Differences between revisions 1 and 2
Revision 1 as of 2006-02-20 20:14:44
Size: 22249
Comment:
Revision 2 as of 2009-03-16 03:30:33
Size: 22249
Editor: anonymous
Comment: converted to 1.6 markup
No differences found!
   1 20:04 < white> ok please say "I'm here"
   2 20:04 -!- JeromeForb [n=jerome@ANancy-153-1-14-49.w83-196.abo.wanadoo.fr] has joined #debian-edu
   3 20:04 < white> I'm here
   4 20:04 < finnarne> I'm here
   5 20:04 < zobel> finnarne: yes, but white told me 1900 CET, not 1900 GMT
   6 20:04 < white> zobel: anyway we start now :)
   7 20:04 < white> Werner_: ping
   8 20:04 < white> jemtland: ping
   9 20:04 < ThierryForb_> French team is here too :)
  10 20:04  * zobel here as well
  11 20:05 < zobel> h01ger: ping!
  12 20:05 -!- finnarne changed the topic of #debian-edu to: RC1 of Sarge-based Debian-edu is out || irc meeting in progress, please look at http://wiki.debian.org/DebianEdu/Meeting/20060220
  13 20:05 < white> anyway starting now maybe someone will come later
  14 20:05 < white> finnarne: can you please summarize the current state of the development?
  15 20:06 < finnarne> I'm not sure if the daily build is installable, I see no reason why it should not be installable
  16 20:06 < h01ger> zobel, pong but #d-c meeting over there...
  17 20:06 < white> anyone familiar with the current daily build?
  18 20:06 < white> has someone tested it?
  19 20:06 < finnarne> but I haven't tested since the image from saturday
  20 20:07 < pere> is the new ltsp package working?
  21 20:07 < finnarne> not sound
  22 20:07  * pere fixed the remaining issues, but did not see test results.
  23 20:07 < finnarne> and not gcompris
  24 20:07 < pere> finnarne: no?
  25 20:07 < pere> well, I hoped sounds was working, and am not surprised gcompris isn't working.
  26 20:07 < finnarne> when I log in on the client, I get sound on the server, if I log in as root
  27 20:08 < finnarne> if I log in as a normal user I get access denied
  28 20:08 < pere> what give you access denied?
  29 20:08 < finnarne> sound device
  30 20:09 < white> ok we have two bugs about ltsp
  31 20:09 < pere> finnarne: any bug number I can use to see the exact error messages?
  32 20:09  * pere lack test bench to test sound.  qemu is not useful there.
  33 20:10 < white> finnarne: can you fill in a bug and point us later to the bugreport?
  34 20:10 < finnarne> well, remove yourself from the audiogroup, and set permission 0660, and you wil se the error
  35 20:10 < finnarne> nope, I dont have a client here now.
  36 20:10 < white> finnarne pere: do we have a possible fix for this bug?
  37 20:10 < finnarne> and I have to test ldap/samba, and upgrade some servers, so I'm kind of busy
  38 20:11 < pere> finnarne: hm, that does not make sense, as the programs running on the server should connect to esd on the client, and thus not need memership in the audio group.
  39 20:11 < finnarne> pere: but it's not iusing esd on the client
  40 20:11 < pere> white: I have no idea how file permissions and group membership can affect esd sound, no.
  41 20:11 < finnarne> it tries to use the soundcard on the server
  42 20:11 < pere> finnarne: finnarne oh.  is ESOUND set in the environment?
  43 20:11 < finnarne> as said - if I log in as root on the thin client, I get sound on the server
  44 20:12 < finnarne> I dont know.
  45 20:12 < white> pere finnarne: don't wanna stop you from bug squashing but let's summarize all parts first please keep it in mind ok?
  46 20:12 < pere> finnarne: ok.  if that is the problem, I have an idea.
  47 20:12 < finnarne> good.
  48 20:12 < pere> white: ok.  next ltsp issue is the gcompris issue?
  49 20:12 < white> finnarne pere : there was a second bug about gcompris
  50 20:12 < white> pere: did you see Werner_'s suggestion?
  51 20:12 < pere> is this problem when selecting gcompris from the menu?
  52 20:12 < white> yes
  53 20:12 < finnarne> # 1056
  54 20:13 < pere> would it help to make the menu use full path to gcompris?
  55 20:13 < finnarne> yes, I think so
  56 20:13 < white> pere: what about sourcing /etc/profile ?
  57 20:13 < white> this was Werner_'s workaround
  58 20:13 < finnarne> I think he mentioned using the debian-menu entry worked
  59 20:13 < white> finnarne: yes he did
  60 20:14 < pere> white: not sure about the effect of sourcing /etc/profile.  will need to investigate more before I recommend that.
  61 20:14 < white> pere finnarne: so possible fix for now?
  62 20:14 < finnarne> white: pere will investigate.
  63 20:14 < pere> I would suggest a new gcompris package with propoer command=/usr/games/gcompris.  I believe that is the correct fix.
  64 20:15 < white> pere finnarne: ok keeping it in mind and adding the suggestions to the bug report please
  65 20:15 < white> finnarne: can you start to introduce the ldap stuff please?
  66 20:15 < white> ldap account maintenance is the point on the Agenda
  67 20:15 < finnarne> I wrote an email about it some days ago, did anyone read it ?
  68 20:16 < white> yes
  69 20:16 < white> finnarne: have you verified the new configuration?
  70 20:16 -!- ThierryForb [n=thierry@ANancy-153-1-57-244.w83-196.abo.wanadoo.fr] has quit [Read error: 110 (Connection timed out)]
  71 20:16 -!- ThierryForb_ is now known as ThierryForb
  72 20:17 < finnarne> I've verified that noone but ldap admin can fetch the passwords from ldap
  73 20:17 -!- vagrantc [n=vagrant@207-224-29-13.ptld.qwest.net] has joined #debian-edu
  74 20:17 < white> finnarne: and does the jrpasswd tool work?
  75 20:17 < finnarne> I've verified that using passwd and smbpasswd does work
  76 20:17 < finnarne> the problem about jrpassword is that there is a missing dependency, and it's not even included in our debian-edu-config
  77 20:18 < finnarne> it's only availible from the source
  78 20:18 < white> finnarne: ok but the new slapd configuration is working
  79 20:18 < white> ?
  80 20:18 < finnarne> for non-samba-users yes. not sure about samba-users
  81 20:18 < white> that's the next point
  82 20:18 < white> finnarne: so far good work
  83 20:19 < white> ok slapd configuration was changed and verified thx to finnarne
  84 20:19 < white> next point is samba
  85 20:19 < finnarne> I want to use my own tool to verify that thepassword settings are working like I expect them to.
  86 20:19 < finnarne> Also I want to check changing password using wlus.
  87 20:19 < finnarne> ok , now I'm done with #1061
  88 20:20 < white> finnarne: good please close it later
  89 20:20 < white> now about http://bugs.skolelinux.no/show_bug.cgi?id=1062
  90 20:20  * finnarne is done discussing #1061, not verifying that it's fixed
  91 20:20 < white> finnarne: ok then i got it wrong
  92 20:21 < white> finnarne: means we still have to check it with wlus or what is missing?
  93 20:21 < finnarne> yes something like that
  94 20:21 < white> finnarne: ok we should add it to the report
  95 20:21 < white> ok next point
  96 20:21 < white> samba:
  97 20:22 < white> bug #1062
  98 20:22 < white> have you all read juergen's comments?
  99 20:22 < finnarne> I dont like the amount of messages about samba not working
 100 20:22 -!- Idefix [n=idefix@p54B49354.dip0.t-ipconnect.de] has joined #debian-edu
 101 20:22 < finnarne> not sure how much integration we should have, since it so late.
 102 20:23 < white> have anyone checked juergen's comments?
 103 20:23 < finnarne> Iæve not
 104 20:24 < finnarne> I've not
 105 20:24 < white> ok at first i want to make sure which severity it has
 106 20:24 < white> if juergen's comments are right is it a release blocker?
 107 20:24 -!- Idefix_ [n=idefix@p54B4991D.dip0.t-ipconnect.de] has quit [Read error: 110 (Connection timed out)]
 108 20:25 < finnarne> I think so. I think we should better than our competitors, and have a working solution for their custommers
 109 20:25 < white> so the state is: we still need to test it again and maybe fix it
 110 20:26 < finnarne> yes
 111 20:26 < white> anyone willing to do this job?
 112 20:26 < pere> who can test?  who got time to do the testing?
 113 20:26 < finnarne> I had hoped to do so yesterday, but will try to do it later tonight
 114 20:26 < white> finnarne: thx
 115 20:26 < white> ok assigning finnarne to the bug!
 116 20:26 < finnarne> I have one valid XP installation , and a fresh main-installation
 117 20:27 < white> any other comments about the samba topic?
 118 20:28 < white> ok no one
 119 20:28 < finnarne> well, we still have to se how much we want to include
 120 20:28  * vagrantc has still not tested rc1
 121 20:28 < white> finnarne: when you have checked the configuration can you write a mail to debian-edu@l.d.o.?
 122 20:29 < white> finnarne: or do you think we should discuss it now?
 123 20:29 < finnarne> I will close or update #1061 and #1062
 124 20:29 < white> finnarne: ok thx
 125 20:29 < white> any other comments about major bugs or bugs blocking the release?
 126 20:29 < white> are there any?
 127 20:30 < finnarne> I've noticed one thin about the ltsp-clients
 128 20:30 < white> finnarne: ?
 129 20:30 < finnarne> When booting a client with the 2.4-kernel (default), it takes about 4 1/2 minutes to boot a diskless workstation
 130 20:31 < finnarne> When booting the same client with a 2.6-kernel, it takes ~1 minute
 131 20:31 < pere> wow, you got ltsp working with 2.6?  I thought that was broken.
 132 20:32 < finnarne> it works fine with the sarge-kernel
 133 20:32 < white> pere vagrantc: suggestions why it takes soooo long with the 2.4-kernel?
 134 20:32 < pere> ah, perhaps it was the etch and sid kernel that was the problem.
 135 20:32 < finnarne> sarge 2.6.8 kernel that is
 136 20:32 < pere> I suspect it is using both hotplug, discover and devfs, and that this might be slow.
 137 20:33 < pere> but the only way to find out is to use bootchart to profile the boot.
 138 20:33 < finnarne> I think it's better to use udev, yes.
 139 20:33 < white> can't we use the sarge 2.6 kernel as default?
 140 20:33 < pere> www.bootchart.org, included in debian/etch and sid.
 141 20:33 < finnarne> but it might only be the diskless workstation setup that uses this much time
 142 20:34 < pere> I have nothing against using 2.6 kernel by default, but want to make sure the way we do it work with etch too.
 143 20:34 < finnarne> there is one change needed in debian-edu-ltsp
 144 20:34 < pere> finnarne: did you ever try bootchart?  Very simple.  Just 'init=/sbin/bootchartd' on the kernel command line to get stats in /var/log/
 145 20:35 < white> pere: so using 2.6 kernel as default for us is IMHO not the best solution?
 146 20:35 < finnarne> and we might need kernel-image-2.6-386 on the CD
 147 20:35 < pere> white: well, I like to know what is going on no matter which solution is selected. :)
 148 20:35 -!- |Terje| [n=Terje@128.39.41.58] has joined #debian-edu
 149 20:35 < finnarne> pere: nope, I have not tried bootchart
 150 20:35 < white> ok anyone some time for testing it and getting in contact with pere about a possible solution?
 151 20:35 < pere> kernel images are ~15 MiB on the Cd. :(
 152 20:35 < finnarne> Maybe I could try later this week.
 153 20:36 < white> finnarne: do we need a fix for that before we can release?
 154 20:36 < finnarne> pere: we already have kernel-image-2.6.8-2-386, so adding kernel-image-2.6-386 would not add much extra
 155 20:37 < finnarne> do we have a bootchart-package I could use in sarge ?
 156 20:37 < pere> finnarne: ah.
 157 20:37 < white> finnarne:  ok we can assign you to this bug ok? and you keep in contact about that with pere?
 158 20:38 < finnarne> white: no
 159 20:38 < pere> finnarne: just fetch the one from sid.  it is a shell script.
 160 20:38 < finnarne> I have enough to do this week.
 161 20:38 < white> finnarne: ok
 162 20:38 < white> finnarne: is it a release blocker?
 163 20:38 < finnarne> not for me, it's not that much impact for a thin client.
 164 20:38 < pere> long boot time is perhaps not a blocker, but it is a pain.
 165 20:39 < finnarne> It's a blocker for using diskless workstations, but I know the workaround.
 166 20:39 < white> all agree that this can be fixed during the next weeks but will not affect the release cycle?
 167 20:39 < pere> Oliver in ubuntu has done some work on speeding up the boot, and his approach is to remove scripts from /etc/rcS.d/ and /etc/rc2.d/, and I considider using his approach too.
 168 20:39 < finnarne> but the way ltsp is setting up the kernel, is a pain
 169 20:40 < white> or is anyone willing to work on this during the next week?
 170 20:40 < finnarne> I'm removing a lot of stuff when doing a diskless workstation
 171 20:40 < white> finnarne: can you please fill in a bug report?
 172 20:40 < finnarne> and juggling with runlevels, depending on the network it's booted from
 173 20:40 < finnarne> white: you know almost as much as I do.
 174 20:41 < pere> finnarne: I would be very interested in knowing which scripts you keep.  the ubuntu patch moves all scripts not in the whitelist.
 175 20:41 < finnarne> take alook in ltsp-make-client
 176 20:41 < finnarne> I dont remember
 177 20:41  * white wants to stop this issue now and maybe discuss it later to finish all agenda tasks ok?
 178 20:41 < pere> finnarne: right.  I'll try to remember that.
 179 20:41 < finnarne> but the good news is that diskless workstations is working
 180 20:41 < white> :)
 181 20:42 < finnarne> pere: you also talked about replacing bash with dash on the thin clients, to save some memory footprint
 182 20:42 < finnarne> ?
 183 20:42 < white> ok going to next agenda task
 184 20:42 < finnarne> ok
 185 20:42 < white> the question is do we need an other rc?
 186 20:42 < white> comments?
 187 20:43 < finnarne> I think we need another rc
 188 20:43  * white too
 189 20:43 < finnarne> after the samba/ldap/sound/gcompris is fixed
 190 20:43 < white> agree
 191 20:43 < pere> I agree.
 192 20:44 < white> finnarne: ok after these bugs are fixed i will ping you for rc2 ok?
 193 20:44 < pere> finnarne: yes, but I am not sure it will work.  Oliver said he tested and ran into problems with udev.  not sure if it is only in newer udev or in our version.
 194 20:44 < finnarne> I also found some bashism in one of the ltsp-script
 195 20:44 < white> ok we agreed in having a new release candidate after the three major bugs are fixed (probably during this week)
 196 20:44 < pere> finnarne: right.  I suspect there might be some left.  will need to check carefully before switching.
 197 20:45 < white> finnarne pere: do we need to spend some time discussing ltsp now (i guess 4 min) or shall we go on?
 198 20:45 < pere> just go on.
 199 20:45 < white> ok
 200 20:45 < finnarne> I think usr/lib/ltsp/ltsp_functions
 201 20:45 < white> last point on the Agenda is archive maintaining
 202 20:45 < white> zobel: ping!
 203 20:45 < pere> finnarne: ltsp_functions should be fixed.
 204 20:45 < white> maybe you let me summarize introduce it
 205 20:45 < zobel> white: pong :)
 206 20:45 < vagrantc> pere: i have a patch for your ltsp to get kernel selection working in sid, etch and sarge (as well as ubuntu, i think)
 207 20:45 < pere> I've continued to set up a.s.n.  I can give status on it.
 208 20:46 < white> pere: go ahead please
 209 20:46 < pere> vagrantc: great!
 210 20:46 < vagrantc> and the default is to use 2.6
 211 20:46 < white> vagrantc: now archive maintaining sorry ;)
 212 20:46 < vagrantc> a meeting, right now?
 213 20:46 < white> vagrantc: yes
 214 20:46 < zobel> vagrantc: see topic
 215 20:46 < white> pere will give a status about admin.developer.or
 216 20:46 < white> g
 217 20:47 < white> s/developer/skolelinux/
 218 20:47 < pere> ok, the machine is up and running, and I've configured it to fetch user info from LDAP.  The 'staff' group is allowed to log into it (me, ragnar, werner and tfheen).
 219 20:47 < white> pere: all jobs except dak are configured and running?
 220 20:47 < pere> backup is missing.  I have not had time to try to get dsm working on it yet.  Hope to get the debian packages I use on d.s.n working.
 221 20:48 < white> pere: ok
 222 20:48 < white> shall i give some more notes now?
 223 20:48 < white> pere: ?
 224 20:48 < pere> The LDAP database need to be updates
 225 20:48 < pere> white: I'm not quite done yet.
 226 20:48 < white> ok
 227 20:48 < pere> LDAP need to be synchronized with /etc/passwd, /etc/group and /etc/netgroup on developer, to make the new users visible in ldap.
 228 20:49 < finnarne> should we change all machines to use ldap ?
 229 20:49 < finnarne> also developer ?
 230 20:49 < pere> we probably want to create  a new group for the archive maintainers, or reuse the www group.  This group need to be made visible in ldap as well.
 231 20:50 < pere> finnarne: well, I suggest we wait a bit.  I would like some auto-sync from /etc/ on developer into LDAP, and move the LDAP server from developer to a.s.n.
 232 20:50 < zobel> pere: an own group might be good.
 233 20:50 < pere> at the moment I manually sync file and netgroup to make sure 'staff' is available as both.
 234 20:50 < Gagatan> werner/pere: I'm going to Oslo tomorrow.. and have some time between meetings at usit and before my flight. If we want/need some work done on ldap or anything else, please say so. I've booked a late flight to get some hours to work in forskningsparken if needed
 235 20:50 -!- schueleg [n=schueleg@p54877D2C.dip.t-dialin.net] has joined #debian-edu
 236 20:51 < pere> Gagatan: right.  are you aware that NUUG have a work meeting in forksningsparken from 18:30?
 237 20:51 < Gagatan> nope
 238 20:51 < pere> I've installed 'dak', but have not configured it.  I've also mounted the home directories from developer.s.n.
 239 20:51 < pere> that is about where I am at the moment.  questions?
 240 20:52 < finnarne> any thoughts about the different flavours, and arch ?
 241 20:52 < Gagatan> 'dak' is the new donated server?
 242 20:52 < white> ok Werner, zobel, holger and me want to start configuring dak
 243 20:52 < pere> Gagatan: dak is a package.
 244 20:52 < pere> Gagatan: a.s.n is the new server.
 245 20:52 < zobel> Gagatan: dak == debian archive kit
 246 20:52 < white> Werner_ and me will be the knowledge of dak *inside* debian-edu
 247 20:52 < finnarne> Gagatan: dak = debian archive komplex :)
 248 20:52 < Gagatan> ah :)
 249 20:52 < white> if another one wants to join us please say so
 250 20:53 < white> i think during the next week(s) we will do some tests with configuration and daily dak maintenance to learn about it
 251 20:53 < pere> first step is to get the users into LDAP.  then I can give them access to a.s.n.
 252 20:53 < zobel> white: it might be a good idea if finnarne also is in that "inside" team!
 253 20:53 < white> when we are finish we can start to copy the pool
 254 20:53 < pere> the LDAP database is several years old.  it does not contain all the users in d.s.n:/etc/passwd.
 255 20:53 < finnarne> as of today we have 3 flavours : woody, woody-test and sarge
 256 20:54 < white> zobel: his decision i am not sure if he has enough time he said no
 257 20:54 < finnarne> woody and woody-test are missing some source packages
 258 20:54 < pere> someone said something about postgres being needed?
 259 20:54 < white> i think the future pools will be woody, woody-test, sarge, sarge-test, etch and etch-test and all will produce a cd
 260 20:54 < white> pere: yeah locally
 261 20:54 < zobel> finnarne: i would suggest woody, woody-test/woody-proposed, sarge, sarge-proposed, etch and etch-proposed
 262 20:55 < white> or testing
 263 20:55 < white> zobel: is it needed that we fix the names now?
 264 20:55 < zobel> pere: yes, dak needs a postgresql-server running
 265 20:55 < zobel> white: no.
 266 20:55 < finnarne> I think we could do without -test/proposed for woody and sarge
 267 20:55 < pere> the -test repositories have normally been populated with experimental packages, before we know if the change should go into the normal repository.
 268 20:56 < pere> so I am not sure if -proposed is a good replacement for -test.
 269 20:56 < white> ok at first summarizing who is interested in joining the ftp-team ...
 270 20:56 < finnarne> I think we should have woody, sarge, etch and etch-test or something like that
 271 20:56 < white> finnarne pere: can we at first make sure who will work on it because time is running we can discuss the details right after it ok?
 272 20:56  * h01ger is sorry to be busy with #debconf-team debconf6 preparations... i thought about the clash of meetings for stockholm, but not for myself... :( :)
 273 20:57  * h01ger will read minutes for sure, maybe even backlog :)
 274 20:57  * pere created the 'ftp' group, and will use it for the ftp archive maintainers.
 275 20:57 < white> i know about zobel, holger, Werner_ and me
 276 20:57 < white> pere?
 277 20:57 < white> finnarne?
 278 20:57 < white> anyone else?
 279 20:57 < pere> I can join, but do not know how much time I got to spend on it.
 280 20:57 < finnarne> I think we should produce a etch, with working packages, and then add etch-test to test new packages
 281 20:58 < pere> is zobel missing an account on d.s.n?
 282 20:58 < white> pere: he sent a request
 283 20:58 < zobel> pere: i send a request
 284 20:58 < pere> probably not processed yet.
 285 20:59 < white> do we need an own ML for the ftp-team?
 286 20:59 < finnarne> white: if someone could produce a nice howto on how to handle the archive, then I could join
 287 20:59 < pere> ok, I'll start by making sure the ftp group can log into a.s.n.
 288 20:59 < pere> white: I would like the ftp-team to work on debian-edu to make sure it all is transparent for now.
 289 20:59 < white> finnarne: we will get howtos with learning by doing i guess :)
 290 20:59 < white> pere: ok
 291 20:59 < zobel> white: my idea is to have 2-3 persons i show how to work with dak, and while that, we write an howto
 292 21:00 < white> pere: i just want to mention all questions which can rise up
 293 21:00 < pere> white: we need written howtos, not 'learing by doing'. :)
 294 21:00 < finnarne> zobel: I will be your crash-test-dummy for the howto.
 295 21:00 < white> finnarne: :)
 296 21:00 < zobel> finnarne: okay.
 297 21:00 < white> ok so we have pere Werner_ finnarne holger zobel and me
 298 21:01 < finnarne> are we done ?
 299 21:01 < zobel> so i guess we start with h01ger, Werner_ and white ...
 300 21:01 < white> so far time is over now i stop the official part of the meeting
 301 21:01 < finnarne> next meeting ?
 302 21:01 < white> maybe someone is interested and still wants to discuss the dak
 303 21:01 < white> finnarne: two weeks?
 304 21:02 < white> i guess 06.03. ok?
 305 21:02 < white> but Werner_ said that monday is not the best day
 306 21:02 < white> maybe we should change the day
 307 21:03 < schueleg> hi white, why?
 308 21:03 < white> pere finnarne: can we change the day?
 309 21:03 < white> schueleg: because many others could not attend
 310 21:03 < pere> tuesday are bad for me.
 311 21:03 < h01ger> white, didnt you get my mail from/about lior kaplan (?)
 312 21:03 < finnarne> for me as well
 313 21:03 < white> h01ger: later please
 314 21:03 < zobel> white: maybe rotating the days is a good idea. like debian-boot does.
 315 21:03 < pere> white: if we change day, I suggest we let people wote on a wiki.  let people list the days they prefer, and count votes.
 316 21:03 < white> what about thursday?
 317 21:03 < finnarne> and not wednesday either
 318 21:04 < white> pere: ok i will create a wiki page for that good idea
 319 21:04 < zobel> pere: wiki is a good idea.
 320 21:04 < white> anyone against a wiki page?
 321 21:04  * white will write a wiki page and write a mail to debian-edu@l.d.o. and fix a date for the next meeting in three or four days ok?
 322