IRC chat logs for #ltsp on irc.libera.chat (webchat)


Channel log from 26 April 2008   (all times are UTC)

00:36asac_ has joined #ltsp
00:48asac has quit IRC
00:48asac_ is now known as asac
00:58indradg has quit IRC
01:00indradg has joined #ltsp
01:54mccann has quit IRC
01:56mccann has joined #ltsp
03:16mikkel has joined #ltsp
03:42Q-FUNK has joined #ltsp
03:46BGomes has quit IRC
03:48indradg has quit IRC
03:53Pascal_1 has joined #ltsp
04:30ogra has joined #ltsp
04:40cyberorg_ has joined #ltsp
04:40cyberorg has quit IRC
04:40cyberorg_ is now known as cyberorg
04:58F-GT has quit IRC
05:20F-GT has joined #ltsp
05:25indradg has joined #ltsp
05:54J45p3r has joined #ltsp
05:54K_O-Gnom has joined #ltsp
06:35Q-FUNK has quit IRC
06:36Q-FUNK has joined #ltsp
07:01Pascal_1 has joined #ltsp
07:10K_O-Gnom has quit IRC
07:10K_O-Gnom has joined #ltsp
07:20Pascal_1 has quit IRC
07:38asac_ has joined #ltsp
07:39ace_suares has joined #ltsp
07:39
<ace_suares>
!seen ogra
07:39
<ltspbot>
ace_suares: ogra was last seen in #ltsp 20 hours, 25 minutes, and 28 seconds ago: <ogra> thats why we do LTS :)
07:42
<ogra>
ace_suares, what a wonderful bug :)
07:42* ogra forwards a hug from the whole ubuntu disro team
07:49hersonls has joined #ltsp
07:49J45p3r has quit IRC
07:50
<ace_suares>
hi ogra thx
07:50
this is the upgrade to fast bug :-) ?
07:50
<ogra>
yeah
07:50
the whole team loves it
07:50
<ace_suares>
it is already decided... 'won't fix ' :-)))))
07:50
<ogra>
hehe
07:50
<ace_suares>
:-)))
07:51
I'ts really amazing tough.
07:51
<Q-FUNK>
ogra: sober again? :)
07:51
<ace_suares>
Did you see a new bug ? Less wonderfull
07:51
btw CONGRATULATIONS !
07:51
very happy that you made another release.
07:52
ogra: https://bugs.launchpad.net/ubuntu/+bug/222513
07:53asac has quit IRC
07:53
<ace_suares>
hmmm ltsp-update-image wants a 64bit /opt/ltsp....
07:53
Error: chroot /opt/ltsp/amd64 doesn't exist.
07:53
altough I Installed 32-bit clients
07:54
<ogra>
it defaults to the host arch, use -a with it
07:54
<Q-FUNK>
on a 64-bit server?
07:54
<ogra>
-a i386 should do
07:54
<Q-FUNK>
right
07:54
good to know
07:54
<ace_suares>
that's not in the man page :-P
07:54
<Q-FUNK>
bug the man page on LP :)
07:54asac has joined #ltsp
07:55
<ogra>
yeah
07:55
its in the help output
07:55
seems the manpages are behind
07:55
can be fixed for 8.04.1
07:56
<ace_suares>
https://bugs.launchpad.net/ubuntu/+bug/222524
07:58
ogra: I did ltsp-update-image but won't help.
07:58
On the thin client, rofs can not be mounted, /root/etc can not me mounted (no such dir),...
07:58
<ogra>
did you recreate the client chroot as usualy on upgrades ?
07:59
*usual
07:59* leio idly notes that everything broke with that on his work machine
07:59
<leio>
btw, what fast bug. Linkie? :)
07:59
<ogra>
https://bugs.edge.launchpad.net/ubuntu/+bug/221918
08:05* ogra watches debian dropping ltsp
08:05
<Q-FUNK>
?!
08:05
<yanu>
ace_suares: did this happend for real? or is it just a joke? (i guess a joke)
08:05
<ace_suares>
wtf !?
08:05
<laga>
ogra: ?
08:05
<ogra>
ace_suares, only temporary :)
08:05
<ace_suares>
yanu: well, the Tim use case is of course fake :-)
08:05
ogra: phew...
08:06
<ogra>
apparently you cant change the arches a package is built for in debian unless you remove the package completely first
08:06
<ace_suares>
yanu: but it is kind of hard to get work with linux system adminsitration since there is nothng to do
08:06
<ogra>
just dont tell your customers :P
08:06
<ace_suares>
Just installng single computer with wixp takes 6 hrs or so, with ubuntu, maybe an hour and a half tops
08:07
So yeah, that's not good for those who earn money with system adminsitration...
08:07
ogra: oh, that makes sense.
08:08
ogra: how can I log bootup thin client :? I am tryin g to make a screen shot with my phone now, doesn't work either
08:09
<ogra>
ace_suares, suso mv /opt/ltsp/i386 /opt/ltsp/i386.gutsy && sudo ltsp-build-client --arch i386
08:10
that doesnt work ?
08:10
*sudo
08:10
<Q-FUNK>
ogra: is whoever that filed this bug for real?
08:10* ace_suares filed that bug :-)
08:11
<ogra>
Q-FUNK, yes, i'm talking to him, his LP name is not much different from his nick
08:11
;)
08:12
<Q-FUNK>
ace_suares: you filed the bug against the wrong distro. it should have been filed against CIO Bogux.
08:12
<ace_suares>
ogra: hey I attached a screen shot. Gotta go now, if you have a sec, look at it. This happened after standard upgrade, and ltsp-update-image didn't do anythinh either...
08:12
CIO Bogux ? Linkplease :-)
08:13
<Q-FUNK>
no, really. let the guy do his upgrade in 2 hours and play tetris for the rest of the week.
08:13
it's not a bug, it's a feature. filing bugs likes these is abusing LP.
08:13
<laga>
oh noes! someone made a joke and now the launchpads are exploding!
08:14
(warning: headache-induced sarcasm ;))
08:15
<Q-FUNK>
or filing it agaist sillyx linux might have been a good choice.
08:16Nubae has left #ltsp
08:16Nubae has joined #ltsp
08:16asac_ has quit IRC
08:23bobby_C has joined #ltsp
08:24
<ace_suares>
I proposed a set of solutions, but the distro that has all those solutions built in is not on launchpad :-(
08:27
<ogra>
ace_suares, ltsp-update-image will only compress a new image
08:27
did you rebuild the chroot with ltsp-build-client before ?
08:32bobby_C has quit IRC
08:33generic has quit IRC
08:36
<ace_suares>
yeah before the upgrade i did that. I *had* too, because it installed 64bit clients, remember :-P
08:36* ace_suares hands aspirin to laga
08:38bobby_C has joined #ltsp
08:39
<ace_suares>
shall I just rebuilt opt/ltsp? Let's try that.
08:40
<Nubae>
hey, what's the best way to transfer users from one install to another, just copy relevant shadow, group, passwd?
08:40
<ogra>
ace_suares, thats what i'm talking about since 40min :)
08:41
<ace_suares>
awww... missed that
08:41
o see it now
08:41
o=i
08:41mccann has quit IRC
08:41
<ogra>
<ogra> ace_suares, sudo mv /opt/ltsp/i386 /opt/ltsp/i386.gutsy && sudo ltsp-build-client --arch i386
08:41
:)
08:42
<ace_suares>
ah did a rm -r /opt/ltsp already :-)
08:43
Have you not had any problems after the upgrade ?
08:43
<leio>
doesn't work at all for me. I'm probably just an odd case there though.
08:43
<ogra>
not with the recomended method, no
08:44
<ace_suares>
ehhhh what is the recommended method... isn't that update-manager ?
08:44
<ogra>
leio, see above, did you run ltsp-build-client ?
08:44
<leio>
yes.
08:44
<ogra>
ace_suares, for ltsp ? no ... "sudo mv /opt/ltsp/i386 /opt/ltsp/i386.gutsy && sudo ltsp-build-client --arch i386"
08:44
<leio>
after that I have an image that boots till something, and then I end up with consoles where I can only enter stuff with no feedback. No login, no terminal, no X, no nothing.
08:44
<ogra>
ltsp was never integrated with update-manager (yet)
08:45
leio, and it worked in gutsy ?
08:45
<ace_suares>
ogra: hmmmm that bugs me tough
08:45
ogra: oooooh
08:45
<leio>
well, yes. I did it on hardy 5 days before gold though
08:45
<ace_suares>
ogra: good to know. These things are 'weirdness'...
08:46
<ogra>
leio, where exactly does the boot stop ? do you get a login prompt ? or does it drop you to a busybox shell ?
08:46
<leio>
i get absolutely no prompt of any kind
08:46
as said above
08:47
<ogra>
you said you see consoles
08:47
so you end up with a black screen and flashing cursor on the top left ?
08:48
<leio>
something like that, yes
08:48
<ace_suares>
ogra: works perfectly now.
08:48
<ogra>
do you get to a login prompt hittin alt+f1 ?
08:48
ace_suares, good
08:48
<ace_suares>
ogra: I think canonical should find a way to let me know when i do an upgrade, the I need to do something
08:48
<ogra>
i will try some u-m integration with intrepid
08:49
<ace_suares>
manually too...
08:49
<ogra>
the prob with the chroot is that there are to many things set at build time
08:49
<ace_suares>
ogra: inthese cases i am sorry I can not find the time nor have the skills to do such integration for you
08:49
ogra: and the only thing i can do is just bug you for it :-(
08:49
ogra: thx for being here !!!
08:50
<ogra>
upgrading these things would become a huge messy upgrade.sh script or so would be hard to maintain, but i can probably hook something into u-m that forces the new build of the chroot
08:50
or at least notifies the user to do so manually
08:52
leio, do you get to a login prompt hitting alt+f1 ?
08:52
on a booted client
08:52* ogra suspects the client boots fine but has an issue with the X driver or setup
08:53
<leio>
ogra: no.
08:53
<ogra>
thats a standard ubuntu hardy ?
08:54
<leio>
that was an upgrade to hardy from gutsy about a week before official release
08:54
<ogra>
hmm
08:54
what kind of clients do you have attached there ?
08:55
<leio>
hmm, 1.5 weeks prior I guess, as it was Thursday or Friday, but not last week, so yeah.
08:55
Was just one ThinCan DBE61 or DBE62
08:55
<laga>
ogra: maybe my 'oversight' in the mythbuntu plugin caused this? (i haven't been following the conversation)
08:55
<ogra>
oh
08:55
<ace_suares>
ogra: notification woudl be best
08:55
<ogra>
laga, hmm, that might be, right, i totally forgot about that one and it only existed for one day
08:56
<ace_suares>
ogra: rebuild the chroot, is good but please take care of arch issues and local mirrors (!)
08:56
<ogra>
laga, but thin can also means likely broken X driver, since it has issues with the geode/amd driver
08:56
<ace_suares>
ogra: I closed the bug. https://bugs.launchpad.net/ubuntu/+bug/222513
08:56
<ogra>
leio, ^^^
08:56
<laga>
ogra: right, i rememeber the discussion here
08:56
<ace_suares>
thanx bye ogra !
08:56
<ogra>
leio, Q-FUNK knows what to do with these thingies
08:57
<Nubae>
leio: I had a similar problem to yours, being dropped to busy box without any reason... the answer was my gateway was plugged directly into the hub, and even though dhcp server is turned off in it, conflicted with ubuntu's dhcp
08:57
<ogra>
ace_suares, thanks for pointing to it, i'll put up an upgrade wikipage
08:57
Nubae, he isnt dropped to busybox
08:57
<leio>
ogra: I disable X server in lts.conf.
08:57
<Nubae>
ah
08:57
<ogra>
and there are known issues with the thin can graphics cards
08:57
<leio>
Nubae: I don't see any busybox, I could have debugged something successfully if I would
08:57
<Q-FUNK>
hm?
08:58
<ogra>
Q-FUNK, thin can on hardy ?
08:58
<Q-FUNK>
yeah, what about it?
08:58
<ogra>
ThinCan DBE61 or DBE62, seems leio doesnt get his to work
08:58
<Q-FUNK>
needs an xorg.conf with the Device declared, then it works
08:59
X core itself doesn't appear to know anything about our driver
08:59
<leio>
I didn't get any login shell when lts.conf had the xserver VT configured to shell
08:59
<ogra>
hmm, probably usplash gets in the way here
08:59
<Q-FUNK>
right and that I'm not sure why, since LTSP uses Xorg -configure to generate xorg.conf and that correctly produces an xorg.conf that uses "geode"
08:59* ogra cant remember having usplash added in the nbi images though
09:00
<Nubae>
hmmm, so these thin cans, seem quite nice thin clients, are they expensive?
09:00
<leio>
anyhow, I would need to upgrade host to hardy final and build the image again and see. Not exactly a priority anymore, as the reason I did all this was to ensure things work before Hardy is out. Now Hardy is out and Q-FUNK got things tested at home and nothing we can do about anything anyhow before 8.04.1 really
09:00
<Q-FUNK>
Nubae: the prices appear at the bottom of thincan.com's product info
09:00
<ogra>
Nubae, they are cheap and good, but sadly use the wrong graphics chipset
09:00
<Nubae>
damn, they are really cheap
09:01
<Q-FUNK>
leio: we can still produce patches to fix whatever issue is remaining and have ogra upload them to proposed-updates
09:01
<leio>
That is, LTSP not working on my edubuntu at office after upgrade to Hardy and fresh ltsp image build had nothing to do with xorg-server
09:01
<ogra>
leio, well, getting 8.04.1 tested :) so it surely works :)
09:01
<Q-FUNK>
ogra: I reall wonder why -amd successfully autodetects our hardware inside an LTSP chroot, but -geode doesn't
09:02
<leio>
ogra: wrong graphics chipset....?
09:02
<ogra>
leio, oh, what was wrong
09:02
<leio>
ogra: huh? ;p
09:02
<ogra>
leio, well, upstreams support for the driver isnt very good
09:02
<ace_suares>
Q-Funk: cool products ! I've resaerch thin clients for 4 years and never came along this products.
09:02
<Q-FUNK>
thanks :)
09:02
<ogra>
leio, i mean with the setup that didnt have graphics issues
09:02
<Q-FUNK>
something for every budget, mainly
09:03
<Nubae>
the gigabit thin terminals look great
09:03
<ogra>
i'm still sad you dropped the beautiful can lookalike case
09:03
<Q-FUNK>
ogra: you mean before the rename, when -amd pretty much worked like magic on anything except ION 603?
09:03
<ace_suares>
yeah, that was a neat design.
09:04
<Q-FUNK>
the round can costed too much to produce, unfortunately
09:04
<leio>
ogra: Q-FUNK took over testing after I gave up on LTSP on my edubuntu due to it taking too much time and continued with other work tasks. Then I just helped with ideas what could be wrong and a backported libddc patch
09:04
<ogra>
Q-FUNK, no i mean that it was necessary that gadi hired someone to take care of the issues instead of having proper upstrea support for it
09:04
<leio>
what issues and what someone?
09:04
<ogra>
leio, ah
09:05
<Q-FUNK>
leio: we'd need to setup an actual LTSP server on the intranet and have it serve boot payloads to hosts whose MAC address fits the Artec namespace.
09:05
ogra: ah yes, those mysterious BIOS issues.
09:05
leio: the driver locking the hardware dead on some BIOS
09:06
<leio>
the upstream is the BIOS vendor there, not us.
09:06
<Q-FUNK>
no
09:06
GSW and Award are the problem
09:07
that and X core "upgrading" from vm86 (which always worked flawlessly) to x86emu (which seems to be a reall load of crap that broke many drivers)
09:07
<leio>
remember that Hardy ship with 1.4.0.90 with patches, that still uses vm86.
09:07
afaik.
09:07
<ogra>
leio, issues --> well, usually i doesnt happen that X core chnges stuff without te drivers being fixed alongside ... that didnt happen for -amd in gutsy and caused todays stuation ... nothing anyone can predict but from a distributor POV its hard to handle
09:07
<Q-FUNK>
no
09:07
not just for amd
09:08
until bartman's patches were applied, X also was messy on my old thinkpad's siliconmotion chip, for instance
09:08
<ogra>
leio, someone -> symbiont (gadis company) hired an X developer to look after the -amd driver issues that showed up in gutsy
09:08
<leio>
oh ok, missed the gadi and symbiont connection, sorry
09:08
<ogra>
since upstream was dead/uninterested in fixing it
09:09
in that view thin can uses the wrong graphics chipset (which is indeed not thin cans fault)
09:10
<Q-FUNK>
ogra: mind you, those were all X core issues. the same driver works rock-solid on older X
09:10
<ogra>
indeed
09:11
<Q-FUNK>
the culprit really is the X core team. too many bold redesigns and leaving non-mainstream drivers in the dust
09:11
<leio>
hmm, why did I see vm86 symbols in backtraces in gentoo 1.4.0.90, but apparently things use x86emu on ubuntu, having caused these freezes until patches were applied. Does that mean ubuntu is grabbing x86emu patches?
09:11
<ogra>
and it could have happened to vias *chrome chips as well ... with the difference that there upstream is apparently more active
09:11
<leio>
*chrome upstream did not fix these xorg issues...
09:12
<ogra>
well, my unichrome and openchrome boards dont have any issues
09:12
<Q-FUNK>
leio: debian and ubuntu switched to x86emu by default, because they want x86-32 code to run on x86-64 as needed
09:12
<leio>
then you didn't hit the problematic path between xorg-server and BIOS communication I guess
09:13
<ogra>
(apart from the fact that i have to use different drivers for them that mutually exclude each other)
09:13
.oO(oh why is X such a mess)
09:13
<leio>
Q-FUNK: eh? ;p I run x86-32 X using code just fine on x86-64
09:14
with vm86
09:14
<Q-FUNK>
beats me
09:15
that's the reason that was claimed to me to justify the switch
09:16
ogra: I sort-of understand Jordan keeping his "not my department" attitude. the geode driver was never at fault. all of bartman's patches were to fix a bad x86 emulator running in the X core.
09:16
even worse is how the coreboot team repeatedly offered to comaintain x86emu. apparently, X core never took them on their offer.
09:17
"not made here" attitude.
09:17
<ogra>
Q-FUNK, well, there is a known issue with a known fix and obviously many distros wont allow new upstream but a .1 release ... refusing .1 just hurts your users
09:17
<Q-FUNK>
for this one, indeed, I agree with you.
09:18
you'll notice that when I was in charge of releases, we have manhy small incremental releases, whose changes were easy to track.
09:18
to me, this is the same bullshit as sabdfl's idea that every piece of free software in the universe should release on the same date. BAD idea!
09:19
<laga>
ack
09:19
<ogra>
when did he say that ?
09:19
<laga>
i was under the impression he implied that distros should release on the same date and ship identical versions of software, too
09:20
<Q-FUNK>
right now, we can track changes to one or two pieces of software and notice regressions caused by those two.
09:20
can you imagine having to track regressions over a dozen of major software, because they all released on the same week, and figure out whether bug X is caused by the new gtk2 or by some crap in e.g. firefox or OOo?
09:20
<laga>
ogra: planet.ubuntu.com today or yesterday
09:20
<ogra>
well, kde switched to 6 month cycles, fedora did
09:20
<Q-FUNK>
ogra: read the planet. sabdfl said many time that he'd like the main software (kde, gnome, ooo, kernel, X) to release accoridng to the exact same chedule.
09:20
<ogra>
cant be that wrong if so many follow us
09:21
its enough if every upstram tracks its own issues
09:21
<leio>
that'd be some major drag and pain to distributions that don't lag behind 6 months at a time in stable tree
09:21
<ogra>
well, you dont have to as a distro
09:21vagrantc has joined #ltsp
09:22
<laga>
6 months isn't long enough, IMHO. (but with that attitude, i should probably go use debian stable)
09:22
<ogra>
but it would make life easier for everyone if upstreams had predictable release schedules
09:22
<Q-FUNK>
in theory, nice that everyone upgrades to the latest and greatest at the same time. in practice, nightmare to track regressions when a zillion of libraries all update on the same day.
09:22
<ogra>
and most helpful indeed if everyone lined up
09:22
<Q-FUNK>
they do. they have schedules, but those schedules don't match ubuntu cycles.
09:23
<laga>
when do these projects actually release? two weeks before the ubuntu release or earlier? is that helpful to get all the bugs ironed out?
09:23
<Q-FUNK>
the only 2 pieces of software that should have joint release cycles are kde and gnome. then, no desktop user can say that it was left out because distro X bases its releases around DE X schedule.
09:24
currently, yes. gnoem tends to release during the freeze of ubuntu+1
09:24
<laga>
hum. i guess that works because there are pre-release snapshots of gnome in testing
09:24
<Q-FUNK>
the real answer is not for the world to accomodate sabdfl. the answer is sabdfl shifting the ubuntu release schedule by 3 months, so that releases have the time to stabilize.
09:25* ogra pats poor vagrantc on the back ... no ltsp in debian right now ...
09:26
<Q-FUNK>
but for gnome, kde and xfce to have similar calendars would indeed help.
09:26
<leio>
GNOME has a roughly 7 months preannounced ~6 months cycles, I knew that Ubuntu bases things off of that
09:27
just recently things were shifted a bit earlier with a week or two shorter or longer cycle for GNOME to accommodate GUADEC happening at the right point of the GNOME cycle
09:27
<Q-FUNK>
then kernel and X could have a joitn shcedule too (they are all drivers), but with a 3-motnh offset to the shcedule for DE
09:28
<Nubae>
what is the adm group?
09:28
<Q-FUNK>
so, kernel+X release at cycle + 2 months, desktops at cycle + 4 months. that leaves 2 months to stabilize everything to hit a 6 months release cycle.
09:30
<leio>
Maybe things could get stable early enough if some contribution to upstream would happen...
09:31* ogra now dug to the whole sabdfl blog without finding him anywhere talk about the above
09:31
<laga>
or if more testing happened. (testing is always good, especially mythbuntu could have used more ;))
09:31
ogra: let me get you the url
09:31
<Nubae>
ah, is there a list somewhere that says what ubuntu user group is what? looking for a while on the net now and can't find anything
09:31* leio found a link to theregister interview from someone elses post were this supposedly was brought up
09:32
<vagrantc>
ogra: no ltsp in debian?
09:32
<ogra>
ah, an interview
09:32
vagrantc, well, the removal bug was processed
09:32
<laga>
ogra: http://useopensource.blogspot.com/2008/04/synching-open-source-release-schedule.html
09:32
<vagrantc>
ogra: did they accidentally remove amd64 i386 and powerpc ?
09:32
<laga>
i was referring to this posting. not by the sabdfl himself, but there's a link
09:33
ogra: http://www.theregister.co.uk/2008/04/22/shuttleworth_hardy_heron/
09:33
<ogra>
yup, i read that one, funny that i didnt keep that part in mind at all
09:37
<vagrantc>
looks like they just removed: alpha, arm, armel, hppa, ia64, mips, mipsel, s390, sparc and m68k
09:48
<Nubae>
I'm unsure what the dip and adm groups are, can someone tell me?
09:48tux_440volt has joined #ltsp
09:50
<Nubae>
and should users be in plugdev and video for any particular reason?
09:50
as in users with home directories
09:50tux_440volt has quit IRC
09:50
<leio>
plugdev in some distributions at least allow automounting and such things work for the given user
09:51
<Nubae>
yeah for ubuntu that's fuse, but should I leave all users in plugdev too?
09:53
<leio>
err, fuse?
09:53mccann has joined #ltsp
09:53
<leio>
that technology in relation to GNOME gives a ~/.gvfs for remote mounts to use for legacy apps. fuse and automounting don't really go together in principle
09:53
<Nubae>
yeah must be in the fuse group to mount usb sticks, cdroms, etc
09:54
<leio>
weird
09:54
(I'm not an Ubuntu user, I'm just trying to help as I can with the knowledge I have regarding GNOME and packaging conventions as no-one else is answering right now)
09:56
<Nubae>
:-) thanks
09:56K_O-Gnom has quit IRC
09:57
<leio>
don't you have some descriptions of the groups in users-admin btw?
09:58
I remember a bunch of checkboxes saying things like "Allow user to automount", "Allow user to administer the system", etc
09:58
could observe what those change in the set of groups that user belongs to
10:01
<Q-FUNK>
ogra: start in the planet, where many people are already discussing the implications that such a decision would have
10:01
<Nubae>
users-admin no longer allows editting by default
10:03hersonls has quit IRC
10:03spectra has joined #ltsp
10:03gentgeen__ has quit IRC
10:05
<ogra>
Q-FUNK, i only see the one post from tristan that laga pointed out in the beginning
10:06Nubae has left #ltsp
10:06cyberorg has quit IRC
10:07
<Q-FUNK>
ace_suares: ah. antilles
10:07* ogra goes back to clean the garden furniture
10:08
<Q-FUNK>
ace_suares: are your guys using european power sockets or amercian ones?
10:12alekibango has quit IRC
10:16
<ace_suares>
Q-Funk: we have adapters so we can work with both. But stadard we use 110V/50Hz American socket,.
10:16
<Q-FUNK>
ace_suares: ok. the thing is that we only stock on european power supplies.
10:16
<ace_suares>
no problem we are used to deal with products both from the US (60 hz/110V) and from
10:16
Europe: 220V/50Hz
10:17
<Q-FUNK>
öö... amercan is at 60Hz, isn'it?
10:17
yup
10:17
<ace_suares>
Q-funk: yes, they are, we are not.
10:17
We have the worst of both wordls :-)
10:17
<Q-FUNK>
yikes!
10:17
us voltage and european frequency?
10:17
<yanu>
where is that?
10:17
<ace_suares>
yep
10:18
<Q-FUNK>
go dutch!
10:18
yanu: dutch antilles
10:18
<yanu>
where is that
10:18
<ace_suares>
www.curacao.com
10:19
<Q-FUNK>
:)
10:19
<ace_suares>
Goodplace to hold an LTSP meeting
10:20
<Q-FUNK>
if anyone asks me, the international standard should become 220V/60Hz, but well...
10:20
<ace_suares>
.ee ? estonia ?
10:20
<yanu>
do they speak dutch out there? hartelijk welkom
10:20
<Q-FUNK>
yup
10:20
<ace_suares>
cool
10:21
<Q-FUNK>
ace_suares: depends. are you guys within the open skies agreement?
10:21
<ace_suares>
alomsot had a project there. was in the lower balkans, romanai, albania,, bulgaria
10:22
Q-Funk: some open skies, yes not all I giess.
10:22
<Q-FUNK>
balkans is quite far from here
10:22
<ace_suares>
then got offerend a project in the georgia/lit/est
10:22
but went to curacao ... cold vs warm, my choice is clear :-)
10:22
alomsot = almost
10:23
<Q-FUNK>
ace_suares: I generally avoid flying to countries that are participating in the US-lead war on peaceful travelers.
10:23
<ace_suares>
I generally avoid flying period
10:23
<Q-FUNK>
that's why i had to ask
10:23
<ace_suares>
ecxept that the nearest island is 70 clicks away and you NEED to fly (bonaire).
10:24
I worked for EYFA and ASEED, more then ten years ago.
10:25
a bit OT :-)
10:25
<Q-FUNK>
:)
10:35Nubae has joined #ltsp
10:56DonSilver has joined #ltsp
10:59DonSilver has quit IRC
11:00
<Nubae>
damnit, after an upgrade, my squid no longer accepts ssl requests
11:03artista_frustrad has quit IRC
11:04
<Nubae>
ah no, its webmin of course...
11:16
hey, what kernel do I need to use on i386 to use more than 4 gigs of ram and it be visible?
11:16
<vagrantc>
depends on your linux distro
11:17
<Nubae>
ubuntu
11:18
I was told generic should do it, but it doesnt
11:22
do I have to recompile my kernel, or is there one I can just download?
11:25Q-FUNK has quit IRC
11:29
<lns>
Nubae, i know the linux-image-server kernel uses PAE to get more than 4GB
11:29
but I thought the others did as well
11:31
<Nubae>
I'm getting mixed answers... in #ubuntu they say u must use a 64 bit OS
11:31
but on the web, there are some indications that you can recompile the kernel to support up to 64 gigs
11:31
but I thought there was a kernel package available so I can just to apt-get install
11:33
<laga>
Nubae: -server kernel AFAIK
11:33
<lns>
Nubae, PAE extensions in the i386 kernel will give you more
11:33
I'm using that on 5 servers currently
11:33
but the amd64 distro definitely has native support (no PAE, which gives a big performance hit)
11:34
<Nubae>
so which package?
11:34
<lns>
Nubae, linux-image-server
11:35
<Nubae>
I've been running 64 bit for a long time, but I'm trying i386 now for the flash and java issues
11:35
<lns>
Nubae, ;) I was in your same shoes ~7months ago
11:35
<Nubae>
and since I have quite a few workstations running low fat, thing the server can take the heat now
11:35
<lns>
for java you can use icedtea java
11:36
<Nubae>
yeah, everything is now running perfectly, except the extra memory issue
11:36
<lns>
flash you can use nspluginwrapper (although i don't think anyone's gotten sound to work on thin clients over that)
11:36
<Nubae>
I used gnash
11:36spectra has quit IRC
11:36
<lns>
good for you ;) did it work ok?
11:36
<Nubae>
worked ok, but not as good as flash
11:36
<lns>
yea
11:37
it would be very nice when gnash is more mature...time will tell
11:37
<Nubae>
problem is when entire classes are using flash based educational programs
11:37
server starts slowing and processor usage goes up to 80%
11:37
<lns>
Nubae, what kind of setup/student count do you have? you sound like you're in a very similar situation to me
11:38
Nubae, that will happen with adobe flash too
11:38
just not as bad
11:38
<Nubae>
well, thats why I've installed low fat clients in the mutlimedia lab
11:38
<lns>
low fat?
11:38
<Nubae>
run their own processor and ram
11:39
local apps
11:39
<lns>
ah
11:39
how's that workign for you
11:39
<Nubae>
very well
11:39* laga waits for wholemeal clients
11:39
<lns>
laga, =p
11:39
<Nubae>
except can't always see what's happening on them so well
11:39
<lns>
i'm using 2% myself
11:40
<Nubae>
from a management perspective thin client is still better, but this takes the stress off the server
11:41
at least, in a real world situation, ie, actual school with lots of students and teachers, you see what the real issues are
11:41
<lns>
Nubae, currently i'm running 5 schools with ~200 student accounts each, with a max of around 30-50 students online at the same time
11:42
<Nubae>
sounds like my one school
11:42
:-)
11:42
<lns>
=)
11:42
<Nubae>
5 schools, wow, so u must sit in front of the 5 servers continuasly, or u got help?
11:42
<lns>
Nubae, i have remote access to them but no, no help yet
11:43
looking for it
11:43
<Nubae>
well, I'm going to help with OLPC in July, so if you wanted to take control of yet another school, I've got another one for you here ;-)
12:05spectra has joined #ltsp
12:10artista_frustrad has joined #ltsp
12:19DonSilver has joined #ltsp
12:22Q-FUNK has joined #ltsp
12:27spectra has quit IRC
13:11spectra has joined #ltsp
13:18DonSilver has quit IRC
13:21K_O-Gnom has joined #ltsp
13:25J45p3r has joined #ltsp
13:31BGomes has joined #ltsp
13:33indradg has quit IRC
13:35mistik1 has quit IRC
13:35mistik1 has joined #ltsp
13:38indradg has joined #ltsp
13:42viking-ice has joined #ltsp
13:50BGomes has quit IRC
14:09
<lns>
Anyone wanna give a report about FF3b5 in Hardy LTSP? Anyone experiencing crashes?
14:10vagrantc has quit IRC
14:13Q-FUNK has quit IRC
14:16
<johnny>
lns, you should worry about crashes even outside of ltsp before being concerned about how it performs in ltsp
14:17
i am using beta3 on there atm, til i do the hardy upgrade
14:17
and i havent heard much of crashes
14:17
<lns>
johnny, that's good
14:18
I don't plan on upgrading servers to hardy for another month
14:18
is ff3b5 backported to gutsy?
14:18
<johnny>
i doubt it
14:18
it's prolly better if they don't
14:18
<lns>
you did a src install?
14:18
<johnny>
uhmm.. just got the binary tarball
14:18
<lns>
ah
14:18* lns hasn't done a src install of firefox in a looong time
14:18
<johnny>
didn't see a need to do a real package when i know the real package is coming
14:19
i just unpacked it in /usr/local/firefox
14:19
<lns>
did you do that because of crashes in ff2?
14:19
<johnny>
no.. because of slowness of firefox 2
14:19
i only have 4 machines
14:19
worst problem with firefox 2..
14:19
was the fact it wouldn't remembre it was dead
14:20
and then wouldn't open again
14:20
<lns>
oh yeah, the famous "firefox is already running" thing
14:20
<johnny>
yes..
14:20
<lns>
i got that from a teacher yesterday
14:20
i hope hardy is better at cleaning up dbus proc stuff
14:20
<johnny>
install gnome-watchdog
14:21
that's one idea i've heard
14:21
<lns>
ah, you just reminded me about that
14:21
cool
14:21
<johnny>
i haven't done it yet
14:21
i just run the equivalent of
14:21
pkill -u $user
14:21
<lns>
i still have 3 other school labs to install
14:21
<johnny>
at the end of the night
14:22
the extra proces that stick around don't tend to do any damage
14:22
altho nautilus can make me sad on that
14:22
<lns>
yeah
14:22
i'm just used to looking at a clean ps list
14:23
i haven't had nautilus cpu issues since that upgrade a while back
14:23
<johnny>
not cpu issues
14:23
<lns>
that's real nice not having to deal with that anymore, that was a daily thing for me
14:23
<johnny>
sometimes it just won't redraw
14:23
the desktop will be black
14:23
could be cuz i'm using sabayon tho
14:23
<lns>
whenever the desktop would be black for me i would notice nautilus was taking 100% cpu
14:23
<johnny>
i need to get some time to work on that again
14:24
<lns>
i kinda gave up on sabayon a while back.. didn't seem to work at all in feisty/gutsy
14:24
<johnny>
truly it didn't
14:24
that's why i went to work on it
14:24
there's still one really annoying bug
14:24
<lns>
do you enjoy working on software like that?
14:24
<johnny>
the maintainer is federico mena quintero
14:24
what do you mean?
14:24
<lns>
ah
14:24
<johnny>
but he didn't have time to really work on it
14:25
so i'd just fix what i could fix
14:25
and tell him to do things for me
14:25
<lns>
that's cool
14:25
<johnny>
almost all the bugs fixed during 2.22 were mine
14:25
<lns>
i wish i could hire a programmer to work on ltsp stuff like guis for common tasks and such
14:25
<johnny>
some he committed, but i did the work of figuring them out
14:26
and then he made them fit the coding style
14:26
or rather.. the sabayon design style
14:26sepski has joined #ltsp
14:26
<johnny>
of what should be done where
14:27
i'm still learning python tho
14:27
atually i'm currently evaluating python web frameworks .. in a bid to learn python more
14:35
<lns>
that's great
14:35
if you're interested in ltsp and want to make more tools for it later on let me know
14:36jammcq has joined #ltsp
14:47sepone has joined #ltsp
14:48sepski has quit IRC
15:01sepone has quit IRC
15:16RockTheHouse has joined #ltsp
15:39artista_frustrad has quit IRC
15:40artista_frustrad has joined #ltsp
15:55Q-FUNK has joined #ltsp
16:00viking-ice_ has joined #ltsp
16:03
<RockTheHouse>
Hi
16:03
Cool nick!
16:13
<xachen>
Hi, I'm using Xubuntu and when I click shutdown from the client, it kills the whole server.
16:13
Does anybody know if there is a way to prevent this?
16:14mikkel has quit IRC
16:16viking-ice has quit IRC
16:24bobby_C has quit IRC
16:33
<lns>
xachen, not sure about xubuntu/ltsp, did you google for it?
16:34
<xachen>
yeah
16:34
its because its executing shutdown right on the server
16:34
I guess my best bet is to make it only root executable
16:34
<lns>
there ya go
16:34
or just take it out of the menu all together and replace with a logoff cmd
16:34
<xachen>
Well
16:34
<laga>
to make what root only executable? :)
16:34
it should be possible to lock shutdown using hal
16:35
<xachen>
I'm doing that, but in case some support agent inds a hack and finds it funny to shutdown our server
16:35
<laga>
"should" as in "it'd make sense"
16:35
<lns>
xachen, yeah - shutdown of the server is only root/admin executable anyway
16:35
<laga>
lns: says who?
16:36
<lns>
laga, says logic? ;)
16:36
<xachen>
lns: when I click shutdown as an unpriv'd user it completes the shutdown command
16:36
<laga>
lns:
16:36
dbus-send --system --print-reply --dest=org.freedesktop.Hal /org/freedesktop/Hal/devices/computer org.freedesktop.Hal.Device.SystemPowerManagement.Shutdown
16:36
^^ that should shut down your box
16:36
(don't try it)
16:37
or you can try it, but don't complain if your computer turns off
16:37
<lns>
well if that is the case with xubuntu i'm very surprised because that's kind of bad in an ltsp env
16:37
especially if the menu option appears for everyone heh
16:37
<laga>
lns: i haven't tried it in an xubuntu ltsp environment, but i guess that's what happens.
16:38
lns: on my kde box, i can't run that command if another user is logged in
16:38
hal blocks shutdown then
16:38
(it's still possible to force it from the menu)
16:38
<lns>
laga, still - no unprivilidged user in a network/ltsp env should be able to shutdown the system, that's just bad defaults (real bad)
16:38
<laga>
yeah.
16:39
<lns>
or any linux system really, unless it's a standalone workstation
16:39
<xachen>
i'm curious is this a default for every ltsp setup?
16:39
<laga>
xachen: maybe you can try the command i posted above to check if it's really dbus/hal shutting down the box and not gdm or whatever. of course, it might turn your computer off :)
16:40
<lns>
xachen, not to my knowledge - in ubuntu ltsp, from the client if you do 'shutdown' it shuts down the terminal
16:40
but i usually take all those cmds out besides logoff anyway through gconf
16:41
<xachen>
aah so its xubuntu specific
16:41
gee ok
16:41
<laga>
xachen: wait for ogra, he knows everything. he'll around tomorrow i guess
16:43
<xachen>
hm ok :)
16:43
<laga>
+be*
16:45
<xachen>
also ltsp runs based on SSH. I'm having issue where if power dies on client and they relogin their old processes are still running and creating conflicts
16:45
so if I adjust ssh timeout settings that should fix that I'm assuming
16:45Q-FUNK has quit IRC
16:46
<lns>
xachen, try googling gnome-watchdog
17:07rjune has joined #ltsp
17:15viking-ice_ has quit IRC
17:19
<xachen>
yeah its related to HAL permissions
17:20cliebow has joined #ltsp
17:21viking-ice has joined #ltsp
17:44cliebow has quit IRC
17:54RockTheHouse has left #ltsp
17:57K_O-Gnom has quit IRC
18:20BGomes has joined #ltsp
19:08mccann has quit IRC
19:18mccann has joined #ltsp
20:11BGomes has quit IRC
21:00
<warren>
Hmm, youtube video over the network on ONE client is using 40% of the server's CPU and 20mbit/sec of network traffic
21:00
<rjune>
yes
21:00
<warren>
how is this usable at all?
21:01
<rjune>
warren, I've covered that long ago
21:01
<warren>
?
21:01
<rjune>
hold on a sec
21:01
http://www.bravegnuworld.com/~rjune/ltsp/
21:02
go in to glx/benchmarks.html
21:02
<warren>
what is this?
21:02
<rjune>
basically if you dont' have GLX or xdamage, you get crap
21:02
that's when I started looking into video apps in LTSP
21:02
<warren>
ah
21:03
doesn't that describe pretty much all clients?
21:03
<rjune>
bandwidth is Xres * Yres * colour depth * fps
21:03
it's worse on a term
21:03
<warren>
term?
21:03
<rjune>
terminal
21:04
so 320 * 240 * 16 * 24 ==
21:04
<warren>
it uses like 12mbit/sec even with the video paused
21:04
<rjune>
flash still refreshes
21:04
the above video should use around 28mbps to play
21:05
<warren>
what video?
21:05
<rjune>
xdamage alleviates that I believe
21:05
the hypothetical 320x240 video at 16bit colour and 24 fps
21:05
<warren>
ah
21:06* warren tests it with my other laptop that has a real video card
21:06
<rjune>
yeah, if you wanna do multimedia, you want local apps
21:06
<warren>
yeah
21:07
<rjune>
as I understand it, GLX does an end run around that by issuing commands, for the other side to draw
21:07
<warren>
still, you want localapps
21:08
<rjune>
yup
21:09
<warren>
is any of the localapp stuff working in ltsp upstream?
21:09
<rjune>
Not sure, I'm not familiar with the newer LTSP stuff.
21:09
I suspect it probably works a treat
21:10
assuming you mount /usr via nfs
21:10
<warren>
eh?
21:10
I thought the point of LTSP5 was to install localapps into the client chroot
21:10
<rjune>
I thought LTSP5 did root as a ramdisk, then you mount /usr as nfs.
21:11
we used to do nfsroot to a client chroot on the server
21:11
<warren>
no
21:11
<rjune>
ogra, jammcq, you around?
21:11
no which?
21:11
<warren>
LTSP5 is entire chroot installed from OS packages instead of built from sources
21:11
you mount that over the network via nbd or nfs
21:11
to boot it
21:12
<rjune>
ahh, I didn't know that.
21:12
<warren>
only thing that really changed from LTSP-4.2 is how it builds the chroot
21:12
<rjune>
gotcha
21:13
ok then, so you have to install whatever apps you need.
21:14
at one point I poked into a bunch of different areas of ltsp. I stopped doing it a while ago.
21:14
<warren>
hmm
21:14
<rjune>
Started building routers again, got working too much.
21:14
<warren>
playing youtube on a decent laptop, same bandwidth usage
21:14* rjune shrugs
21:14
<rjune>
yup.
21:14
I believe we predicted that.
21:15
<warren>
and paused is still 12mbit/sec
21:15
<rjune>
flash seems to update every pixel, not just what has changed.
21:15
nor does it do any sort of compression
21:15
which isn't too noticeable locally, but over a network, licks the sweat off a dead mans balls.
21:15
<warren>
actually, the latest flash can use opengl to render
21:16
which makes it possible to play h264 fullscreen with hardware acceleration, locally at least
21:16
<rjune>
that might make a difference.
21:16
<warren>
it doesn't work in nspluginwrapper
21:16
<rjune>
is konq a priority for you? or does something else use it?
21:17
<warren>
I don't use konq
21:17
and nspluginwrapper doesn't work for konq at all
21:17
<rjune>
so what else uses nsplugin wrapper?
21:17
<warren>
nspluginwrapper is used by default in firefox since Fedora 8
21:17
even on i386
21:17
<rjune>
I'm currently drapped in windows, or I would contradict that statement.
21:17
s/drapp/trapp
21:17
<warren>
it isolates the plugins to a separate process so crashes don't take down the entire browser
21:18
<rjune>
a wise decision
21:18
<warren>
(Fedora's browser almost never crashes a result)
21:18
<rjune>
that would explain why konq does so well for me.
21:18
<warren>
konqueror has something with a similar name that does something similar
21:18
<rjune>
ubuntu uses nspluginwrapper
21:18
<warren>
but it is even more broken than nspluginwrapper
21:18
<rjune>
LOL
21:18
when are you guys looking to switch to gnash?
21:18
<warren>
Ubuntu doesn't wrap i386 plugins in firefox
21:19
<rjune>
no it doesn't
21:19
<warren>
we tried shipping it by default but decided against it
21:19
<rjune>
not useful yet?
21:19* rjune hasn't looked at it
21:19
<warren>
because even if you buy all the gstreamer codec plugins from fluendo, it is unusable
21:19
oh wait, there's gnash and swfdec
21:19
two competing projects
21:19
<rjune>
yup
21:19
<warren>
they're both unusable
21:19
<rjune>
neither work yet
21:20
<warren>
here's the problem
21:20
lots of the stuff in flash has well known patents
21:20
<rjune>
figures.
21:20
<warren>
and they're trying to implement a moving target
21:21
<rjune>
yeah
21:22
<warren>
I'm amused by compiz working on the thin client
21:22
too bad there aren't any real thin clients sold with a decent video card
21:22
<rjune>
heh
21:22
<warren>
I can drag around my window with spinning cube and wobble
21:22
over the network
21:22
<rjune>
the eye candy is great, until somebody says, why doesn't youtube play?
21:23
<warren>
it doens't?
21:23
<rjune>
sure, what, three streams?
21:23
<warren>
youtube is working with compiz here
21:24
<rjune>
what kind of bandwidth and usage?
21:24
<warren>
wow, it lowered my CPU usage on the server
21:24
but bandwidth is double!
21:24
<rjune>
compiz did?
21:24
LOL
21:24
<warren>
yeah, compiz lowered the CPU usage on the server
21:24
it makes sense, direct rendering
21:24
I mean
21:24
indirect rendering
21:25
rjune: Fedora 9 has experimental DRI2... indirect direct rendering
21:25
rjune: opengl rendering on the side of a spinning cube
21:26
<rjune>
cool
21:27
<warren>
rjune: constnatly spinning the cube to the thin client, < 5% CPU usage on one core of server, about 400KB/sec bandwidth
21:27
<rjune>
that's not bad at all
21:27
<warren>
yeah
21:27
<rjune>
brb\
21:27
<warren>
compiz actually IMPROVES performance of the thin client
21:27
due to indirect rendering
21:29
<rjune>
that's a good thing
21:29
it cost more bandwidth?
21:29
<warren>
less
21:30
ah, compiz + youtube had no effect on bandwidth
21:30
I was looking at a different youtube video
21:30
it seems to be # of pixels
21:31
<rjune>
'k
21:59indradg_ has joined #ltsp
22:14J45p3r has quit IRC
22:18indradg has quit IRC
23:21johnny_ has joined #ltsp
23:27
<johnny_>
hardy dread
23:27
i knew the upgrade would break one of my systems
23:38dtrask has joined #ltsp
23:41dtrask has quit IRC