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


Channel log from 6 February 2012   (all times are UTC)

01:06yalu has left IRC (yalu!~yalu@12.18-200-80.adsl-dyn.isp.belgacom.be, Ping timeout: 240 seconds)
01:09Parker955_Away is now known as Parker955
02:00adrianorg_ has joined IRC (adrianorg_!~adrianorg@177.18.173.128)
02:03adrianorg__ has left IRC (adrianorg__!~adrianorg@187.115.111.198, Ping timeout: 256 seconds)
02:17adrianorg_ has left IRC (adrianorg_!~adrianorg@177.18.173.128, Ping timeout: 272 seconds)
02:23mistik1 has left IRC (mistik1!mistik1@unaffiliated/mistik1, Ping timeout: 240 seconds)
02:26mistik1 has joined IRC (mistik1!mistik1@unaffiliated/mistik1)
02:42VectorX has joined IRC (VectorX!~knight@unaffiliated/vectorx)
03:29Damianos has left IRC (Damianos!~Damianos@adsl-070-145-074-043.sip.cha.bellsouth.net, Read error: No route to host)
04:01Parker955 is now known as Parker955_Away
04:03cyberorg has left IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg, Ping timeout: 276 seconds)
05:06alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
06:06map7 has left IRC (map7!~map7@teksup41.lnk.telstra.net, Ping timeout: 252 seconds)
06:07map7 has joined IRC (map7!~map7@teksup41.lnk.telstra.net)
07:22marjus is now known as zamba
07:43killermike has left IRC (killermike!~killermik@2.26.100.200, Ping timeout: 252 seconds)
08:14loather-work has left IRC (loather-work!~khudson@wsip-98-175-250-115.sd.sd.cox.net, Quit: This computer has gone to sleep)
08:39dobber has joined IRC (dobber!~dobber@213.169.45.222)
08:49freedomrun has joined IRC (freedomrun!~quassel@BSN-142-163-238.dial-up.dsl.siol.net)
08:56bobapple has joined IRC (bobapple!~apfel@focit.de)
09:11Steve_The_Pirate has joined IRC (Steve_The_Pirate!~Gary@ip-80-238-8-128.bskyb.com)
09:45bakyt has joined IRC (bakyt!~ba@airport.kg)
09:46
<bakyt>
hi! can you tell me how much disk space consume each fat client working through nbd?
09:48
<alkisg>
None, all fat clients use the same image
09:49
<bakyt>
alkisg, but I have found out that once I restart the DHCP+TFTP (i.e ltsp server) it clients are showing SQUASHFS errors
09:49
so they require some connection to ltsp server?
09:50
<alkisg>
They require a constant NBD connection to the LTSP server
09:50
They are using a network disk served by NBD or NFS, so you can't just cut their access to their disk and expect them to work
09:50
That disk is read-only for them
09:51
Any changes are temporary and are stored locally in the client RAM
09:51
<bakyt>
alkisg, sure, I just wonder what kind server hardware would I need to server 100 clients
09:51
<alkisg>
100 fat clients?
09:51
<bakyt>
alkisg, so it's a readonly right?
09:51
alkisg, yes 100
09:51
<alkisg>
Yes the NBD export is read-only
09:51
<bakyt>
alkisg, yes, fat clients (actually it doesn't matter, thin or fat in this case I think)
09:51
<alkisg>
It matters a lot
09:51
For fat clients you don't need any server CPU at all
09:52
For thin clients, you'd need a 24 core monster etc
09:52
So for fat clients you can use any low-end PC with 2 gigabit NICs
09:52
Even single core will do, just add 2 gb of ram for caching
09:52
<bakyt>
alkisg, I mean it it doesn't matter in terms of NBD
09:52
but in my case htey are fat clients
09:53
<alkisg>
You mean the protocol?
09:53
Weren't you asking about the server?
09:53
To decide what server to buy, you need to know if you want to run thin or fat clients, it makes a lot of difference
09:53
<bakyt>
alkisg, NBD runs on a server so eventually I mean server of course
09:53
alkisg, fat clients
09:53
<alkisg>
OK, I answered above
09:54
You don't care about the server cpu, you need a couple of gb ram for the nbd disk caching, and you need good network
09:54
<bakyt>
alkisg, right.
09:55
alkisg, btw when I change (for example upload some file to the fat client) it would store it in the RAM of the client or on the server's RAM?
09:55
<alkisg>
Client RAM
09:55
<bakyt>
alkisg, thank you! And the last question...is it possible to somehow load balance the ltsp server? including DHCP+TFTP server's
09:56
<alkisg>
Sure
09:56
<bakyt>
alkisg, or, at least, can I use another NBD server? like one server for DHCP+TFTP and dedicated NBD server?
09:56
alkisg, is there any documentations already about this?
09:56
<alkisg>
You can use as many nbd servers as you like
09:56
I don't know, check the ubuntultsp wiki
09:56
<bakyt>
alkisg, thanks!
09:56
<alkisg>
You're welcome
09:57
<bakyt>
alkisg, oh sorry :D one last thing :D is it possible to change the splash screen image? :D
09:57
alkisg, during the boot
09:57
<alkisg>
Yes, but that's not ltsp related, you can find info for it in google
09:57
<bakyt>
alkisg, ok
09:57* bakyt going to use dedicated nbd server
10:01
<VectorX>
alkisg so i just went in #mac and asked why a mac is so supirior to any other linux distro and they kick and ban me
10:01
wtf is up with that
10:01
<alkisg>
Dunno, noone here uses macs, I don't know their elitist culture :D
10:02
<VectorX>
im trying to get a new mac as a sub ltsp serving thing, and ask a simple question in there
10:02
they are a bunch of freebsd copying retards by the look of it, only good at bashing windows
10:02adrianorg_ has joined IRC (adrianorg_!~adrianorg@177.18.168.15)
10:03
<VectorX>
i ask whats the diffrent between a mac and any other linux distro and an op calls me a troll
10:03
<bakyt>
VectorX, Mac users mostly are snobs and histerious :D
10:03
VectorX, lol :D
10:03
<VectorX>
alkisg you for one know ive been trying to get this thing to work for atleast the last couple of days
10:04
god
10:06alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
10:08bakyt has left IRC (bakyt!~ba@airport.kg, Quit: Leaving)
10:08
<Hyperbyte>
VectorX, asking a question like that in #mac won't get you any unbiased answers anyways. :)
10:09
<VectorX>
Hyperbyte i just asked a simple question, you got to be really gay to kick someone for it, im trying to setup a low budget computer network with different OS's for those who need, i need to come up with explanations for justifying why i need to buy a new $1500+ mac, i think i warrent some explanations
10:11
<Hyperbyte>
Why don't you Google a review?
10:12
It's a very well-known fact that a discussion on the internet about one platform vs another, nearly always derails into a flamewar.
10:12
<VectorX>
well i can understand if i asked whats so special about a mac vs windows pc
10:12
but thats not what i asked
10:12
<Hyperbyte>
You can't really blame them from trying to prevent that in #mac. They probably even have a channel policy against it which you didn't read. :)
10:13
I'm done being devil's advocate now. :P
10:13
<VectorX>
:)
10:13
<Hyperbyte>
Type 'mac vs linux' in Google.
10:13
You'll get loads of results.
10:15
<VectorX>
yeah there are a lot of reviews, but i dont think they are saying what i really need to know, which is a open question, why would i use a mac instead of any other linux distro
10:16
so i get answers from people who have some sort of practical use for it
10:16
ie why would someone prefer using photoshop on a mac vs photoshop on windows ?
10:17
its the same exact program
10:17
the hardware is pretty much identical or you can make it
10:18
so why do i need to buy a mac for my network when you can already use everything on a windows pc
10:18
or since its running ubuntu, photoshop on wine
10:26ogra_ has joined IRC (ogra_!~ogra@p5098ed03.dip0.t-ipconnect.de)
10:33
<Hyperbyte>
VectorX, http://lmgtfy.com/?q=Photoshop+Mac+vs+Windows
10:33
<VectorX>
seriously
10:33
<Hyperbyte>
Well the answer isn't clear-cut.
10:33
You need to do a lot of research. :P
10:34
Why would you use OpenOffice rather than Microsoft Office?
10:34
They're two different programs.... the answer to that differs from person to person... it's not science.
10:35
I have this friend who is in love with his iBook, iPad and iPhone. Takes them everywhere, and for him they're superiour to anything else.
10:35
I have the same thing, with my HTC Desire running Android...
10:36
You can't make a feature comparison to justify why, it's just personal preference.
11:01ogra_ has left IRC (ogra_!~ogra@p5098ed03.dip0.t-ipconnect.de, Excess Flood)
11:01
<VectorX>
Hyperbyte well its not really a feature comparison, coz to me that seems like a level playing field, we are not comparing apples to oranges, the example was why use photosop on a mac vs windows vs linux, we can even take out windows from the example
11:02
so if i can for instance use Microsoft Office on linux why would i pay $200+ to get windows sort of thing
11:03ogra_ has joined IRC (ogra_!~ogra@p5098ed03.dip0.t-ipconnect.de)
11:04
<VectorX>
im trying to understand it from a users point of view, not from a admin view which is what most of the reviews talk about
11:04
<Hyperbyte>
VectorX, it's never about one program is it? It's the whole user experience that matters.
11:04
<VectorX>
well for instance the HTC Desire is from my point of view on par or even better than and iphone 4, so how can you convince your friend it is
11:04
free apps for once
11:04
once/one
11:05
hey look i didnt have to pay for angry birds and you did
11:07
Hyperbyte user experience wise i dont really see much differcne between say ubuntu and a mac
11:08
although id just prefer to stay with windows if using the gui or linux for cli, but thats personal preference
11:10
but anyway this conversations here nor there, lets just hope greece gets its financial bailout today
11:10
<Hyperbyte>
Hah
11:11
<VectorX>
and hope that in the future nations dont spend money they dont have to host the olympics
11:13
go greece
11:17freedomrun_ has joined IRC (freedomrun_!~quassel@89.142.161.121)
11:17freedomrun has left IRC (freedomrun!~quassel@BSN-142-163-238.dial-up.dsl.siol.net, Ping timeout: 244 seconds)
11:33VectorX has left IRC (VectorX!~knight@unaffiliated/vectorx, *.net *.split)
11:33Mava has left IRC (Mava!~Mava@ip-45-224.dhcp.opintanner.fi, *.net *.split)
11:33||cw has left IRC (||cw!~chris@phpgroupware/cw, *.net *.split)
11:33alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
11:34VectorX has joined IRC (VectorX!~knight@unaffiliated/vectorx)
11:34Mava has joined IRC (Mava!~Mava@ip-45-224.dhcp.opintanner.fi)
11:34||cw has joined IRC (||cw!~chris@phpgroupware/cw)
11:36artista_frustrad has joined IRC (artista_frustrad!~fernando@200.247.43.2)
11:36sviesusisalus has left IRC (sviesusisalus!~sviesusis@46.251.62.244, Read error: Connection reset by peer)
11:36sviesusisalus has joined IRC (sviesusisalus!~sviesusis@46.251.62.244)
11:51sviesusisalus has left IRC (sviesusisalus!~sviesusis@46.251.62.244, Read error: Connection reset by peer)
11:51sviesusisalus has joined IRC (sviesusisalus!~sviesusis@46.251.62.244)
11:53Q-FUNK has joined IRC (Q-FUNK!~q-funk@ubuntu/member/q-funk)
11:53
<Q-FUNK>
howdy!
11:54
am I the only one for whom NBD upgrade from Lucid to Precise fails and breaks LTSP in Ubuntu?
11:59
<alkisg>
Not sure how many people tested upgrading...
12:00* alkisg did a clean precise installation, everything worked ok
12:00
<alkisg>
Error messages?
12:00
<Q-FUNK>
bug 924614
12:01
and 924867
12:02
925310 will also be an issue here
12:03VectorX has left IRC (VectorX!~knight@unaffiliated/vectorx, Ping timeout: 240 seconds)
12:07khildin has joined IRC (khildin!~khildin@ip-80-236-193-104.dsl.scarlet.be)
12:07
<Q-FUNK>
alkisg: if there's any info missing, please let me know.
12:08
<alkisg>
Q-FUNK: err I won't be going through those bugs soon, I just thought it would be something small to help with...
12:09
OK if there are filed bugs the troubleshooting + fixing can continue there
12:09
<Q-FUNK>
alkisg: yes, but time is running short before Precise is released.
12:10
<alkisg>
Indeed, it would be nice to have them fixed in the precise cycle
12:11
<Q-FUNK>
alkisg: not just nice to have. if these are not fixed, lots of schools upgrading from LTS to LTS+1 will experience serious breakage.
12:11
<alkisg>
Locally here we're using a PPA because we need a customized ltsp anyway...
12:12
We're upgrading in Septembers, with lots of months of testing after the LTS releases
12:12
<Q-FUNK>
alkisg: as such, either they get fixed, or Ubuntu will have to be honest enough and inform the user base that LTSP has essentially become unsupported.
12:12
<alkisg>
But yeah for everyone else that won't help :)
12:12
Q-FUNK: I'm not the correct person to tell all these though :)
12:12
I'm just a volunteer here
12:13
<Q-FUNK>
ok :)
12:13
is anyone else taking care of LTSP there asides from stgraber nowadays?
12:14
<alkisg>
vagrantc maintains the debian version of ltsp
12:14
<Q-FUNK>
yes but doesn't Ubuntu maintain its own fork?
12:14
<alkisg>
Not much
12:14
LDM is completely in sync
12:14
LTSP is a bit forked, but it will converge in the future too
12:15
<Q-FUNK>
ok
12:15
<alkisg>
Others here help too, but not with a distro maintainer status. Except for cyberorg which isn't much around lately
12:15
Ah and knipwim does a great job for gentoo
12:16
<Q-FUNK>
ok
12:17
well, I would gladly help at Ubuntu, but since the packages are in main, my powers are limited.
12:17
<alkisg>
Q-FUNK: if you can provide patches, it'd be easy to push them
12:18
<Q-FUNK>
alkisg: most of what I would propose as fixes would involve reverting recent changes.
12:18
<alkisg>
Such as?
12:18
<Q-FUNK>
the PAE kernel change
12:18* alkisg didn't read or test that one...
12:19
<alkisg>
But I think stgraber plans today to be an ltsp bug fixing day
12:19
<Q-FUNK>
ltsp-build-client now pulls generic-pae instead of -generic.
12:19
<alkisg>
Maybe it would help if you were around later on
12:20
<Q-FUNK>
according to changelog, it should fall back to -pae only if -generic is not available and yet it currently pulls -pae regardless.
12:20
sure.
12:20
when does stgraber usually return?
12:21
<alkisg>
Not sure... in 3-4 hours?
12:21
<Q-FUNK>
ok
12:40
<muppis>
How /etc/udev/rules.d/70-persistent-net.rules -file is/was handled in nfs mounted ltsp?
12:41
<alkisg>
Since udev doesn't start while running ltsp-build-client, that file should be empty, right?
12:42
So without looking at the code, a rw bind-mount should be enough...
12:44
<muppis>
But while udev starts at boot it write to that file if it founds a new nic during boot causing connectiong error if booting multiple nfs clients.
12:44
<alkisg>
Are you using ltsp?
12:45
<muppis>
Not really. :) Just trying to figure out how to solve that.
12:45
<alkisg>
RW bind mounts make the writes to go to the client RAM
12:46
It's not just one file, there's a long list of files and dirs that need to be writeable
12:47
Don't try to boot multiple clients from the same read/write location, whether it is nfs, nbd, real disk or whatever
12:47Parker955_Away is now known as Parker955
12:49
<muppis>
So I should really mount root somewhere else and bind it to / ?
12:49
<alkisg>
Yes, use some copy-on-write mechanism, tmpfs, nbd rw, and aufs to bind your read-only disk with the writeable location together
12:49
Essentially if you spend some months dealing with all the problems, you'll reimplement ltsp
12:50
<muppis>
:)
12:53Parker955 is now known as Parker955_Away
12:54khildin has left IRC (khildin!~khildin@ip-80-236-193-104.dsl.scarlet.be, Quit: I'm gone, bye bye)
12:54khildin has joined IRC (khildin!~khildin@ip-80-236-193-104.dsl.scarlet.be)
13:40mgariepy has joined IRC (mgariepy!mgariepy@ubuntu/member/mgariepy)
13:41brunolambert has joined IRC (brunolambert!bruno@nat/revolutionlinux/x-xaxusyqvzuxulntm)
14:01dobber has left IRC (dobber!~dobber@213.169.45.222, Read error: Connection reset by peer)
14:02dobber has joined IRC (dobber!~dobber@213.169.45.222)
14:07xsl has joined IRC (xsl!~silence@unaffiliated/xsl)
14:09
<xsl>
hi all
14:10
<alkisg>
Hello
14:11Gremble has joined IRC (Gremble!~Ben@cpc10-aztw24-2-0-cust114.aztw.cable.virginmedia.com)
14:31sviesusisalus has left IRC (sviesusisalus!~sviesusis@46.251.62.244, Read error: Connection reset by peer)
14:31Gremble has left IRC (Gremble!~Ben@cpc10-aztw24-2-0-cust114.aztw.cable.virginmedia.com, Quit: I Leave)
14:31sviesusisalus has joined IRC (sviesusisalus!~sviesusis@46.251.62.244)
14:32Q-FUNK has left IRC (Q-FUNK!~q-funk@ubuntu/member/q-funk, Quit: Leaving.)
14:32Gremble has joined IRC (Gremble!~Ben@cpc10-aztw24-2-0-cust114.aztw.cable.virginmedia.com)
14:40Gremble has left IRC (Gremble!~Ben@cpc10-aztw24-2-0-cust114.aztw.cable.virginmedia.com, Quit: I Leave)
14:57dgroos has joined IRC (dgroos!~dgroos@s199-177.mpls.k12.mn.us)
15:02Q-FUNK has joined IRC (Q-FUNK!~q-funk@ubuntu/member/q-funk)
15:03Q-FUNK has left IRC (Q-FUNK!~q-funk@ubuntu/member/q-funk, Quit: Leaving.)
15:06
<stgraber>
ok, fixing that pae issue ... well, trying to find a way to fix it
15:12laurei has joined IRC (laurei!3a069039@gateway/web/freenode/ip.58.6.144.57)
15:13
<laurei>
wow found something funny, the autologin feature doesnt actually allow a person to log out, because it just keeps logging back in, is this a feature?
15:13
<stgraber>
yes, it's what you want on kiosk clients
15:14
otherwise, you most likely want either the guestlogin feature or you want to use the timeout
15:15
<laurei>
and there is no way to circumvent this otherwise?
15:16
the idea is to have fat clients that operate as DLNA renderers until someone decides to use it as a PC
15:17
timeout sounds like it could work
15:22
yes, yes, timeout is just what i needed thank you
15:22
stgraber.
15:23
<stgraber>
np
15:35khildin has joined IRC (khildin!~khildin@ip-80-236-193-104.dsl.scarlet.be)
15:42killermike has joined IRC (killermike!~killermik@2.26.113.218)
15:44alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
15:50dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174)
16:15staffencasa has joined IRC (staffencasa!~staffenca@128-193-150-248.oregonstate.edu)
16:25ogra_ has left IRC (ogra_!~ogra@p5098ed03.dip0.t-ipconnect.de, Read error: Operation timed out)
16:26dgroos_ has joined IRC (dgroos_!~dgroos@x-134-84-69-128.uofm-secure.wireless.umn.edu)
16:28ogra_ has joined IRC (ogra_!~ogra@p5098ed03.dip0.t-ipconnect.de)
16:29Q-FUNK has joined IRC (Q-FUNK!~q-funk@ubuntu/member/q-funk)
16:29dgroos has left IRC (dgroos!~dgroos@s199-177.mpls.k12.mn.us, Ping timeout: 256 seconds)
16:29dgroos_ is now known as dgroos
16:30Steve_The_Pirate has left IRC (Steve_The_Pirate!~Gary@ip-80-238-8-128.bskyb.com, Read error: Operation timed out)
16:31darkpixel_ has joined IRC (darkpixel_!~darkpixel@65.100.44.217)
16:31darkpixel_ has joined IRC (darkpixel_!~darkpixel@curetheitch/staff/darkpixel)
16:32Steve_The_Pirate has joined IRC (Steve_The_Pirate!~Gary@ip-80-238-8-128.bskyb.com)
16:33loather has joined IRC (loather!~khudson@wsip-98-175-250-115.sd.sd.cox.net)
16:42freedomrun_ has left IRC (freedomrun_!~quassel@89.142.161.121, Read error: Connection reset by peer)
16:47dgroos has left IRC (dgroos!~dgroos@x-134-84-69-128.uofm-secure.wireless.umn.edu, Quit: dgroos)
16:48[GuS] has joined IRC ([GuS]!~MysT@unaffiliated/gus/x-663402)
16:49dobber has left IRC (dobber!~dobber@213.169.45.222, Remote host closed the connection)
16:49
<[GuS]>
Hi guys, I am having a problem building LTSP client (latest LTSP) in Gentoo: http://pastebin.com/ZCuyxb8h
16:52loather has left IRC (loather!~khudson@wsip-98-175-250-115.sd.sd.cox.net, Quit: This computer has gone to sleep)
16:54
<knipwim>
[GuS]: building with 5.2.18 i see :)
16:54
<Q-FUNK>
re
16:56Steve_The_Pirate has left IRC (Steve_The_Pirate!~Gary@ip-80-238-8-128.bskyb.com, Quit: Leaving)
16:56
<[GuS]>
knipwim: indeed
16:56
knipwim: 5.2.19 fails in other part
16:56
<knipwim>
i saw your bug
16:56
<[GuS]>
Yep...
16:57
knipwim: i am modifying the profile... maybe i can re emerge that package tehre...
16:57
there8
16:57
*
16:57Trixboxer has joined IRC (Trixboxer!~Trixboxer@115.124.115.71)
16:57
<knipwim>
the profile in the ltsp overlay?
16:57
<[GuS]>
knipwim: of quickstart
16:58
<knipwim>
that won't work until 5.2.20
16:58
check
16:58
if you emerge it before the packages it emerge, it will work
16:58
<[GuS]>
knipwim: is what i am doing :)
16:59sviesusisalus has left IRC (sviesusisalus!~sviesusis@46.251.62.244, Ping timeout: 248 seconds)
16:59
<knipwim>
it's a gentoo issue btw: http://forums.gentoo.org/viewtopic-t-197754.html
17:00sviesusisalus has joined IRC (sviesusisalus!~sviesusis@46.251.62.244)
17:00
<knipwim>
it's the first time i see it though
17:00
<[GuS]>
knipwim: yes... noticed..
17:01
<Q-FUNK>
stgraber: is today your ltsp bug squashing day, by any chance? :)
17:01
<knipwim>
[GuS]: i will look at your bug later this week
17:02
<[GuS]>
knipwim: oks, no problem. I will try to do this quick fix meanwhile
17:02
for the 5.2.18..
17:02
<stgraber>
Q-FUNK: not really, though I fixed the PAE stuff you mentioned earlier
17:02
well, fixed in ltsp-trunk, so will be there in the next release
17:03
<Q-FUNK>
stgraber: next upload, you mean? on time for precise?
17:04
stgraber: my main concern is that NBD reconfiguration completely failed on a Lucid->Precise upgrade. Instead of getting LDM, I get dropped into BusyBox.
17:06
<stgraber>
Q-FUNK: what did you upgrade exactly? (chroot upgrades have never been supported, so the only way to "upgrade" is by upgrading the server, removing /opt/ltsp and running ltsp-build-client)
17:07
<Q-FUNK>
stgraber: the host, silly :)
17:07
the chroot was obviously regenerated
17:08
<stgraber>
and you ran: "/etc/init.d/nbd-server restart" after you did the ltsp-build-client?
17:09
<Q-FUNK>
I rebted the host
17:09
<stgraber>
k, that should have done the trick :)
17:09
<Q-FUNK>
Bug 924867
17:14
the nbd daemon's config fails to get upgraded
17:20
that's why it didn't do the trick.
17:20
and that was even before the switch to PAE kernel.
17:26
bug #924867 and #924614 actually.
17:30
this happens at the end of whatever ltsp-build-client does.
17:31
<stgraber>
yeah, that's like update-nbd-config (or whatever its name is) that's failing. Luckily we no longer need it on Precise, so I'll just make the change to drop it
17:37hfranco has joined IRC (hfranco!~Adium@206.169.71.12)
17:38
<Q-FUNK>
stgraber: do we still need nbd at all though? how would you handle the config upgrade?
17:38
stgraber: and is that what makes the client drop into BusyBox rather than show me LDM?
17:40
without any upgrade, I get this:
17:40
Restarting Network Block Device server: Stopping Network Block Device server: nbd-server.
17:40
** Message: No configured exports; quitting.
17:40
nbd-server.
17:41
stgraber: so we probably need something like update-nbd-config after all.
17:46
<stgraber>
we no longer need the script because we don't have to touch the config anymore
17:46
the current version comes with a config.d kind of thing
17:47
so we can just dump a file in there without ever modifying the existing conffile
17:47
<[GuS]>
knipwim: question, in which step of the quickstart you recomend me to put the fix for the XML-Parser?
17:49
<knipwim>
[GuS]: i would say in the pre_install_extra_packages()
17:49
before emerging udev
17:49
<[GuS]>
Ok
17:49
knipwim: i just spawn the emerge of that package?
17:50
<knipwim>
yeah, like the udev
17:50
spawn_chroot "emerge XML-Parser"
17:50
<[GuS]>
knipwim: ok, i did that.. and failed
17:51
mmm let me see...
17:51
<knipwim>
what, you added that line, and did an ltsp-build-client?
17:52
<[GuS]>
knipwim: nothing, waitinf there... but i think by mistake i've added the -a param to emerge :P
17:52
<knipwim>
nice one :)
17:52
<[GuS]>
haha yeah...
17:52* [GuS] damn me!
17:53
<[GuS]>
I have to sleep more hours nex ttime..
17:56
<knipwim>
looking at the bug now
17:56
i don't have a clue yet
17:58
you have used a recent stage3 file?
18:00
<Q-FUNK>
stgraber: ok, but how do we handle upgrades from the previous LTS? right now, the server fails to start.
18:01
stgraber: the config that's in /etc/nbd-server/conf.d is precisely what fails to be configurated
18:01
stgraber: or actually, no customized config got dropped there at all. the failed upgrade concerns the default config.
18:03
<[GuS]>
knipwim: yep
18:04
knipwim: not exactly today, but yes
18:08Steve_The_Pirate has joined IRC (Steve_The_Pirate!~Gary@cpc16-haye17-2-0-cust169.haye.cable.virginmedia.com)
18:10
<xsl>
is there a recomended partitioning scheme for LTSP servers?
18:10
or just standard best practices
18:14
<||cw>
xsl: depends on use case
18:15
if clients are going to login to on X on the ltsp server, treating like a desktop is probably best
18:16
also, if it's a testing system and the client images are going to updated often make sure to have tons of free space on the location the images are saved, as clients are not swapped to the new image until they are rebooted, so the old image inodes remain "active" but hidden, eating free space
18:17
<xsl>
||cw: yes that happened to me some times
18:17
do we have a command to delete those "inodes"
18:17
<||cw>
I had 5 images active at once one time
18:17
"what do you mean out of space? I have 10GB!"
18:17
<xsl>
hehe
18:17
<||cw>
well, that crashes the ndb instance
18:18
more or less forcing the clients to reboot if anyone tries to do anyting that isn't in ram
18:21
<xsl>
yes
18:27
<[GuS]>
knipwim: ok, that "fix" failed anyways, still the XML-Parser error.
18:29alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
18:40xsl has left IRC (xsl!~silence@unaffiliated/xsl, Quit: Connection reset by fear)
18:41vagrantc has joined IRC (vagrantc!~vagrant@c-76-115-60-19.hsd1.or.comcast.net)
18:47
<alkisg>
Hey vagrantc, if you got any time for epoptes today, tell me to upload the changelog for the new version
18:52alexqwesa has left IRC (alexqwesa!~alex@alexo-veto.broker.freenet6.net, Ping timeout: 245 seconds)
18:54sviesusisalus has left IRC (sviesusisalus!~sviesusis@46.251.62.244, Quit: Leaving)
18:55freedomrun has joined IRC (freedomrun!~quassel@BSN-61-23-20.dial-up.dsl.siol.net)
18:56alexqwesa has joined IRC (alexqwesa!~alex@alexo-veto.broker.freenet6.net)
18:57
<vagrantc>
alkisg: yeah, probably in about an hour i'll be ready
18:57
<alkisg>
Perfect
19:01[GuS] has left IRC ([GuS]!~MysT@unaffiliated/gus/x-663402, Remote host closed the connection)
19:02brunolambert has left IRC (brunolambert!bruno@nat/revolutionlinux/x-xaxusyqvzuxulntm, *.net *.split)
19:02mgariepy has left IRC (mgariepy!mgariepy@ubuntu/member/mgariepy, *.net *.split)
19:02Hyperbyte has left IRC (Hyperbyte!jan@middelkoop.cc, *.net *.split)
19:02Hyperbyte has joined IRC (Hyperbyte!jan@middelkoop.cc)
19:02brunolambert has joined IRC (brunolambert!bruno@nat/revolutionlinux/x-gqldoahwyeigdynp)
19:02mgariepy has joined IRC (mgariepy!mgariepy@ubuntu/member/mgariepy)
19:03Q-FUNK has left IRC (Q-FUNK!~q-funk@ubuntu/member/q-funk, Read error: Operation timed out)
19:04
<alkisg>
Suppose we want to support xvnc4viewer, ssvnc and vinagre. And suppose that the user has all of them installed. Which one would we use? Should we check /etc/alternatives/vncviewer? Or offer an option in /etc/default/epoptes or in ~/.config/epoptes?
19:34rthomson has joined IRC (rthomson!~rthomson@mars.pet.ubc.ca)
19:53Parker955_Away is now known as Parker955
19:55
<vagrantc>
alkisg: eesh.
19:55
<alkisg>
Yup?
19:55
<vagrantc>
alkisg: /etc/alternatives/vncviewer might support all sorts of vnc clients which may or may not have the features we need...
19:56
<alkisg>
Of course I'll check and only use the supported ones, else I'll fall back to e.g. xvnc4viewer
19:56
<vagrantc>
alkisg: how will youu check ... check the symlink and what it points to?
19:57
alkisg: guess that makes sense ... default to whatever /etc/alternatives/vncviewer uses as long as it's supported, else fall back to a "sane" default.
19:58
alkisg: /etc/default/ is mostly for /etc/init.d scripts ...
19:58
alkisg: although maybe that's changed ...
19:58
alkisg: but yeah, overriding the default in /etc and ~/.config/epoptes makes sense.
19:58
alkisg: /etc/epoptes/epoptes.conf ?
19:59
<alkisg>
vagrantc: err we don't have that currently
19:59
We only have /etc/default/epoptes, and ~/.config/epoptes/*
19:59
<vagrantc>
stgraber: how's LTSP testing going?
19:59
alkisg: oh.
19:59
alkisg: ok, go with what we already have then :)
20:00
alkisg: what reads /etc/default/epoptes ?
20:00
<alkisg>
vagrantc, btw, I'm thinking of putting the LTSP_CLIENT_MAC to LTSP so that we can slowly get rid of it in epoptes
20:00
vagrantc: the epoptes daemon and the epoptes UI
20:00
<highvoltage>
hey vagrantc
20:00
<alkisg>
So, I was about to export it from ltsp_nbd, but then I thought about nfs ... :D Where should I put it?
20:00
I want to write it to ltsp_config from the initramfs,
20:01
<highvoltage>
vagrantc: I have two ltsp related packages that need some review, would you have a few minutes some time to look at it?
20:01
<alkisg>
and export it later to the user session from an ldm/rc.d/X* script
20:01
<stgraber>
vagrantc: found a few issues, though I then got involved into netcfg debugging ;) will try to do some more LTSP tonight
20:01
<highvoltage>
alkisg: is that for the guest users on live ltsp?
20:02
<alkisg>
highvoltage: nope... if you want that, I can have it ready in half an hour
20:02
It's just a small daemon in python + an RCFILE_01 line in lts.conf
20:02
No chroot modiciations necessary (to dynamically create users based on client hostname)
20:02
<highvoltage>
alkisg: hehe, no rush :)
20:03
<alkisg>
highvoltage: I mean, the last time you didn't answer me if you wanted it at all, or not...
20:03
<highvoltage>
alkisg: oh sorry, I thought I did
20:03
<alkisg>
I guess we need stgraber's opinion too, when he has some time after the netcfg thing :)
20:04
Here's the daemon part: http://docs.python.org/library/socketserver.html#socketserver-tcpserver-example
20:04
<highvoltage>
alkisg: yep, that's basically what I said. I think we should run it by stgraber, I know he's *very* careful of what goes into LTSP for in 12.05 since it needs to be supportable for 5 years
20:04
<alkisg>
plus an adduser call
20:04
An the RCFILE_01 line just calls nc with the hostname as a parameter
20:04
<highvoltage>
alkisg: but imo I think it should get in there at some point even if it's for 12.10
20:05Mava has left IRC (Mava!~Mava@ip-45-224.dhcp.opintanner.fi, Quit: leaving)
20:05
<stgraber>
are we talking about running a daemon by default which lets anyone remotely creating an account on the server?
20:05
if so, that's a definite no
20:05
<alkisg>
stgraber: when the user click to run ltsp-live
20:05
<stgraber>
ok, if only for ltsp-live, then maybe we can talk ;)
20:05
(considering we already don't care about any kind of security at this point)
20:07
so what's the benefit of this exactly? it only sounds like to be we'd be saving a few hundred accounts that don't really take space anyway, no?
20:07
<alkisg>
stgraber: it takes a long time to create 255 accounts
20:07
highvoltage: remember how much time that takes?
20:07
<stgraber>
2s-ish (here)
20:07
<vagrantc>
alkisg: as far as where to put the LTSP_CLIENT_MAC stuff ... in the initramfs-scripts hooks, no?
20:07
<stgraber>
I get ltsp-live fully running in less than 20s
20:07
<alkisg>
vagrantc: yes, but ltsp_nbd would be nbd-specific...
20:07
<vagrantc>
highvoltage: URLs for review?
20:08
alkisg: right, use the initramfs-scripts stuff we wrote in october- i'll be using that in the next debian upload and it works with either NBD or NFS
20:08
<stgraber>
alkisg: I believe I limited it to only the number of accounts we really need (based on netrange)
20:08
<highvoltage>
vagrantc: http://mentors.debian.net/package/ltsp-cluster-agent
20:08
<alkisg>
stgraber: Ah, I thought it needed more than 10 secs. Also it helps if the netmask is not 255.255.255.0, where the 255 accounts wouldn't be useful, but that's not a big deal
20:08
<vagrantc>
highvoltage: oh-ho! ltsp-cluster
20:08
<highvoltage>
vagrantc: http://mentors.debian.net/package/ltsp-cluster-agent-weblive
20:09
<stgraber>
alkisg: the netmask is 255.255.255.0 but dnsmasq is configured to only give 50 IPs or something like that
20:09
<highvoltage>
vagrantc: isn't it about time some of this makes it into debian? :)
20:09
<alkisg>
stgraber: I mean, the client hostname would be ltsp123456 and the account would be ltsp123
20:09
<stgraber>
alkisg: and since 11.10 we only create these 50 accounts instead of 255 as we used to
20:09
<alkisg>
Sounds like there's no problem to solve then :)
20:09
<vagrantc>
highvoltage: suppose so. i've never used it or tried it
20:10
<highvoltage>
vagrantc: ah ok. the weblive stuff is really nice? have you seen it?
20:11
<vagrantc>
highvoltage: nope
20:11
<highvoltage>
you need a java plugin to use it (free java like iced-tea works fine) then you can access a remote desktop session from within your browser: http://edubuntu.org/weblive
20:12
<alkisg>
stgraber: that initramfs-scripts stuff won't make it for precise, right? I'd like to expose LTSP_CLIENT_MAC in the user's environment, to use it from epoptes, can I put it somewhere in the initramfs script today and have it included in precise?
20:12
<highvoltage>
stgraber hosts the edubuntu weblive images on his server.
20:12
but it would also be nice to allow people to create debian instances
20:12
<vagrantc>
the initramfs-scripts isn't going into precise?
20:13
<alkisg>
No idea, last I heard about them was Gadi saying that he wanted them to be optional in case stgraber didn't want to include them
20:13
Let's hear what Stephane will say :)
20:13
<vagrantc>
in my next debian packaging, they'll pretty much be mandatory ...
20:13
<alkisg>
(because of little testing they received so far)
20:14
<vagrantc>
alkisg: so what do i need to do to test the vnc issues on amd64 you mentioned yesterday? i'm just getting started on projects today and making a TODO list, apparently...
20:14
<stgraber>
alkisg: in theory the initramfs-scripts stuff is going to be in Precise, if people help testing it (see my mail to -devel)
20:15
<alkisg>
Nice! vagrantc, run `xvncviewer -listen` in an amd64 installation, and `x11vnc -connect amd64pc` from any other pc
20:16khildin has left IRC (khildin!~khildin@ip-80-236-193-104.dsl.scarlet.be, Quit: I'm gone, bye bye)
20:16
<alkisg>
If that crashes, then try to include that patch to debian:
20:16
https://bugs.launchpad.net/ubuntu/+source/vnc4/+bug/845855
20:17
<vagrantc>
alkisg: ok.
20:19Trixboxer has left IRC (Trixboxer!~Trixboxer@115.124.115.71, Quit: "Achievement is not the end, its the beginning of new journey !!!")
20:34
<vagrantc>
alkisg: can the same machine with xvncviewer -list also be the one to run x11vnc -connect ?
20:35
alkisg: if so, it works fine on debian testing
20:36
alkisg: and has the same version in sid
20:37
<alkisg>
vagrantc: and dpkg --print-architectur == amd64?
20:37
Let me try if it fails on the same pc in precise without the patch...
20:38
<vagrantc>
alkisg: yup.
20:39
alkisg: xvnc4viewer 4.1.1+X4.3.0-37 and x11vnc 0.9.13-1
20:39adrianorg__ has joined IRC (adrianorg__!~adrianorg@201.22.230.238)
20:39
<alkisg>
That's strange... maybe the gcc stack protection isn't activated somehow
20:39
And the bug is there, and doesn't appear so easily
20:40* alkisg reboots to precise to better test things...
20:42adrianorg_ has left IRC (adrianorg_!~adrianorg@177.18.168.15, Ping timeout: 252 seconds)
20:43alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
20:47alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
20:48
<vagrantc>
alkisg: so, it should be fine to connect from the same machine, yes?
20:48
<alkisg>
vagrantc: I believe so, but let me try it... once I find an unpatched version of the package..
20:58
vagrantc: yes, the unpatched package crashed for me here even locally. Let me try the debian package in my xubuntu amd64 vm... sid?
21:00
Yup worked fine here too
21:00
<vagrantc>
alkisg: maybe http://bugs.debian.org/509949 fixed it?
21:00
alkisg: oh, that was ancient, nevermind
21:01
xvnc4 hasn't seen changes since march of 2010
21:01
<alkisg>
Maybe a compilation option, we could compare sources to pinpoint it, but I don't know if it's worth it now that it's fixed in precise too
21:02freedomrun has left IRC (freedomrun!~quassel@BSN-61-23-20.dial-up.dsl.siol.net, Remote host closed the connection)
21:03
<vagrantc>
ubuntu defaults to using hardening flags, maybe?
21:03
<alkisg>
vagrantc: currently I'm using dpkg-query to find the epoptes-client version... I'd like to avoid that to be more cross-distro. So I'd e.g. put VERSION=xxx inside the epoptes-client source. But how would I update it automatically, based on the changelog? From inside debian/rules? From a bzr scrript?
21:04
<vagrantc>
alkisg: oh, the dpkg-query call was the wrong call anyways
21:04
or at least, it was at one point
21:04
<alkisg>
Really? I didn't notice... and we base our compatibility test on that
21:04
<vagrantc>
(i.e. it returned the version of a previously installed package if it was no longer installed)
21:05Q-FUNK has joined IRC (Q-FUNK!~q-funk@host-94-101-11-3.igua.fi)
21:05Q-FUNK has joined IRC (Q-FUNK!~q-funk@ubuntu/member/q-funk)
21:06
<vagrantc>
alkisg: dunno how you want to handle the releases.
21:06
<alkisg>
Currently we don't have anything automated
21:06
So any suggestions are most welcome
21:06
<vagrantc>
i've been rolling my own simplified version of mkdst
21:07
<alkisg>
So I could just write a small script and run that manually before releases? Sounds easy...
21:08
<vagrantc>
i could patch my changes into mkdst, and then we could use mkdst again
21:09
but what i've been doing is parsing the upstream version out of debian/changelog, exporting a bzr repo to annother dir, dumping the changelog from bzr into that dir, and making a tarball of it.
21:10
alkisg: but yes, a VERSION file in /usr/share/epoptes and/or /usr/share/epoptes-client would be simpler.
21:11alexqwesa has left IRC (alexqwesa!~alex@alexo-veto.broker.freenet6.net, Quit: Хана X'ам !!!)
21:11
<vagrantc>
alkisg: you could just ship a version file in bzr, and simply as a release, increment that.
21:12
<alkisg>
Epoptes does have a version number in __init__.py that we already have to (manually) sync with debian/changelog
21:12
I think it's customary in python to do that
21:12
<vagrantc>
alkisg: should probably use mkdst, and just fix it to support some of the things i prefer.
21:13
<alkisg>
So maybe I could just put VERSION=xxx inside epoptes-client, and make a small script that updates those files from debian/changelog
21:13
<vagrantc>
alkisg: well, if you've got that already, why not have it pull that version during build time?
21:13
<alkisg>
How?
21:13
<vagrantc>
alkisg: surely there's a way to query the version
21:14
<alkisg>
No I mean what build-time-hook could I use to do that?
21:14
Call a script from debian/rules?
21:14
<Q-FUNK>
stgraber: I finally got the NBD server running again after purging it, deleting the old configs and re-installing it, but this definitely isn't the proper way to upgrade from LTS to LTS+1.
21:15* vagrantc gently pats NFS on the head
21:15
<alkisg>
Hehe
21:15
<vagrantc>
alkisg: probably just parse it from the debian/changelog in debian/rules
21:15* vagrantc rumages around for code
21:16
<Q-FUNK>
yeah, you cannot shake the filesystem elderlies too much.
21:16
<alkisg>
vagrantc: that would be the best
21:16
<vagrantc>
alkisg: DEB_VERSION := $(shell dpkg-parsechangelog | egrep '^Version:' | cut -f 2 -d ' ')
21:16
alkisg: probably add a | cut -d - -f 1
21:16Parker955 is now known as Parker955_Away
21:17
<vagrantc>
alkisg: to get the upstream version
21:17
<alkisg>
Maybe make a script and call: shell update-my-changelog ?
21:17
<vagrantc>
alkisg: and then during build, echo that to a file.
21:17
alkisg: oh, you want debian/changelog to get auto-updated?
21:17
<alkisg>
No, to have epoptes/__init__.py auto-updated from debian/changelog
21:17
And epoptes-client/epoptes-client too
21:18
<vagrantc>
alkisg: but that really should be updated in bzr, not in debian/rules.
21:18
<alkisg>
So, a bzr push hook?
21:18
<vagrantc>
alkisg: i just don't see what's the evil in manually updating that.
21:18
<alkisg>
Or just a script that I run manually?
21:19
<Q-FUNK>
stgraber: after this, things worked again once I configured gnome-fallback as the session in lts.conf
21:19
<vagrantc>
alkisg: and then having a build-time check to make sure it's consistant.
21:19
<alkisg>
OK, that sounds easiest :D
21:19
<vagrantc>
alkisg: i.e. fail to build
21:19
<Q-FUNK>
stgraber: the only bug remaining is that ntpdate still cannot fetch time from the LTSP server and so the CMOS batteryless terminal remains stuck at an old date. that bug remains since Hardy.
21:19
<alkisg>
Can I leave that last part to you? Or at least for a later day, it's getting late here...
21:19
<vagrantc>
alkisg: we've done fine without it so far :)
21:20
<alkisg>
Hehe... ok so should we upload what we have now? There are some fixes that would increase stability...
21:20
(the already committed ones, I don't have anything in my current TODO list )
21:22
<vagrantc>
alkisg: should be able to do that today
21:22
<alkisg>
Nice, ok, I'm uploading the changelog...
21:22
Thanks!
21:22
<vagrantc>
ltsp itself keeps get putting off till later...
21:23
<alkisg>
Tomorrow I'll try to find some time to test ltsp
21:23
...and commit that LTSP_CLIENT_MAC thing
21:35artista_frustrad has left IRC (artista_frustrad!~fernando@200.247.43.2, Quit: Leaving)
21:42
<alkisg>
'night all :)
21:42alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
21:42
<Q-FUNK>
vagrantc: do we still need the three inetd.conf stanzas for NBD in the current LTSP or do we specify all that in NBD's own configs?
21:44
<vagrantc>
Q-FUNK: i use and debian defaults to NFS, so you're asking the wrong person.
21:44
i haven't tested an NBD setup in quite some time
21:45
<Q-FUNK>
ok
21:45Mava has joined IRC (Mava!~Mava@ip-45-224.dhcp.opintanner.fi)
21:49
<||cw>
Q-FUNK: you could always brute-force it with an rcscript
21:50
<Q-FUNK>
||cw: force what?
21:57adrianorg_ has joined IRC (adrianorg_!~adrianorg@187.113.251.14)
21:58adrianorg__ has left IRC (adrianorg__!~adrianorg@201.22.230.238, Read error: Connection reset by peer)
21:59Steve_The_Pirate has left IRC (Steve_The_Pirate!~Gary@cpc16-haye17-2-0-cust169.haye.cable.virginmedia.com, Quit: Leaving)
22:08
<||cw>
ntp
22:28
<Q-FUNK>
||cw: I'm not sure I follow you.
22:30klausade has joined IRC (klausade!~klaus@84.215.157.180)
22:57
<||cw>
does ntpdate not work at all, or just not at boot?
22:57brunolambert has left IRC (brunolambert!bruno@nat/revolutionlinux/x-gqldoahwyeigdynp, Quit: brunolambert)
22:59hughessd has joined IRC (hughessd!~hughessd@173-164-117-109-Oregon.hfc.comcastbusiness.net)
23:05
<Q-FUNK>
||cw: not at boot, at least. wouldn't know if it works afterwards, since it's started in the chroot image.
23:13adrianorg_ has left IRC (adrianorg_!~adrianorg@187.113.251.14, Ping timeout: 252 seconds)
23:16Q-FUNK has left IRC (Q-FUNK!~q-funk@ubuntu/member/q-funk, Quit: Leaving.)
23:26adrianorg_ has joined IRC (adrianorg_!~adrianorg@186.215.21.123)