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


Channel log from 28 May 2012   (all times are UTC)

00:06uskerine has left IRC (uskerine!~uske@237.Red-95-122-134.staticIP.rima-tde.net, )
00:42Lns has left IRC (Lns!~Lns@pdpc/supporter/professional/lns, Quit: Leaving)
01:40risca_ has joined IRC (risca_!~risca@81-233-43-131-no18.tbcn.telia.com)
01:41risca has left IRC (risca!~risca@81-233-43-131-no18.tbcn.telia.com, Ping timeout: 265 seconds)
01:56Phantomas has left IRC (Phantomas!~Phantomas@unaffiliated/phantomas)
02:06F-GT has joined IRC (F-GT!~phantom@ppp121-44-97-140.lns20.syd6.internode.on.net)
02:48Parker955 is now known as Parker955_Away
03:02adrianorg_ has left IRC (adrianorg_!~adrianorg@187.58.154.153, Ping timeout: 256 seconds)
05:15bauerski has joined IRC (bauerski!~witekb@frodo.psp.opole.pl)
05:41bauerski has left IRC (bauerski!~witekb@frodo.psp.opole.pl, Quit: Leaving.)
05:41bauerski has joined IRC (bauerski!~witekb@frodo.psp.opole.pl)
06:52alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
07:03Phantomas has joined IRC (Phantomas!~Phantomas@unaffiliated/phantomas)
07:32risca_ has left IRC (risca_!~risca@81-233-43-131-no18.tbcn.telia.com, Quit: Lämnar)
07:36highvoltage has left IRC (highvoltage!~highvolta@ubuntu/member/highvoltage, Read error: Connection reset by peer)
07:38highvoltage has joined IRC (highvoltage!~highvolta@ubuntu/member/highvoltage)
07:41Phantomas has left IRC (Phantomas!~Phantomas@unaffiliated/phantomas, Quit: Leaving.)
08:28garymc has joined IRC (garymc!~chatzilla@81.138.225.164)
08:29
<garymc>
Hi guys i need some serious help
08:29
I have forgot or for some reason my root password has changed
08:29
I cant add any new users :S
08:36
<muppis>
Boot it in recovery mode.
08:36dobber has joined IRC (dobber!~dobber@213.169.45.222)
08:37
<muppis>
If you don't have sudo for some user. Then you can use sudo su or sudo bash to gain root shell.
08:46risca has joined IRC (risca!~risca@81-233-43-131-no18.tbcn.telia.com)
08:48mealstrom has joined IRC (mealstrom!~mealstrom@46.63.71.254)
09:02garymc has left IRC (garymc!~chatzilla@81.138.225.164, Ping timeout: 244 seconds)
09:12Gremble has joined IRC (Gremble!~Ben@cpc10-aztw24-2-0-cust114.aztw.cable.virginmedia.com)
09:19garymc has joined IRC (garymc!~chatzilla@81.138.225.164)
09:21garymc has left IRC (garymc!~chatzilla@81.138.225.164, Client Quit)
09:27risca has left IRC (risca!~risca@81-233-43-131-no18.tbcn.telia.com, Quit: Lämnar)
10:02komunista has joined IRC (komunista!~slavko@adsl-195-168-237-046.dynamic.nextra.sk)
10:25Trixboxer has joined IRC (Trixboxer!~Trixboxer@115.124.115.71)
10:43
<alkisg>
Hyperbyte: I can't find a tool to convert man pages to mediawiki format, could you do something with the html output that man supports, i.e. to put them outside the wiki?
10:48
E.g.: man --html=firefox doc/ltsp-update-kernels.8 (you need to have the groff package installed for that to work)
10:50alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
11:10mealstrom has left IRC (mealstrom!~mealstrom@46.63.71.254)
11:10mealstrom has joined IRC (mealstrom!~mealstrom@46.63.71.254)
11:10
<mealstrom>
hi
11:17sysadmin has joined IRC (sysadmin!~sysadmin@237.Red-95-122-134.staticIP.rima-tde.net)
11:17
<sysadmin>
hi
11:17sysadmin is now known as Guest96020
11:17
<Guest96020>
i have set up a LTSP but it is terrible slow for 1920x1080 resolutions
11:18
i guess i should move chrome as a local application (fat client)
11:18
and potentially also libre-office too
11:18
can those applications be moved to local?
11:18
any recommended reading or website on how to do it?
11:19
i am using dns proxy (dnsmasq) so the thin clients have internet access through the gateway
11:20
also, is there any possibility to make things faster? like using rdesktop instead of ssh?
11:20Guest96020 is now known as uskerine
11:21
<uskerine>
any help is really appreciated
11:27
<mealstrom>
are you using thin or fat client ?
11:27
<uskerine>
right now thin client
11:27
but my "thin clients" are 1Gb Ram with intel atom, they can do fat
11:27
<mealstrom>
any videocard there ?
11:28
<uskerine>
intel g950
11:29
<mealstrom>
you can try to work with local_apps_menu and local_apps_menu_items at lts.conf
11:29
to lunch some programs localy
11:30
<uskerine>
users can only use google chrome and libreoffice
11:30
plus a sip phone i did not install yet
11:30
but i have read that for fat clients, the whole image needs to be built as fat
11:30
<mealstrom>
look at this variable: LOCAL_APPS_WHITELIST
11:30
<uskerine>
what would you recommend?
11:31
<mealstrom>
if you try lastone -- users will be able to run only whitelisted application and they would run on client
11:32
<uskerine>
seems easier
11:32
and i guess what it is making things slow is actually chrome
11:33
https://help.ubuntu.com/community/UbuntuLTSP/GetMoreFromLocalApps#A10.04
11:34
i am running xubuntu 12.04
11:34
any specific web page for that release?
11:37
https://wiki.ubuntu.com/LTSPLocalAppSetup will that one work?
11:40alexqwesa_ has left IRC (alexqwesa_!~alex@109.172.12.47, Ping timeout: 252 seconds)
11:42alexqwesa_ has joined IRC (alexqwesa_!~alex@109.172.12.47)
11:49Parker955_Away is now known as Parker955
11:54Parker955 is now known as Parker955_Away
12:03
<mealstrom>
!unity2d
12:03
<ltsp>
mealstrom: Error: "unity2d" is not a valid command.
12:03
<mealstrom>
!help
12:03
<ltsp>
mealstrom: (help [<plugin>] [<command>]) -- This command gives a useful description of what <command> does. <plugin> is only necessary if the command is in more than one plugin.
12:04
<mealstrom>
!help unity
12:04
<ltsp>
mealstrom: Error: There is no command "unity".
12:04
<mealstrom>
:(
12:08adrianorg_ has joined IRC (adrianorg_!~adrianorg@187.58.154.153)
12:09[GuS] has joined IRC ([GuS]!~MysT@unaffiliated/gus/x-663402)
12:09
<mealstrom>
how to set on ubuntu 12.04 thin client to use gnome2d default session?
12:10
gnome-session --session=ubuntu-2d?
12:14uskerine has left IRC (uskerine!~sysadmin@237.Red-95-122-134.staticIP.rima-tde.net, Quit: Leaving)
12:22uskerine has joined IRC (uskerine!~sysadmin@237.Red-95-122-134.staticIP.rima-tde.net)
12:23
<mealstrom>
!LDM_SESSION="
12:23
<ltsp>
mealstrom: Error: "LDM_SESSION="" is not a valid command.
12:23
<mealstrom>
!LDM_SESSION
12:23
<ltsp>
mealstrom: Error: "LDM_SESSION" is not a valid command.
12:23
<mealstrom>
!help
12:23
<ltsp>
mealstrom: (help [<plugin>] [<command>]) -- This command gives a useful description of what <command> does. <plugin> is only necessary if the command is in more than one plugin.
12:32
<mealstrom>
!ubuntu-2d
12:32
<ltsp>
mealstrom: ubuntu-2d: To select ubuntu-2d as your default session, put this line in your lts.conf: LDM_SESSION="gnome-session --session=ubuntu-2d"
12:32
<mealstrom>
yes!
12:34litlebuda has joined IRC (litlebuda!~litlebuda@204.0.166.178.rev.vodafone.pt)
12:36lefteris_nik has joined IRC (lefteris_nik!968c0ee3@gateway/web/freenode/ip.150.140.14.227)
12:40
<mealstrom>
what i need to install on client or server 12.04 ubuntu to be able to login with session ubuntu-2d ?
12:53
<uskerine>
why /opt/ltsp/i386/etc/network/interfaces is empty?
12:53
I mean, there is no eth0 definition, only local loopback
12:54
i have installed google chrome as local app but I can not browse
12:57mgariepy has joined IRC (mgariepy!mgariepy@ubuntu/member/mgariepy)
12:57bauerski has left IRC (bauerski!~witekb@frodo.psp.opole.pl, Quit: Leaving.)
13:11uskerine has left IRC (uskerine!~sysadmin@237.Red-95-122-134.staticIP.rima-tde.net, Quit: Leaving)
13:11Phantomas has joined IRC (Phantomas!~Phantomas@unaffiliated/phantomas)
13:16litlebuda has left IRC (litlebuda!~litlebuda@204.0.166.178.rev.vodafone.pt, Remote host closed the connection)
13:17bengoa has joined IRC (bengoa!~bengoa@2001:1291:229:2:216:cbff:feab:6cc9)
13:23uskerine has joined IRC (uskerine!~sysadmin@237.Red-95-122-134.staticIP.rima-tde.net)
13:24
<mealstrom>
!FSTAB
13:24
<ltsp>
mealstrom: Error: "FSTAB" is not a valid command.
13:24
<mealstrom>
!FSTAB_01
13:24
<ltsp>
mealstrom: Error: "FSTAB_01" is not a valid command.
13:28
<mealstrom>
does FSTAB_[09] requires CONFIGURE_FSTAB=YES ?
13:31msignor has joined IRC (msignor!~matt@24.115.67.19.res-cmts.sesp.ptd.net)
13:39Phantomas has left IRC (Phantomas!~Phantomas@unaffiliated/phantomas, Ping timeout: 244 seconds)
13:47rhorstkoetter has joined IRC (rhorstkoetter!~rhorstkoe@opensuse/member/rhorstkoetter)
13:49
<rhorstkoetter>
hi. I'm looking for an idea what could go wrong with my setup. I installed an LTSP server on top of ubuntu 12.04, configured network, built the client environment and it boots up, awesome.
13:49
but
13:49
after login in with ldm I get a terribly distorted screen with both a KDE and razor-qt session
13:50
may you provide any hints why this is/could be?
13:50
cleint is PXE booting atom n450 with integrated intel graphics
13:51
s/cleint/client
13:53Gremble has left IRC (Gremble!~Ben@cpc10-aztw24-2-0-cust114.aztw.cable.virginmedia.com, Quit: I Leave)
13:54Phantomas has joined IRC (Phantomas!~Phantomas@unaffiliated/phantomas)
14:03Phantomas has left IRC (Phantomas!~Phantomas@unaffiliated/phantomas, Ping timeout: 260 seconds)
14:13alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
14:18Phantomas has joined IRC (Phantomas!~Phantomas@unaffiliated/phantomas)
14:24
<rhorstkoetter>
I think I'm getting somewhere. Haven't searched for the right search terms obviously and managed to dive deeper into the client configuration with lts.conf. I'll get back to you if it still doesn't work. thanks for your time. see you
14:25
<alkisg>
rhorstkoetter: a screenshot might help
14:26
<rhorstkoetter>
alkisg: from the client machine?
14:26
<alkisg>
Yes, to see what distrortion are you talking about
14:26
<rhorstkoetter>
not sure how to achieve this. it's hardly useable at all
14:27
<alkisg>
A modile phone maybe?
14:27
<rhorstkoetter>
don't have one with a camera (honestly)
14:27
<alkisg>
rhorstkoetter: is that only _after_ ldm login?
14:27
<rhorstkoetter>
I may describe the distortion quite well though
14:27
yes, after ldm. ldm displays perfectly
14:28
<alkisg>
Then it might be related to composition, search for "disable kde composition" on the net
14:28
Please describe it as well :)
14:28
<rhorstkoetter>
lol. I try
14:29
the screen basically is upside down, purple, green, yellow, what not. it's too small. the KDE menu is mirrored upsaide down and displays very fancy charcters (greek?)
14:30
<alkisg>
Are you using nvidia proprietary drives on the server?
14:30
<rhorstkoetter>
I do
14:30
<alkisg>
That's one reason why the problem might exist. Also, try this in lts.conf: X_SMART_COLOR_DEPTH=False
14:31
Do you know how to create an lts.conf?
14:31
sudo kate /var/lib/tftpboot/ltsp/i386/lts.conf
14:31
[Default]
14:31
X_SMART_COLOR_DEPTH=False
14:31[GuS] has left IRC ([GuS]!~MysT@unaffiliated/gus/x-663402, Quit: Konversation terminated!)
14:31
<rhorstkoetter>
mine is at /opt/ltsp/i386/etc
14:31
<alkisg>
...this might give you an upside-down screen with no other problems, and you may need to completely uninstall the nvidia drivers in order to fix the upside-down part
14:31
No, that's the chroot, lts.conf goes in tftp
14:32
Do it as I wrote above
14:32
(kdesu better than sudo though)
14:32
<rhorstkoetter>
ok, I see. I guess I need to read WAY more docu to actually get behind all LTSP magic
14:33
<alkisg>
Yeah I hope we can simplify it a lot more in the future
14:33
<rhorstkoetter>
I'll use nano/vim anyway
14:33
<alkisg>
But we can't make anything for the nvidia libs problems, they overwrite the normal ones
14:33
*can't do
14:33
<rhorstkoetter>
I understand that. binary bomb
14:35
I just did a quick test install of ubuntu 12.04 (don't want to mess my primary) and ubuntu installer pulled proprietary nvidia in
14:35
<alkisg>
The installer? Really? You didn't get a prompt _after_ the installation?
14:35
<rhorstkoetter>
nope. I was surprised as well
14:35
I checked 3rd party and that's it
14:35
no jockey
14:36
opened jockey after fresh install and nvidia current has been activated by default
14:37
to be percise, it has been kubuntu 12.04 installer
14:37
if that makes any difference
14:37
maybe kubuntu decided to do so while ubuntu does not by default
14:38
I don't know
14:38
<alkisg>
It might, haven't tried any nvidia installations recently though
14:39
<rhorstkoetter>
me neither. the machine I installed the test setup on is my old desktop machine
14:39
good to mess around with ;)
14:40
<alkisg>
mealstrom: it's FSTAB_1, not 01
14:41
<mealstrom>
alkisg: thanks.
14:41
<rhorstkoetter>
alkisg: to sum up (I'm eager to learn something) … the problem is most likely a server one due to nvidia binary drivers, not a client side one
14:42
<alkisg>
rhorstkoetter: a server one (nvidia drivers) and a KDE one (kwin not working in 16 bit color modes)
14:42
Possibly...
14:42lefteris_nik has left IRC (lefteris_nik!968c0ee3@gateway/web/freenode/ip.150.140.14.227, Quit: Page closed)
14:42
<rhorstkoetter>
that would mean switching to nouveau drivers may also solve the problem
14:42
<alkisg>
Half of it
14:42
If I guess right
14:43
<rhorstkoetter>
ah, you know what? I'm using kwin in razorqt as well
14:43
that would support your guess
14:43
will try that later on. unfortunately my thin client is away for a moment (gf has been going out stealing my thin client)
14:44
.P
14:44
:p
14:45
<alkisg>
virtualbox is your friend :)
14:46
<rhorstkoetter>
the thin client will return in a few :)
14:46
I just disabled 3d effects on the server side
14:46
as a first step
14:47
<alkisg>
It's a per-user setting, should make no difference unless you use that user to login on the client (and of course logoff from the server first)
14:49
<rhorstkoetter>
I only have one user on the server so far and I try to login with that particular user in ldm
14:49
certainly logged out locally from KDE
14:49
locally=server side
14:50[GuS] has joined IRC ([GuS]!~MysT@213-117-16-190.fibertel.com.ar)
14:50[GuS] has joined IRC ([GuS]!~MysT@unaffiliated/gus/x-663402)
14:51
<rhorstkoetter>
or is there anything else to truly disable compositing, instead of disabling desktop effects? for a quick test I mean
14:51andygraybeal has joined IRC (andygraybeal!~andy@173.245.135.246)
14:53andygraybeal_ has joined IRC (andygraybeal_!~andy.gray@obsidian.casanueva.com)
14:55
<rhorstkoetter>
alkisg: man … you're awesome!!!
14:55
disabled kwin desktop effects and it works
14:55
<alkisg>
(05:51:25 μμ) rhorstkoetter: or is there anything else to truly disable compositing, instead of disabling desktop effects? for a quick test I mean ==> google it, I don't use KDE
14:56
Did you remove the nvidia drivers as well?
14:56
<rhorstkoetter>
nope
14:56
<alkisg>
Nice, and did you specify X_SMART_COLOR_DEPTH?
14:56
<rhorstkoetter>
nothing
14:56
<alkisg>
Nicer :)
14:56
<rhorstkoetter>
just disabled 3d effects in kde settings and that's it
14:57
<alkisg>
There's probably some way to do it system-wide
14:57
If so, mention it here, to create a factoid for it
14:57
Like this one:
14:57
!compiz
14:57
<ltsp>
alkisg: compiz: the default window manager in gnome is gnome-wm, which automatically chooses compiz if it thinks that the card supports it. Compiz is causing login problems to some clients (LP #673072). To disable it, see !disable_compiz. To restore it, see !restore_compiz
14:57uskerine has left IRC (uskerine!~sysadmin@237.Red-95-122-134.staticIP.rima-tde.net, Remote host closed the connection)
14:58
<rhorstkoetter>
I'll do some more research and share my insights
14:58
I mean within this IRC channel
14:59
first hiccups aside … I love to be logged in for the very first time to LTSP … love it
15:00
now, I'll try to understand what specific I may do with this setup
15:00* rhorstkoetter smells a lot of tinkering :p
15:00
<alkisg>
If you're using a new laptop as an ltsp client, you might also be interested in ltsp fat clients
15:00
!fatclients
15:00
<ltsp>
alkisg: fatclients: You may find some info about the Ubuntu/LTSP implementation of fat clients at https://help.ubuntu.com/community/UbuntuLTSP/FatClients
15:00
<alkisg>
Then 3d effects would also probably work
15:03
<rhorstkoetter>
I read about fat clients before. in fact I'm planning to install a bigger lab in some local school
15:03
and currently I try to do my very first steps (so to spek baby steps) with LTSP
15:04
just an old desktop as a server and a n450 netbook as a client machine
15:04
next step is to try how that beast scales
15:05
and my impression so far (with the on client now working with a proper graphics display) .. man, this is awesome stuff. I'm really excited
15:05
<alkisg>
How many clients do you expect to have in the school?
15:05
Also,
15:05
!ldm_directx
15:05
<ltsp>
alkisg: Error: "ldm_directx" is not a valid command.
15:05
<alkisg>
!directx
15:05
<ltsp>
alkisg: directx: If you want to turn off encryption to the workstation, which will speed up performance at the cost of security, you need to specify LDM_DIRECTX=True in your lts.conf file. Type !docs in IRC to get a link to the current documentation
15:06
<rhorstkoetter>
not sure yet. this is all theory so far. I also read about ltsp-cluster
15:06
you see .. I'm at 3% knowledge maybe
15:06
and there's a lot to learn still
15:07uskerine has joined IRC (uskerine!~sysadmin@237.Red-95-122-134.staticIP.rima-tde.net)
15:07
<uskerine>
hi
15:10
where is the lightdm configuration (theme) applied in LTSP?
15:10
is it possible to limit the preferences menu so only language can be choosen?
15:11
<alkisg>
uskerine: lightdm isn't used in LTSP. LDM is used instead.
15:11
<uskerine>
i thought it was the same!
15:11
:!
15:11
so
15:11
<alkisg>
No, they're completely different products
15:11
<uskerine>
where can i configure LDM theme
15:11andygraybeal has left IRC (andygraybeal!~andy@173.245.135.246, Quit: Ex-Chat)
15:12
<alkisg>
Have a look at /opt/ltsp/i386/usr/share/ldm/themes/
15:12
<uskerine>
and how could I limit preferences menu so only language and shutdown can be choosen? -to avoid the user choosing session-
15:12
ok i will do it
15:12
<alkisg>
The preferences menu can't be easily disabled, you'd need to edit its code
15:12
For the user session, check LDM_SESSION in lts.conf manpage
15:12
<rhorstkoetter>
alkisg: let me thanks you for your patient and kind support. I'll try to dive deeper into the LTSP docs now and I'll get back to you if I find out something new about kwin's global compositing conf and share that information to the channel for a factoid
15:13
have a good day
15:13
<alkisg>
You too
15:15
<rhorstkoetter>
thanks
15:15rhorstkoetter has left IRC (rhorstkoetter!~rhorstkoe@opensuse/member/rhorstkoetter, Quit: Leaving.)
15:19
<uskerine>
alkisg should the alkisg, thanks, i see that it will be easy to change the look & feel in that directory, should I modify the one that it is currently being used or should I add a new one and point to it somewhere in other config file
15:19
currently it is clearly being used
15:19
"/opt/ltsp/i386/usr/share/ldm/themes/ubuntu
15:19
"/opt/ltsp/i386/usr/share/ldm/themes/xubuntu
15:19
<alkisg>
Best to modify it. There's a xubuntu ldm theme available.
15:19
LDM_THEME=xxx in lts.conf if I remember well
15:20
<uskerine>
so i can add my own theme
15:20
and then point using LDM_THEME in tls.conf
15:20
lts.conf
15:20
?
15:20
<alkisg>
Yes
15:20
<uskerine>
thanks
15:20
what does "session" under preferences menu exactly means
15:20
?
15:20
what happens if user select other than "Default"
15:24
<alkisg>
...I'm not sure I understand, you want to disable it without knowing what it does?
15:24
It's the same as in all display managers, it selects the x session
15:38dobber has left IRC (dobber!~dobber@213.169.45.222, Remote host closed the connection)
15:39vagrantc has joined IRC (vagrantc!~vagrant@c-76-115-60-19.hsd1.or.comcast.net)
15:39vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc)
15:39
<uskerine>
hi alkisg, i want to ensure that they only have one session
15:39F-GT has left IRC (F-GT!~phantom@ppp121-44-97-140.lns20.syd6.internode.on.net, Ping timeout: 252 seconds)
15:40
<alkisg>
uskerine: hi, I answered that above
15:40
<uskerine>
alkisg, you mentioned LDM_SESSION, if i put there a value am i forcing the user to always use that one?
15:40
<alkisg>
Yes
15:40
<uskerine>
ok thanks a lot
15:40
i will try it
15:41* alkisg waves to vagrantc... I committed an example lts.conf, if there's anything you don't like about it we can comment it out or change it
15:41
<uskerine>
btw alkisg, i finally included xrandr commands to fix up the resolution as I99 script under xinitrc.d
15:41
<alkisg>
uskerine: which ubuntu version?
15:41
<uskerine>
xubuntu 12.04
15:41
<alkisg>
You could use XRANDR_COMMAND_0 instead
15:41
<uskerine>
soudns like much more cleaner
15:42
i will fix that up too
15:51alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
15:52
<vagrantc>
alkisg: oh, sure now you leave.
15:59mealstrom has left IRC (mealstrom!~mealstrom@46.63.71.254, Quit: Leaving.)
16:15
<uskerine>
vagrantc, should I worry about " or ' while specifying a command through XRANDR_COMMAND_0?
16:15
xrandr --newmode "1920x1080" 148.5 1920 2008 2052 2200 1080 1089 1095 1125 +hsync +vsync
16:16
<vagrantc>
uskerine: try and find out ...
16:16
if it requires the ", you'll probably have to escape them.
16:17
<uskerine>
it he " are not required, do i have to put everything between "?
16:17
it the " are not required, do i have to put everything between "?
16:17
<vagrantc>
i don't think so, no.
16:17
<uskerine>
i mean like ="xrandr --newmode ....... "
16:17
ok
16:17
thanks i will try it
16:18
<vagrantc>
but there's nothing like experimentation :)
16:19
<uskerine>
i think we both would agree that there are much more interesting stuff to spend time with :)
16:20
<vagrantc>
answering questions that could be resolved through simple testing? dunno.
16:21
<uskerine>
hehe
16:21
nothing is simple here, i have been several days with this stuff :)
16:37shogunx has left IRC (shogunx!~shogunx@rrcs-67-79-182-232.se.biz.rr.com, Remote host closed the connection)
16:41shogunx has joined IRC (shogunx!~shogunx@rrcs-67-79-182-232.se.biz.rr.com)
16:48alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg)
16:57
<alkisg>
15:52<vagrantc>alkisg: oh, sure now you leave. ==> hehe back now
17:04andygraybeal has joined IRC (andygraybeal!~andy@173.245.135.246)
17:04
<vagrantc>
alkisg: actually, since the semi-recent initramfs-tools switch, using NFS will still use lts.conf from tftp if present
17:04
<alkisg>
vagrantc: ah ok please do change whatever comments are inappropriate
17:06
<vagrantc>
alkisg: what's hard is to accurately explain what's acctually going to happen
17:06* alkisg goes on to edit lts.conf.xml with xxe...
17:07
<alkisg>
"man lts.conf for details :P"
17:07
Btw when knipwim has a man page ready in the new ltsp site, we should change the link to ubuntu's manpage
17:07
(or Hyperbyte)
17:08
<vagrantc>
essentially, it will download from tftp (which may be /srv/tftp/... /var/lib/tftpboot/... or /tftpboot/...) if the file is present, and use that, otherwise, it'll read it from /etc/lts.conf in NFS ...
17:08
<alkisg>
vagrantc: are you shipping nbd-exclude.conf by default?
17:08
<vagrantc>
alkisg: not yet
17:08
alkisg: should i?
17:08
<alkisg>
OK, it's supposed to be there though, I removed the /proc checks
17:08
Yup
17:09
It'll make the image smaller and it's a bit safer about proc being mounted etc
17:09
And for ltsp-update-image -c /, it also adds some security
17:09
<vagrantc>
alkisg: also, ltsp-update-image breaks NBD images created with create-fs-image ... not sure what to do about that.
17:09
well, it will overwrite them
17:10
<alkisg>
Err why? If it doesn't find /opt/ltsp/<arch>/bin/true, does it continue?
17:10
(and why would one run ltsp-update-image if he uses loopback images in the first place?)
17:11
<vagrantc>
because the internet is full of documentation that implies the source of all your problems is that you haven't run ltsp-update-image
17:11
<alkisg>
Haha
17:11
OK, but still, it shouldn't overwrite them if a chroot isn't present...
17:11
<vagrantc>
finally fixed that issue for NFS more or less... but now other NBD use-cases still have the issue.
17:11
alkisg: and if a chroot is present?
17:11
<alkisg>
Well then it's supposed to overwrite them and keep a backup
17:12
One should use different names in this case
17:12
How would you expect to handle /opt/ltsp/i386 and /opt/ltsp/images/i386.img if they are _different_ installations?
17:12
<vagrantc>
ok, so to update an ext2 image, i manually mount the image to /opt/ltsp/i386 ... it's what makes sense after all ...
17:13
but they're the same installation, it just shouldn't be obliterated by ltsp-update-image
17:13
alkisg: what about if the chroot is a mountpoint, bail?
17:14
<alkisg>
So to hit the "problem" a user would need to mount /opt/ltsp/images/i386.img to /opt/ltsp/i386 and then run ltsp-update-image? I don't think anyone would be so dumb to do that, and I don't think he'd find something similar in the net...
17:14
...or am I missing something?
17:15* vagrantc has done it several times already...
17:15
<alkisg>
But sure a "check if the target image is a mountpoint" check wouldn't hurt
17:16
Also the automated backup would help a bit in this case
17:16
<vagrantc>
ltsp-build-client --create-fs-image mounts /opt/ltsp/images/i386.img at /opt/ltsp/i386 ... if that fails to unmount on failure, there you are.
17:16
<alkisg>
ltsp-build-client --create-fs-image shouldn't run ltsp-update-image at all in the end, but yeah the user might run it manually after ltsp-build-client crashes
17:17
<vagrantc>
although if ltsp-build-client crashes, you've debateably got a borked image anyways
17:18
<alkisg>
So ltsp-update-image would rename it to .old, and create a new borked one :P
17:18
OK, if ! grep -qs /opt/ltsp/images/i386.img /proc/mounts ...
17:18
<vagrantc>
why not use: mountpoint /opt/ltsp/FOO ?
17:19
<alkisg>
It's a valid use case to mount /opt/ltsp/images from NFS or .vdi and then run ltsp-update-image
17:19
So we need to check the image, not the dir
17:19
<vagrantc>
alkisg: it shows up as /dev/loopN, not /opt/ltsp/images/i386.img
17:19
<alkisg>
Hmm
17:21
mount does list it
17:21
<vagrantc>
at least today :)
17:23* vagrantc checks a bit more
17:23
<alkisg>
I can't think of a better way to do it... and if mount doesn't list it in some cases, we just don't prevent users from shooting themselves in the foot, it's not that bad...
17:25
<vagrantc>
alkisg: so, check the output of mount for an /opt/ltsp/images/FOO.img and bail?
17:25
<alkisg>
It's the best I can think of now
17:25
<vagrantc>
alkisg: what about checkinng the filesystem type of the image?
17:25
i suppose you want to add support to ltsp-update-image to support writeable filesystems...
17:25
(if you haven't already)
17:25
<alkisg>
The user might have the squashfs image mounted to /opt/ltsp/i386 so...
17:26
Checking the filesystem type shouldn't catch all cases
17:26
<vagrantc>
update a squashfs to the same squashfs?
17:26
<alkisg>
We're still talking about what dumb things users can do, right? :P
17:26
It's not much worse than trying to create a squashfs image from an ext image
17:27
(08:25:38 μμ) vagrantc: i suppose you want to add support to ltsp-update-image to support writeable filesystems... ==> do you mean with rsync?
17:28
Or to create an ext image instead of a squashfs one?
17:28khildin has joined IRC (khildin!~khildin@5ED3EFC1.cm-7-4d.dynamic.ziggo.nl)
17:28
<alkisg>
I thought about it but I had problems calculating the image size
17:28khildin has left IRC (khildin!~khildin@5ED3EFC1.cm-7-4d.dynamic.ziggo.nl, Remote host closed the connection)
17:28
<alkisg>
So I postponed it
17:29
First I'd like to add support for rsync-style updates if the user has manually created a writeable image
17:29
But let's leave that for the next release :)
17:30
Maybe nbd will gain native support for syncing in the future and will save us the turoble
17:30
trouble
17:39
<vagrantc>
ok, off to test the latest and greatest
17:43* vagrantc ponders TODO list
17:54
<muppis>
How I can get remote X via ssh?
18:00
!factoids vnc
18:00
<ltsp>
muppis: Error: The "Factoids" plugin is loaded, but there is no command named "vnc" in it. Try "list Factoids" to see the commands in the "Factoids" plugin.
18:00
<muppis>
18:02
Whoa, I managed it.
18:02andygraybeal has left IRC (andygraybeal!~andy@173.245.135.246, Ping timeout: 245 seconds)
18:03Trixboxer has left IRC (Trixboxer!~Trixboxer@115.124.115.71, Quit: "Achievement is not the end, its the beginning of new journey !!!")
18:17
<vagrantc>
alkisg: nbd-excludes.conf seems might long
18:17
<alkisg>
vagrantc: which ones would you remove?
18:18
<vagrantc>
alkisg: oh, not long in the sense that it should be trimmed, just long as an observation :)
18:18
<alkisg>
Ah, ok... it took me some time to find the most security sensitive paths...
18:18
<vagrantc>
so, ltsp-update-image really just generates images ... they're not inherrantly NBD images, per se.
18:19
so... nbd-excludes.conf ?
18:19
could concievably export those images through another protocol
18:21
what about ltsp-update-image.excludes ?
18:22
or ltsp-update-image-excludes.conf ?
18:22
<alkisg>
Both sound fine, np with the name
18:23
But the format of the file is specific to nbd
18:23
E.g. to reuse it with rsync, it needs a few changes
18:23
<vagrantc>
it isn't specific to mksquashfs?
18:24
it seems specific to the tool used to generate the image, not the protocol over which it's delivered
18:25
<alkisg>
True
18:25
But it can be reused for rsync with only a few automated changes done by a script
18:25
<vagrantc>
ltsp-update-image-mksquashfs-excludes.conf
18:26
that's getting unweildly ...
18:26
<alkisg>
I liked ltsp-update-image.exlcudes better
18:27
<vagrantc>
if we can handle the excludes implementations (rsync, mksquashfs, etc) behind the scenes, yeah, something generic sounds better.
18:27
haven't yet tried ltsp-update-image --export /
18:28andygraybeal has joined IRC (andygraybeal!~andy@173.245.135.246)
18:29
<alkisg>
ltsp-update-image --cleanup /. It works fine for me, although I didn't test it extensively, it produces a fat chroot from the server disk in just 10 minutes, while keeping all the apps installed on the server
18:30
<vagrantc>
"while keeping all the apps instalaled on the server" ?
18:30
<alkisg>
erm, I meant that any apps or /etc settings on the server are exported to the fat chroot without requiring the user to do them again in the chroot
18:31
<vagrantc>
i should give it a whirl soonish
18:31
<alkisg>
vagrantc: I think we have some bug in debian/rules that deletes locapps/missing and nbd-proxy/missing before calling the final `make clean`, which makes debuild error
18:32
If I run autogen.sh or mkdst before debuild, all work fine
18:32
But if I just run debuild, it doesn't work
18:32
<vagrantc>
"works for me"
18:32
<alkisg>
Did you try copying debian/ in the trunk and a plain debuild?
18:32
I'm using your debian/rules, and it doesn't...
18:33
<vagrantc>
hmmm.... what i actually typically do is: debuild -us -uc -S ; sbuild ../ltsp_*.dsc
18:33
haven't tried a straight debuild in a while
18:34
so i build the source package, and then send that through a build machine.
18:34
<alkisg>
It makes it harder to build it on launchpad directly from trunk without downloading/uploading anything
18:35
(with a recipe that uses your packaging tree)
18:35
<vagrantc>
i'll see if i can get a straight debuild to work.
18:35
<alkisg>
So, I'm thinking that I can use the debian ltsp packaging for greek schools without modifications
18:36
<vagrantc>
other than that issue?
18:36
alkisg: the debian ltsp packaging on ubuntu?
18:36
<alkisg>
Yup
18:36
Except maybe for a couple of dependencies like busybox-static
18:37
<vagrantc>
alkisg: could explore using ${foo:Depends} type stuff and detect distro at build time
18:38
<alkisg>
I'll start testing now and send you the minimal diff required
18:45
vagrantc: does it sound sane to put code in debian/rules that does distro detection, and if it's Ubuntu, it copies all the Ubuntu/* dirs over the Debian/* ones? (e.g. init-ltsp.d/Ubuntu)
18:45
<vagrantc>
heh. i didn't even have the build-depends for ltsp installed :)
18:45
alkisg: it could be implemented sanely.
18:46
alkisg: i don't see a fundamental problem with it, at least.
18:46
<alkisg>
Cool, so basically only debian/control is the "problem"..
18:46
<vagrantc>
and i think we could fix that with ${foo:Depends} type stuff
18:47
i've never actually figured out how to do it, but i know it's possible.
18:48
alkisg: alternately, we could generate control from control.in ... but we need to be careful about that one.
18:51
alkisg: debuild worked fine for me.
18:51
alkisg: you sure you have the appropriate build-deps defined? are there bugs on dh-autoreconf?
18:51
<alkisg>
vagrantc: and you didn't run autogen.sh manually before that? ... :(
18:51
<vagrantc>
alkisg: nope.
18:52* vagrantc double-chekcs
18:52
<alkisg>
I do have the build-deps
18:52
vagrantc: verify that localapps/missing is missing before debuild -b -tc
18:52
If it's there, everything works fine
18:53
<vagrantc>
alkisg: i'm running "debuild -us -uc"
18:53
<alkisg>
autogen.sh creates it, but it's deleted again in the clean phase before make clean is executed, so make clean complains that it doesn't find a Makefile
18:53
<vagrantc>
maybe "debuild -b" doesn't work.
18:54
alkisg: what's -tc ?
18:54hughessd has joined IRC (hughessd!~steve@173-164-117-109-Oregon.hfc.comcastbusiness.net)
18:55
<alkisg>
-tc Clean the source tree (using gain-root-command debian/rules clean) after the package has been
18:55
built.
18:56
(that's the part that fails, the cleanup)
18:56monteslu_ has joined IRC (monteslu_!~monteslu@ip68-109-174-213.ph.ph.cox.net)
18:57
<vagrantc>
oh
18:59* vagrantc tries with -tc
18:59monteslu has left IRC (monteslu!~monteslu@ip68-109-174-213.ph.ph.cox.net, Ping timeout: 252 seconds)
19:00
<alkisg>
vagrantc: no pulseaudio in the ltsp-client-core dependencies?
19:00
<vagrantc>
alkisg: it's a dependency of ltsp-client
19:01
<alkisg>
pulseaudio-esound-compat,
19:01
No plain pulseaudio, is that enough?
19:01
<vagrantc>
or, it should be a dependency of ltsp-client
19:01
do we even support esound anymore? :)
19:01
<alkisg>
I think there's some broken code about it somewhere... :P
19:02
<vagrantc>
alkisg: ok, so debuild -b succeeds, but yes, debuild -b -tc fails
19:03
<alkisg>
I don't mind too much about it, but it would make building from trunk + your packaging possible without any downloads/uploads involved
19:04* vagrantc takes a peek at the cleanup rules
19:05
<vagrantc>
seems like dh-autoreconf clean is running in the wrong order
19:06hughessd has left IRC (hughessd!~steve@173-164-117-109-Oregon.hfc.comcastbusiness.net, Quit: hughessd)
19:07
<vagrantc>
hm...
19:07
<alkisg>
So for ltsp-client-core, ubuntu has these additional dependencies: busybox-static, udhcpc, usbutils (I don't mind about that last one)
19:07hughessd has joined IRC (hughessd!~steve@173-164-117-109-Oregon.hfc.comcastbusiness.net)
19:09
<alkisg>
For ltsp-client, there's dbus, some fonts, plymouth, language-pack-en, laptop-detect, numlockx, wget. I don't think I mind for any of those. Cryptsetup would be nice to get in recommends instead of depends though, as it takes some ram...
19:09
<vagrantc>
i haven't yet been convinced of the need for udhcpc ... no idea about usbutils
19:09
<alkisg>
We have code in udhcpc that handles some of the ipappend 3 part and some parts of nbd too
19:09
<vagrantc>
busybox-static would be a change
19:10
alkisg: speaking of cryptsetup ... the lts.conf defaulted to LDM_DIRECTX=True ?
19:11
<alkisg>
Yeah I wanted to talk about that
19:11
Since the default is False, I thought it would be a good compromise to set the default to true if one decides that he needs an lts.conf
19:11
But if you don't like it np
19:12
<vagrantc>
is there a specific reason why ubuntu needs busybox-static?
19:12
(and is LTSP the right place to define that, if it is?)
19:13
<alkisg>
I think ubuntu splits busybox in a different way from debian, let me check again...
19:14
<vagrantc>
debian's ltsp-client dpends on numlockx
19:15
where's wget used?
19:15
<alkisg>
Hmmm yeah busybox-static might not be needed, I'll do a test without it
19:15
<vagrantc>
regarding dbus, fonts, plymouth, language-pack-en ... are those needed for Ubuntu's LTSP?
19:16
<alkisg>
Ah, in ltsp-cluster, so no worries about it
19:16
I think ltsp-cluster should become a separate package
19:17
I don't know why dbus is a dependency
19:17
<vagrantc>
highvoltage was working on getting that stuff into debian, i think, and i've never mustered the time to give it a try :(
19:17
<alkisg>
We don't use it in the ltsp code
19:17
<vagrantc>
(and why it would hard-depend on language-pack-en?)
19:18
<highvoltage>
ah yes I should probably spend some time on that before debian hits freeze
19:18
<alkisg>
language-pack-en depends on language-pack-en-base which has a lot of en_*/MESSAGES/*.mo files, but I wonder why those are needed...
19:19
<vagrantc>
highvoltage: sorry i haven't had a chance to review
19:19
alkisg: especially, say, if you had a non-english steup
19:19
<highvoltage>
vagrantc: and sorry that I haven't had a chance to poke you more :)
19:19
<vagrantc>
main thing is just getting a test server up and running
19:20
don't want to mess with my regular LTSP test environments
19:20
<highvoltage>
the parts I sent you are mostly for weblive though (not full ltsp-cluster)
19:20
<vagrantc>
highvoltage: you have ITPs filed?
19:21
<highvoltage>
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=655358
19:21
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=655359
19:21
<alkisg>
vagrantc: ...so basically I think I only want udhcpc, and other Ubuntu users would want plymouth. The rest differences don't seem very significant.
19:23
plymouth might already be a dependency of some other package
19:24
mountall depends on plymouth
19:24
upstart depends on mountall, so we're ok
19:25
<vagrantc>
wow. plymouth is essentially mandatory on ubuntu now?
19:25
i haven't even tried splash on debian
19:25
at least, not for years ... probably should give it a whirl
19:27
<alkisg>
Wait wait we don't need udhcp either
19:27
Since the udhcp ltsp script does handle ipappend 3 even without it
19:28
:)
19:28* alkisg tests directly with the debian packaging without any changes at all...
19:29
<vagrantc>
exciting!
19:30
weird, installing plymouth didn't cause a splash screen to start, but it did break LDM.
19:30
<highvoltage>
it's funny I've noticed some stuff creeping in on /etc/init on debian recently
19:30
(well, on my laptop at least, /etc/init/anacron.conf, but I've seen some on other machines too)
19:32
<alkisg>
Ah I forgot the Ubuntu/* overrides that may go in debian/rules
19:33
vagrantc: if you have time to do that, and put pulseaudio and ltsp-update-image.excludes, I can do some extensive tests then
19:33
Or I could try to propose a patch for debian/rules...
19:34
<vagrantc>
alkisg: to install the Ubuntu/* stuff?
19:34
<alkisg>
Yeah
19:34
I.e. just copy the Ubuntu/share dirs over the Debian/ ones
19:34
(with distro detection)
19:36
<vagrantc>
alkisg: client/share and server/share ?
19:36
<alkisg>
I believe so, let me check...
19:37
<vagrantc>
what was i even testing just now...
19:38
<alkisg>
Actually just client/Ubuntu, there's no server/Ubuntu
19:38
<vagrantc>
oh yes, the update-kernels thing that makes it run the appropriate script
19:38
<alkisg>
Also put ltsp-config.8
19:39Phantomas has left IRC (Phantomas!~Phantomas@unaffiliated/phantomas)
19:39andygraybeal_ has left IRC (andygraybeal_!~andy.gray@obsidian.casanueva.com, Quit: Ex-Chat)
19:39
<vagrantc>
alkisg: there's all of two scripts in client/Ubuntu ...
19:40
alkisg: wonder if it would be feasible to just install them on Debian at the moment.
19:42litlebuda has joined IRC (litlebuda!~litlebuda@204.0.166.178.rev.vodafone.pt)
19:42
<vagrantc>
alkisg: ones an upstart script, which will only get used if upstart is installed ... and the other is update-motd stuff?
19:42* alkisg checks how the upstart script gets installed...
19:43
<vagrantc>
alkisg: and apparently, the update-motd stuff could just be set to rm -f || true some files ...
19:43
<alkisg>
$ rm -f /nothing; echo $?
19:43
0
19:44
<vagrantc>
sure, thee || true is extra
19:44Steve_The_Pirate has joined IRC (Steve_The_Pirate!~Gary@cpc16-haye17-2-0-cust169.haye.cable.virginmedia.com)
19:44
<alkisg>
So it shouldn't hurt on debian as is
19:45
<vagrantc>
somehow i missed the -f
19:45
so... we could fix the diff upstream
19:45
<alkisg>
cp $(CURDIR)/client/initscripts/ltsp-core $(CURDIR)/debian/ltsp-client-core.init
19:45
<vagrantc>
alkisg: what package in ubuntu hasthe update-motd stuff?
19:45
<alkisg>
I wonder if it would be easier to drop it in the debian/ dir instead
19:46
<vagrantc>
alkisg: that just allows dh_installinit to do it's thing.
19:46
<alkisg>
base-files
19:47
http://packages.ubuntu.com/quantal/i386/base-files/filelist
19:47
Those /etc/motd files don't exist on debian, http://packages.debian.org/sid/i386/base-files/filelist
19:47
<vagrantc>
alkisg: could go through and manually install ltsp-client-core directly. or not ship it upstream ... but i settled on that approach as it allowed to easily rename it appropriately, as well as use the standard initscript installation mechanisms.
19:47
alkisg: yes, i know, that's why i asked where they come from :)
19:48
<alkisg>
vagrantc: I think that if both upstart and sysvinit scripts exist on debian/, then ubuntu prefers the upstart one, while debian would use the sysvinit one, and everyone would be happy... no?
19:48
So you could copy them both from debian/rules, or put them in debian/, your call
19:49
<vagrantc>
alkisg: exactly.
19:49
<alkisg>
Cool
19:49* alkisg is anxious to test all that
19:49
<vagrantc>
and if someone installs upstart on debian, i *think* it should behave appropriately, then.
19:50
presuming you're not doing crazy there.
19:51
probably need the "plymouth quit --retain-splash" bits in my init script to get that to work
19:52* alkisg doesn't care about plymouth at all, it makes 64mb ram clients unbootable
19:52
<vagrantc>
well, i'd like to be able to get it to work if it's installed
19:52
though i don't want it to derail my focus :)
19:53
wonder if i should have default MKFS_OPTS for btrfs
19:53
or no, it's a mount option that i need to compress it, eh?
19:54
alkisg: so far, ltsp-trunk works ok with NFS, NBD-squashfs, and i'm now testing NBD-ext2
19:54
<alkisg>
compression is a mount option, yes
19:55
<vagrantc>
are there mkfs options i should default to?
19:55
<alkisg>
I haven't used any in my btrfs loopback images and they work fine
19:56
http://packages.ubuntu.com/search?keywords=kbd-chooser ==> only available on natty, ltsp-client-builder wouldn't be installable on ubuntu
19:56
I don't care though, that's only used in the alternate cd, right?
19:57
<vagrantc>
alkisg: yes, only used in the alternate CD ... i should see if it works on Debian, too.
19:57[GuS] has left IRC ([GuS]!~MysT@unaffiliated/gus/x-663402, Quit: Konversation terminated!)
19:57
<vagrantc>
it's been a long time since i've tested that
19:58
<alkisg>
Why is it a dependency? What do you use from it?
19:59
Could it be a recommends: instead?
19:59
<vagrantc>
alkisg: it's a way to place what needs to be setup in the installer ... udebs are a little odd.
19:59
alkisg: might be a better way to do that no.
19:59
now
20:00
<alkisg>
OK, I don't think anyone would want to generate an Ubuntu alternate CD from the Debian LTSP package so noone would mind about that...
20:01
<vagrantc>
alkisg: ah, so you're not actually planning on pushing this to ubuntu?
20:01
alkisg: oor you're putting it out in a test deployment first?
20:01
<alkisg>
I plan to upload it to my PPA for greek schools to use it, and for anyone else that wants the ltsp-update-image --cleanup / functionality in 12.04
20:01
But I'm in no position to affect ltsp packaging in Ubuntu, it's stgraber's area :)
20:03
I'll stop testing Ubuntu's LTSP though in the future, as I'll be only using the Debian packaging
20:03
<vagrantc>
alkisg: so /etc/ltsp/nbd-excludes.conf (or ltsp-update-image.excludes) is only used with --cleanup ?
20:03
<alkisg>
No
20:03
It's used by ltsp-update-image every time, and it should be a good thing to do so
20:04
<vagrantc>
what's --cleanup ?
20:04
<alkisg>
It would avoid problems with /proc, remove chroots logs and sensitive data etc
20:04
The --cleanup runs: chroot $chroot /usr/share/ltsp/ltsp-cleanup
20:04
...while mounting the chroot in a cow environment first,
20:04
<vagrantc>
alkisg: i'm getting a lot of errors with //sys/.... creating empty file.
20:04
alkisg: should /sys/* be added to excludes?
20:04
<alkisg>
and that removes users, regenerates ssh keys etc
20:06
vagrantc: cleanup mounts --bind / to a temp dir, so it shouldn't have anything in /sys or /proc
20:06
But yeah /sys should be added too
20:06
(mount --bind doesn't also mount submounts)
20:06
Not /sys, but /sys/*, to retain the dir
20:07
<vagrantc>
right
20:07
<alkisg>
vagrantc: should I commit the sys + ltsp-update-image.excludes changes?
20:07* vagrantc wonders where it mount --bind's /
20:08
<alkisg>
In a $(mktemp -d) dir
20:08
<vagrantc>
alkisg: i can do that, or if you'd rather
20:08
<alkisg>
Sure, please go ahead
20:08* alkisg continues with lts.conf.xml then
20:08
<vagrantc>
hmmmm....
20:08
alkisg: i'm running ltsp-update-image --cleanup / ... and i don't seem any dirs in /tmp
20:09
<alkisg>
vagrantc: do you have /opt/ltsp/i386/usr/share/ltsp/ltsp-cleanup ?
20:09
<vagrantc>
alkisg: nope
20:09
alkisg: oh, i also don't have ltsp-client installed.
20:09
<alkisg>
Ah sorry yeah
20:10
<vagrantc>
hopefully it isn't rm -rf'ing my system :)
20:10
<alkisg>
Not /opt/ltsp/i386, just plain /usr/share/ltsp/ltsp-cleanup
20:10
Haha no I was extra careful to avoid such things
20:10
<vagrantc>
no ltsp-cleanup, no.
20:10
<alkisg>
You need ltsp-client installed in order to run ltsp-update-image --cleanup /
20:10hughessd has left IRC (hughessd!~steve@173-164-117-109-Oregon.hfc.comcastbusiness.net, Quit: hughessd)
20:11
<vagrantc>
this is a bit nervous-making ... but here we go.
20:11
alkisg: ltsp-client-core should be enough, no?
20:12
or ltsp-client-core + ldm
20:12
that's probably why it was trying to hit stuff in /sys
20:13
alkisg: ah, it assumes overlayfs.
20:14
<alkisg>
Or aufs, but they should be loaded from the initramfs once you install ltsp-client(maybe -core)
20:14
<vagrantc>
oh, no, it just spits out warnings...
20:14
<alkisg>
I think I have some code to automatically load them if the user hasn't rebooted yet
20:14
(after installing ltsp-client)
20:14
<vagrantc>
rebooted?
20:15
now i get segfaults ... hmmm...
20:15
<alkisg>
When you install ltsp-client, a new initramfs is generated, and it loads aufs or overlayfs
20:15hughessd has joined IRC (hughessd!~steve@173-164-117-109-Oregon.hfc.comcastbusiness.net)
20:15
<alkisg>
But I have code to load them even if the user hasn't rebooted (which would have them loaded automatically)
20:16
vagrantc: pastebin the output?
20:16
(all of it, to check for warnings..)
20:16
<vagrantc>
not much to go on... but...
20:17
alkisg: paste.debian.net/171632
20:17
after openbsd restarts...
20:17
<alkisg>
Now it should be preparing for mksquashfs
20:17
That takes a few seconds if you have a lot of data in /
20:18
ps aux should show mksquashfs running
20:19
Also I'm not sure if I've put a policy-rc.d in my system manually
20:19
<vagrantc>
alkisg: it segfaults and quits
20:20
<alkisg>
Which one? mksquashfs?
20:20
<vagrantc>
oh, my paste was bunk
20:20
<alkisg>
Ah
20:20
Hmm --cleanup should generate a policy-rc.d temporarily while running all the cleanup stuff
20:20
<vagrantc>
alkisg: it was followed by three "Segmentation fault" lines.
20:20
<alkisg>
By which programs?
20:21
<vagrantc>
didn't say
20:21
<alkisg>
Can you try putting a policy-rc.d with exit 101 temporarily?
20:21
In the server's /usr/sbin...
20:22markit has joined IRC (markit!~marco@88-149-177-66.staticnet.ngi.it)
20:22
<vagrantc>
alkisg: same thing.
20:23
<markit>
hi, I'm trying to see vm output through spice, but the guest X process is eating a lot of CPU when I move windows... KDE
20:23
<vagrantc>
alkisg: the segfaults come immediately after the restarting openssh
20:23
<markit>
ops
20:23
wrong channel :))
20:23
hi you all :)
20:24
<alkisg>
vagrantc: putting policy-rc.d to alternatives might be necessary too
20:24
Or even a start-stop-daemon wrapper
20:24
I'll put those to cleanup.d
20:24
Hi markit
20:25
<markit>
alkisg: hi, I want to produce screencasts about ltsp installation, so trying to run kubuntu 12.04 in a vm and grab the output... KVM is driving me crazy :(
20:25
<vagrantc>
alkisg: i may have rm -rf'ed some of this... eeyk.
20:26
<markit>
if you try A, then B is broken... if you try C, then D is broken... if you try E, then B and D are broken
20:26
<alkisg>
vagrantc: hey, careful there! I don't want bad karma for cleanup.d! :)
20:27
<vagrantc>
alkisg: you bind-mounted it read-only, yes?
20:27
<alkisg>
Yes
20:27
<vagrantc>
ok, should be safe, then
20:27
<alkisg>
And a cow/aufs on top of it
20:29
vagrantc: alternatively, to test, you might remove cleanup.d/50-openssh-server
20:29
It would show if the problem is ssh trying to restart without /proc, and segfaulting
20:29
I'll push a policy-rc.d fix asap
20:30
<vagrantc>
alkisg: something borked run_parts function
20:30
alkisg: it should exclude .orig files...
20:31
<alkisg>
vagrantc: I'm using my own function that supports /etc/ overrides
20:31
It doesn't exclude .orig by default
20:31
<vagrantc>
alkisg: you're evil :P
20:32
<alkisg>
Hehe, I'm planning ahead for the future :P
20:32
<vagrantc>
our run_parts or run_parts_list functions should support such things
20:33
i used it in ltsp-build-client recently.
20:33
alkisg: so, the segfaults appear to be on userdel calls
20:34
<alkisg>
Urm
20:34
Any service users from uid=500 to 1000?
20:35litlebuda has left IRC (litlebuda!~litlebuda@204.0.166.178.rev.vodafone.pt, Remote host closed the connection)
20:36
<vagrantc>
alkisg: it's segfaulting on "userdel vagrant" which is what it should be doing ...
20:36
<alkisg>
Why? It's only deleting it on a cow copy...
20:36
It never segfaulted for me
20:37
<vagrantc>
you've tried with aufs?
20:37
<alkisg>
No, just overlayfs
20:37
Let me check aufs
20:37
<vagrantc>
if i run without --cleanup, it doesn't mount the tempdir...
20:38
just goes straight into mksquashfs
20:38
<alkisg>
Yes, the tmpfs/aufs etc stuff is only there for the cleanup phase
20:39
But ltsp-update-image / without --cleanup would cause problems on the clients, as some users would exist in /etc/passwd
20:39
<vagrantc>
so .... service ignores policy-rc.d
20:39
if you want policy-rc.d you have to use invoke-rc.d
20:40
<alkisg>
We're not using `service` anywhere in ltsp-update-image, are we?
20:40
<vagrantc>
dunno.
20:40
just an observation
20:41
<alkisg>
OK so my TODO list for tomorrow is: policy-rc.d wrapper for cleanup.d, and test aufs, and also test the debian packages
20:48
<vagrantc>
i need to figure out how to get debuild -b -tc to work, i guess?
20:50
so, other than ltsp-update-image --cleanup ... ltsp-build-client with NFS, NBD+ext2 and NBD+squashfs are all working fine for me.
20:50* vagrantc will try btrfs for good measure.
20:51
<vagrantc>
oh, i also wanted the kiosk mode to autocreate the kiosk user/dir if not present...
20:52
then the admin only needs to create them if they want to change defaults...
20:53
<alkisg>
(11:48:11 μμ) vagrantc: i need to figure out how to get debuild -b -tc to work, i guess? ==> not really a priority
20:54
The debian/rules part with Ubuntu/* would be first, I guess
20:54
And ltsp-update-image.excludes, and ltsp-chroot.8
20:54
Yeah the kiosk plugin needs some love
20:58
Meh I keep forgetting that you pulled the Ubuntu's init-ltsp.d stuff so the Ubuntu/* part isn't very important now except for the upstart job
20:59
<vagrantc>
alkisg: and even that seems appropriate to move over
20:59
alkisg: ltsp-update-image.excludes implemented.
20:59
<alkisg>
We may just go ahead and delete the client/Ubuntu dir completely
20:59
Cool, ty
21:00
<vagrantc>
alkisg: what about ltsp-chroot.8 ?
21:01
<alkisg>
Just put it in the debian/manpages
21:01
<vagrantc>
ltsp-config.8 ?
21:01
<alkisg>
Or change it to install *.8
21:01
Yes
21:01markit has left IRC (markit!~marco@88-149-177-66.staticnet.ngi.it, )
21:01
<alkisg>
server/doc/*.8
21:01
server/doc/*.1
21:01
etc
21:05
<vagrantc>
went with the *.N approach
21:05
saves repackaging later as we add stuff.
21:05rhorstkoetter has joined IRC (rhorstkoetter!~rhorstkoe@opensuse/member/rhorstkoetter)
21:07
<rhorstkoetter>
hi. I have some questions regarding my general understanding of the client environment
21:08
<muppis>
Hit us.
21:08
<rhorstkoetter>
that is, assumed I'm running an ubuntu 12.04 ltsp server machine … in what regard differs the client environment from that built on a server machine running kubuntu/kde?
21:09
I noticed the default ldm wallpaper/background
21:09
is there anything else?
21:10
I mean, I hope this is understandable, what impact has the actual DE config of the server on the client enviroment being built?
21:10
<vagrantc>
rhorstkoetter: typically you log into the server and run whatever the server is running.
21:11
rhorstkoetter: the login manager is LTSP specific
21:11
<rhorstkoetter>
vagrantc: that is totally obvious
21:11
<vagrantc>
rhorstkoetter: ok, so your question is not understandable, then :P
21:12
<rhorstkoetter>
but it doesn't answer my question about differences in the client environment depending on DE server configuration. I mean, the most visible (maybe, and that's the question, only) change is another wallpaper in ldm depending on server's UI
21:13
<vagrantc>
rhorstkoetter: that's about the difference, since you're running the desktop environment on the server.
21:13
<alkisg>
You mean what differences would go in the chroot if you had gnome instead of kde?
21:13
<rhorstkoetter>
alkisg: exactly
21:13
<alkisg>
It doesn't affect thin clients at all. You can install ldm-kubuntu-theme manually if you want something that looks KDE-ish
21:14
<rhorstkoetter>
alkisg: so the wallpaper is the only change? I mean this is what's visible to me and thus I wondered if there's more
21:15
<alkisg>
Yup
21:15
<rhorstkoetter>
ltsp-client-built does exactly this and that was a bit confusing to me
21:15
ok, thanks for explaining
21:16
just for understanding. I'm eager to learn actually ...
21:17
how the hell does ltsp-client-built decide which wallpaper to choose and why is this?
21:18
according to the default UI. what would it do if I had unity and kde installed?
21:18
<alkisg>
vagrantc: nbd-excludes.conf needs to be renamed too, want me to do it?
21:18
<rhorstkoetter>
I know this is a very minor thing but I'd like to get behind the scenes
21:19
<vagrantc>
alkisg: no, i was just noticing that
21:19
<alkisg>
rhorstkoetter: cat /usr/share/ltsp/plugins/ltsp-build-client/Ubuntu/030-artwork
21:20
<rhorstkoetter>
alkisg: great. I'll look into that file after our conversation (server is off atm) < in the basement w/o WOL :p
21:21
<alkisg>
rhorstkoetter: http://bazaar.launchpad.net/~ltsp-upstream/ltsp/ltsp-trunk/view/head:/server/share/ltsp/plugins/ltsp-build-client/Ubuntu/030-artwork
21:21
<rhorstkoetter>
great. one sec please
21:22* rhorstkoetter looks
21:24
<rhorstkoetter>
ok, that would mean several themes will be installed (one for every -desktop meta-package) available at the server
21:24
<alkisg>
Nope
21:24andygraybeal has left IRC (andygraybeal!~andy@173.245.135.246, Ping timeout: 245 seconds)
21:24
<alkisg>
Just the first one
21:24
There's a "break" there
21:24
So if you had gnome and kde installed, it would install the kde one only
21:25
<vagrantc>
alkisg: regarding the debuild -b -tc issue ... is it possible to work around the recipe by running the appropriate bits?
21:25
<alkisg>
vagrantc: sure, if we e.g. put ./autogen.sh when debian/rules starts...
21:25
<rhorstkoetter>
alkisg: oh certainly, I see
21:25* vagrantc wonders if this is generally a problem with dh-autoreconf stuff that doesn't ship things upstream?
21:26
<vagrantc>
doesn't ship autofoo upstream
21:26
<alkisg>
I don't know how many packages use autoreconf, but I haven't had any problems with debuild -b -tc except for the ltsp package
21:26
<vagrantc>
or if i'm somehow using dh_autoreconf_clean wrong
21:26
<rhorstkoetter>
alkisg: ok, this anyway is a very minor thing but good to understand
21:27
my second question is regarding ldm-sessions
21:28
assumed I build the chroot with just the unity desktop installed at that time and later install e.g. lxde … would ldm provide the new session instantly or is there a need to rebuild the client enviroment
21:28
?
21:29
<alkisg>
You'd just need to reboot the clients
21:29bengoa has left IRC (bengoa!~bengoa@2001:1291:229:2:216:cbff:feab:6cc9, Quit: Leaving.)
21:29
<rhorstkoetter>
I mean apps (FF, chrome, whatever) would certainly be available without reconfiguration/rebuilding of the chroot but what about the sessions available to ldm?
21:29
ah ok
21:29
awesome stuff
21:31
alkisg: btw, after disabling kwin compositing (not sure yet how to do it systemwide) login worked as previously worked out due to your help
21:31andygraybeal has joined IRC (andygraybeal!~andy@173.245.135.246)
21:31
<alkisg>
Cool
21:32
<rhorstkoetter>
I still have minor graphics distortions though clicking on menus
21:32
kind of purple, yellow, green graphics errors
21:32
<vagrantc>
actually, you probably just need to restart ldm
21:32
login/logout
21:33
<rhorstkoetter>
vagrantc: is that an answer to me?
21:33
<vagrantc>
or at least, it used to re-ask at every time it displayed the login
21:33
<alkisg>
rhorstkoetter: try the X_SMART_COLOR_DEPTH=False lts.conf directive
21:33
<vagrantc>
rhorstkoetter: both you and alkisg
21:33
<rhorstkoetter>
vagrantc: ah, you meant the sessions
21:34
<vagrantc>
restarting the thin client will work, of course, but i think you just need to get ldm to restart as it asks about sessions every time
21:34
<rhorstkoetter>
alkisg: ah yes, will try that
21:34
vagrantc: I see
21:34
<vagrantc>
though nothing like experimentation
21:34
<rhorstkoetter>
respectively I'll try it
21:34* rhorstkoetter likes experiments
21:35
<vagrantc>
revenge of X_SMART_COLOR_DEPTH
21:35
<rhorstkoetter>
first of all I'm going to reinstall the whole rig one more time. installing kubuntu was a bad idea (I actually try it once a year or so just to cry "GTK please" after 5 min :p)
21:37
<vagrantc>
alkisg: just as an experiment ... i tried installing a user in a thin client and running "userdel user" on the booted thin client with aufs over NBD and all that fun ...b and it worked fine.
21:37
<rhorstkoetter>
btw, as I'm dealing with ltsp pros here … I installed http://www.epoptes.org/ just out of curiousity. is there any other management software I may find useful? i.e. you find very useful and recommend
21:37
?
21:38
<vagrantc>
oh yeah, i need to retest epoptes again... it's been a while.
21:38
<alkisg>
rhorstkoetter: there's italc, but it had problems so we developed epoptes :)
21:39
vagrantc: are you sure it was the userdel segfaulting and not ssh?
21:39
<vagrantc>
alkisg: yes.
21:39
<alkisg>
k, will test tomorrow
21:39
<rhorstkoetter>
alkisg: I see. I read that there is some sort of web interface for ltsp-cluster. is it possible (recommended) to use that on top of "std ltsp"?
21:39
<vagrantc>
alkisg: put set -x in ltsp-cleanup and that's what was segfaulting ... one for each user.
21:40
<rhorstkoetter>
or is there any advantage over using epoptes?
21:40
<alkisg>
rhorstkoetter: dunno, never used ltsp-cluster
21:40
rhorstkoetter: ltsp-cluster has a completely different goal than epoptes
21:40
<vagrantc>
alkisg: and then it would fail to generate the image.
21:40
alkisg: or rather, it seemed to stop after running ltsp-cleanup
21:40
<alkisg>
It's to manage multiple ltsp servers, while epoptes manages the screens of ltsp clients
21:40
<rhorstkoetter>
alkisg: ok
21:40
<vagrantc>
suppose i could || true it to see what it does.
21:41
<alkisg>
vagrantc: that would be due to set -e, but the real problem is why it was segfaulting, being in a cow etc...
21:41
vagrantc: or you could remove the userdel script from cleanup.d
21:41
<rhorstkoetter>
btw, I don't know if that is desired or not … it seems to be impossible to login to ldm with a user w/o passwd set
21:41
<vagrantc>
alkisg: right, i'd like to see it work without
21:41
<rhorstkoetter>
I figured that out some hours ago
21:42
or better with a blank passwd
21:42
<vagrantc>
alkisg: actually, when i commented out all the userdel lines, it still stopped before creating the image.
21:42
<alkisg>
Maybe due to groupdel?
21:42
<vagrantc>
and after commenting those out too.
21:44
alkisg: adding || true to the userdel/groupdel lines at least moves on to mksquashfs
21:46* vagrantc also would like to implement a copy-nbd-to-ram mode
21:46
<vagrantc>
suppose you could also do it with NFS and rsync
21:47
<alkisg>
Yeah if the file system is to go to RAM, there's no need for nbd-client after the initramfs
21:49
<vagrantc>
yeah, to save ram. :)
21:50
alkisg: the one big thing we lost switching away from nbdrootd was robustness with NBD image reboots
21:50
<alkisg>
How so?
21:50
<vagrantc>
when you restard nbd-server, it kills off all your NBD connections, no?
21:50
i.e. crashes all your thin clients
21:51
whereas before, each client had it's own nbd-server instance
21:51
<stgraber>
that's why we have -persistent and nbd-proxy for ;)
21:51
<alkisg>
vagrantc: ah, I thought you meant client reboots. Yeah, nbd-server restarts crash the clients
21:51
stgraber: -persistent doesn't work currently
21:52
<vagrantc>
though overall, it's much more elegant and seems much more responsive to detecting dead clients.
21:52
<alkisg>
vagrantc: but nbd-server restarts isn't something one does regularly, is it?
21:52* alkisg really disliked the "one port per chroot" part
21:52
<vagrantc>
yes, the named exports are excellent.
21:53
<stgraber>
alkisg: true, it never quite did anyway ;) I hear that nbd-proxy works quite well nowadays but I guess it still uses too much ram for you but might be fine for others
21:54
<alkisg>
The clients would get a different nbd swap anyways, so they'd still crash
21:54
<vagrantc>
it deletes on disconnect?
21:54
<stgraber>
right, you have to disable nbdswap if you don't want your client to die in case of server disconnect
21:54
<alkisg>
...nbd-server restart? It would assign a new swap file then
21:55
<vagrantc>
right
21:55
<stgraber>
no-nbdswap + nbd-proxy with a clustered nbd is what Revolution Linux uses IIRC and that should work quite well (if you have enough memory)
21:55
<alkisg>
Firefox can use up to 500mb ram for pixmaps so that would require very new clients :(
21:56
I'd prefer to fix it on the nbd-server side
21:56
<stgraber>
we'll also get zram turned back on in 12.10 (likely for everyone as there's supposedly no more downside in using it, compared to the mess compcache was doing)
21:57hughessd has left IRC (hughessd!~steve@173-164-117-109-Oregon.hfc.comcastbusiness.net)
21:57
<alkisg>
For 14.04 I'll try to help Wouter properly implement reconnections for both root and swap
21:58
<vagrantc>
alkisg: so, i managed to get a working image using "ltsp-update-image --cleanup /", but it deleted all the kernels from the image's /boot ... so there are no kernels for ltsp-update-image
21:58jammcq has joined IRC (jammcq!~root@74.198.150.94)
21:58
<jammcq>
ping stgraber
21:59
<alkisg>
Hi jammcq :)
21:59
<jammcq>
hey alkisg
21:59
<alkisg>
vagrantc: ...it deleted the kernels from /boot?!
21:59* vagrantc noticed ramzswap is gone from Debian's kernel :(
21:59
<jammcq>
i'm in canada at Scotty's cabin
21:59
and we're beginning our planning of BTS-2012
21:59
<vagrantc>
alkisg: not the server's /boot, but the image's
21:59
<alkisg>
I don't think there's code to delete anything from /boot
21:59
<stgraber>
hey jammcq
22:00
<jammcq>
vagrantc: howdie
22:00
<vagrantc>
alkisg: well, i've got an image without anyhting in /boot :(
22:00* vagrantc waves to jammcq and sbalneav, so so far away
22:00
<alkisg>
Ah, different /boot partition
22:00
<jammcq>
stgraber: any word yet on UDS for the fall?
22:00
<alkisg>
vagrantc: mount --bind doesn't mount submounts
22:00
<vagrantc>
alkisg: that'd do it.
22:01
<stgraber>
jammcq: dates are very likely to be 29th October to the 2nd of November, we just don't know on what side of the atlantic we'll be...
22:02
<jammcq>
ah
22:02
well... if we try to go JUST before UDS, then it'll be a problem if it's in europe
22:03
<vagrantc>
could we just change which sea it's by?
22:03
<stgraber>
fall was usually in north america but we had our spring UDS in California so that kind of messed up everything :)
22:03
<jammcq>
if we go a whole week before or after, then we likely won't get the "uds funding effect"
22:04
vagrantc: not easily cuz many of us drive there
22:06
<vagrantc>
alkisg: what's the nearest sea to you?
22:07
<alkisg>
Adriatic, between Greece and Italy :P
22:07
<jammcq>
stgraber: any idea how soon UDS info will be available?
22:07* alkisg is afraid he won't make it to BTS 1012
22:07
<alkisg>
*2012
22:08
<jammcq>
and is it possible that it's already been decided, but the info just hasn't propagated yet?
22:08
<stgraber>
jammcq: not really, I'd expect us to know by end of July at least, otherwise it becomes a nightmare on Canonical's side to book everything
22:08
<vagrantc>
alkisg: ok, so other than those glitches, "ltsp-update-image --config /" worked for me :)
22:08
<stgraber>
jammcq: I know that 3 weeks ago (past-UDS) I heard of a few options and they were spread on multiple continents...
22:08
<vagrantc>
alkisg: interestingly enough, it actually deleted the users.
22:09
<alkisg>
Meh
22:09
I hope that I can reproduce it here, I haven't ever seen userdel segfaulting so far...
22:09
<vagrantc>
alkisg: so, a mounted /boot and the segfaulting userdel ...
22:10
<alkisg>
And the policy-rc.d wrapper
22:10
<jammcq>
ok, we'll be discussing various dates. we'll probably come up with something tentative, and then make a firm descision when we know more about uds
22:10
<vagrantc>
ah yes, that would be a good addition
22:11jammcq has left IRC (jammcq!~root@74.198.150.94, Quit: leaving)
22:11
<vagrantc>
alkisg: oh, also ssh_known_hosts wasn't created.
22:12
<alkisg>
vagrantc: let me see if the aufs options are wrong and mount the tmpfs as read-only...
22:13Phantomas has joined IRC (Phantomas!~Phantomas@unaffiliated/phantomas)
22:14
<alkisg>
...they seem ok
22:14
<vagrantc>
alkisg: it seems to make other edits just fine.
22:15Phantomas1 has joined IRC (Phantomas1!~Phantomas@unaffiliated/phantomas)
22:17
<alkisg>
Ah we need an ltsp-update-sshkeys manpage, the current one doesn't document any options
22:17F-GT has joined IRC (F-GT!~phantom@ppp121-44-97-140.lns20.syd6.internode.on.net)
22:18
<vagrantc>
alkisg: it's very impressive to see in action, though!
22:18Phantomas has left IRC (Phantomas!~Phantomas@unaffiliated/phantomas, Ping timeout: 250 seconds)
22:18
<vagrantc>
alkisg: good work!
22:18
<alkisg>
vagrantc: I need to add support for ltsp-update-sshkeys, it doesn't exist
22:18
ty! :)
22:19andygraybeal has left IRC (andygraybeal!~andy@173.245.135.246, Ping timeout: 245 seconds)
22:19
<alkisg>
(...in cleanup.d, with BASE=/ and CHROOT=/)
22:20
<rhorstkoetter>
alkisg: you know what? < we talked/wondered about that previously … ubuntu 12.04, just as kubuntu 12.04, installs proprietary nvidia without asking
22:20
it enabled right after a fresh install
22:20
just as a follow-up
22:21
<vagrantc>
through ubuntu-restricted-extras?
22:21
<rhorstkoetter>
jockey shows the driver as installed and activated at first boot after installation
22:21
<alkisg>
Good to know, that might cause problems in some installations..
22:21
<rhorstkoetter>
vagrantc: nope, stock install
22:21
<vagrantc>
that seems like a license violation
22:21
<rhorstkoetter>
vagrantc: alkisg and I were wondering why that happened with kubuntu and ubuntu install does the same
22:23
<alkisg>
vagrantc: it downloads them from the internet if one check the [x] add 3rd party programs
22:24
<vagrantc>
installing "I Agree" licensed stuff without the user ever even having a chance to read the license ... suspicious.
22:25
<rhorstkoetter>
alkisg: most likely it's the 3rd party checkbox, yes. I haven't seen any license agreement though
22:25
not at all
22:26
just install and voila, proprietary driver installed and in use by default
22:26
bummer :p
22:26F-GT has left IRC (F-GT!~phantom@ppp121-44-97-140.lns20.syd6.internode.on.net, Ping timeout: 260 seconds)
22:27Phantomas1 is now known as Phantomas
22:29
<rhorstkoetter>
does nouveau support 3d?
22:29
<vagrantc>
somewhat
22:29
<rhorstkoetter>
it could be a reason why it's enabled by default cause of unity
22:30
maybe devs would want to avoid major headaches of users installing ubuntu/unity on nvidia hardware and first screen that comes up looks somewhat … cause 3d isn't working
22:30
I don't know
22:32
a bit suspicious in any case (if not critical from a license perspective - without any disclaimer I mean). I cannot remember what the text says below "3rd party" option though. maybe that's enough already from a legal perspective in regard of eula
22:32
<vagrantc>
depends on the license
22:33
<rhorstkoetter>
as of the reason, I vote for seamless unity for newcomers
22:34
nothing else would explain that to me, although it still isn't an excuse, just a possible explanation
22:36
while this is off-topic, http://linuxlibrary.org/wp-content/uploads/2011/11/ubuntu-11-10-4.jpg shows relevant section of the installer in 11.10
22:37
gray, I'd say, deep gray ;)
22:38
<vagrantc>
"give blanket permission to install whatever"
22:39komunista has left IRC (komunista!~slavko@adsl-195-168-237-046.dynamic.nextra.sk, Quit: Leaving.)
22:44andygraybeal has joined IRC (andygraybeal!~andy@173.245.135.246)
22:45
<vagrantc>
alkisg: ok, so we're sticking with "ltsp-update-image --config-nbd" to enable nbd configuration if NFS is detected and it calls "ltsp-config nbd-server" ?
22:45
alkisg: that's what appears to happen now ... any designs to change that? :)
22:46Steve_The_Pirate has left IRC (Steve_The_Pirate!~Gary@cpc16-haye17-2-0-cust169.haye.cable.virginmedia.com, Quit: Leaving)
22:46
<alkisg>
vagrantc: isn't it good enough? I can't think of anything more appropriate....
22:47
<vagrantc>
alkisg: it seems to work for me.
22:47
<alkisg>
Let's leave it as is then
22:47
<vagrantc>
alkisg: we've toyed around with that idea so much, just wanted to make sure it was essentially settled :)
22:48
<alkisg>
I wrote the man pages too so I wouldn't like to change it without reason :)
22:48
<vagrantc>
alkisg: i'd say it's pretty much ready for upload
22:48
<alkisg>
vagrantc: did you do the Ubuntu/* stuff?
22:49
<vagrantc>
alkisg: i was thinking for now i'd just add individual files
22:49
<alkisg>
Cool, then I'll be able to test with the debian's .debs directly
22:49
<vagrantc>
presuming libc is compatible.
22:49
and other libs are compatible
22:51
the two Ubuntu files i see seem to be safe ... or should i move them into Debian upstream?
22:51Phantomas has left IRC (Phantomas!~Phantomas@unaffiliated/phantomas, Ping timeout: 252 seconds)
22:52
<alkisg>
If not, you'd need to install the upstart script from debian/rules
22:52
So maybe in debian upstream is better
22:52
<vagrantc>
alkisg: how/where should the upstart script go?
22:53
alkisg: does it need any special handling?
22:53
i.e. postinst code
22:53
<alkisg>
vagrantc: I think putting it in debian/ltsp-client-core.upstart would do it
22:53
<vagrantc>
then it'd make sense to put it in debian/rules
22:53
<alkisg>
But you'd need to put the ltsp-client-core.init there too, for debian to select that one instead
22:53
(for dh_installinit)
22:54
<vagrantc>
oh, dh_installinit supports either?
22:54
<alkisg>
I think so
22:54
<vagrantc>
but not both?
22:54
<alkisg>
It automatically selects the correct one
22:54* vagrantc experiments.
22:55
<vagrantc>
correct seems only determinable at runtime
22:55
<alkisg>
I think it's "if ubuntu, select upstart, otherwise sysvinit"
22:55
Dunno what happens for debian+upstart
22:56
Maybe postinst code is added
22:56
<vagrantc>
the manpage says "and/or"
23:03
would be really great if we could get ltsp in sync in Debian/Ubuntu for the first time ever :)
23:03
only thing i still have left t work on, i guess, is the differing dependencies?
23:03
(presuming this last build works)
23:05Parker955_Away is now known as Parker955
23:06
<alkisg>
For my needs I don't think we have to change any dependencies
23:06
I don't think the fonts, udhcpc etc are important
23:07Phantomas has joined IRC (Phantomas!~Phantomas@unaffiliated/phantomas)
23:07* vagrantc wants to learn how to handle build-time dependency variables anyways
23:07
<alkisg>
One easy way would be to overwrite control with control.Ubuntu from debian/rules, right?
23:07
Or is there a more formal way?
23:08
or just have #ubuntu-start / #ubuntu-end sections and sed through them...
23:08
<vagrantc>
alkisg: we can add custom variants like ${misc:Depends}
23:08
<alkisg>
Aaah
23:09
${ubuntu:depends}, cool!
23:09
<vagrantc>
alkisg: so ${ltsp_client:Depends} and set those values differently depending on which distro it's building on
23:09
you'd want it on a per-package basis, and set the values in debian/rules
23:10
(or something called from debian/rules)
23:10
as long as it's not an insane list of differences, we should be fine
23:10
<alkisg>
Why on a per-package basis? Can't we decide the package name from debian/rules?
23:10
Like, misc:depends isn't the same for all packages, is it?
23:11
We'd need a debian:depends too, for nfs-kernel-server
23:11
(I don't mind having that either, I'm exporting /home with nfs)
23:11
<vagrantc>
man deb-substvars
23:12
ah, true enough, yes, we could implement them individually
23:16staffencasa has joined IRC (staffencasa!~staffenca@128.193.8.220)
23:21
<vagrantc>
alkisg: ltsp-config's install_examples function doesn't correctly detect existance of languageless files
23:21
<alkisg>
Hmm ok one more for tomorrow... 'night for now :)
23:22
<vagrantc>
sure :)
23:22alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.)
23:25staffencasa has left IRC (staffencasa!~staffenca@128.193.8.220, Ping timeout: 244 seconds)
23:26
<rhorstkoetter>
do you know - out of experience - how to define/modify the default desktop settings for new user accounts
23:28
for an ltsp setup, I imagine it an opportunity to define defaults not are non-standard applied to all new ltsp users
23:29
do you know of any customization scripts/tools etc to achieve this without the pita to do it for every single new user created. simple things like 3d on/off, icons, wallpaper etc
23:29
<vagrantc>
depends on the application ... gconf, dconf, update-alternatives ...
23:30
<rhorstkoetter>
dconf isn't per-user?
23:30
<vagrantc>
there are ways to set it globally, from what i hear.
23:30
!dconf
23:30
<ltsp>
vagrantc: I do not know about 'dconf', but I do know about these similar topics: 'docs'
23:30
<rhorstkoetter>
update-alternatives certainly is not but I'm wondering about default desktop settings (while this certainly is DE-dependent)
23:31* vagrantc remembers someone extolling the virtues of dconf over gconf in here in recent weeks
23:31
<rhorstkoetter>
ok, I need to research here a bit more. I just thought you know already as a std task for a ltsp admin
23:31
<vagrantc>
apparently, they found it's global configurations much easier
23:32
<rhorstkoetter>
ok, need to look at it
23:32
good to know that it's at least possible
23:33andygraybeal has left IRC (andygraybeal!~andy@173.245.135.246, Ping timeout: 245 seconds)
23:33
<rhorstkoetter>
I came to that question earlier on while talking with alkisg about compiz
23:33
<vagrantc>
ok, fixed the ltsp-config issue with install_example
23:33
<rhorstkoetter>
thing is, disabling compiz (alltogether) is, for example a systemwide setting while disabling kwin 3d effects is not
23:34
it's per-user and doing it (per-user) for an almost countless amount of pupils seems like a major pita to me :p
23:35
<vagrantc>
true enough.
23:35
worst case is you could write an LDM script to run at login
23:35
before kwin even starts
23:36
<rhorstkoetter>
that would be an opportunity, good idea
23:37
btw, let me take the opportunity to appreciate your hard work with ltsp. this actually is my very first day with that beast and I think it's awesome stuff. kudos to you all
23:38
<vagrantc>
glad it's of use! :)
23:40
<rhorstkoetter>
honestly, I'm running my very first test-env and I'm more than excited to see how WELL it actually works
23:41
i.e. I try to get familiar with it in order to be able to deploy it in a larger scale at some point in time, i.e. when learning curve is further down the road
23:42
my gf haven't even noticed at first that her netbook (serving as a test-client) isn't actually running her local ubuntu install but an ltsp session
23:42
and, I mean, what better prove of good result would you want?
23:42
:p
23:42
<vagrantc>
that's long been one of the goals ... that people don't even notice :)
23:48rhorstkoetter has left IRC (rhorstkoetter!~rhorstkoe@opensuse/member/rhorstkoetter, Ping timeout: 244 seconds)
23:53rhorstkoetter has joined IRC (rhorstkoetter!~rhorstkoe@opensuse/member/rhorstkoetter)