Differences between revisions 1 and 2
Revision 1 as of 2008-11-05 20:41:57
Size: 26184
Editor: ?DanielHess
Comment:
Revision 2 as of 2009-03-16 03:31:26
Size: 26184
Editor: anonymous
Comment: converted to 1.6 markup
No differences found!

Time in UTC

nick

message

19:00:23

< vagrantc>

#startmeeting

19:00:23

< meetbot>

Meeting started at 19:00 UTC. The chair is vagrantc.

19:00:23

< meetbot>

Commands Available: #TOPIC, #IDEA, #ACTION, #AGREED, #LINK

19:00:33

< h01ger>

\o/

19:00:35

* danielsan

= Daniel Hess

19:00:42

* h01ger

= Holger Levsen

19:00:43

< vagrantc>

well, why not identify yourself with /me == Name

19:00:52

* vagrantc

= Vagrant Cascadian

19:00:59

* pere

= Petter Reinholdtsen

19:01:39

< h01ger>

#topic agenda at http://wiki.debian.org/DebianEdu/Meeting - please add your topics now and tell vagrantc :)

19:02:06

* danielsan

thinks we could skip #1 :)

19:02:24

< vagrantc>

anyone mind me just taking moderatorship?

19:02:39

< danielsan>

vagrantc: no, please go ahead and thanks

19:03:00

< h01ger>

vagrantc, no, please go aheads & thanks!

19:03:01

* klausade

= Klaus Ade Johnstad

19:03:21

* itais

=José L. Redrejo

19:03:26

< vagrantc>

#topic 2.1 Next (Lenny based) release : Name and version number

19:03:27

-!-

Irssi: Topic: -: Skolelinux | free software system for schools | http://wiki.debian.org/DebianEdu/Documentation/Etch/ | Next irc meeting: Wed. 5th November 19:00 UTC | Agenda: http://wiki.debian.org/DebianEdu/Meeting | http://wiki.debian.org/DebianEdu/Status/Etch is 3.0r1 "Terra" and stable

19:03:27

-!-

Irssi: Topic: +: 2.1 Next (Lenny based) release : Name and version number

19:03:27

-!-

meetbot changed the topic of #debian-edu to: 2.1 Next (Lenny based) release : Name and version number

19:03:51

* pere

believe we should stick to planets in the solar system if we are going to pick a name.

19:03:55

< h01ger>

http://doodle.com/participation.html?pollId=ycn7rw4rtiasbmy6 says that everybody who participated in the poll was favoring "5.0r0"

19:04:15

< h01ger>

but then vagrantc hat the great idea of "5.05r0~edu.r0"

19:04:17

< h01ger>

(or such)

19:04:26

< itais>

shit

19:04:29

< vagrantc>

or +eduN

19:04:35

< itais>

you must be kidding

19:04:45

< h01ger>

http://doodle.com/participation.html?pollId=cvt5bi5qz5x2iu7a says that everybody who participated in the poll was favoring not to use codenames, except white.

19:04:51

< danielsan>

5.0+edu1 no?

19:05:17

< vagrantc>

my proposal was to simply tack on +eduN to whatever version debian uses.

19:05:19

< itais>

well, that's easier to read, but I can not imagine how to say "5.05r0~edu.r0" to the users

19:05:42

< danielsan>

five dot zero plus edu one?

19:05:47

< h01ger>

5.0r0+edu1 was vagrantcs proposal

19:05:48

< itais>

what's wrong with Debian Edu 5.0r0?

19:06:03

* h01ger

had bad memory

19:06:13

< h01ger>

itais, we also have point upgrades

19:06:18

< vagrantc>

itais: it doesn't give room for debian-edu to make a sub-release outside of debian's cycle.

19:06:39

< itais>

why? you always can say 5.0r1, 5.0r2, etc.

19:06:42

< danielsan>

and does not show that we are almost debian

19:06:44

< h01ger>

5.0r2+edu4 is not that hard, IMHO

19:06:53

< h01ger>

itais, no, those are debian numbers

19:07:27

< itais>

h01ger, it's not hard for me, but think of telling to a school teachers, hey, I'm going to install here 5.05r0~edu.r0

19:07:48

< h01ger>

itais, not 5.05r0~edu.r0 - 5.0r0+edu0

19:08:02

< pere>

~ is not a well known characters outside hardcore unix communities.

19:08:28

< pere>

ran into a windows user the other day that was convinced the /~pere/ part in my URL was a typo.

19:08:30

* h01ger

nods, but vagrantc didnt suggest ~ in the thread where people agreed

19:08:39

< h01ger>

5.0r0+edu0

19:08:44

< danielsan>

one question is, whould it be 5.0r0+edu1 and then 5.0r1+edu1 or 5.0r1+edu2?

19:08:56

< h01ger>

debian will release 5.0r1

19:08:57

< itais>

:) good question

19:08:58

< h01ger>

so

19:09:08

< vagrantc>

i think it's really more of a + anyways, as it's actually additional stuff above and beyond plain debian at this point

19:10:47

< pere>

will the rest of the debian community react negatively if we use a string too close to the official one?

19:11:07

* h01ger

doubts it as we are debian for more than 99.9%

19:11:15

< danielsan>

if we prefix something, why should one?

19:11:18

< h01ger>

and its just a version number

19:11:19

< vagrantc>

#link http://doodle.com/participation.html?pollId=cvt5bi5qz5x2iu7a

19:11:29

< vagrantc>

#link http://doodle.com/participation.html?pollId=ycn7rw4rtiasbmy6

19:11:55

< vagrantc>

h01ger: what was poll 1 of 3 ?

19:12:02

< h01ger>

itais, the name to tell other people is "Debian Edu lenny" anyway, not "Debian Edu 5.0r0+edu1"

19:12:06

< danielsan>

or suffix :)

19:12:31

< h01ger>

vagrantc, pollId=ycn... was poll1

19:12:47

< itais>

so, at the end, I will tell Debia Edu lenny, and Debian Edu lenny r1, and so on, so who cares about the name we use internally?

19:12:51

< vagrantc>

h01ger: ah yes, 3 of 3 ?

19:13:09

< h01ger>

http://doodle.com/participation.html?pollId=bcb76h6f2cep9hud is poll three and there everybody except two people said we should release when its ready.

19:13:21

< h01ger>

itais, yes, i'd say so

19:14:35

< h01ger>

so lets go for 5.0r0+edu0 ?

19:14:51

< h01ger>

(and communicate "Debian Edu lenny")

19:15:59

* danielsan

likes the suffix thing

19:16:24

* itais

has no problem with that

19:16:56

< vagrantc>

any outstanding concerns?

19:17:48

< h01ger>

we have 2 more polls to discuss :)

19:17:53

< h01ger>

and 43min left

19:18:00

< h01ger>

(and other topics)

19:18:01

< vagrantc>

alright, dare i say:

19:18:18

< vagrantc>

#agreed use 5.0r0+edu0 for initial release

19:18:31

< h01ger>

:)

19:18:52

< pere>

get the spec for the next version too.

19:18:58

< pere>

is it edu1 or r1?

19:19:12

< h01ger>

r1 is what debian uses

19:19:14

< h01ger>

so its edu1

19:19:27

< h01ger>

vagrantc, please add this to #agreed

19:19:40

< danielsan>

ro ...r1+edu1?

19:19:45

< danielsan>

s/ro/so/

19:19:55

< vagrantc>

pere: you just want clarity, or actually have concerns about it?

19:20:00

< h01ger>

danielsan, we increase rX whenever debian does a pointrelease..

19:20:03

< pere>

vagrantc: I want clarity.

19:20:44

< h01ger>

pere, solved?

19:20:49

< danielsan>

h01ger: right :)

19:21:02

< pere>

h01ger: assuming it make it into the summary, sure.

19:21:15

< vagrantc>

i think the eduN should only be incremented when debian-edu makes a new release.

19:21:16

* danielsan

will take care

19:21:40

< h01ger>

so, codename? agreed to drop it?

19:21:41

< danielsan>

vagrantc: if we don't release noone would know that we incerement it :)

19:22:00

< vagrantc>

danielsan: indeed.

19:22:29

< danielsan>

we already have the codename of the debian version we base on

19:23:01

* vagrantc

sees multiple codenames as confusing.

19:23:45

< pere>

vagrantc: confusing for whom? edu users know which code name they are using (like terra), but not which name it was based on.

19:24:10

< danielsan>

pere: at least in the sources.list there is still lenny

19:24:10

* pere

do not care much about code name, but believe we should stick to the planet theme if we keep it.

19:24:25

< h01ger>

pere, you voted against the codename. everybody except white did. and no-one came up with good reasons for it..

19:25:22

< vagrantc>

h01ger: are we looking at the same poll? http://doodle.com/participation.html?pollId=cvt5bi5qz5x2iu7a

19:25:40

< h01ger>

yup

19:26:02

< pere>

h01ger: as I remember it, I voted against luna.

19:26:28

< vagrantc>

there's luna, no codename, mars and pluto as choices

19:26:40

< h01ger>

<nitpick>pere, but you voted for pluto, which is not a planet :-D</nitpick> - the votes are in the url vagrantc just repeated

19:27:23

< pere>

are we going to discuss my votes, or something else?

19:27:43

< pere>

I suggested having all the unused planet names as options, but did not spend time on pressing the issue.

19:28:44

< danielsan>

well we could talk about the code name for a while, but as (almost) anybody seems ok with no codename why don't we go for it?

19:29:08

< danielsan>

wasn't it sarge who also has not had a codename?

19:29:29

< h01ger>

12 voted. 11 in favor of "no codename". why are we discussing this at all?

19:29:47

< h01ger>

also noone came up with a good reason _for_ a codename, except nostaligy

19:30:03

< h01ger>

while two people said the codename is causing confusing, and thus hurts

19:30:25

< h01ger>

we had 3 releases so far. 1 without codename (sarge), 2 with (woody+etch)

19:30:56

< pere>

sarge could be seen as a anomality.

19:31:20

< danielsan>

how about another one? :)

19:32:26

< vagrantc>

it seem slike there is an overwhelming vote in favor of no codename, but i hear some outstanding ... cornerns about simply moving forward ?

19:32:27

< itais>

I'm with h01ger: we should respect the result of the votation, 11 against 1 is quite clear, and we should pass to the next item

19:33:54

< vagrantc>

pere: you've been the voice in support of codenames, do you have outstanding concerns with "no codename" ?

19:34:25

< pere>

vagrantc: not really. Suspect it might help with marketing, but do not have any strong opinion on the topic.

19:34:39

< vagrantc>

lets move on then.

19:34:52

< h01ger>

.oO( "lenny" helps with marketing. "lenny mars" confuses )

19:35:00

< h01ger>

the third poll... "release when"...

19:35:05

< vagrantc>

#agreed we will simply use lenny as a codename

19:35:13

< h01ger>

"...its ready!"

19:35:41

* h01ger

would like to define what "ready" means. unfortunatly i didnt manage to put all the issue on http://wiki.debian.org/DebianEdu/Status/Lenny into bugs

19:35:44

< h01ger>

yet

19:35:49

< h01ger>

but what are our blockers?

19:35:58

< vagrantc>

seems the result of the vote is to release when ready ...

19:36:03

< danielsan>

tasksel

19:36:08

< vagrantc>

#link http://doodle.com/participation.html?pollId=bcb76h6f2cep9hud

19:36:09

< pere>

there are several blockers.

19:36:11

< h01ger>

do we need a fixed lwat? (ie for editing dhcp entries with it)

19:36:29

< pere>

h01ger: yes, probably.

19:36:42

* h01ger

suggests to split the issues in http://wiki.debian.org/DebianEdu/Status/Lenny into "normal issues" and "blocking issues" now

19:36:56

< pere>

should we change topic?

19:37:25

< vagrantc>

#topic 2.2 Next (Lenny based) release : Current state of our Lenny

19:37:25

-!-

Irssi: Topic: -: 2.1 Next (Lenny based) release : Name and version number

19:37:25

-!-

Irssi: Topic: +: 2.2 Next (Lenny based) release : Current state of our Lenny

19:37:25

-!-

meetbot changed the topic of #debian-edu to: 2.2 Next (Lenny based) release : Current state of our Lenny

19:37:45

< vagrantc>

how is lenny looking for debian-edu ?

19:38:00

< pere>

the package installation is broken because I tried to add desktop and language specific packages.

19:38:19

< pere>

when it was working, ltsp clients did not boot via pxe, at least when I used qemu.

19:38:29

< h01ger>

mostly ready (bug wise), but we lack features.

19:38:44

* h01ger

would add like to stop adding features so we can concentrate on the bugs

19:39:09

< danielsan>

some important ones

19:39:12

< pere>

the partition sizes generated by the installer are too small, which is worked around by extending partitions during installation, but this lead to fsck on first boot because of a misfeature in online ext3 resizing.

19:39:38

< pere>

a fix is to make sure the minimum partition sizes are large enough to avoid resizing during normal installs.

19:39:40

* h01ger

adds what pere just said to the status page

19:40:04

< pere>

the edu-expert profile question is messed up in text mode.

19:40:16

< h01ger>

pere, for all languages?

19:40:22

< pere>

h01ger: not sure.

19:40:23

< itais>

and the ldapify is not done yet

19:40:37

* h01ger

suggests to either drop the sugar profile from expert mode or upload jonas' sugar backports to our repo

19:40:42

< pere>

yes, dhcp uses ldap for its info, but we do not have any tools to maintain it.

19:41:10

< pere>

h01ger: I agree to the dropping sugar. we can make a sugar CD instead if we got time.

19:41:15

< itais>

and not for pdns, nor fill the ldap tree with the right configs

19:41:30

< pere>

itais: yeah, DNS config is not (yet) in LDAP.

19:41:50

< pere>

italc or the controlaula stuff should probably be configured to work out of the box.

19:41:54

* h01ger

asks everyone to update the status page with known issues (or the state of them). i would also suggest to really define the blockers for our lenny release based on bug numbers in the next irc meeting (which we should have sooner again) - i would also promise to file bugs for each issue on the status page

19:42:19

< pere>

we should get the OCS Inventory stuff configured out of the box.

19:42:22

* h01ger

adds "dropping sugar" to the page

19:42:52

< itais>

for the client it's pretty easy, for the server I don't see it possible unless a good amount of work is done via cfengine (for OCS inventory)

19:43:16

< pere>

we should fix PXE install setup to work also for DVD installs (the d-i PXE images are missing in the archive, currently downloaded on the fly).

19:43:28

< pere>

itais: yes, the server part is the problem.

19:43:42

< vagrantc>

seems like the topic has naturally moved on to the next point:

19:43:44

< itais>

another possibility is making some inventory.skolelinux.org server and not installing the inventory in the school, only configurating the client to send the data there, as popcon

19:43:59

< vagrantc>

#topic 2.3 Next (Lenny based) release : bugs+features that need fixing before we can release

19:43:59

-!-

Irssi: Topic: -: 2.2 Next (Lenny based) release : Current state of our Lenny

19:43:59

-!-

Irssi: Topic: +: 2.3 Next (Lenny based) release : bugs+features that need fixing before we can release

19:43:59

-!-

meetbot changed the topic of #debian-edu to: 2.3 Next (Lenny based) release : bugs+features that need fixing before we can release

19:44:20

< h01ger>

#info please update http://wiki.debian.org/DebianEdu/Status/Lenny

19:44:25

< pere>

itais: I would be very mad if my hosts submitted that amount of info out of the site.

19:44:49

< pere>

it would be nice if we could get the roaming laptop stuff working.

19:44:49

< h01ger>

itais, why not setting up tjener as an ocs inventory server?

19:45:04

< pere>

h01ger: because it is hard to do automatically at install time.

19:45:12

< h01ger>

ah, that bug

19:45:13

< itais>

h01ger: because the deb package can not be preseeded and it's not easy to configure

19:45:23

* h01ger

looks that bug up

19:45:26

< pere>

when installing the main-server on a laptop, network-manager is installed, messing up the network configuration.

19:45:39

< itais>

pere: about the roaming stuff,it's not hard, maybe we might do the same with all the pcs, not only laptops

19:45:44

< pere>

this is actually because the debian laptop task is installed by default.

19:46:06

< pere>

itais: yes, I believe it should be the default for all installs, if it work.

19:46:27

< itais>

pere: why does network-manager mess the configuration?

19:46:49

< itais>

it should not do it, in fact it does not do it in the laptops I've set up

19:47:03

< pere>

itais: because it block network setup from happening at boot. it only happen after the user log in, and no user can log in because the network isn't configured.

19:47:08

< vagrantc>

as a time check, we have 14 minutes left, and 2-3 remaining agenda items...

19:47:23

< itais>

pere: it should not be that way

19:47:32

< pere>

I guess my point is that there is a lot of stuff left to fix in the lenny version.

19:47:32

< vagrantc>

(usbstick images, other business, next meeting)

19:47:41

< itais>

network-manager gets an ip as soon as a dhcp server is available when using a cable connection

19:47:42

< pere>

who did a lenny install recently? on real hardware?

19:47:48

< itais>

what you say happens when using wireless

19:47:55

< itais>

pere: me, today

19:48:08

< pere>

itais: with kde?

19:48:18

< itais>

not :'(

19:48:24

< h01ger>

pere, if ocs-inventory needs mysql running (which is not the case in d-i..), one can use a firstboot script to configure it. (a script that runs upn the first boot)

19:48:32

< pere>

btw, I suspect kdm isn't able to handle expired passwords properly. this is annoying.

19:48:45

* h01ger

can confirm the kdm issue

19:49:04

< pere>

h01ger: sure, or start mysql from d-i to configure it, or something. someone just need to do the work.

19:49:06

* vagrantc

suspects ldm doesn't handle all cases of password expiry either

19:49:13

< h01ger>

definitly for 3.5.5, not fully sure about 3.5.9

19:49:23

< klausade>

pere: did ltspserver install yesterday, also no kde.

19:49:26

< h01ger>

pere, right :)

19:49:43

< pere>

there is also the iceweasel SSL cert issue. it refuses to connect to our web site.

19:49:47

< danielsan>

pere: don't forget start-stop-deamon does not work during install

19:49:59

< pere>

danielsan: I did not say it was easy. :)

19:50:46

< h01ger>

pere, i dont think the iceweasel ssl issue is (or should be) release critical. we can say: "you have to use konqueror for the internal server"

19:51:04

< pere>

also, I suspect we should extend the IP range of the backbone network from 10.0.2.0/23 to something/22 or perhaps even /16.

19:51:11

< vagrantc>

so, we have a list of issues reported in irc, can someone go through the logs and make sure they're appropriately added to the lenny status page, and/or file bug reports?

19:51:42

< h01ger>

best those people who said them. because they probably know these issues best

19:51:59

< pere>

I also believe we should rewrite the squid configuration to avoid the hardcoded IP address and instead use all subnets listed in LDAP.

19:52:10

< h01ger>

and please file bug instead. using the wiki page doesnt work well, or at least i learned from the etch release that using bugzilla works _much_ better

19:52:12

< danielsan>

maybe we could do a package that creates a ca on tjener, that can be used with tinyca. but still there is the problem that how to get the ca onto the workstations

19:52:14

* pere

doubt he will find time to do it, and thus prefer others to look into it.

19:52:23

< itais>

there's still a lot of remaining work, so volunteers?

19:52:28

< pere>

I got one our off today, but reporting all these bugs will take a while.

19:53:11

< klausade>

h01ger: it's rather easy to include sites to exlude from ssl checks, it's just a small hack to the global firefox/profile. it defeats all security ...

19:53:12

< pere>

please do test installs, fix stuff, and more test installs. also, test on real hardware, to discover hardware issues. I'm told that X.org vesa driver do not work in lenny. it is probably fatal for us.

19:53:18

* h01ger

thinks we should stop adding features (like rewrite squid conf) and rather get a bug free release out the door. squeeze will come/freeze in 18 months (approx), so we should not shorten our squeeze development cycle too much

19:53:51

< pere>

h01ger: well, some of us are not happy with all the hardcoded IP addresses.

19:54:03

< h01ger>

klausade, see.. if this was an issue tracked as a bug it would be very easy to add the info there :) (of course its also easy to add to the status page, but...)

19:54:34

< vagrantc>

with 7 minutes left, shal we postpone the discussion on usbticks and other business, and move towards hashing out the next meeting time?

19:54:39

< h01ger>

pere, its in 3 files or so. not that many. i believe, more people would profit from an earlier release than from that feature :)

19:54:39

< pere>

hardcoded IP addresses only belong in the DHCP and DNS config, I believe.

19:54:44

< h01ger>

vagrantc, yes. please move on

19:54:56

* h01ger

thinks its important to define a next meeting

19:54:58

< vagrantc>

#topic Next meeting?

19:54:59

-!-

Irssi: Topic: -: 2.3 Next (Lenny based) release : bugs+features that need fixing before we can release

19:54:59

-!-

Irssi: Topic: +: Next meeting?

19:54:59

-!-

meetbot changed the topic of #debian-edu to: Next meeting?

19:55:13

< pere>

I might be able to make every three weeks.

19:55:15

< h01ger>

#info no agreement on release blockers found.

19:55:33

< danielsan>

wednesdays seem to work best

19:55:46

< danielsan>

#link http://wiki.debian.org/DebianEdu/Meeting/VoteForDate

19:55:51

< pere>

#info several potential release blockers mentioned.

19:55:52

-!-

MarioB [~user@125.33.1.81] has quit [Quit: Leaving.]

19:55:52

* h01ger

would rather prefer every two weeks but prefers peres participation more :)

19:56:09

* pere

have not had time to do test installs since the devcamp. :(

19:56:17

< itais>

pere: I've removed all the hardcoded ip from my squid conf, but I remember you told me you needed them due to some license issues

19:56:24

< vagrantc>

wednesdays work good for me

19:56:25

< itais>

it's that right?

19:56:36

* h01ger

sighs

19:56:45

< vagrantc>

andreas made a request for a different time

19:56:51

< pere>

thursday are busy for me.

19:56:59

< danielsan>

hour later, yes

19:57:12

< itais>

wednesdeays are ok for me, and an hour later is not a problem

19:57:20

< danielsan>

that would make the meeting end at 22:00 localtime for most people

19:57:41

< h01ger>

or 8am for white, iirc. which maybe could even work :)

19:57:42

< danielsan>

at least until march :)

19:58:24

< pere>

22:00 is a bit late for me.

19:58:32

* pere

is getting up rather early these days.

19:58:50

< pere>

itais: did you find time to look at the tasksel issue, btw?

19:58:50

< itais>

ah, I thought you did not even go to bed ;)

19:58:51

< vagrantc>

3 weeks from now would be november 26th

19:59:04

< h01ger>

i wont be able to make that date

19:59:08

< itais>

pere: not, I even forgot it buried in a pile of work

19:59:15

< h01ger>

i will be traveling to extremadura that date :)

19:59:22

< h01ger>

(+time)

19:59:40

< pere>

itais: if you get it working, you could install the gnome version by adding desktop=gnome on the Cd boot propmt.

19:59:46

< vagrantc>

2 weeks from now, november 19th

19:59:54

< vagrantc>

and 4 weeks from now, december 3rd

20:00:07

< itais>

pere: then I should find some time slot to work on it :)

20:00:33

< itais>

as the lenny release is closer I think we should do it in 2 weeks

20:00:41

* vagrantc

concurs with itais

20:00:45

* pere

will try to show up.

20:00:45

< itais>

to avoid people begin to relax

20:00:54

* h01ger

too

20:01:00

< pere>

should we plan another devcamp? we got the funds for it. :)

20:01:10

< vagrantc>

november 19th, 19:00 UTC ?

20:01:23

* itais

agrees

20:01:29

< danielsan>

ok for me

20:02:17

* pere

will do its best.

20:02:19

< danielsan>

and as andreas did not show up yet, one hour seems not to be enough :)

20:03:08

< vagrantc>

ok, i'll call it:

20:03:25

< vagrantc>

#agreed next meeting, november 19th, 19:00 UTC

20:03:32

< vagrantc>

thanks everyone!

20:03:33

< danielsan>

\o/

20:03:36

< vagrantc>

#endmeeting

20:03:36

< meetbot>

Meeting finished at 20:03.