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


Channel log from 17 December 2008   (all times are UTC)

00:00ltsppbot has quit IRC
00:03ltsppbot has joined #ltsp
00:11
<pr0cess0r>
is it normal that firefox dont detect the adobe flash player ?
00:12
<rjune_>
it is if you don't have the flash player installed
00:12
<pr0cess0r>
i installed it from adobe web site un DEB package
00:13
but when i load a page it show me the plug-in thing
00:14
<rjune_>
sounds like firefox can't find it
00:14
it's been a long time since I've had to think about flash
00:14
ubuntu has it packaged up and working
00:15
<johnny>
maybe firefox is using the swfdec or gnash plugin by default instead
00:15pasmen has joined #ltsp
00:15
<pr0cess0r>
i try gnath and it dont wokr on the site i try
00:16
i wil remove all packages for flash and re install the player v10 from .deb
00:16
<rjune_>
it's a distro problem. so #debian is probably the best place to start
00:16plamengr has joined #ltsp
00:17
<pr0cess0r>
yes sorry im in the wrong place for this
00:17plamengr has quit IRC
00:18
<rjune_>
no problems, just less likely to find the help you need here.
00:18
<pr0cess0r>
i remove all package with flash in the name lol
00:20ltsppbot has quit IRC
00:20
<pr0cess0r>
i think its not linked to firefox like you say
00:23ltsppbot has joined #ltsp
00:27pr0cess0r has quit IRC
00:34Chidder has left #ltsp
00:39ltsppbot has quit IRC
00:40klausade has joined #ltsp
00:42ltsppbot has joined #ltsp
01:01ltsppbot0 has joined #ltsp
01:07ltsppbot has quit IRC
01:14Shingoshi has quit IRC
01:15Shingoshi has joined #ltsp
01:18ltsppbot0 has quit IRC
01:19ltsppbot has joined #ltsp
01:19gate_keeper_ has joined #ltsp
01:35ltsppbot has quit IRC
01:36ltsppbot has joined #ltsp
01:46otavio has quit IRC
01:53ltsppbot has quit IRC
01:56ltsppbot has joined #ltsp
02:02rcy`` has quit IRC
02:08jaja has quit IRC
02:12ltsppbot has quit IRC
02:13ltsppbot has joined #ltsp
02:21otavio has joined #ltsp
02:30ltsppbot has quit IRC
02:31ltsppbot has joined #ltsp
02:31EXP2 has left #ltsp
02:44jstephan has quit IRC
02:44jstephan has joined #ltsp
02:48ltsppbot has quit IRC
02:48hanthana has joined #ltsp
02:51ltsppbot has joined #ltsp
03:08ltsppbot has quit IRC
03:09ltsppbot has joined #ltsp
03:14chrisinajar_ has joined #ltsp
03:22alkisg has joined #ltsp
03:24mikkel has joined #ltsp
03:25ltsppbot has quit IRC
03:26ltsppbot has joined #ltsp
03:27chrisinajar has quit IRC
03:27brendan0powers has quit IRC
03:27brendan0powers has joined #ltsp
03:43ltsppbot has quit IRC
03:44ltsppbot has joined #ltsp
04:00ltsppbot has quit IRC
04:03ltsppbot has joined #ltsp
04:14kriegaffe has joined #ltsp
04:14
<kriegaffe>
Good morning.
04:15
I'm experiencing a problem to mount a usb stick on my ltsp server (xubuntu 8.04lts with ltsp 5.0.40)
04:16
I followed the ubuntu usb troubleshooting guide: https://wiki.ubuntu.com/DebugLocalDev
04:16
then at the very last line when I try to mount it with: ssh -X -S /var/run/ldm_socket_vt7_10.1.0.1 10.1.0.1 "/usr/sbin/ltspfsmounter /tmp add"
04:17
I get the following output:
04:17
Wrote 7bd41d281e3055fcbe4d7ad00c506372 size 32, waiting
04:17
Error: /media/gast04 is not mounted
04:17
gast04 being the user that logged in on that thinclient.
04:18
Does anybody have an idea? I would like the link to the usb to automagically show up on the thin client desktop.
04:19
I can see this in /var/run/ltspfs_fstab on the thinclient: Wrote 7bd41d281e3055fcbe4d7ad00c506372 size 32, waiting
04:19
sorry, this is in the fstab file: /dev/sdb1 /var/run/drives/STORE_N_GO vfat defaults 0 0
04:20
which corresponds exactly with the output in dmesg.
04:20ltsppbot has quit IRC
04:20
<kriegaffe>
I can put files in that folder.
04:21
so my bottomline: the usb stick seems to work but it doesn't show up automatically on the desktop.
04:21ltsppbot has joined #ltsp
04:23jstephan has quit IRC
04:26artista-frustrad has quit IRC
04:26artista_frustrad has joined #ltsp
04:27
<kriegaffe>
looks like this is maybe a bug in ltspfs? My version is the newest available for xubuntu 8.04 0.5.0~bzr20080109-3ubuntu3
04:30
also, the folder /media/gast04 to which it should be mounted looks strange on the server:
04:30
d????????? ? ? ? ? ? STORE_N_GO
04:37ltsppbot has quit IRC
04:38ltsppbot has joined #ltsp
04:38jstephan has joined #ltsp
04:55ltsppbot has quit IRC
04:58ltsppbot has joined #ltsp
05:14ltsppbot has quit IRC
05:15ltsppbot has joined #ltsp
05:15otavio has quit IRC
05:18
<kriegaffe>
does anybody have the strange permission problem when using usb disks in thin clients? the folder gets: d????????? ? ? ? ? ? STORE_N_GO
05:27otavio has joined #ltsp
05:28randra has joined #ltsp
05:32ltsppbot has quit IRC
05:33ltsppbot has joined #ltsp
05:36hanthana_ has joined #ltsp
05:39polytan has quit IRC
05:42dirigeant has joined #ltsp
05:49ltsppbot has quit IRC
05:51alkisg has quit IRC
05:52ltsppbot has joined #ltsp
05:53hanthana has quit IRC
05:59F-GT has quit IRC
06:00johnny has left #ltsp
06:04ekin has joined #ltsp
06:06
<ekin>
Hey all. I have a problem related to K12 Linux.
06:07
I have set up a thin client system following the steps at https://fedorahosted.org/k12linux/wiki/RHEL5Server but we cannot change the screen resolution on the thin clients.
06:08F-GT has joined #ltsp
06:08dirigeant has quit IRC
06:09ltsppbot has quit IRC
06:09hanthana_ has quit IRC
06:09
<ekin>
Specifiying screen resolution in lts.conf does not effect the thin clients. Some work in 1024x768 some in 800x600 even though both are capable of 1024.
06:10
the command we used is: X_MODE_0 = 1024x768
06:10ltsppbot has joined #ltsp
06:11
<Appiah>
did you put it under [Default] or where?
06:12alkisg has joined #ltsp
06:14
<ekin>
default didn't work. i tried it under a specific mac adress for the client and that didn't work either.
06:14dirigeant has joined #ltsp
06:14johnny has joined #ltsp
06:15johnny is now known as Guest80746
06:27ltsppbot has quit IRC
06:30ltsppbot has joined #ltsp
06:33Shingoshi has quit IRC
06:34Shingoshi has joined #ltsp
06:39gentgeen__ has joined #ltsp
06:40Shingoshi has quit IRC
06:41Shingoshi has joined #ltsp
06:46ltsppbot has quit IRC
06:49ltsppbot has joined #ltsp
06:53warren has quit IRC
06:54bobby_C has joined #ltsp
06:56
<kriegaffe>
:guni ?
06:57loather-work has quit IRC
06:57xmagixx has quit IRC
06:58xmagixx has joined #ltsp
07:02
<gate_keeper_>
guys
07:03
does anyone had setup ltsp5(ubuntu) + linux ha + dhcp failover :-/
07:03vagrantc has joined #ltsp
07:06ltsppbot has quit IRC
07:07ltsppbot has joined #ltsp
07:08* ogra sighs about the debian fontforge package ... there are days where i doubt the mental sanity of DDs
07:08
<ogra>
whoops .... ECHAN ...
07:10* vagrantc forwards the comment on to all DDs
07:11
<ogra>
heh
07:11
did you ever look at fontforge ?
07:11
its insane
07:12
<vagrantc>
doesn't look like anything i would be inclined to look into if i could get away with it :)
07:13
<ogra>
orig.tar.gz contains four tarballs ... debian/rules unpacks them, copies custom configure.in files into the different source dirs, runs eachs autofoo proggy separate etc etc
07:13
and indeed it isnt compatible at all with the latest libtool so explodes on each of these runs ...
07:15
<vagrantc>
ah, you whet my appetite!
07:15
<ogra>
heh
07:15
<rjune_>
ogra!
07:15
how was the booze?
07:15
<ogra>
rjune, hey
07:16
i loved the beer (thanks from david as well)
07:16
<rjune_>
how about the wine?
07:16
Glad he liked it too. I didn't hear back from him yesterday, was thinking of dropping a line today
07:17
<ogra>
i found the wine a bit to sweet, i know met in a slight more bitter way
07:17
was tasty as well though
07:18
<rjune_>
i know met?
07:18
<ogra>
i shared it with a lot of people on the last evening party ... my japanaes colleagues loved it since it tasted a bit like their plum wine
07:18
<rjune_>
That was a sweet wine, I have some stuff going now that's less so
07:19* ogra fires up dict ...
07:19
<rjune_>
wonderful, glad everything was enjoyed
07:19
<ogra>
ah, you call met -> mead
07:19
<rjune_>
ahh.
07:21
anyway, I have to head to work
07:23ltsppbot has quit IRC
07:26hanthana_ has joined #ltsp
07:26ltsppbot has joined #ltsp
07:31
<alkisg>
Ehm... I've just created a launchpad team for Greek teachers using LTSP, and I've named it "ltsp.sch.gr". Are there any copyright issues with the LTSP word?
07:33
<vvinet>
from ltsp.org I could find: LTSP is a registered trademark of DisklessWorkstations.Com, LLC
07:34
but I am no expert on ltsp or copyright
07:37
<alkisg>
Hmmm... I should ask before using it! :)
07:43ltsppbot has quit IRC
07:43alkisg has quit IRC
07:46ltsppbot has joined #ltsp
07:48alkisg has joined #ltsp
07:49cliebow has joined #ltsp
07:57loather-work has joined #ltsp
08:02ltsppbot has quit IRC
08:06
<gate_keeper_>
# lsmod | grep snd
08:06
# nothing .....
08:07dmaran has joined #ltsp
08:08six2one has joined #ltsp
08:09brendan0powers has quit IRC
08:13
<pscheie>
is this channel logged/archived/recorded somewhere?
08:14* ogra points pscheie to /topic
08:14davidj has joined #ltsp
08:22alkisg has quit IRC
08:22warren has joined #ltsp
08:25hanthana_ has quit IRC
08:25six2one has quit IRC
08:26six2one has joined #ltsp
08:27Gadi has joined #ltsp
08:35japerry_cat has joined #ltsp
08:40rjune has quit IRC
08:43rjune has joined #ltsp
08:44Egyptian[Home] has quit IRC
08:45Egyptian[Home] has joined #ltsp
08:46chrisinajar_ is now known as chrisinajar
08:51japerry has quit IRC
09:01Guest80746 is now known as johnny
09:07rjune__ has joined #ltsp
09:07rjune has quit IRC
09:08rjune__ has quit IRC
09:08rjune has joined #ltsp
09:10
<pscheie>
warren, ping
09:12
<warren>
pscheie: pong
09:15
<pscheie>
new quick start guide files at peterscheie.fedorapeople.org/k12linux/quick-start-guide
09:15
I fixed the path in the .desktop
09:15
etc.
09:16
did not change the icon
09:17
warren, I also responded to some of Fabrice's concerns on bugzilla ticket
09:17CAN-o-SPAM has joined #ltsp
09:20
<warren>
pscheie: not a big deal, but in the future please include a blank line between each changelog entry
09:21
<pscheie>
warren, oh, ok,
09:21
warren, for readability?
09:21
<warren>
pscheie: yes
09:21
pscheie: yes
09:21
oops
09:21
GenericName[en_US]=
09:21
why blank?
09:22
<pscheie>
I removed the Name[en_US] line because I thought it was redundant just like Fabrice suggested Icon[en_US] was
09:23
<johnny>
can somebody please try to see if namecheap.com works for them?
09:23
<pscheie>
I figured they were initially created by my right-clicking on the desktop to create the .desktop file, figured it wasn't really necessary
09:24
johnny, I can get to namecheap.com, if that's what you're asking
09:24
<johnny>
weird..
09:26
<warren>
I wouldn't trust a name like that.
09:26
<pscheie>
warren, my bad, GenericName[en_US] is in the source .desktop file
09:26
warren, I missed it, should be removed
09:26
<warren>
highbrowregistrar.com I would trust
09:27
<pscheie>
warren, I'll fix it, but I probably won't be able to get to it until tomorrow night at the earliest
09:27
<vagrantc>
warren: do you use server/services/ltsp-dhcpd.init ?
09:27
<warren>
pscheie: if I get around to spinning today I might just fix it
09:27
<vagrantc>
warren: in ltsp-trunk
09:27
<pscheie>
warren, that's fine
09:27
<warren>
vagrantc: checking
09:28
<vagrantc>
i get some .pot related warnings with it on build
09:28
<warren>
vagrantc: seems so
09:28
<vagrantc>
../server/services/ltsp-dhcpd.init:78: warning: the syntax $"..." is deprecated due to security reasons; use eval_gettext instead
09:29
<warren>
vagrantc: hmm, every initscript in Fedora is that way.
09:29
vagrantc: what tool is saying that?
09:30* vagrantc looks around for exactly where it happens
09:31
<vagrantc>
warren: when i run "make" from ltsp-trunk/po
09:32
<warren>
vagrantc: ok, so msgmerge
09:32
<vagrantc>
more specifically, make ltsp.pot
09:32
scratch that
09:32
<kriegaffe>
Hello folks, I know I've asked this before but I didn't find a solution yet. my usb stick is not automounted on the desktop of the thinclient and the permissions on hte server look like this: d????????? ? ? ? ? ? STORE_N_GO
09:32
<vagrantc>
ah yes, that is it
09:32
after removing it
09:33
warren: so msgmerge what?
09:33
<warren>
vagrantc: I think it is safe if you control the content of the script it is doing msgmerge on
09:33
<vagrantc>
warren: ok, i just noticed the warning and figured i'd mention it
09:34
<warren>
vagrantc: my people are saying that is just how initscripts are, and changing it will break it
09:34
msgmerge might just be overly paranoid
09:35
ok, I got the ppc to netboot, but when X runs it gets stuck, and I can't VT switch.
09:35
how could I log into it from the network?
09:35
<vagrantc>
ah, it's actually xgettext that's issuing the warning
09:36
<warren>
kriegaffe: the usb stick sounds like it has proprietary or non-standard stuff on it
09:36
kriegaffe: (it behaves like not a normal usb stick)
09:38
<kriegaffe>
ah ok
09:38
it worked with ltsp4 but ofcourse that is a completely different story
09:39
in fact, this usb stick is just basic vfat partitioned thing and it is empty.
09:39
I could go to var/run/drives/... and then in the usbstick.
09:39
<ogra>
and you get these permissions in /media/<devicename>
09:39
<kriegaffe>
when I take a closer look, on the server in /media I see folders with each accounts name
09:40
and therein, I also see that floppy has these strange permissions.
09:40
<ogra>
right, and below the account logged in to the client using that key there should be a device entry for your USB key
09:40
<kriegaffe>
yes
09:41
<ogra>
and the user owning it should be able to access it
09:41
<kriegaffe>
example of a folder in /media: "drwxr-x--- 3 root gast03 4,0K 2008-12-17 13:53 gast03"
09:42
so the permissions on this folder are wrong indeed
09:42
<ogra>
looks fine
09:42
no, the are right
09:42
<kriegaffe>
owner is root and group is gast03
09:42
<ogra>
right
09:42
<kriegaffe>
should the group have write?
09:42
because the logged on client is the user gast03 in this case
09:42
<ogra>
yes, how else would gast03 be able to write to the key ?
09:43
<kriegaffe>
indeed, but I get instead r-x for the group and the owner is root instead of the user
09:43
<ogra>
right
09:43
<kriegaffe>
Is this a wrong config somewhere?
09:43
<vagrantc>
would anyone object if i moved the screen-session.d/XS* scripts to use a wider range ... makes it hard to insert stuff in the middle of things
09:43
<ogra>
no
09:43
its all correct
09:44
the mounter tool on the server needs to be suid root so the dir thats created is owned by root
09:44
the user permissions are handled through the users group permissions
09:44
<kriegaffe>
ok, I understand that
09:45
this is how the floppy looks under the gast03 folder: "d????????? ? ? ? ? ? floppy0"
09:45
<ogra>
even if oyu look *as* guest03 ?
09:45
<kriegaffe>
same for any other device in any other gastXX folder
09:45
hmm, did not try that
09:45
<ogra>
try it :)
09:46
<kriegaffe>
ah, like you said: as the user, it looks ok: "drwxr-xr-x 2 gast03 gast03 40 2008-12-17 14:00 floppy0"
09:47
so this is exactly as it should be then.
09:48
<ogra>
right
09:48
<kriegaffe>
Doesn't the link to the folder pop up automatically on the gnome desktop for a reason? Do I need to config gnomevfs or anything to get that behavior?
09:48
at the moment, it doesn't pop up
09:48
<ogra>
fuse somehow hides permissions on that weird questionmark syntax for non owners
09:49
<kriegaffe>
it was confusing for me, I could not find it with google... (maybe my bad search) but this explains a lot to me
09:49
thanks
09:49
<ogra>
gnomevfs monitors /media so it should just pop up automatically
09:50
<kriegaffe>
but it doesn't so that is a gnomevfs problem then and not ltsp, ok. I'll google some info on that.
09:50
thanks again
09:57gate_keeper_ has quit IRC
10:07kharloss has joined #ltsp
10:07hanthana has joined #ltsp
10:08ekin has quit IRC
10:12Chidder has joined #ltsp
10:20kharloss_ has joined #ltsp
10:22kharloss_ has quit IRC
10:25Ahmuck has joined #ltsp
10:27kharloss has quit IRC
10:30
<Gadi>
is anyone near a test thin client that they have shell access to and can reboot to help me test something?
10:30
<warren>
Gadi: what's up?
10:31
<vagrantc>
possibly
10:31
<Gadi>
I just need some one to reboot without touching mouse/keyboard, then log into the shell, and cat /proc/sys/kernel/random/entropy_avail
10:31
<warren>
you need to touch the keyboardd to do that
10:32
<Gadi>
right
10:32
touch keyboard for that part
10:32
but dont move mouse
10:32
<warren>
Gadi: entropy is known to be non-existent after booting, there are no sources of entropy
10:32
<Gadi>
that's improtant
10:32
<warren>
upstream kernel doesn't use network as a source
10:32
<Gadi>
warren: yeah - I am finding that
10:32
<warren>
Gadi: could do better, just cat /proc/sys/kernel/random/entropy_avail into a log file at the end of bootup
10:32
Gadi: what are you trying to do with entropy?
10:33
Gadi: I've been forced to use /dev/urandom intead of /dev/random because of this
10:33* ogra was ondering the same
10:33
<Gadi>
I tracked down the most interesting weirdness yesterday
10:33* warren already hit related weirdness
10:33
<Gadi>
the ICA client uses /dev/random (not urandom) for SSL
10:33
<johnny>
i wouldn't call that weirdness..
10:33
<Gadi>
so, if you up the encryption level on a thin client,
10:33
the citrix client won't start until you move the mouse
10:34
its not weirdness
10:34
just took me a bit of thinking to find the root cause
10:34
to a user, its a bit odd
10:34
and looks like the session fails to come up
10:34
<vagrantc>
one might even say, real weird.
10:34
could you seed the randomness from the server somehow?
10:35
<Gadi>
warren: forced wheere?
10:35
vagrantc: yeah, I can
10:35
<vagrantc>
it would be imperfect randomness ...
10:35
<Gadi>
I just want to make sure that our LDM is not affected
10:35
(as we use ssh)
10:35
<vagrantc>
surely it uses urandom
10:35
<Gadi>
I have had times where when I use LDM across NAT
10:36
LDM_DIRECTX=False
10:36
where the display does not update until I move the mouse
10:36
(like with progress bars and such)
10:36
and I am thinking it may be realted
10:36
*related
10:36
<warren>
it is likely related yes.
10:36
Gadi: if we REALLY want to be sure there is entropy, we could have LDM not display anything until someone has moved the mouse around
10:36
<Gadi>
warren: in what case and how did you move to urandom?
10:37
from what I can tell, openssl > 0.9.7 uses urandom
10:37
<warren>
Gadi: before I used /dev/random to generate only 16 characters for mcookie xauth
10:37
<Gadi>
ah
10:37
<warren>
Gadi: booting a client, it coudln't generate even 16 characters from /dev/random
10:38* vagrantc reads RANDOM(4)
10:38
<Gadi>
openssh uses openssl, correct?
10:38
<warren>
Gadi: if entropy is REALLY desired then a "mouse gesture" requirement in the greeter would be good.
10:38
<vagrantc>
"As a general rule, /dev/urandom should be used for everything except long-lived GPG/SSL/SSH keys.
10:38
"
10:38
<warren>
vagrantc: yeah, but even /dev/urandom sucks if you have zero entropy
10:39
<johnny>
patch the kernel..
10:39
get your network entropy..
10:39
<warren>
our distro wont do that
10:39
<johnny>
is there a reason such a patch isn't in the kernel?
10:39* Gadi wonders if he should just inject /dev/urandom with a seed on boot
10:39
<Gadi>
even if it is the same seed
10:39
<vagrantc>
seeding it from the server over the network?
10:39
<warren>
johnny: it was judged to be not cryptographically safe
10:39
<johnny>
aha
10:39
<Gadi>
or if that defeats the security
10:39
<warren>
Fedora ships some userspace entropy generator
10:40
it gets it from random quantum noise of a webcam or microphone
10:40
<vagrantc>
i guess there's no way to securely transfer entropy... without a sufficient existing entropy pool :)
10:40
<Gadi>
I just need a one-time injection I think
10:41
I dont think I need a daemon
10:41
<warren>
well, no guarantee that clients have microphones or webcams
10:41
so that can't be depended on anyway
10:42
<vagrantc>
if it has it, it probably wouldn't hurt to use (at least the mic)
10:42
<Gadi>
Yeah, my entropy_avail for this client (after logging in with keyboard) is like 257
10:42
<ogra>
isnt there a sysctl way to switch it on ?
10:42
<Gadi>
way too low
10:42
<ogra>
or bump it
10:43Chidder has left #ltsp
10:43
<Gadi>
bump it how?
10:44
<ogra>
no idea ... via a sysctl variable probably
10:44
i have no idea if thats possible
10:44
<Gadi>
but what do you mean bump it?
10:44
you need to generate entropy from env sources
10:44
<warren>
ogra: is network entropy in the upstream kernel?
10:44
<ogra>
no idea
10:44
<vagrantc>
using a predictable seed gives the illusion of security ...
10:45
<ogra>
i'm luckily no kerne guy :) but i can ask
10:45
<warren>
normal linux machines are able to get entropy from disk timings
10:45
<Gadi>
right
10:45
sucks to have no disks
10:45
:)
10:45
I suppose I can tftp a random seed file from the server
10:46
and have the server generate that file periodically
10:46
<vagrantc>
i wonder if that hasn't been causing other slowdowns with X over SSH all along...
10:46
<Gadi>
like every minute or so
10:46
<warren>
wikipedia links to https://admin.fedoraproject.org/pkgdb/packages/name/audio-entropyd
10:46
If your hardware HAS a sound card, it might be able to generate something with audio-entropyd
10:47
without a microphone plugged in, background radiation still creates analog noise
10:47
if you boost the gain you can hear it
10:47
<Gadi>
right - but my environments that care about security also disable audio
10:47
:)
10:47
in the BIOS, I mean
10:48
network would definitely be the way to go
10:48
<warren>
oh
10:48
<vagrantc>
there's also randomsound ...
10:48
<warren>
vagrantc: eh?
10:48
<vagrantc>
warren: http://packages.debian.org/randomsound
10:48dmhardison has quit IRC
10:48
<ogra>
<apw> i think there is some way to ask the kernel which interrupts are contributing to entropy, and to add new ones if you know ones which are cool for that
10:48
from one of our kernel guys
10:49
<warren>
vagrantc: same problem, no device to input from
10:49
ogra: you can do that in the upstream kernel?
10:50
<ogra>
i assume so, i asked explicitly about upstream
10:50
he is digging for me ...
10:50
if we could just base on the NIC interrupt that should help
10:50japerry_cat has quit IRC
10:50
<warren>
vagrantc: randomsound is userspace?
10:51
<vagrantc>
warren: i just did apt-cache search ... if you've read the URL you know as much as i do :)
10:51
<warren>
http://lwn.net/Articles/283103/ relevant article on this subject
10:52* ogra doesnt like the idea to rely on HW that might not be present for this
10:53
<warren>
ogra: agreed
10:53
"But network interrupts are seen as a dubious source of entropy because they may be able to be observed, or manipulated, by an attacker. In addition, as network traffic rises, many network drivers turn off receive interrupts from the hardware, allowing the kernel to poll periodically for incoming packets. This would reduce entropy collection just at the time when it might be needed for encrypting the traffic."
10:55
<ogra>
well, still better than a stuck value at around 250
10:55* vagrantc would say the same of hardware that may or may not be present
10:56
<ogra>
well, a NIC is definately present on a thin client :)
10:56
<johnny>
too bad there isn't onboard hardware to help..
10:56* Gadi wonders if disk access includes ramdisk
10:59
<warren>
Gadi: http://egd.sourceforge.net/
10:59
Gadi: doesn't help to feed /dev/random but it acts as a replacement...
11:01
<Gadi>
warren: but it only gathers entropy on a busy system
11:01
<warren>
Gadi: hmm, is there any X_* lts.conf option I could use to turn off DRI?
11:01
<Gadi>
which does not help when waiting for a login
11:02* johnny awaits evolution-rss 0.1.2 for fedora10
11:02
<warren>
Gadi: mouse gesture might be the only option
11:02
"Wiggle mouse to continue"
11:02
johnny: is someone actually building it?
11:02
johnny: you could just join as a fedora developer and do it
11:03
<johnny>
how many packaging systems am i supposed to know :)
11:03
hehe
11:03
<warren>
We apparently have no standards of what is acceptable as an update...
11:03
<ogra>
so recommendation of our kernel tem is to actually use IRQF_SAMPLE_RANDOM
11:03
from the NIC driver
11:03
<warren>
ogra: NIC drivers have been ripping that out
11:04
ogra: http://lwn.net/Articles/283103/
11:04
<ogra>
yes i saw that
11:04
<johnny>
warren, it seems like it is updates-testing.. now that i can tell
11:05
go go packagekit :)
11:06
i just installed it directly from updates-testing..
11:07
warren, .. still no problems with the dbus issue since the packagekit-glib dep issues were solved
11:07
<warren>
that was only the most visible
11:07
<johnny>
sure.. i have yet to see any other issue ;)
11:08
now.. if only evolution,bansee,f-spot,tomboy,rhythmbox and the like could use packagekit to discover addons..
11:08
cross platform dreams make it more difficult tho :(
11:09
perfect..
11:10
fixed my problem with adding ebay search result rss feeds
11:10
it also fixes issues when using dark themes..
11:10jstephan has quit IRC
11:10
<johnny>
i'd still like to use email for rss at some point.. but there's no gui to manage subscriptions..
11:10
i installed some ruby sript to do it.. but i don't feel like editing config files..
11:11
also.. evolution makes me sad.. with it's lack of managesieve support..
11:11* johnny <3 sieve
11:11
<johnny>
works with every client..
11:11
luckily squirremail has a plugin..
11:12
<ogra>
vagrantc, how about rng-tools ?
11:12
warren, ^^^ does FC have something like that ?
11:13
<johnny>
rng-utils
11:13
<ogra>
ah, cool
11:13
that would at least help on clients where the HW is available
11:13
<johnny>
[root@falling ~]# repoquery -l rng-utils
11:13
/sbin/rngd
11:13
/usr/bin/rngtest
11:14
is that what you were expecting?
11:14
<ogra>
rngd
11:14* johnny wins
11:14
<ogra>
:)
11:14
<johnny>
at least knowing debian and ubuntu is similiar enough
11:14
<Nubae>
hmmm, u know what would be really nice for ltsp wishlist...
11:15
<johnny>
i'm just getting familiar with rpm tools..
11:15
<Nubae>
some sort of collaboration across apps
11:15
<warren>
what does rngd do?
11:15
<johnny>
Nubae, huH?
11:15
<Nubae>
like gobby, but for more than just gobby
11:15
<ogra>
it uses the hw_random module to request/set entropy
11:15
<johnny>
is that really ltsp related?
11:15
<Nubae>
well, maybe, maybe not
11:15
<johnny>
abiword has jabber support
11:15
so does inkscape
11:15
<Nubae>
oh they do?
11:15
<johnny>
and then there's cochinella
11:15
<Nubae>
nice to know....
11:16
<johnny>
well.. i don't know if ubuntu builds those in
11:16
but they are options on gentoo
11:16
<Nubae>
well, then doing that would be so awsome for schools
11:16
set up ejabberd, alongside support in various apps
11:17
<johnny>
Nubae, check out chesspark sometime
11:17
to see what folks are doing on top of xmpp
11:17
<Nubae>
what is it?
11:17
<johnny>
guess..
11:17
<Nubae>
ok
11:17
<johnny>
lol
11:18
<Nubae>
xmpp is real neat, especially considering google decided to use it
11:18
<johnny>
google doesn't do a good job..
11:18
<Nubae>
but there is no combined user experience for collaboration through the desktop manager
11:18
<johnny>
i need to find proof whether they still .. or ever did filter IQ packets
11:18
<Nubae>
or the apps
11:18
well, it seems fine to me, more stable than msn for sure
11:19
<ogra>
warren, amd-rng, geode-rng, intel-rng or via-rng are the respective modules
11:19
<johnny>
i heard that they filter iq packets that their official client does not support
11:19
so.. even if you were to use a client that did support them
11:19
<warren>
ogra: oh, hardware?
11:19
<ogra>
warren, yes
11:19
<warren>
ogra: not all hardware has hardware sources though
11:19
<ogra>
where available
11:19
<Nubae>
which is their official client, u mean web based?
11:19
<johnny>
googletalk and webclient
11:20
both
11:20
<Nubae>
ah, ok, never used googletalk, I just use pidgin
11:20
<ogra>
warren, well, we could make a recommendation ... "if you use citrix with ltsp, buy this or that HW"
11:20
<johnny>
and they still have a nasty bug in their code where if you join a muc room and change your nick
11:20
<ogra>
makes it at least better than "wiggle your mouse"
11:20
<johnny>
it starts off some endless cycle
11:20
<Nubae>
oh, that sounds bad indeed
11:21
<johnny>
they aslo dont' support tls1
11:21
<Gadi>
ogra: I am not convinced that LDM is unaffected here, too
11:21
<Nubae>
that aint no biggy
11:21
<Gadi>
even if it uses urandom
11:21
<johnny>
Nubae, not for you as a user..
11:21
but for coders it is
11:21
<ogra>
Gadi, well, my kernel guys would patch the NIC randomness back in if i begged very hard ... but that wont help warren ... or debian ... or gentoo
11:22
<Gadi>
right
11:22
<ogra>
but i got such an offer ...
11:22
<Gadi>
it would be nice if we could use memory interrupts or some such
11:22
<warren>
how do you automatically load the right rng module?
11:22
you need the rng module + daemon?
11:22
<Gadi>
and just read from memory periodically
11:22
<ogra>
yeah
11:22
i would expect udev to work and in hand with the daemon
11:23
if thats not the case thats a bug to be fixed :)
11:23japerry has joined #ltsp
11:23
<warren>
i'm seeing if we have that package
11:23
<ogra>
warren, according to johnny you do
11:23
warren, rng-utils
11:23
<warren>
via really needs a daemon? I thought via padlock already does this.
11:23
<Gadi>
ogra: on my server, I needed to modprobe the module
11:24
for intel-rng
11:24
<ogra>
Gadi, with rng-tools installed ?
11:24
<Gadi>
hmm
11:24
ah, no
11:24
<ogra>
right, the deamon should trigger udev ...
11:24
if not we should fix that
11:25
<Gadi>
/etc/modprobe.d/rng-tools
11:25
<ogra>
though i'm not sure if the module provides a device ...
11:25
<Gadi>
thats how it does it
11:25
<ogra>
ah
11:25
that looks debianish
11:25
<Gadi>
alias char-major-10-183 hw_random
11:25
alias /dev/hwrng hw_random
11:25
alias /dev/misc/hwrng hw_random
11:25
thats weird
11:25
<johnny>
you look debianish today ogra
11:25
hehe
11:25
new word
11:25
<ogra>
/dev/hwrng
11:25
thats what we want a udev rule for
11:25
<Gadi>
I wonder if it would load the correct module from hw_random
11:26
<johnny>
it might..
11:26
magic like that happens often
11:26
<Gadi>
will rngd take bits from hwrng and push to random?
11:27
or do apps need to be told to use hwrng?
11:27
<ogra>
ogra@osiris:~/Devel/packages/fontforge-0.0.20080429$ grep hwrng /etc/udev/rules.d/*
11:27
/etc/udev/rules.d/20-names.rules:KERNEL=="hw_random", NAME="hwrng"
11:27
well, intrepid udev seems to have it
11:27
according to the package description rngd sets the entropy based on hwrng
11:27
<Gadi>
which comes from the udev pkg
11:28
<ogra>
right
11:28
where it belongs
11:28
<Gadi>
right
11:28
<warren>
Gadi: hmm, I used XSERVER=fbdev and it failed to create or use an xorg.conf
11:28
<Gadi>
hmm... now I need to find out which of my thin clients have hwrng
11:28
<warren>
Gadi: ltsp-5.1.39
11:28
<ogra>
if the kernel finds a hw_random device it will load the module
11:28
<Gadi>
warren: did you look in /var/run/ltsp-xorg.conf
11:29
or whatever that is
11:29
<warren>
Gadi: it didn't create one, and X was run without -config
11:30
<Gadi>
is CONFIGURE_X=False?
11:30
ah - is /var/run/ltsp-xorg.conf in your rwtab?
11:31
<vagrantc>
/var/run surely is
11:32
<CAN-o-SPAM>
gadi/warren/vagrantc: is there any interest in another summer hackfest aside from brazil? *i have a feeling some wont be able to attend brazil
11:32
<warren>
CAN-o-SPAM: I likely can't do more than one event during the summer, and Brazil is looking unlikely given I have no budget
11:32
<CAN-o-SPAM>
& ogra ...
11:33
<warren>
Gadi: /var/run itself is writable
11:33
Gadi: I've never actually used X_* options ever before now
11:33
<vagrantc>
CAN-o-SPAM: i'm in a similar boat as warren
11:34
<ogra>
brazil depends on my company ...
11:34
<CAN-o-SPAM>
so would that mean were looking for an affordable summer event for everyone?
11:34
<vagrantc>
CAN-o-SPAM: i could probably host another event at freegeek
11:34
<johnny>
i can also host in baltimore
11:34
<warren>
I could host in Westford, but it sucks in Westford
11:35
<vagrantc>
heh
11:35
<pscheie>
having something in a major city makes transportation cheaper usually
11:35
<warren>
hotels are expensive and there's no public transport
11:35
<johnny>
i can find housing
11:35
<warren>
Boston there are some hostels and I could probably get MIT to host for free
11:35
I like Oregon
11:36
<johnny>
i like having an excuse to go elsewhere of course
11:36
<CAN-o-SPAM>
jammcq and i will be visiting a place in Ann Arbor, MI which has somewhat offered to provide free hosting, only cost would be transport/hotel
11:36* ogra likes oregon too, but baltimore is nearer
11:36
<CAN-o-SPAM>
maybe google could show some interest? MIT would be a nice place as well
11:36
<warren>
I doubt Google would care.
11:36
<ogra>
we just had our UDS at google
11:36
<warren>
hey, lets visit Norway!
11:36
<johnny>
lol
11:36
<pscheie>
since BTS is on east coast, should it be west coast?
11:36
<ogra>
and it really sucked a bit
11:36* pscheie likes Norway
11:36
<johnny>
really sucked.. a bit?
11:36
<warren>
ogra: how so?
11:37
<CAN-o-SPAM>
norway = brazil cost
11:37
<ogra>
no photos to be taken inside buildings etc
11:37
<johnny>
how could it really suck.. a bit :)
11:37
<ogra>
security guards all over watching you
11:37
<johnny>
it either sucks.. or it doesn't :)
11:37
sounds sucky
11:37
i don't want to go to google anyways
11:37
<ogra>
well, it was ok, but compared to two years ago its not really a pleasure anymore to do a conf at google
11:38Nubae has quit IRC
11:38
<pscheie>
vagrantc, you're in Oregon, correct?
11:38
<vagrantc>
pscheie: yeah
11:38
<warren>
Gadi: hmm, nothing in logs indicating why it didn't try to create an xorg.conf
11:38* warren gets food
11:38
<johnny>
warren becomes food
11:39nubae has joined #ltsp
11:39
<johnny>
and the cycle of life continues
11:39
<vagrantc>
i'll likely be heading to spain mid-july, and possibly midwest or east-cost early july or early august
11:39
<johnny>
so.. no july for freegeek
11:39
<ogra>
spain !
11:40
<johnny>
might as well go to brazil as spain for me
11:40
<ogra>
that sounds like a good opportunity ...
11:40
<vagrantc>
ogra: debian conference ... extremadura ... 4,000 LTSP *servers* deployed on debian :)
11:40
<johnny>
not a wholt lot of difference
11:40* Gadi just requests someplace near an airport
11:40
<Gadi>
:)
11:40
<pscheie>
spain = norway = brazil cost
11:40
<ogra>
pfft
11:40
you US guys
11:40
<Gadi>
or the Tri-state area
11:40
:)
11:40
<ogra>
always whining if you have to leave your country
11:40
<johnny>
lol
11:40
<Gadi>
we could all og to the Jersey shore
11:40
*go
11:40
<johnny>
ogra, i would love to leave.. as long as you're paying!
11:41* vagrantc hopes to avoid swimming with the fishes ... or the toxic waste
11:41
<johnny>
i have no business interests in ltsp
11:41
so i can't justify the costs of such travel for it
11:41
well except for our store..
11:41
but my entire budget is $50 every 2 weeks
11:41
<Gadi>
fine. virginia beach
11:41
<johnny>
i know the area
11:41
<Gadi>
thats closer to johnny
11:41
or cape may
11:42
<johnny>
Gadi, i lived in virginia beach for 20 years
11:42
<Gadi>
see
11:42
:)
11:42* vagrantc has no business sense, so can hardly manage any un-sponsored travel
11:42
<ogra>
Gadi, cape ? as in captown ? :)
11:42
<johnny>
cape may..
11:42
<ogra>
*capetown
11:42
<johnny>
it's a place
11:42
<Gadi>
florida's always cheap in the summer
11:42
<nubae>
ogra, u going to 25C3 in Berlin?
11:42
<johnny>
Gadi, but in baltimore.. i have an entire church :)
11:42
<CAN-o-SPAM>
ogra: anyword from the HW guy?
11:42
<johnny>
http://redemmas.org/2640/
11:43
<ogra>
CAN-o-SPAM, he said he'd mail you ... yesterday
11:43
<pscheie>
Gadi, there's a reason florida is cheap in the summer
11:43
<johnny>
don't mind the crappy page.. we're getting a new one soon
11:43
cuz it's soo damn hot
11:43plamengr has joined #ltsp
11:43
<CAN-o-SPAM>
the maine lobster market has been yielding much cheaper prices, so we must take advantage of lobster around the world.
11:44
<warren>
CAN-o-SPAM: dude, you didn't eat lobster... in Maine
11:44
<ogra>
spain ! :)
11:44
<CAN-o-SPAM>
haha warren neither did you
11:44
<warren>
I don't like lobster.
11:45
<nubae>
me either tastes bitter
11:45
<CAN-o-SPAM>
but i dont think the lobster will ever part with LTSP for some reason, eat it or not
11:45* vagrantc imagines a cheesy lobster mascot for LTSP
11:45
<pscheie>
should LTSP have a lobster as mascot?
11:45
<vagrantc>
heh
11:46* pscheie prefers monty python allusions
11:46
<pscheie>
is there a lobster sketch?
11:46* vagrantc infers monty python delusions
11:47* vagrantc ponders tagging ltsp-trunk
11:47
<CAN-o-SPAM>
DisklessWorkstations plans to do an LTSP logo contest and give out some free Thin Clients, and other things as prizes for places 1 - 3 ... so get your ideas going. we also need to complie a panel of judges, if your interested let me know
11:47
<warren>
LTSP's mascot is named Clampy
11:47
Clampy appears on the screen "You appear to be doing something stupid. Would you like to purchase support?"
11:48
<ogra>
CAN-o-SPAM, do an announcement o the distro artwork mailing lists
11:48
*to
11:48
<warren>
CAN-o-SPAM: dude... plenty of artists would do it without promise of a reward
11:48
<Gadi>
vagrantc: you tested all of the moving files around stuff?
11:48* vagrantc still needs to figure out where to get the debian artists to do a debian ldm theme
11:48
<vagrantc>
Gadi: about to :)
11:48
<Gadi>
hehe
11:48
<CAN-o-SPAM>
i will, i'm putting all the details together now, i'll let u know when were going to release
11:48pscheie has quit IRC
11:48
<ogra>
vagrantc, there is no debian-art ML ?
11:48
<vagrantc>
Gadi: i made sure to preserve the order
11:49
<Gadi>
ah, ok
11:49
<warren>
CAN-o-SPAM: fedora-art-list did the new ldm background image that I added last month.
11:49* ogra cant belive that
11:49
<CAN-o-SPAM>
i'm thinking a 2 month time-frame for submissions, do u guys think that is too short or too long?
11:49
<warren>
It is vastly superior to the old one.
11:49
<vagrantc>
ogra: haven't found one so obvious as that ... there's a debian-desktop group which owns some of the packages
11:49
<ogra>
CAN-o-SPAM, 2month should be fine
11:49
<vagrantc>
CAN-o-SPAM: just trying to put a face to a nick ... did i meet you in michigan back in 2006?
11:49
<ogra>
(ubuntu is going with the default gdm theme anyway though)
11:50
<CAN-o-SPAM>
i'm getting everything together, and thinking we will release Jan 1st 09
11:50
vagrant, i dont think we've had the chance to meet yet, unless you met me with jam or ron outside of an LTSP event
11:50
<ogra>
... but we have about 150 good artists on the ubuntu-artwork ML
11:51
<vagrantc>
CAN-o-SPAM: probably not, then
11:53* vagrantc found a mipsel machine to test LTSP with
11:53
<vagrantc>
although ... it's headless
11:53
<ogra>
heh
11:53
i doubt you find many mipsel clients
11:53
<warren>
CAN-o-SPAM: dude, I don't mean to seem too arrogant here in knowing what is best for you, but giving hardware for this particular purpose is not a particularly necessary.
11:53
<vagrantc>
vs. plain mips ?
11:53
<ogra>
though i finally got my beagleboard ... i'll look into that for ltsp as supported client
11:54
<warren>
CAN-o-SPAM: you're much better off getting hardware yourself to airlied for example
11:54
<ogra>
vagrantc, vs. *anything* ? :)
11:54
<vagrantc>
seems like armel will be a more promising future for low-powered thin clients
11:54
<ogra>
yeah
11:54
beagle has the disadvantage of not having an onboard NIC though
11:54
<CAN-o-SPAM>
warren: i figured some freebies would be motivating ... good artists typically would like to work on projects that pay them, no?
11:55
<vagrantc>
but i guess mips* thin clients do exist... somewhere.
11:55BrunoXLambert has joined #ltsp
11:55
<ogra>
but there are likely nettops with ARM coming soom
11:55
<warren>
CAN-o-SPAM: typically, except some art communities seem to get a lot done by reputation alone
11:55
<ogra>
given they will be faster than the x86 ones you get and demand less power that should be a perfect thin client
11:56
<warren>
Given that the monitor uses far more power than the client, how much of a difference does 50 vs 55 watts make?
11:56
<ogra>
vagrantc, yeah, SGI might have built some a century ago
11:56
<EAG>
generally speaking, is there any problems with using a 64-bit client with 32-bit server?
11:56
<ogra>
warren, you mean 3.5 vs 55
11:56
<warren>
EAG: pain in the ass to setup the server side for that
11:56
<EAG>
ok
11:56
<warren>
EAG: and no real benefit performance wise
11:56
<ogra>
there is a *massive* difference in power consumption between x86 and ARM
11:57
<vagrantc>
between x86 and dang near everything
11:57
<warren>
ogra: via or geode x86?
11:57
<ogra>
warren, doesnt matter
11:57
<warren>
thincan uses max 5w?
11:57
<ogra>
ARM will be lower
11:57pscheie has joined #ltsp
11:57
<ogra>
the ARM CPU itself uses 0.1-0.5W
11:58
depending at which speed you run it (OMAP3 CPU which is the most recent)
11:58
<EAG>
"ati-driver-installer-8-12-x86.x86_64.run" this smells 64-bit right?
11:58
<ogra>
ATOM is at 5-10W atm
11:58hanthana has quit IRC
11:59
<EAG>
I am trying to figure out how to load fglrx into the client... but it doesnt work
11:59
<ogra>
intel still has a long way to go until they match the ARM specs
11:59hanthana has joined #ltsp
11:59nubae has left #ltsp
12:03
<EAG>
somehow I always find myself pretty lonely with my problems ;) nothing even pops up on a google search...
12:05
<ogra>
on ubuntu you only need to install xorg-driver-fglrx in the chroot ... that should give you everything needed
12:06
not sure what to do on other distros
12:06
<EAG>
that didnt do the trick... unfortunately
12:06
I am using ubuntu
12:06
<ogra>
though your prob will be that both proprietary drivers (nvidia as well as fglrx) replace the mesa libs
12:06
which adds discrepancy between server and client Xorg
12:07
<EAG>
ok
12:07
<ogra>
xorg-driver-fglrx didnt install fglrx in your chroot ?
12:07
<EAG>
I cant seem to get xvideo to work with this ati-chip...no matter what I do
12:07
yes it is installed... but the module doesnt load
12:08
<ogra>
it should pull in fglrx-kernel-source which builds athe module using dkms
12:08
*the
12:09
if thats built yu need to regenerate the initramfs and run ltsp-update-kernels
12:09
that should get you the module on boot
12:10* ogra really doesnt see the usecase for such stuff on thin clients though
12:10
<EAG>
how do I regenerate initramfs?
12:10
<ogra>
update-initramfs -u in the chroot
12:10
<EAG>
ah, ok
12:10
<ogra>
though dkms should have done that for you
12:11
<EAG>
well, the reason for this mess is that I wanted xvideo to work
12:11
<ogra>
ltsp-update-kernels is more important, so the bootkernel/initramfs knows aboutteh new module
12:11
<EAG>
video playback seemed to work a lot better with that enabled
12:11
<ogra>
after you are done with everything you also need to run ltsp-update-image ... since the actual module will be in /lib/modules in the chroot
12:11
<johnny>
as long as all your clients have an ati card..
12:12
<EAG>
I have tried both ltsp-update-kernels and -image (in that order) after installing the driver
12:12
I only have one client
12:12
<ogra>
if you go to a shell on a booted client, do you see the module under /lib/modules in the tree ?
12:13
<EAG>
I will have to reboot and check
12:16
<warren>
vagrantc: might any of those post 5.1.39 changes help my "doesn't seem to try to create an xorg.conf" issue?
12:17
vagrantc: Gadi: oh! Is CONFIGURE_X=true required?
12:17
<Gadi>
warren:
12:17
no
12:17
CONFIGURE_X=False is required
12:17
<vagrantc>
warren: maybe ltsp-trunk 1013 ?
12:18
<warren>
Gadi: CONFIGURE_X is not in the parameters.txt
12:18
<ogra>
Gadi, ?
12:18
<Gadi>
well, it will go away
12:18
<ogra>
how did you guys break that ?
12:18
<Gadi>
the default should be False
12:18
from what you guys tell me
12:18
<ogra>
the default should be unset
12:18
<warren>
Gadi: I mean, I set only XSERVER=fbdev and it doesn't seem to be doing anything
12:19
am I missing something else?
12:19
<Gadi>
warren: do you change the default for CONFIGURE_X in fedora?
12:19
<warren>
Gadi: no
12:19
<Gadi>
ok, then that should be it
12:19
<ogra>
so it should be unset as well
12:20
<EAG>
ogra: yes, fglrx is in there under /lib/modules
12:20
<ogra>
XSERVER=fbdev will only fire if you use CONFIGURE_X=True
12:20
<Gadi>
# Only assemble a xorg.conf is CONFIGURE_X=False
12:20
if ! boolean_is_true "${CONFIGURE_X}"; then
12:20
<ogra>
??
12:20
<vagrantc>
ogra: we've been making crazy changes.
12:20
<Gadi>
ogra: no
12:20
not in the newest stuff
12:20
<warren>
vagrantc: oh yes... /var/run/ltsp-xorg.conf doesn't exist prior to that
12:20
<vagrantc>
we, as in, mostly not me :)
12:20
<ogra>
vagrantc, Gadi thats insane
12:20
<warren>
vagrantc: is it expected that it exists?
12:20
<ogra>
why did you break it
12:20
<vagrantc>
warren: it should create it
12:21
<Gadi>
ogra: not to worry
12:21
<ogra>
Gadi, CONFIGURE_X is the variable that tells the system to create an xorg.conf ...
12:21
why would you want it to be set to false to create one ?
12:21
<Gadi>
because we have not yet removed configure-x.sh
12:21
<ogra>
and ?
12:21
<Gadi>
and it would conflict
12:22
<ogra>
just dont call configure-x.sh
12:22
<vagrantc>
or the ability to specify a custom CONFIGURE_X_COMMAND
12:22
<ogra>
eeek
12:22
we shouldnt need any of that stuff anymore at some point
12:22
<vagrantc>
kind of, sort of.
12:22
<ogra>
why do you guys start to boat it in all possible directions now ?
12:22
*bloat
12:22
thats just crap
12:23
<vagrantc>
ogra: because we like getting you all riled up :P
12:23
<ogra>
remove the old broken method
12:23
and add fixed configuration modules piece by piece
12:23
there is definately no need to keep configure-x.sh in any way
12:23
<Gadi>
agreed
12:24* vagrantc thinks ogra undervalues sentimentality
12:24
<ogra>
so just rip it out *right now*
12:24
<Gadi>
hehe
12:24
<ogra>
and dont do such insane things with the vars :)
12:24
<Gadi>
u bet
12:24
<ogra>
vagrantc, i wrote configure-x.sh and i dont undervalue embarassement :)
12:24
<vagrantc>
ogra: heh :)
12:25
<ogra>
really, make it just go away asap ... dont try to do nifty things
12:25
<Gadi>
u got it
12:25
no more configure-x.sh
12:25
<vagrantc>
but i don't want to kill off CONFIGURE_X_COMMAND
12:25
<warren>
DIE!!!
12:25
<ogra>
we can live with a gap for a while until ll bits and pieces are added to the new system
12:25brendan0powers has joined #ltsp
12:25
<ogra>
warren, +++++
12:25
:)
12:26* Gadi tries not to think of ogra's bits and pieces
12:27
<warren>
OHHHH
12:27
vagrantc: hahaha
12:27
vagrantc: I've failed to add screen-session.d to be installed at all and my stuff worked
12:27
<vagrantc>
well, that's no wonder :)
12:27
<ogra>
vagrantc, CONFIGURE_X_COMMAND wont work with the new configuration system as i understand it
12:28
<johnny>
so.. what's the word on fixing ldm default them to use clearlooks or something else?
12:28
so i don't have to dep on ubuntulooks anymore
12:28
<ogra>
johnny, thats gone already
12:28
<vagrantc>
ogra: it will work just fine.
12:28
<johnny>
oh.. what does it default to now?
12:28
<ogra>
murrine and a dep on human in ubuntu
12:29
<johnny>
can i just dep on gtk-engines then?
12:29
<warren>
johnny: what exactly did that change do? I've been seeing the ubuntulooks error message forever
12:29
<vagrantc>
warren: clearlooks
12:29
<warren>
vagrantc: where is screen-session.d supposed to be installed?
12:29
<johnny>
murrine ? seriously? why not just clearlooks?
12:29
or whatever comes in gtk-engines
12:30
<vagrantc>
warren: /usr/share/ltsp/screen-session.d
12:30
<johnny>
warren, ldm would never even come up for me until in installed ubuntulooks
12:30
<ogra>
johnny, because the ubuntu default theme uses murrine and we use the human gtkrc for ubuntu ... the ltsp theme uses clearlooks since forever
12:30
<johnny>
since forever? not since i started
12:30
<ogra>
well, since vagrantc changed it
12:30
<warren>
johnny: odd, it worked for me all this time.
12:31
vagrantc's change isn't tagged yet, I think
12:31
vagrantc: btw you forgot to tag ldm yesterday.
12:31
<ogra>
johnny, it changed in ubuntu when i split out the ubuntu themes from ldm
12:31
<johnny>
sure.. i'm not talking about ubuntu tho.. i'm talking about upstream :)
12:31
<vagrantc>
warren: oops!
12:32
<ogra>
if you dont install the ldm-themes-ubuntu package only clearlooks will be used and it will show the ltsp theme
12:32
<warren>
vagrantc: I did it
12:32
<vagrantc>
warren: thanks
12:32
<ogra>
johnny, well, since when do you use plain upstream ? :) i made the change in ubuntu in hardy ... before it was indeed depending on ubuntuloks
12:32
<vagrantc>
warren: you got any pending pushes for it?
12:33
<ogra>
and i think vagrantc made the clearlooks change for upstream shortly before that
12:33
<vagrantc>
ogra: it's been in upstream for ages...
12:33
<johnny>
ogra, i always use plain upstream..
12:33
<ogra>
vagrantc, a year i think
12:33
<warren>
vagrantc: which?
12:33
<vagrantc>
ogra: actually, i just made the change to clearlooks yesterday
12:33
<ogra>
johnny, ?? before you developed on gentoo ?
12:33
<vagrantc>
warren: ldm-trunk ... otherwise the new tags don't get pushed
12:33
<johnny>
no.. certainly not on ubuntu
12:34
i use your packages on ubuntu..
12:34
<warren>
vagrantc: no pending pushes for ldm
12:34
working on ltsp now
12:34
<ogra>
johnny, right
12:36
<warren>
brb
12:40
<vagrantc>
tagged and pushed ltsp-trunk 5.1.40 ...
12:40
<ogra>
with al the redesign of the x configuration, how about moving to 5.2 if thats done ?
12:41
to set a visible milestone for the bigger change
12:42
<vagrantc>
well, we've already implemented the changes in 5.1 ...
12:43
<ogra>
but not fuly done and debugged yet
12:43
<vagrantc>
we really haven't made any sort of feature-based versioning
12:43
<EAG>
regarding that "configure-x.sh"... thats an error I end up with when I boot the client
12:43
<ogra>
i was thinking if CAN-o-SPAM wants to make a trademark release in jan. we might move along with the versioning
12:43
<vagrantc>
i mean, i would like to do more of that- actually incrementing the version when a major change happens
12:44
<EAG>
"/usr/lib/ltsp/configure-x.sh: line 27: 3920 aborted etc..."
12:44
what is this?
12:44
<ogra>
looks liek a bug
12:45johnny has quit IRC
12:45
<EAG>
maybe I should uninstall everything related to ltsp and start over
12:45johnny has joined #ltsp
12:45
<EAG>
its pretty screwed up now :(
12:46
<ogra>
well, mv /opt/ltsp/i386 /opt/ltsp/i386.bak; sudo ltsp-build-client --copy-sourceslist :)
12:46
no need to uninstall anything
12:46
<EAG>
true
12:48
<ogra>
line 27 sounds like Xorg -configure doesnt work at all
12:49
<EAG>
I would really like this to work... I liked the responsiveness when I used it
12:49moquist has joined #ltsp
12:49
<ogra>
you should see "Xorg failed to autodetect this card" in your servers syslog
12:49
<EAG>
but xorg shouldnt autodetect the card
12:49
I use a xorg.conf
12:50
<ogra>
with the X_CONF= parameter in lts.conf ?
12:50
<EAG>
X_CONF=/etc/X11/xorg.conf in lts.conf
12:50
thats the weird part
12:50
<ogra>
change that to be /etc/X11/xorg.conf-special or something
12:50
<EAG>
ok
12:50
<johnny>
you never set X_CONF=/etc/X11/xorg.conf
12:50
<ogra>
and try to set CONFIGURE_X=False
12:50
<johnny>
i wonder if that is worth detecting..
12:51
<EAG>
yep CONFIGURE_X=False is set
12:51
<ogra>
though it should be unset (at least in intrepid)
12:51
<johnny>
he's using hardy
12:51
i can tell
12:51
<EAG>
hehe
12:51
<johnny>
/usr/lib/ltsp
12:51
not /usr/share/ltsp
12:51
<EAG>
yes I am using hardy
12:51
<ogra>
ah, right
12:51
<warren>
vagrantc: we could always pick a cerremonial point to go to 5.2.x
12:51
<ogra>
intrepid is a lot better ...
12:51
<vagrantc>
warren: yeah
12:51
<warren>
vagrantc: however I have no interest in helping in a 5.1.x branch
12:51
<EAG>
I should upgrade my server anyway
12:52
<ogra>
warren, i would be all for doing it alongside CAN-o-SPAM doing the new logo etc
12:52
<vagrantc>
i don't see any need to branch off of ltsp-trunk as it stands
12:52
<warren>
ogra: well, if it is only for marketing purposes, nobody would see the new logo because distros have their own
12:52
<ogra>
no, we should just bump the version and have a party
12:52
<johnny>
warren, except me.. who uses all usptream :)
12:52
<vagrantc>
everything so far appears to be compatible with debian lenny, so i'm not worrying too much
12:53
<warren>
vagrantc: modulo bugs that need to be fixed
12:53
<ogra>
warren, well, we could all incorporte the upstream logo in our themes for celebration for one release :)
12:53
<warren>
ogra: my artist will be upset if I switch away...
12:53hanthana has quit IRC
12:53
<ogra>
not swithc away
12:53plamengr has quit IRC
12:53
<ogra>
just drop a button in a corner :)
12:54
<warren>
Let's treat 5.2 as the 10-year anniversary PR edition
12:54
<ogra>
yeah
12:54
<warren>
we'll get linux.com to interview all of us
12:54
<ogra>
that sounds good
12:54
and pobably a fixed heise article
12:54
that crap they wrote is so odd
12:55
it hurts
12:57
<warren>
horray
12:57
XSERVER=fbdev worked
12:58
<ogra>
what kind of client is that, that you require fbdev ?
13:04
<warren>
ogra: extremely old ppc imac
13:04
<ogra>
ah
13:04
<warren>
400MHz 768MB RAM R128 video
13:04
<ogra>
didnt they all have ati ?
13:04
<warren>
the R128 video isn't working
13:04
<ogra>
hmm
13:04
weird
13:05
<rjune>
ogra, do you know the name of the script used to convert sobby sessions to a web page?
13:05
<ogra>
did my dhcpd.conf work btw ?
13:08
rjune, nope ... i know our sysadmins have some experience with sobby (since they have to handle gobby.ubuntu.com which has to comep with several thousand users during UDS)
13:08
*cope
13:08
<rjune>
heh
13:09
A couple of us here are looking at it, I've seen reference to this magical script. but little to no actual documentation for sobby
13:09
<Gadi>
does anyone know if ssh pulls bits from /dev/urandom for each packet or whether it only does it once upon starting the session?
13:10
<ogra>
pkern is the upstream author of gobby and sobby, he is usually around in #ubuntu-motu once a week or so ... just keep your eyes open for him :)
13:11
Gadi, ask cjwatson in #ubuntu-devel ;) he does a lot of upstream ssh work alongside his package maintenance
13:11
though he might be gone for the day
13:13
<Gadi>
ogra: thx
13:14randra has quit IRC
13:14
<Gadi>
seems the whole world is gone for the day
13:14
<ogra>
heh
13:15
well, most canonical people are going on christmas holiday this week ...
13:15
<Gadi>
hmm... I thought I stole that holiday
13:16
bah, humbug
13:16
<ogra>
its a bridge week ... we had UDS until friday ...
13:16
<Gadi>
how waas it?
13:16
<ogra>
great from the productivity and evening events ...
13:16
<rjune>
Gadi, I don't think the jews stole christmas
13:16
<johnny>
that was me
13:16
sorry
13:17
<ogra>
not so gret to see how moch google has turned into a real business company since we were there last time
13:17
<Gadi>
well, I cannot speak for the rest
13:17
ogra: times are tough
13:17
<ogra>
they seem to have really developed some kind of paranoia alongside chrome and android
13:17
<Gadi>
cannot be all segues and vacation pods
13:18
<ogra>
well, it was very nice and open two year ago
13:18
it felt rather corporate this time
13:19
but beyond that it was the most productive UDS we ever had imho
13:19
<highvoltage>
I guess tough financial times can do that to a company.
13:20
<ogra>
you can watch the videos of the sessions soon :)
13:20
<highvoltage>
I see some of them are up already though
13:20
<johnny>
i don't think it has to do with financial times..
13:20
<ogra>
all discussions were taped and had live streaming audio
13:20
<johnny>
they are just hardening up..
13:20
<ogra>
so people could easily follow and attend via IRC
13:21rjune has quit IRC
13:26
<Gadi>
hmm... strace seems to suggest that it only opens it once
13:26
gets what it needs
13:26
and closes
13:26cyberorg has quit IRC
13:28
<stgraber>
vagrantc: do you have a url for 5.1.40 ?
13:30
<vagrantc>
stgraber: still testing it
13:30
stgraber: seems like your groups change doesn't work as well
13:31
stgraber: the gids don't match between the server and client, and i'm pretty sure the old way handled that ... but switching to using the "groups" command doesn't seem to.
13:32
<warren>
vagrantc: sigh, syntax difference between dhcpd.conf in dhcpv3 and dhcpv4
13:33
<vagrantc>
gah.
13:34
maybe it was broken then, too ...
13:35* Gadi is pretty sure the old way did the right thing
13:35* Gadi looks at the new way...
13:37
<vagrantc>
i don't really see *why* it would cause a problem
13:37
<warren>
The upstream maintainer of dhcp is telling me the syntax we have been using for if and else in dhcpd.conf is wrong.
13:37
and that woudl explain why my ppc detection isn't working
13:37
http://bazaar.launchpad.net/~ltsp-upstream/ltsp/ltsp-trunk/annotate/head:?file_id=dhcpdk12linux.conf-20080129180034-aiosbwg7a1bwv42c-1
13:37
<vagrantc>
it's worked for PXEClient vs. etherboot
13:38
<Gadi>
ah - yeah this way with groups doesn't handle groups with two words
13:38
<warren>
yes it did
13:38
<Gadi>
like: Domain Users
13:38
<warren>
if substring (option vendor-class-identifier, 20, 3) = "ppc" {
13:38
62 617 option root-path "172.31.100.254:/opt/ltsp/ppc";
13:38
vagrantc: this part isn't working
13:40
<Gadi>
warren: I think they dropped IP from root-path
13:40
<ogra>
warren, FC is shipping dhcpdv4 ?
13:40
<warren>
ogra: for a while now
13:40
<vagrantc>
Gadi: no, it was just broken in initramfs-tools in debian and ubuntu for a long time (ip in root-path)
13:40
<ogra>
intresting, i didnt know it was released
13:40
<warren>
ogra: since F-9
13:41
Gadi: no, that isn't the issue
13:41
Gadi: it always goes to the else no matter what
13:41ltsppbot has joined #ltsp
13:41
<ogra>
vagrantc, long time ... like fixed a year ago though
13:41
<Gadi>
warren: try creating a class instead
13:41
<warren>
Gadi: yeah, that's what he suggested
13:41
<stgraber>
well, we need to use "groups" as with the previous way of doing it, groups specified using /etc/security/groups.conf didn't appear
13:42
<Gadi>
class "LTSP" {
13:42
match if substring (option vendor-class-identifier, 0, 4) = "LTSP";
13:42
}
13:42
<vagrantc>
ogra: yes, i think i even am on a first name basis with the person who wrote the patch :)
13:42
<Gadi>
like that
13:42
<stgraber>
so if you use ldap and then add audio, fuse and video using pam_group that won't work
13:42
<warren>
Gadi: does that work in v3?
13:42
<Gadi>
yeah
13:42
I use classes all the time
13:42
and then allow/deny the class in IP pools
13:43
<vagrantc>
stgraber: actually, looks like the old way of setting groups was just as broken
13:43
<Gadi>
you can just add another line to the class with the root-path option
13:43
<vagrantc>
though i really don't understand how it would be broken.
13:43pmatulis has joined #ltsp
13:43
<ogra>
vagrantc, you talk to yourself nowadays ?
13:43
<warren>
Gadi: wait... why does this redefine the vendor class identifier?
13:43* warren has no friggin clue where this came from
13:43
<vagrantc>
ogra: when have i not?
13:43
<ogra>
heh
13:43
<Gadi>
romm@bart:~$ /usr/bin/groups | tr ' ' ','
13:43
Domain,Users,fuse,admin,sambashare,Domain,Admins
13:43
that is yucky
13:44
<vagrantc>
well, ok, it didn't have that breakage.
13:44
<Gadi>
warren: anything with that identifier will belong to that class
13:44
<vagrantc>
but it, for reasons i don't yet understand, doesn't properly handle when there's a gid mis-match between the server and the thin-client
13:45rjune has joined #ltsp
13:47
<vagrantc>
gah. i think i missed the cut-off for uploading ...
13:48
another 12 hours or so ...
13:49
or i made it exactly...
14:04
<Gadi>
vagrantc: if the chroot has a group with the same name as the server, it will use the chroot's group (as these are system groups)
14:05
for groups that the server has additionally, it will use the server's group
14:05
which gids are you comparing?
14:06
warren: I think the dev is objecting to the if/else syntax
14:07
not to the option line
14:07
stgraber: why does pam_group not honor getent?
14:07
that sounds like a bug in pam_group
14:08
<stgraber>
Gadi: yeah, it may be an upstream bug but I need it to work, you know :)
14:09
<Gadi>
either way, I would expect more users to use AD or smbldap than pam_groups
14:09
<ogra>
stgraber, did you talk to slangasek ?
14:09
<Gadi>
can we change the code to do both for now?
14:12
<warren>
Gadi: no
14:12
Gadi: found the real cause
14:12
Gadi: the dhcp client in the initrd is doing dhcp again with a different vendor class identifier
14:12
<Gadi>
ah
14:13* Gadi has always advocated specifying the VCI ourselves in our initramfs
14:13* Gadi already does in his
14:13
<Gadi>
we should set it to: LTSP-$arch
14:16Ahmuck has quit IRC
14:16
<warren>
Gadi: will defining a class allow it to "remember" the same machine and not give a different answer?
14:16
or not really?
14:17nubae has joined #ltsp
14:18CAN-o-SPAM has quit IRC
14:22
<Gadi>
remember the machine?
14:22Ahmuck has joined #ltsp
14:22
<Gadi>
to give the same IP?
14:24
<ogra>
it should get the same IP ... but it still wont send the same srting
14:26
<warren>
yeah, there's no way to do this apparently.
14:26
ogra: does your initrd do its own DHCP request with hits own vendor class identifier?
14:26* Gadi is unclear as to the goal
14:26
<warren>
its*
14:26
<Gadi>
mine does
14:26
<warren>
OH!
14:27
<Gadi>
I set VCI = LTSP
14:27
<ogra>
warren, it has to
14:27
the first request comes out of your HW
14:27
<Gadi>
the upstream one and ubuntu one do not
14:27
<warren>
Gadi: I'm the only one setting the root-path to mount during initrd from dhcp
14:27
<ogra>
i.e. the PXE protocol
14:27
<warren>
Gadi: I have no root-path hard coded in initrd image
14:27
<Gadi>
right
14:27
neither do we
14:27
<ogra>
the second one comes from ipconfig
14:27
<warren>
huh?
14:27
<Gadi>
I set a VCI
14:28
<warren>
oh
14:28
<Gadi>
not a root path
14:28
so, I know what VCI my client is sending
14:28
<warren>
And your dhcp server is then able to tell which root-path to use?
14:28
<Gadi>
and I can distinguish between LTSP clients and fat clients
14:28
what I proposed a long time ago was that we set our VCI's to LTSP-$arch
14:29
that way, we can use it to pass the appropriate root-path
14:29
(currently I only support i386)
14:29
<cliebow>
Gadi!!god ol vci
14:29
<Gadi>
*I* set the VCI because I set different ip pools whether it is a thin client or fat client
14:29
and I can deny all thin clients this way
14:29cliebow has quit IRC
14:29
<warren>
oh
14:29
<Gadi>
or deny all fat clients
14:30
thats why I have:
14:30
class "LTSP" {
14:30
match if substring (option vendor-class-identifier, 0, 4) = "LTSP";
14:30
}
14:30
<ogra>
warren, for me in ubuntu its a bit different, since we dont use root-path at all with nbd
14:30
<Gadi>
as well as:
14:30
class "PXE" {
14:30
match if substring (option vendor-class-identifier, 0, 9) = "PXEClient";
14:30
max-lease-time 1;
14:30
}
14:30
class "Etherboot" {
14:30
match if substring (option vendor-class-identifier, 0, 9) = "Etherboot";
14:30
max-lease-time 1;
14:30
}
14:31
<warren>
Gadi: what kind of section is that within?
14:31
within a shared-network?
14:31
<Ahmuck>
yesterday i had a high school student come in and try our test lab. this was his/her second login, and some key sequence threw them into language selection screen. this is an example of what we were disucssing aobut locking the options on the login screen. it was interesting to see this happen so quickly
14:31
<warren>
Gadi: withing a subnet?
14:31
<Gadi>
warren: global
14:31
I define classes globally
14:31
then, I have:
14:31
pool {
14:31
allow members of "PXE";
14:31
allow members of "Etherboot";
14:31
allow members of "LTSP";
14:31
within a subnet decl
14:32
(makes for cleaner code, too)
14:33jammcq has joined #ltsp
14:33
<jammcq>
hello friends
14:33* Gadi waves to jammcq
14:33
<ogra>
jammcq, !
14:34
<jammcq>
ogra: hey, welcome back
14:34
<ogra>
i lost my inner clock somewhere over the atlantic though ... still trying to find it
14:35
<jammcq>
hmm, damned inner clock issues
14:35
you'll have to get that fixed
14:37
<ogra>
yeah
14:37
:)
14:37* ogra had one of his most scary flight experiences the way home
14:38elisboa_off is now known as elisboa
14:40
<rjune>
ogra, what happened?
14:40
<ogra>
at 14000ft the captain told us they have issues to get the gear back in
14:41
so we flew to nevada and dumped fuel over the desert for 2h ...
14:41
<rjune>
aie!
14:41
<ogra>
then returned to SFO
14:41
and it wasnt really clear if the front gear was ok
14:41
<warren>
jammcq: ping
14:41
jammcq: http://wiki.ltsp.org/twiki/bin/view/Ltsp/CpuArchitectures#LTSP_on_Latter_Day_Macintosh_G3
14:41
jammcq: it says that you figured this out before
14:42
<ogra>
warren, i think cliebow is the G3 guy
14:42
<warren>
ogra: it mentions Jim
14:42
<ogra>
might be but i see cliebow working on ppc stuff since years :)
14:42
<warren>
jammcq: ppc imac -> yaboot downloads kernel and initrd, initrd itself does DHCP again with a vendor client identifier of "nash", dhcp server thinks it is a different request so it has no way of knowing it is a mac.
14:42
<ogra>
he should be your best resource of info
14:43
<jammcq>
warren: cliebow and I worked on it together. I don't have any of the hardware. if I recall correctly, we were able to figure out exactly what the mac needed in the vendor encapsulation. once we stuffed that into the dhcpd.conf file, things started working
14:43
<ogra>
if substring (option vendor-class-identifier, 0, 9) = "AAPLBSDPC" {
14:43
filename "/ltsp/yaboot";
14:43
option vendor-class-identifier "AAPLBSDPC";
14:43
option vendor-encapsulated-options
14:43
01:01:02:08:04:01:00:00:01:82:05:69:6d:61:63:34;
14:43
}
14:43
<warren>
jammcq: what does # OS and # hostname mean on that page?
14:43
<ogra>
that one i guess
14:43
<jammcq>
where?
14:44
<warren>
jammcq: http://wiki.ltsp.org/twiki/bin/view/Ltsp/CpuArchitectures#LTSP_on_Latter_Day_Macintosh_G3
14:44
<ogra>
since thats what i got from you ages ago for mac support
14:44
<warren>
ogra: where is that from?
14:44
<ogra>
warren, cliebow
14:44nicoAMG has joined #ltsp
14:44
<warren>
ogra: fun thing... OFW on this mac cannot download tftp from a directory
14:44
<ogra>
yeah, you said so yesterday
14:44
<warren>
jammcq: why does it set option vendorclass-identifier within it?
14:44
<ogra>
just change the filename directive to something that matches
14:46
<vagrantc>
gah. my upload of ltsp to experimental missed the deadline to get uploaded by about 35 seconds.
14:46
<jammcq>
warren: '69:6d:61:63:34' = 'imac4'
14:47
I'm thinking that's the hostname that gets assigned to the workstation
14:47
<warren>
jammcq: hmm, you haven't had the problem of initrd doing dhcp again itself?
14:47
<jammcq>
ummmm
14:48* Gadi desperately needs entropy
14:48
<jammcq>
I'm wondering who wrote that page, cuz they spelled my name wrong
14:49
<warren>
jammcq: what is the purpose of option vendor-class-identifier "AAPLBSDPC"; ?
14:49
jammcq: that looks like setting instead of detecting
14:50Sarten-X[bikinis has joined #ltsp
14:50
<jammcq>
warren: sorry, I really don't remember how all that stuff works
14:52
<warren>
jammcq: I think I ran into a limitation of our particular initrd
14:52
jammcq: back then you did it, it didn't actually use the root-path from dhcp server right?
14:53
<jammcq>
ummmm
14:53
dunno
14:53
<ogra>
sure it did
14:53
at least for me it did back when we used nfs in ubuntu
14:54
<warren>
jammcq: any idea where those vendor encapsulated options came from? I see many different examples of that online and they are all different
14:55
<jammcq>
I think we sniffed them on the wire by watching a mac boot from a mac server
14:55
<warren>
ogra: but how can it get the correct root-path if it ran dhcp request again?
14:55
ogra: with the incorrect VCI
14:56
<ogra>
warren, iirc because it sends the vendor-encapsulated-options string
14:56
<EAG>
*sigh*
14:56
<warren>
I think that's actually unrelated
14:56
<ogra>
but thats really two or more years ago for me ... i dont remember completely
14:56
<EAG>
where does the xorg.conf come from which the client automagically seem to load?
14:56
<ogra>
we dont support ppc in ubuntu since a while anymore
14:56
<EAG>
this is hillarious
14:56
<ogra>
EAG, it creates it
14:56dmaran has quit IRC
14:57Sarten-X has quit IRC
14:57
<warren>
EAG: if you run X without a xorg.conf it tries to autoconfig itself
14:57
EAG: LTSP has an override thing that writes its own xorg.conf based upon lts.conf options
14:57
<EAG>
it wont boot properly without my own xorg-ltsp.conf-file, but when it has booted it still uses some other xorg.conf... according to Xorg.6.log
14:58
I use X_CONF=/etc/X11/xorg-ltsp.conf in my lts.conf
14:58
<ogra>
EAG, you set CONFIGURE_X=False
14:58
right ?
14:58
<EAG>
yes
14:58
<ogra>
and you dont have /etc/X11/xorg.conf
14:58
<EAG>
and if I let it autoconfig it wont find my monitor properly (its fucked, edid doesnt work)
14:58
<ogra>
in your chroot
14:59
<EAG>
I have one, but it doesnt contain anything worth speaking of... if I look into the xorg.conf while on the client though, then it is a completely different xorg.conf
14:59
<ogra>
remove it
15:00
<EAG>
allright
15:00
<ogra>
and dont forget ltsp-update-image
15:00
<EAG>
what I really would love to know is why it thinks I have a synaptics touchpad
15:03Ahmuck has quit IRC
15:04* rjune kicks Gadi
15:04
<Gadi>
ow
15:05
<rjune>
there's some entropy
15:05
you weren't expecting that were you? ;-)
15:05
<Gadi>
entropy is not about expectations
15:05
<EAG>
btw... module-detect.c seems to fail to detect sound now...
15:05
weird
15:05
<Gadi>
its about not cleaning up your room
15:05
<EAG>
nothing has changed really
15:05
<rjune>
heh
15:06rjune has quit IRC
15:10warren has quit IRC
15:11
<EAG>
hmm nah.. the client still thinks it should autoconfig a xorg.conf of its own
15:12warren has joined #ltsp
15:12
<EAG>
no xorg.conf exists in /etc/X11 now, but it has been created when I look in the client
15:12
and in Xorg.6.log
15:12
any ideas?
15:34johnny has left #ltsp
15:38cyberorg has joined #ltsp
15:40CAN-o-SPAM has joined #ltsp
15:43
<vagrantc>
stgraber: http://incoming.debian.org/ltsp_5.1.40.orig.tar.gz
15:47
<stgraber>
vagrantc: thanks
15:48
<vagrantc>
35 seconds... *sigh*
15:55
<ogra>
??
15:55warren has quit IRC
15:55warren has joined #ltsp
15:56
<vagrantc>
ogra: i have to wait till the next upload run in about 11 hours ... missed this run by about 35 seconds...
15:56
<ogra>
wow, you have long sequences in debian
15:59rjune has joined #ltsp
15:59
<Gadi>
Ryan52: dont suppose you want to add a --progress to ldm-dialog :)
16:00* Gadi decided to go the route of having the user supply the entropy and has a dialog showing the amount of entropy relative to the total required
16:02ogra has quit IRC
16:02* Ryan52 wonders what Gadi is talking about
16:02alkisg has joined #ltsp
16:02ogra has joined #ltsp
16:03BrunoXLambert has quit IRC
16:03
<Gadi>
oh, just a --progress
16:03
like in zenity
16:03
:)
16:04
<Ryan52>
no, I mean, why do you need it?
16:04
<Gadi>
oh
16:04
thats a bit of a long story
16:04* Ryan52 is just curious
16:04
<Gadi>
hehe
16:04
do you know what entropy is?
16:04
<Ryan52>
ya
16:05
<Gadi>
so, basically, /dev/random - which is used as the RNG for Citrix client - is blocking until it has sufficient entropy
16:05
our thin clients have very little entropy on boot
16:05
because the only sources of entropy it has are by user intervention
16:05
like moving mouse, typing
16:06
so, if you try to use an encrypted ICA connection and there is not enough entropy, the application will simply wait
16:06
until there is
16:06
so, I have a dialog that has a meter of how much entropy the system has
16:07
and it tells the user to move the mouse until the system has sufficient entropy
16:07
(not in those words, of course)
16:07
<Ryan52>
okay, I might do it later then...it wouldn't be that hard :)
16:07
gotta go, be back later.
16:07
<Gadi>
cool
16:07johnny has joined #ltsp
16:10
<alkisg>
CAN-o-SPAM: I don't know if you're the right person to ask this, but then again, I don't know who else to ask! :) I've registered a launchpad team with a name of "ltsp.sch.gr". I selected this to reflect that it will host GReek SCHool LTSP - related projects. Can I use the LTSP name, or is it a violation of the copyright, and I should change it?
16:18
<CAN-o-SPAM>
alkisg: could you please send me an e-mail @ alexc@disklessworkstations.com ? I don't initially see it as a problem, but i need to review what your doing and don't have time to at the moment
16:18
<alkisg>
CAN-o-SPAM: will do, thanks
16:18
<CAN-o-SPAM>
thanks for askign.
16:19
<jammcq>
alkisg: it's not a copyright issue. it's a trademark issue
16:20
<alkisg>
jammcq: my english is not so good, and my law-knowledge is worse, so I can't tell the difference! :)
16:20
<jammcq>
no prob
16:21
<CAN-o-SPAM>
hey jam, the guy heading this event tomorrow (dug song), suggested that where we are going might be a good place to host a free hackfest
16:22
might keep our eyes open for that
16:22
<jammcq>
hmm
16:22
CAN-o-SPAM: do you have the instructions of where we are going?
16:23
<CAN-o-SPAM>
you need address? i don't know about this korean place, but the event is at: Cottage Inn Pizza, 2nd Floor, 512 E William St, Ann Arbor
16:23
<jammcq>
ooh
16:23
bring that address with you
16:24
<CAN-o-SPAM>
will do
16:24six2one has quit IRC
16:24
<CAN-o-SPAM>
gotta jet, c u 2morrow
16:24nubae has left #ltsp
16:24CAN-o-SPAM has quit IRC
16:26nicoAMG has quit IRC
16:30
<EAG>
why does the client have a Xorg.1.log?
16:30
<Gadi>
EAG: if you set CONFIGURE_X=True, it will have that
16:31
<EAG>
I have one no matter what
16:31
<Gadi>
because it runs an X -configure :1
16:31
<johnny>
that's cuz CONFIGURE_X=True is default on hardy
16:32
<EAG>
but I have one even if I set it to false
16:32
<johnny>
who cares..
16:32
lol
16:32
<EAG>
its just like as if the lts.conf isnt read
16:32
I care?
16:32
I cant get this to work as I want it to...
16:32
<Gadi>
do you have a syntax error in your lts.conf?
16:32
if you do, it won't read any of it
16:32
paste your lts.conf
16:33
!pastebot
16:33
<ltspbot`>
Gadi: "pastebot" is The LTSP pastebot is at http://pastebot.ltsp.org. Please paste all text longer than a line or two to the pastebot, as it helps to reduce traffic in the channel. A link to the content will be pasted in the channel.
16:33
<johnny>
Gadi, it'd be nice if it yelled louder about a badly parsed lts.conf
16:34
<Gadi>
that should be an easy fix
16:35
<ltsppbot>
"EAG" pasted "[DEFAULT] CONFIGURE_X = True L" (5 lines) at http://pastebot.ltsp.org/142
16:36
<Gadi>
try putting double quotes around /etc/X11/xorg-ltsp2.conf
16:36
<EAG>
ok
16:36
<Gadi>
or, even better - if you have a shell on the client, type: getltscfg -a
16:36
<ogra>
LOCALDEV = True
16:36
SOUND = True
16:36
both not needed
16:36
they are default values
16:37
<EAG>
okay
16:37
<ogra>
and the default section might be case sensitive
16:37
<johnny>
isn't it lowercase default?
16:37
yeah
16:37
<ogra>
i.e. [default]
16:37
the vars definately arent ... but the sections might be
16:40
<EAG>
maybe I can focus on getting fglrx to load properly instead... from Xorg.6.log it seems like that the autoconfig finds all hardware etc...
16:42
<Gadi>
EAG: u mean for compiz?
16:43
or opengl?
16:43
<alkisg>
About [default]/[Default]: both are working.
16:43
<Gadi>
alkisg: and DEFAULT?
16:43
<alkisg>
Gadi: I don't know, but could it really be different?
16:44
<Gadi>
depends how jammcq coded it way back when
16:44
<alkisg>
OK :)
16:46
<ogra>
Gadi, for xvideo
16:46
<Gadi>
he needs to install the mesa stuff in the chroot
16:46
<warren>
hmm
16:46
<ogra>
thats installed
16:47
GL works out of the box with intel HW
16:47
<warren>
I need to find an explanation of the vendor-encapsulated-options
16:47
I want it to boot like the example on http://wiki.ltsp.org/twiki/bin/view/Ltsp/CpuArchitectures#LTSP_on_Latter_Day_Macintosh_G3 without setting a hostname
16:47
<Gadi>
well, compiz checks the server's /var/log/Xorg.0.log to see if it is capable
16:47
he can try copying his client's Xorg log file to the server
16:48* ogra didnt talk about compiz :)
16:48
<Gadi>
right - but mat be similar?
16:48
<ogra>
i talked about GL support
16:48
i dont even think you need any mesa stuff for composite
16:48
<Gadi>
well, it is also driver dependent
16:48nubae has joined #ltsp
16:49
<Gadi>
he may not get opengl with acceleration
16:49tinpardo has quit IRC
16:49
<Gadi>
(which might be what xvideo needs)
16:49
<warren>
hmm, simply removing the length and hostname made it boot
16:49* Gadi shrugs
16:49
<warren>
so this might work...
16:50
Gadi: do you have any idea how to match in dhcpd.conf on a range of mac addresses?
16:50
<Gadi>
a range of mac addresses?
16:50
just make a host decl for each mac addy
16:50
<loather-work>
yeah, i don't think you can do a range of macaddrs
16:50
<Gadi>
in a separate group
16:51
<warren>
I want a wide range, like everything that begins with XX:XX:XX
16:51
<loather-work>
it'd be kinda nice if it had something ala bind's $GENERATE syntax
16:51
<Gadi>
you might be able to create a class that matches on mac addres
16:52
like:
16:52
<warren>
if binary-to-ascii (16, 8, "-", substring (hardware, 0, 4)) = "1-0-16-3e" {
16:52
Not exactly what I want
16:52
I want an actual range
16:52
like number wise
16:53
<Gadi>
mathc if (substring (fixed-address, 0, 8) = "aa:bb:cc");
16:53
oops
16:53
I meant hardware
16:54
<loather-work>
yeah. that might actually work.
16:54
<Gadi>
you may need "hardware ethernet"
16:54
if you specify by octets
16:55
<warren>
oooh
16:55
this isn't too bad
16:55
<loather-work>
yeah, otherwise it has the ethertype prefix associated with it
16:55
<warren>
Apple has only a few MAC prefixes
16:55
http://pastebin.com/m6cc2ed07
16:55
crap... this matches against Intel Mac's too I bet.
16:56
<loather-work>
you know what i would do that'd make it even easier? stick the wonky terminals on their own VLAN
16:56chrisinajar is now known as chrisinanoffice
16:56
<loather-work>
then use a helper address on the switch to handle dhcp
16:56* Gadi doesnt know why you dont just set your own VCI
16:56
<Gadi>
and call it a day
16:56
<warren>
Gadi: can't
16:56
<Gadi>
why not?
16:57
<warren>
Gadi: can't modify the initrd tool from 3 years ago
16:57
<Gadi>
u have no control over the initrd?
16:57
<warren>
gotta go now, commuting
16:57
<Gadi>
doesnt the tool pull in scripts?
16:57
<vagrantc>
Gadi: do you use klibc's ipconfig as a dhcp client, or some other client?
16:57
<warren>
Gadi: nope
16:57
<Gadi>
vagrantc: ipconfig
16:57warren has quit IRC
16:57
<Gadi>
I kinda patched the source, because I was too lazy to see if it had a flag
16:57
<vagrantc>
Gadi: how do you do it? :)
16:57
<Gadi>
:)
16:58
<vagrantc>
ah.
16:58
<Gadi>
changed the default in the source
16:58
and I only did that, because it had another bug that I needed to fix
16:58
<vagrantc>
that's a feature we should really get upstreamed...
16:58
well, configureable, that is
16:58
<Gadi>
it may have it
16:58
it was hard to find good docs
16:58
bec it is klibc
16:58
:P
16:59
<vagrantc>
klibc is sparse with the docs, yes.
16:59
that's how it gets to be so tiny!
16:59
<Gadi>
if my C were better
16:59
...
16:59
;)
16:59
<vagrantc>
if my C were better ...
16:59
<rjune_>
you both have wonderful C
17:00* vagrantc is a cargo cult C coder
17:00
<rjune_>
LOL
17:00
<vagrantc>
or maybe i'm a shaman, with an inkling of actual understanding
17:01
<alkisg>
Gadi: if you even need help with klibc ipconfig, I spend some weeks over its source, so I may help.\
17:01
<loather-work>
the code comes out of spirit walks with vim
17:01
<Gadi>
alkisg: cool - can you pass a flag to it to modify the VCI?
17:02
(like every other dhcp client)
17:02
<alkisg>
Gadi: and VCI stands for.... ?
17:02
<Gadi>
vendor-class-identifier
17:02
<alkisg>
Ah... sure, pretty easy. A command line option?
17:02
<Gadi>
it is a field that is modifiable by the vendor
17:02
yeah
17:02
the option may already exist
17:02
and be undocumented
17:02
:)
17:02
<alkisg>
Gadi: no, I didn't see any option
17:03
<Gadi>
ah, ok
17:03
<alkisg>
But pick a letter! Is "v" ok?
17:03
<Gadi>
then, I would do "i"
17:03
if available
17:03
<alkisg>
Let me see...
17:03
<Gadi>
that is usually the flag
17:03
or -I
17:03
(capital)
17:05
btw: debian/ubuntu guys - I have some code that uses mii-tool/ethtool in the ltsp-nbd script to use the interface that is connected (if there are multiple NICs in the client) - is that of use?
17:05
<alkisg>
Gadi: these are used: c = getopt(argc, argv, "c:d:i:onp:t:r:");
17:05
<Gadi>
ah
17:05
<alkisg>
c : d : i
17:05
<Gadi>
so they already use i
17:05
<alkisg>
Yeap
17:05* Gadi wonders what for?
17:06
<alkisg>
Let me find the help page...
17:06
<Gadi>
good luck
17:06jammcq has quit IRC
17:07
<alkisg>
Gadi: http://codtech.com/wiki/index.php/Ipconfig
17:07
Hehe
17:07
<vagrantc>
Gadi: would be useful to see, yes.
17:07
<alkisg>
There *is* a vendor identifier... :)
17:07
<Gadi>
beautiful
17:07
:)
17:07
then, we should modify ltsp-nbd with that, as well
17:07
IMHO
17:09
<vagrantc>
yes!
17:10
now, how to get it into upstream initramfs-tools and such ...
17:10
<Gadi>
vagrantc: we dont need to
17:10
we are upstream
17:11
we call ipconfig in lts-nbd
17:11
<vagrantc>
not in debian
17:11
still use NFS
17:11
but i've had good luck getting patches into initramfs-tools
17:11
<Gadi>
ah - and you dont have your own nfs script?
17:11
<vagrantc>
i'd like to avoid it, yes.
17:12
and debian also uses the initramfs-tools scripts for NBD from nbd-client for some of the possible ways to configure NBD in debian
17:13
ltsp_nbd is used when NBD+squashfs+aufs is used, though
17:13* alkisg still wonders why isn't uclibc preferred over klibc...
17:14
<rjune_>
dunno
17:15
<ogra>
Gadi, is mii-tool in the initramfs already ?
17:15
<rjune_>
uclibc is pretty damn slick
17:15
<ogra>
no option in ubuntu
17:15
<vagrantc>
alkisg: i don't think it's even in debian ... my guess would be liscensing issues, since it's been around so long
17:15
<alkisg>
rjune_: much, much, much more stable / bugfree / readable code
17:15chrisinanoffice has quit IRC
17:15brendan0powers has quit IRC
17:16
<alkisg>
vagrantc: gnu lgpl... ???
17:16
<ogra>
and puts extra maintenance on your team if you dont have it already ... the advantage of klibc is that it is maintained inside glibc
17:17
so if you have a glibc maintainer you get klibc for free ...
17:18
doesnt uclibc also requires special linking and rebuilds of everything you want to use with it ?
17:18
<alkisg>
ogra, in my opinion, klibc has many more bugs than uclibc, and that alone makes more difficult to maintain... Anyway, I'll stop talking since I haven't done any linux programming yet. :)
17:19
<vagrantc>
alkisg: sometimes, if authors didn't properly state their liscensing, and they can't or work correct their mistakes, it may interfere with the ability to distribute it ... like 1 out of 10 files has it's liscense, and one just doesn't contain the appropriate information
17:19
<alkisg>
I see... that would be a serious problem, yeap
17:19
<vagrantc>
ogra: klibc requires custom builds for everything
17:19
but it ships everything
17:20
<ogra>
alkisg, well, it might not be the lib itself but the overhead you produce through rebuiling the world
17:20
<vagrantc>
alkisg: i'm purely speculating on why...
17:20
<Gadi>
ogra: no, I add a hook for it
17:20
and ethtool
17:20
<ogra>
vagrantc, right, but klibc is binary compatible with glibc afaik
17:20
<Gadi>
(some network drivers are broken with one or the other)
17:20
<EAG>
Gadi: no, I am not using compiz, but I want xvideo to work
17:20
<ogra>
so you easily can use ethtool in initramfs without rebuilding it :)
17:20
<Gadi>
thats cool
17:21
<ogra>
Gadi, so go ahead :)
17:21
<rjune_>
binary compatble is it's big feature?
17:21
<vagrantc>
i'm pretty sure glibc is still shipped in the initramfs, and that's why you don't have to rebuild
17:21
<rjune_>
admittedly, that's a big feature.
17:21
<johnny>
uclibc.. go ..
17:21
<ogra>
vagrantc, but i dont think you have to afaik
17:22
<Gadi>
ok - gotta run
17:22* vagrantc can't see any reason for shipping glibc if you didn't actually have to
17:22
<Gadi>
night all
17:22Gadi has left #ltsp
17:22
<rjune_>
night Gadi
17:22
vagrantc: depends on your reqs
17:22
<ogra>
vagrantc, what rjune says
17:22
<vagrantc>
i suppose
17:22
<alkisg>
johnny: my USR9108 router uses uclibc, AFAIK... :)
17:22
<ogra>
there are apps that use specific glibc stuff
17:22
<rjune_>
for a full distro, use glibc
17:23
for embedded stuff, uclibc + busybox looks attractive
17:23
<ogra>
and initramfs-tools is designed in a way that you can add everything
17:23
thats its big advantage
17:24
<rjune_>
glibc you mean?
17:24
<ogra>
initramfs-tools i mean
17:24
<rjune_>
ah, ok
17:25
<ogra>
the thing that builds our initramfs in ubuntu and debian
17:26mikkel has quit IRC
17:26Ahmuck has joined #ltsp
17:29bobby_C has quit IRC
17:39
<vagrantc>
stgraber, ogra: other than the DHCPPORT feature, i think you could replace much of ltsp_nbd with the configure_network function in initramfs-tools
17:39
and DHCPPORT and VCI stuff we should try to get upstreamed
17:40
<ogra>
upstreamed to ?
17:40
<vagrantc>
initramfs-tools
17:42
<ogra>
ah
17:42
yeah
17:42
<alkisg>
What good thing are you cooking with VCI? Maybe force ipconfig to only accept answers from the ltsp dhcp server?
17:43
<ogra>
well, gadi's stuff works the other way round
17:43
<ltsppbot>
"vagrantc" pasted "use configure_networking from initramfs-tools" (52 lines) at http://pastebot.ltsp.org/143
17:43
<ogra>
and sends a LTSP VCI
17:43
<vagrantc>
alkisg: making it easier to identify which clients are LTSP clients
17:43
alkisg: and possibly architecture
17:44
<alkisg>
vagrantc, and what's wrong with the existing "Linux ipconfig"? - ah, ok...
17:44
<vagrantc>
alkisg: we have no idea if it's LTSP or not...
17:44
<ogra>
vagrantc, aaah, i was wondering how configure_networking replaces the nbd and squashfs mounting :D
17:44
yeah, that looks good
17:44
<vagrantc>
you'd loose DHCPPORT, at least until you got it into initramfs-tools
17:45
<ogra>
well, there are only some weird guys that use DHCPPORT ... alkisg is one of them iirc :P
17:45
<vagrantc>
heh
17:45
<alkisg>
ogra, not since I found out about IPAPPEND 3! :P
17:45
<ogra>
ah !
17:45
so we can drop it :)
17:45
<vagrantc>
IPAPPEND rocks the house.
17:45
<alkisg>
And got rid of all the ipconfig bugs...
17:46
<vagrantc>
and i'll work on getting the VCI stuff into initramfs-tools... somehow...
17:46
<ogra>
could we have a translation function DHCPPORT-> IPAPPEND
17:46
?
17:46
<vagrantc>
IPAPPEND happens in pxelinux
17:47
so there's only a single DHCP request
17:47
<ogra>
oh, i thought that was IPCONFIG
17:47
<alkisg>
ogra, but some may use an external dhcp server
17:47
<vagrantc>
the DHCPPORT handling in ltsp_nbd looks a little weird ...
17:47
<alkisg>
pxelinux passes the ip as a kernel parameter, and then ipconfig gets it and doesn't ask for a new dhcp request
17:48
<ogra>
vagrantc, well, ask laga ... if he ever shows up again
17:48
he needed it in teh mythbuntu thinclient
17:48
but in the end copied ltsp_nbd anyway
17:49
<vagrantc>
i'll try and get VCI and the DHCPPORT stuff into initramfs-tools in debian
17:49
and either patch around it in debian, or commit it upstream ...
17:49
in ltsp-trunk
17:50
<ogra>
sounds ok to me
17:51
i dont personally use VCI anywhere and dont know anyone who does ...
17:51
apart from the person who implemented it
17:51
and the same goes for DHCPPORT
17:52
<alkisg>
I only use it to enable my dhcp server to give IPs to ltsp clients, but not to windows clients. So the default "Linux ipconfig" is fine for me...
17:52
<vagrantc>
i can see some believeable (though unfortunate) use cases for both options
17:52
<ogra>
sure
17:52
<alkisg>
vagrantc, aren't there better ways to check for the client hardware than to patch ipconfig?
17:52
<ogra>
but neither is widely documented anyway
17:53
<vagrantc>
alkisg: no patches to ipconfig needed
17:53
at least on debian lenny ... can't imagine ubuntu is behind much
17:53
<alkisg>
Oh, yeap
17:53
ipconfig -i, forgot already...
17:53
<vagrantc>
just need to patch initramfs-tools to make use of it
17:53
alkisg: you're the one who discovered it! :)
17:54
<alkisg>
It'll a little late, and after 4 beers my head is not thinking quite straight... :)
17:54
<ogra>
not even 1am
17:54
come on
17:54
<alkisg>
1:54 :)
17:54
<ogra>
and its greek beer, isnt it ?
17:54
<alkisg>
You're one hour behind! :)
17:54
<ogra>
pfft
17:54
<alkisg>
2 of them were
17:54
<ogra>
and two beer behind :)
17:55brendan0powers has joined #ltsp
17:55
<alkisg>
Ok, German beers are larger, I'll give you that!
17:55* vagrantc forgets the rudiments of git
17:55* ogra doesnt want to remember any rudiments of git
18:02pmatulis has quit IRC
18:04johnny has left #ltsp
18:04
<alkisg>
Wow... I'm the only one on the internet that checks for substring(option vendor-class-identifier, 0, 14) = "Linux ipconfig"...
18:19warren has joined #ltsp
18:21
<ogra>
well, you are just smarter than the others :)
18:24
<alkisg>
:P
18:25alkisg has quit IRC
18:32
<vagrantc>
ok, i think i have a patch to add support for DHCPPORT and DHCPVCI to initramfs-tools in debian ...
18:34
<ltsppbot>
"vagrantc" pasted "initramfs-tools support for setting DHCPPORT and DHCPVCI" (49 lines) at http://pastebot.ltsp.org/144
18:38
<ogra>
looks good to me
18:40
<warren>
Anyone have any Intel Macs?
18:56
<rjune_>
warren: no, but I'll take one. ;-)
18:56
ogra: speaking of beer, did you think the stouts I sent were particularly strong?
18:57
<ogra>
average i'd say
18:57
but then i'm used to german beer
18:57
for american they were strong :)
18:58hersonls has joined #ltsp
18:59
<rjune_>
it was very strong for the mass market american stuff
18:59
<ogra>
yeah
18:59
<rjune_>
don't be fooled though, US microbrew is pretty good.
18:59
<ogra>
yeah, i know
18:59
though i'm usually fine with sam adams or anchor steam if i have to pick a bottled one
19:00
<rjune_>
ugh, sam adams?
19:00
I would tell you to at least have a Becks, but I think you get that at home. :-)
19:01
<ogra>
yes and i dont drink it here either :)
19:01
<rjune_>
LOL
19:01
<vagrantc>
dark.
19:02
<ogra>
dark ?
19:02
<rjune_>
I like becks dark
19:02
<ogra>
ah
19:02
thats only exported, you dont get it in germany
19:02
<vagrantc>
y'all talking about drinks, all i can say is "dark."
19:02
<rjune_>
oh, sorry to hear that
19:02
<ogra>
i dont think it matches the regulations
19:03
<vagrantc>
beer purity laws of 18somethingorother?
19:03
<ogra>
rather 16somethingorother :)
19:04
or even 14somethingorother
19:04
its pretty old
19:04
if there is sugar in it you cant call it beer ;)
19:04
hop, malt, water ...
19:07
<rjune_>
all beer has sugar in it.
19:08
<ogra>
as a base ingredient ?
19:08
german beer doesnt
19:08
<rjune_>
yes it does.
19:08
<ogra>
most american beer i know does though
19:08
<rjune_>
what do you think the yeast eats to make alcohol?
19:08
malting the barley turns the starches in the seeds to sugar
19:08
<vagrantc>
the ambient sugars in the initial ingredients?
19:09
<ogra>
right
19:09
glucosis from the malt
19:09
glucose ?
19:09
the boiling sets it free
19:10
<rjune_>
and becks claims that all their beer's adhere to Reinheitsgebot.
19:10
not the boiling, steeping is more like it
19:10
~ 150F for a while
19:11
the boiling comes after
19:11
<ogra>
wow
19:11
i just looked it up
19:12
it exists since 1156
19:12
<rjune_>
yeah, you germans take your beer *VERY* seriously.
19:12
<ogra>
though written as a law in 1487
19:12
and established german wide in 1493
19:14
so its malt, hop, yeast and water
19:15
<rjune_>
and amended to allowed cultured yeast in the late 1900s
19:15
specifically barley malt
19:15
the Reinheitsgebot killed off a lot of brewing traditions.
19:15
<ogra>
there are sorts that allow sugar apparently ... the wheat beers may have a certain minimal amount
19:16
<rjune_>
that's technically not compliant though
19:17
<ogra>
the yeast used in these sorts requires it
19:17johnny has joined #ltsp
19:17
<ogra>
but in the common german plsener no sugar is allowed
19:17
<rjune_>
yup.
19:17
but no wheat is allowed either.
19:18
<ogra>
indeed :)
19:18
<rjune_>
I've got most of the stuff to do a batch of beer from grain.
19:18johnny is now known as Guest99531
19:18
<ogra>
its intresting that all the sorts they list in the sugar allowed sorts are fully local
19:18
Kölsch, Alt, Weizenbier, Bönnsch, Degraa, Gose, Dampfbier, Berliner Weiße, Ale, Lüttje-Lage-Bier
19:19
the last one is from where i come from
19:19
and inst used anywhere outside the area
19:19
same goes for Kölsch in germany
19:19
alt is also very local
19:20
but you at least find it somewhere else
19:20
i never heard of "Bönnsch, Degraa, Gose, Dampfbier" :)
19:20
and have never seen german ale in germany
19:20* stgraber just installed Jaunty on his laptop, now the boring task of reinstalling only the useful bits of my previous install
19:20
<rjune_>
heh
19:21Guest99531 is now known as johnny
19:22
<ogra>
Berliner Weiße is pretty special ... you only drink it with woodruff or strawberry syrup
19:26japerry has quit IRC
19:30Nubae-laptop has joined #ltsp
19:30
<Nubae-laptop>
/join #edubuntu
19:31
oops
19:39nubae has quit IRC
19:45japerry has joined #ltsp
19:47J45p3r__ has joined #ltsp
19:49J45p3r__ has left #ltsp
20:01Ahmuck has quit IRC
20:35
<dberkholz>
warren, ogra, vagrantc: seen this initramfs-related thing, dracut? http://lwn.net/Articles/311890/
20:36
<warren>
dberkholz: yes
20:36
dberkholz: entirely new design, hopefully will be worked on at kernel.org
20:36
<dberkholz>
heh, of course warren's seen it.
20:36
Jeremy Katz <katzj@redhat.com>
20:36
<warren>
I haven't seen it, I only heard about it.
20:37* dberkholz mentions it to our releng guy
20:50
<johnny>
yay.. drop the mkinitrd!
20:53
<warren>
he's been thinking about how to redo it from scratch for years
20:55
<johnny>
gentoo's initramfs is particularly bad.. it loads all the modules in sequence.. which make the boot take forever..
20:55
it needs to do it in parallel
20:55
using udev of course.. seems the smart way. except for certain devices..
20:56
<warren>
Ryan52: progress of what?
20:57
<Nubae-laptop>
johnny: u tried empathy?
20:57
<Ryan52>
warren, I still don't know...:)
20:57
<Nubae-laptop>
I installed from intrepid ppa and its very nice, foundations of good collaboration
20:58
<warren>
I obviously haven't tried empathy. =)
20:58
<johnny>
no.. empathy has very poor xmpp support
20:58
i use gajim still
20:58
nobody seems interested in programming a power users version
20:58
and writing python is faster
20:59
altho it seems to me
20:59
that cross platform apps are going to start disappearing from linux..
20:59
start disappearing from common usage that is.. by linux users
20:59
<warren>
wait, why?
20:59
<johnny>
because of the amazing amount of integration possible
21:00
with gnome-shell, dbus, hal, packagekit, etc
21:00
<warren>
pidgin hasn't disappeared
21:00
<johnny>
not yet
21:00
<warren>
and it uses some of that
21:00
<Nubae-laptop>
well actually telepathy is now integrated into gnome
21:00
<johnny>
cuz empathy still sucks compared to it :)
21:00
<warren>
does empathy use libpurple?
21:00
<Nubae-laptop>
and empathy uses that, so its very integrated already
21:00
<johnny>
for telepathy-haze
21:00
transport
21:00
<Nubae-laptop>
no
21:01
it uses all of them
21:01
<johnny>
telepathy-haze is what it is called
21:01
<warren>
so pidgin isn't going away
21:01
<johnny>
libpurple isn't going away
21:01
at least for now
21:01
give it a few years..
21:02
<Nubae-laptop>
it uses most imporantly salut for avahi based coms
21:02
<johnny>
link local messaging
21:02
<Nubae-laptop>
and all the major coms protocols
21:02
yep
21:02
<johnny>
it's really sad tho..
21:02
<Nubae-laptop>
sends files through it too
21:02
sad why?
21:02
<johnny>
because telepathy is too lcd
21:02
<Nubae-laptop>
lcd?
21:03
<johnny>
lowest common denominator
21:03
between the various protocols
21:03
<Nubae-laptop>
well, its just very open and used a lot
21:03
is that a bad thing?
21:03
<johnny>
well.. supporting more xmpp features would be nice :)
21:03
<Nubae-laptop>
its the heart of collab software, the reason we can get abiword and inkscape and co sharing
21:04
they are seriously working on that... just takes time
21:04
<johnny>
inkscape had support before telepathy
21:04
<Nubae-laptop>
should replace pidgin in jaunty
21:04
<johnny>
i think people will be upset over that
21:04
<Nubae-laptop>
it uses telepathy for its support
21:04
<johnny>
not enough plugins..
21:04
not before telepathy existed :)
21:04
or was really usable
21:05
<Nubae-laptop>
well, then it used something non standard
21:05
<johnny>
it was standard
21:05
<Nubae-laptop>
dont u think its good telepathy is in gnome now?
21:05
<johnny>
it was just an xmpp lib..
21:05
we'll see
21:05
when i can replace gajim :)
21:05
<Nubae-laptop>
ok, so u want them to replicate the wheel every time
21:05
;-)
21:05
<johnny>
no i don't
21:05
gajim is great
21:05
i wish it was a telepathy client instead
21:05
and worked on fixing telepathy in process
21:06
<Nubae-laptop>
well, it can still do that, no?
21:06
<johnny>
but.. cross platform prevents such thing
21:06
thus why pidgin won't use telepathy.. but telepathy will use libpurple ..
21:06
<Nubae-laptop>
loads of apps are working on using telepathy to share stuff and collaborate, I think its great
21:07
especially on the mobile side
21:07
<johnny>
hmm.. i wonder if fedora's abiword
21:07
has the support built in
21:07
<Nubae-laptop>
I think so yes
21:08
<warren>
https://fedoraproject.org/wiki/Initrdrewrite
21:08
I added some of the debian/ubuntu requirements
21:08
<Nubae-laptop>
in fact fedora is leading the way cause of olpc
21:08
the xos run on Fedora
21:09
<warren>
why is this because of OLPC?
21:09
maybe we're leading because we suck?
21:09
Ubuntu can release it first and take the credit.
21:09
<johnny>
nonsense warren
21:09
<warren>
I'm just joking.
21:10
<Nubae-laptop>
well its a nice marriage... though I still think Ubuntu is more straight forward in its ways :-p
21:12
sound and media apps are now doing collab stuff, so u can share media easily across networks and the net, I think thats pretty neat
21:20
<warren>
Ryan52: I don't understand what those patches do
21:20
Ryan52: and it will likely make it difficult to redo the ldm layout patch?
21:20
<Ryan52>
which patch? I made 2 separate commits in the last hour..
21:23
what do you mean the ldm layout patch? the k12linux one? it doesn't even touch any of that code..
21:24
<warren>
Ryan52: oh
21:24
ok
21:24
Ryan52: the penultimate patch
21:25
<Ryan52>
that's on ldm-dialog
21:25
add's the same functionality of "zenity --progress"
21:28
<warren>
oooh
21:34davidj has quit IRC
21:39hersonls has quit IRC
22:01CaScAdE^FarAway has joined #ltsp
22:18CaScAdE^1arAway has quit IRC
22:28
<Ryan52>
does anybody know if launchpad is case sensitive with the email address?
22:29
I've been using my email with caps for a while, but then when I learned that email addresses (technically) are case sensitive I switched to all lowercase to be less confusing, but I'm scared to change my launchpad/bzr..
22:29
<johnny>
i've used lowercase the entire time.. so no idea..
22:37rcy has joined #ltsp
22:51warren has quit IRC
22:53warren has joined #ltsp
22:56alkisg has joined #ltsp
23:04wiscados has joined #ltsp
23:05
<wiscados>
is a thin-client like a virtual machine running on the server?
23:05
<johnny>
no
23:06
<wiscados>
does the OS run on the local machine?
23:06
<johnny>
no
23:06
well .. AN os runs on the local machine
23:06
to get it to the point where it can log into the server
23:06
the the users on the server.. are the same as any other users
23:06
just as if they were to be logged in directly to the machine
23:07
<wiscados>
oh, I see
23:07
like a mainfram
23:07
like a mainframe
23:07
<johnny>
uhmm.. sorta i guess
23:07
standard unix stuff here..
23:07
multi user system
23:07
i don't consider my laptop a main frame :)
23:07
<wiscados>
hehe, no. I guess not
23:08
<johnny>
my laptop could be a thin client server
23:08
it'd be slow..
23:08
but it would work
23:09
<wiscados>
is it possible to log in on one thin-client, and then login with a different computer and see the same desktop?
23:09
<johnny>
no.. gnome nor firefox likes that..
23:09
well.. i guess you could SEE the same desktop
23:09
with something like vnc
23:09
you wouldn't be logging in as that user per se..
23:10
<wiscados>
ok
23:10
<johnny>
but a similiar technology is at work with italc
23:10
used in conjunction with ltsp usually
23:10
if you had a class full of students.. you could remote control their mouse and stuff
23:11
and see what they see
23:11
<wiscados>
could you login as the same user twice, but with different x-sesions?
23:11
<johnny>
why are you asking?
23:11
<wiscados>
yeah, I recently saw iTALC, it's pretty cool
23:11
just curious
23:11
<johnny>
focus on the goal you're trying to achieve instead of how to get there
23:11
firefox would not like it at all..
23:11
as one example of a program
23:12
but if you were using vnc/italc and the like..
23:12
you could certainly start a firefox for them
23:12
or something like that
23:12
i've never personally used it
23:12
i use ltsp in an internet cafe environment
23:12
not in an educational one
23:12
<wiscados>
I'm just trying to see what the possibilities are with a thin-client are. I have so many ideas
23:13
<johnny>
most things that people do on their computers.. can be done with a thin client
23:13
except sadly.. cd burning (atm at least)
23:13
<wiscados>
I'm think about things you can't do on a normal computer
23:13
<johnny>
like what?
23:14
btw.. ltsp uses standard unix stuff everywhere..
23:14
ltsp is just glue
23:14
to make it all work in conjunction
23:14
<wiscados>
Like, for example if one computer was hooked up to a projector, and than students could display someting on the projector from their thin.clients
23:14
<johnny>
sure.. but you wouldn't even need thin clients for that
23:15
<wiscados>
yeah, you probably could just use vnc
23:15
now that I think about it
23:15
<johnny>
lots of options really
23:15
like shared whiteboarding ..
23:15
<wiscados>
What about colaborating on a document, is that possible
23:15
<johnny>
sure.. gobby can do it
23:16
or inkscape..
23:16
or abiword
23:16
inkscape for pictures that is..
23:16
<wiscados>
opensource is so great...
23:16
everything is already done
23:16
<johnny>
there is very little specific about ltsp in ltsp :)
23:17
most of the work was done for us
23:17
with standard tools
23:17
the things that are ltsp specific
23:17
<wiscados>
hehe, thanks to unix heritage, I guess
23:17
<johnny>
is a generated client image so they can boot
23:17
which uses apt,kickstart, or whatever for that distro
23:17
that includes ltspfs.. which is a fuse system that allows local devices to work
23:18
jetpipe to connect to local printers
23:18
and ldm, which is a display manager that is basically a thin front end to X ssh forwarding or xdmcp
23:19
<wiscados>
ok, so there is somethin in the ltsp package...
23:19
<johnny>
sure..
23:19
as i said.. glue :)
23:19
ties all the tools together
23:19
<wiscados>
so, ltsp is made from horses then
23:19
<johnny>
yuck
23:19
not that kind of glue i ihope
23:19
the nice ethical glue :)
23:20
<wiscados>
code-glue
23:20
<johnny>
and yeah.. it's all cuz of unix
23:20
<wiscados>
god bless those hairy hackers..
23:20
<johnny>
truly
23:36
<Ryan52>
warren, what are you talking about? My gdm asks me if I want to change my default..
23:36
<warren>
Ryan52: what version of gdm?
23:37
Ryan52: you probably have an old version
23:37
<Ryan52>
2.20.7
23:37
<warren>
gdm-2.24.0-12.fc10.x86_64 here
23:37
<Ryan52>
ah.
23:37
<johnny>
warren, .. isn't fedora the only distro shipping 2.24 ???
23:37* Ryan52 growls at Debian for being behind
23:37
<johnny>
Ryan52, don't growl too hard. you should be at 2.22.. but not 2.24
23:38
<Ryan52>
oh :)
23:38
<johnny>
ubuntu still ships 2.22
23:40
gentoo as well
23:40
not sure about suse
23:41
Ryan52, big story behind it..
23:41
<dberkholz>
depends on what you mean by shipping, i guess
23:41
<johnny>
lol
23:41
<dberkholz>
2.24 has been in testing for a while
23:41
<johnny>
really?
23:41
must have missed that
23:41
<dberkholz>
gnome-2.24.0: Thu Oct 23 08:45:28 2008
23:41
<johnny>
aha.. really recent!
23:41
like.. waay after.. i wonder how many patches they had to apply
23:42
oh.. wait.. i'mc razyin
23:42
i'm thinking about april
23:42
dberkholz, .. gdm?
23:42
i was talking about gdm.. not gnome
23:42
dberkholz, .. so i went ahead and started using funtoo stage 3s
23:42
unlike gentoo releng..
23:42
<dberkholz>
oh, that's been chillin' in the gnome overlay
23:43
<johnny>
daniel will actually create the symlinks i askedo
23:43
asked for*
23:43
or gentoo infra.. whatever
23:43
<dberkholz>
johnny: fyi, i have a releng announcement waiting to get posted on the homepage.
23:43
<johnny>
what's that?
23:43
see.. daniel will do this for me
23:43
<dberkholz>
something about the automated builds starting
23:43
stuff is on the mirrors already
23:44
<johnny>
http://www.funtoo.org/linux/x86/stage3-x86-current.tar.bz2
23:44
which is really useful for ltsp
23:45
he makes stages for all many gentoo targets..
23:45
<dberkholz>
i know
23:48
<johnny>
dberkholz, can you get some symlinks like that?
23:49
<dberkholz>
is it a good idea to be pointing to files we haven't tested?
23:49
<johnny>
for now yes
23:49
until there are tarballs
23:49
i think it is fine to use while still using bzr
23:50
how about you persuade the genkernel guys to include a patch so i can provide an external modules_load .. so we can actually release something :)
23:50
i seem to be unsucessful
23:50
all i get is.. robbat did a patch.. but we want to do it another way
23:52
if stuff is marked stable tho.. shouldn't it be stable?
23:53
if gentoo can provide stage3 that are more up to date, and also include portage 2.2 and openrc.. then we might actually be able to pin a stage
23:53
otherwise i have to hardcode in manual blocker resolutions
23:54
<dberkholz>
johnny: 2.1.6 has pretty much everything from 2.2 but sets
23:54
it's set to go stable soon
23:54
<johnny>
really? didn't know that
23:54
neat
23:54
<dberkholz>
i pinged agaffney about your genkernel thing
23:54
<johnny>
ok.. well lemme know when that tarball goes stable
23:54
<dberkholz>
if he will give details somewhere about the way they want it, we can get a patch together
23:54
<johnny>
err a tarball with that.. goes stable
23:54
but we still need openrc
23:55
and that will complete everything
23:55
<dberkholz>
stages are gonna be ~weekly so it'll be around a week after
23:55
who knows irt openrc. weird crap going on there.
23:55
<johnny>
really?
23:55
for example?
23:55
maybe gentoo should switch to upstart :)
23:55
hehe
23:56
<dberkholz>
upstream no longer uses gentoo, for the most part
23:56
<johnny>
hasn't it been that way since the beginning?
23:56
well.. not THE beginning
23:56
for for the past year
23:56
<dberkholz>
it's getting worse
23:56
<johnny>
oh.. well isn't that super..
23:56
:(
23:56
any chance on upstart then?
23:57
<dberkholz>
it's a combination of him being a little weird and the gentoo maintainer having some issues dealing with people in a polite way
23:57
<johnny>
lol.. cardoe..
23:57
<dberkholz>
weird irt just not understanding the needs of distros despite working in gentoo for years
23:58
<johnny>
it seems smart for gentoo to start using thinigs that seem to be heading towards standardization
23:58
instead of inventing everything
23:58
<dberkholz>
gentoo wants to use the best solution
23:58
being a standard doesn't make it good
23:59
<johnny>
except when being the best gets in the way..
23:59
the perfect being the enemy of the good and all
23:59
<dberkholz>
i didn't say anything about an exception. =P
23:59
<johnny>
lol