and a lot of other rdiff/slbackup bugs
19:58 < pere> how often should we do IRC meetings? one per month,
every 2 weeks, or ever 2 months?
19:58 < ThierryForb> we'll speak about that in Erkelenz
19:58 * Werner_ votes for every two weeks
19:58 * white agrees to Werner_
19:58 * finnarne to - until we have a Skolelinux 2.0
19:58 < Werner_> I have updated
http://wiki.debian.org/DebianEduMeeting20060119 a bit
during the meeting
19:58 < [-oskar-]> pere, We are working with Raphael and I'm working
with sven so I think we will be able to do something ;)
19:58 * pere votes for once per month, and want us to coordinate the
rest on irc and mail.
19:59 < Werner_> finnarne: whatever we call it :)
19:59 * jemtland also votes for every 2 weeks until release
19:59 < jemtland> But, could we doit one hour later ....
19:59 * white recommends to make a first release date at the meeting
after erkelenz
19:59 < white> so that we have a goal now
19:59 < k4x> german team has irc meeting every week wednesday 22
o'clock CET
20:00 < pere> jemtland: 1900GMT, you mean?
20:00 < jemtland> pere: yes.
20:00 < finnarne> jemtland: not sure my wife agrees
20:00 < pere> ok, I am outvoted. every two weeks it is. I guess we
should stay away from wednesdays. any prefered day?
20:00 < white> jemtland: otherwise you are unable to attend???
20:00 < finnarne> Thursday is fine by me
20:00 < white> no wednesday and no weekend
20:01 < white> i prefer
20:01 < finnarne> and no tuesday (unless after 2000GMT)
20:01 < k4x> Thursday is ok by me
20:01 < pere> thursday and tuesday are not too good for me. I would
prefer monday.
20:01 < jemtland> I'm putting my son to bed after childrens tv. It ofthen
takes more then 20 minutes...
20:01 < jemtland> But I'll manage. I'll just be 5-10 minutes late...
20:01 < white> monday also sounds good
20:01 < pere> any protests for monday?
20:01 < k4x> Monday is ok by me too
20:01 < white> finnarne?
20:02 < white> Werner?
20:02 < white> jemtland?
20:02 < jemtland> monday ok for me to
20:02 < finnarne> monday is fine I guess (except for the next meating -
some may be on they way back from erkelenz?)
20:02 < k4x> ThierryForb: ?
20:02 < laprag> he wanted protests, not accepts
20:02 < ThierryForb> ok for me toot
20:02 < pere> next meeting monday 2006-01-30 or 2006-02-06?
20:02 < pere> -30 is a bit close to the german meeting.
20:02 < k4x> 2006-02-06
20:02 < Werner_> monday is fine
20:02 < laprag> 30 jan is too soon after next conf
20:03 < finnarne> lets have a slow start 2006.02.06
20:03 < ThierryForb> the 30 in france there is Solution linux
20:03 < pere> ok, -02-06 it is. thank you all for attending. white,
when will you have the summary ready?
20:03 < ThierryForb> linux expo in Paris
20:03 < finnarne> ok - next meeting monday 2006.02.06 at 1900GMT
20:03 < white> finnarne: i need the logs and i will start the summary now
20:04 < white> pere: writing it now i hope in 1-2 hours
20:04 < pere> I consider the meeting closed and step down from the
moderator chair. :)
Reply to: