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


Channel log from 19 October 2008   (all times are UTC)

00:00indradg has quit IRC
00:32nubae1 has quit IRC
00:33nubae has joined #ltsp
00:33makghosh has joined #ltsp
00:55mikkel has quit IRC
00:58makghosh has quit IRC
01:07basanta has joined #ltsp
01:10Subhodip has joined #ltsp
01:22nubae1 has joined #ltsp
01:23nubae has quit IRC
01:32captain_magnus has quit IRC
01:33captain_magnus has joined #ltsp
01:36captain_magnus has quit IRC
01:38captain_magnus has joined #ltsp
02:31indradg has joined #ltsp
02:36Subhodip has quit IRC
02:42captain_1agnus has joined #ltsp
02:48alkisg has joined #ltsp
02:48indradg_ has quit IRC
02:54captain_magnus has quit IRC
02:55captain_1agnus is now known as captain_magnus
02:56Tshital has quit IRC
02:57captain_magnus has quit IRC
02:57captain_magnus has joined #ltsp
02:58kamudhar has joined #ltsp
03:22kamudhar has quit IRC
03:26Q-FUNK has joined #ltsp
03:49Hyperbyte has quit IRC
04:34plamengr has joined #ltsp
04:42basanta has quit IRC
05:00ccherrett has quit IRC
05:08ccherrett has joined #ltsp
05:22ccherrett has quit IRC
05:30ogra has quit IRC
05:33siki has joined #ltsp
05:34ogra has joined #ltsp
05:39ccherrett has joined #ltsp
05:42Hyperbyte has joined #ltsp
06:02ccherret1 has joined #ltsp
06:12ccherrett has quit IRC
06:44warren has joined #ltsp
06:48warren has quit IRC
06:49K_O-Gnom has joined #ltsp
06:50alkisg has quit IRC
06:56warren has joined #ltsp
06:56plamengr has left #ltsp
07:03siki has quit IRC
07:21siki has joined #ltsp
07:27indradg has quit IRC
07:28indradg has joined #ltsp
07:41alkisg has joined #ltsp
07:46petre has joined #ltsp
07:46
<petre>
morning all
07:47
warren, ping
07:50Barbosa_ has joined #ltsp
07:54
<jammcq>
hey petre
07:54
<petre>
morning jammcq
07:54
<jammcq>
Im wondering if Warren will be here this morning. I think he's still in Japan
07:55
or travelling back
07:55
<warren>
i'm not here
07:55
<petre>
I think he's still there, but I don't know what time it is
07:55
<jammcq>
ah
07:55
<warren>
I return from Japan Oct 28th
07:55
<jammcq>
warren: you still in Japan?
07:55
wow
07:55
<warren>
it is almost 10PM here
07:55
<jammcq>
9 more days?
07:55
how's it going there?
07:56Q-FUNK has quit IRC
07:56
<warren>
very well, the big presentation was yesterday, it went really good
07:57
<jammcq>
oh, that's great
07:57
<warren>
prefecture and local government, educators from university and others were impressed
07:57
prefecture is roughly equivalent to "state"
07:57
<petre>
warren, was that for a school?
07:57
<warren>
petre: for ~60 schools
07:57
<petre>
wow
08:03Barbosa has quit IRC
08:05cyberorg has quit IRC
09:07mikkel has joined #ltsp
09:11cyberorg has joined #ltsp
09:14nubae1 has quit IRC
09:26nubae has joined #ltsp
10:03gooddoogles has joined #ltsp
10:04
<gooddoogles>
K12LTSP 5EL Will K12Linux support SATA drives? I'm thinking of upgrading.
10:14
<johnny>
hmm?
10:14
support sata drives where?
10:14
on the server?
10:14
sure
10:14
the ide and sata layers are the same in recent kernels
10:15
i'm not sure about the uuid labeling in /etc/fstab tho.. i don't know if you have to manually change them or not
10:15
that's just standard stuff, not at all related to ltsp
10:20
<petre>
gooddoogles, yes, it will support sata drives
10:21
both K12LTSP EL5 and K12Linux support sata drives
10:30
<johnny>
all el5 supports sata drives
10:34
<gooddoogles>
Sounds good.
10:34vagrantc has joined #ltsp
10:35
<gooddoogles>
When I installed K12LTSP 5EL the installer would not recognize the drives. I tried having it load all of the drivers that seemed to relate to SATA drives, but with no success. I ended up going out and buying an IDE drive to get my server to work.
10:36
<jammcq>
maybe an unsupported chipset or something. I've been using SATA for several years
10:36
<gooddoogles>
One of the computers my school gave to me to use as a server has spaces for two drives and I have two SATA drives. I'd like to set them up in RAID so that, if and when a drive fails, my system will keep going.
10:37
<jammcq>
good plan
10:37
using software raid? or does the motherboard have raid?
10:38
<gooddoogles>
I think the motherboard has some kind of RAID option.
10:38
I've seen it in the BIOS.
10:38
<jammcq>
my experience with the raid on motherboards has been bad
10:38
<gooddoogles>
Oh.
10:38
<jammcq>
Iv'e done software raid and I've used lots of 3ware raid controllers
10:39
the 3ware controllers are VERY well supported by Linux
10:39
<gooddoogles>
Should I set up software RAID when I partition the disks for the K12Linux install?
10:39
If the 3ware controllers are inexpensive, I might be able to get the school to buy me one.
10:39
<jammcq>
i've not installed k12 in a very long time, I don't know what options you have during install. but I'd certainly give it a try
10:40
<gooddoogles>
I've heard that hardware RAID is superior to software RAID.
10:40
I have another server computer also. Two weeks ago I set up a second server computer using dhcpd failover. However, it doesn't seem to be speeding up the lab. In fact, sometimes computers in the lab freeze for a few seconds. I think there might be some kind of misconfiguration in my network.
10:40
<johnny>
gooddoogles, there will be real clustering for ltsp soon
10:41
it exists for ubuntu, but not yet for redhat distros
10:41
<jammcq>
well, i've heard that software raid performs better, but I like hardware raid because I like the idea of abstracting that away from the operating system
10:41
<johnny>
jammcq, are you an apm guy then?
10:41
lol
10:41
<jammcq>
also, hardware raid allows for hot swapping
10:41
heh
10:42
<gooddoogles>
Because the network is 100 Mbps for clients (with 1 Gbps uplinks), I suspect that I would do better with one optimized server than two servers.
10:42
<jammcq>
depends on the number of clients
10:43
<gooddoogles>
There are three reasons I configured two servers (1) I wanted to see if I could do it, (2) I want to improve performance, and (3) I want to be able to quickly recover (within minutes) from a server failure.
10:43
Right now I have about 14 clients. I would like to eventually have about 30.
10:44
One of the things I want to have the kids do it Blender.
10:44
I mean "is" Blender.
10:44
<johnny>
jammcq, hot swapping works with sata in general
10:45
<jammcq>
johnny: yeah, it's part of the spec, but will Linux do the right thing when a new drive is plugged in?
10:46
<gooddoogles>
In case you are interested, here are pics of my lab: http://picasaweb.google.com/CramptonFamily/CramptonOpenSourceLab
10:46
I'm teaching at a middle school in Jacksonville, FL. It's a magnet school for math, science, and technology.
10:46
<jammcq>
wow, very impressive
10:46
<johnny>
jammcq, it should now
10:47
<rjune_>
gooddoogles: magnet?
10:47
<gooddoogles>
We have a lot of poor kids. A lot of them use improper English, but can factor reasonable large numbers in their heads. An interesting population.
10:47
<nubae>
gooddoogles: what are the client specs?
10:48
<johnny>
jammcq, i haven't had the hardware to test it myself, since my power supply doesn't have he the proper support or enough actual sata power connectors
10:48
i'm using a ide -> sata adaptor atm
10:48
<rjune_>
interesting, gooddoogles state school?
10:48
<johnny>
jammcq, you can atually hot plug ram with the right hardware now
10:48
<gooddoogles>
A magnet school means that the parents had to apply to send their kids. There are no admission requirements; it's done through a lottery. However, all of the kids at least have parents who want them to go to a better school.
10:48
<nubae>
gooddoogles: if the client specs are high enough u could use local apps or fat client
10:49
<rjune_>
yeah, I just looked it up on wp
10:49
<nubae>
to run heavy duty stuff like blender
10:49
<gooddoogles>
The current clients are about 3-year-old Lenovo machines.
10:50
I'm not sure what the specs are exactly, but they boot up in about 30 seconds on the network. I disconnected the hard drives.
10:50
<jhutchins_lt>
I installed ltsp on a headless debian server and don't have a desktop environment available on my client. I added kde to the server, but there's still nothing avialble for the client.
10:50
<nubae>
how much ram, that'd be the important factor
10:50
<jhutchins_lt>
How do I make kde available?
10:50
<gooddoogles>
I'll have to look into that.
10:51
I'd like to replace the current thin clients with the one pictured that sits behind the LCD monitor.
10:51
It's a Nohrtec Microclient, Sr. It cost about $200 and has 512 MB of RAM. It's very unobtrusive and energy efficient.
10:52
<nubae>
hmm, well that would probably be good as a fatclient
10:52
or to run local apps
10:52
<gooddoogles>
Yes, it seems a waste to not use the clients' own processing power.
10:53
I tried using Edubuntu, but for some reason the system ran very very slowly after I added more than about 4 clients.
10:53
<nubae>
I see more advantage of running fatclient right now though... www.nubae.com there's an article there with a plugin you can download
10:53
<gooddoogles>
I did all kinds of troubleshooting and tried different settings and tried to reconfigure the network switch.
10:53
<nubae>
it has blender, kino, vlc, flash, scribus, etc preinstalled to the fat chroot
10:53
<gooddoogles>
Eventually, I installed K12LTSP and it "just worked" out of the box.
10:54
<nubae>
ah k12ltsp...
10:54
u'd need to modify the script to work on that... right now it is a Ubuntu plugin...
10:55
<gooddoogles>
Oh, cool. I "StumbledUpon" one of your webpages a while ago. I don't remember which one, but I gave it a thumbs up.
10:55
<nubae>
thanks... one of my next tasks is getting the plugin to work on other distros...
10:56
if only there were more hours in the days =)
10:57
<gooddoogles>
The Computer Club at my school is interested in building a supercomputer, similar to the "Budget Supercomputer" that someone built with four motherboards.
10:57
I was thinking, perhaps, I could attach it to my network and use it for rendering the Blender artwork that students create.
10:58
<nubae>
u might want to look at cluster-ltsp, though it is more for load balancing
10:58
<gooddoogles>
I already had the students try running Blender and the terminals seemed to do fine just rotating and zooming around the design scene. We haven't tried any rendering yet, however.
10:58
<nubae>
https://wiki.stgraber.org/LTSP-Cluster/Documentation/TestSetup
10:59
<alkisg>
gooddoogles, in one of my schools I also have 100Mbps switch. I saw great performance increase using link aggregation (3 NICs on the server all connected to the same switch => load balancing).
11:00
<nubae>
alkisg: u have that written up some place?
11:00
<alkisg>
nubae, yes, but in Greek! :(
11:00
<gooddoogles>
That sounds like a good idea.
11:00
<rjune_>
anybody in southwest, GA?
11:00
<alkisg>
I'm thinking of starting an english blog, to gather whatever new I tested...
11:00
<nubae>
sure do it... its not only helpful but fun
11:01
<gooddoogles>
I figure that, if I get the RAID configuration going, the likelihood of server downtime will be small.
11:01
I could take the NIC out of the other server (as well as the RAM) to make one more powerful server.
11:02
Then, if something happened to the server motherboard, I could move the RAM and hard drives to the other server.
11:02
That would probably put the system down for 10-15 minutes max.
11:03
rjune_, I'm in Jacksonville, FL.
11:03
<rjune_>
yeah, I know you're close
11:03
I may be looking at moving to Albany, GA soon
11:03
<gooddoogles>
I don't know this area real well. My wife, kids, and I moved here 6 months ago.
11:04
<alkisg>
nubae, I do have a forum where many greek teachers talk, so it'll be more of a burden than fun... but I suppose it may help others, so I may do it anyway.
11:04
<gooddoogles>
We are here because my sister (and her young family) moved here about 3 years ago.
11:04
<alkisg>
gooddoogles, 2 servers are always better than one!
11:04
<gooddoogles>
Why's that, alkisg?
11:04
<jhutchins_lt>
Well, I guess I'll remove ltsp and reinstall it.
11:05
<gooddoogles>
I figure that administration is so much easier with only one server.
11:05
<alkisg>
well, load balancing would be much more effective. Administration would be simpler, yes...
11:06
<gooddoogles>
The school district is paranoid about putting my new lab on their network, so I'm currently running an isolated network. One of the things that really irks me is I have to bring the server(s) home to upgrade them.
11:06
Because I'm isolated, I have to supply /home for the kids.
11:07
One of my biggest challenges was figuring out how to sync /home between the servers. I ended up writing a script that does it with rsync.
11:07
<vagrantc>
jhutchins_lt: before you re-install ...
11:07
<jhutchins_lt>
vagrantc: Yes?
11:08
I mean, I'm three minutes into the reinstall already, but...
11:08
<vagrantc>
jhutchins_lt: could you run /usr/sbin/ldminfod and paste the output to the pastebot:
11:08
!pastebot
11:08
<ltspbot>
vagrantc: "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.
11:08
<vagrantc>
jhutchins_lt: reinstalling ltsp isn't likely to actually fix your problem.
11:08
<jhutchins_lt>
vagrantc: Lovely.
11:08
vagrantc: What should I do?
11:09Subhodip has joined #ltsp
11:09
<jhutchins_lt>
I mean, I was thrilled when I saw the PXE boot take off and load the login manager. I stare at those stupid things all day at work, building systems.
11:09
However, there is nowhere to go from the login page.
11:10
<vagrantc>
jhutchins_lt: please paste the output "/usr/sbin/update-alternatives --display x-session-manager ; /usr/sbin/update-alternatives --display x-window-manager" to the pastebot.
11:10
jhutchins_lt: what do you mean "there is nowhere to go" ?
11:10
<jhutchins_lt>
vagrantc: Well, like I said, now five minutes into reinstalling, so it'll be a few minutes.
11:10
<vagrantc>
jhutchins_lt: re-installing the whole server?
11:10
<jhutchins_lt>
I mean there is no GUI/Desktop Environment available.
11:11
vagrantc: aptitude install ltsp-server-standalone
11:11
vagrantc: Having un-installed it and removed /opt/ltsp
11:11
<vagrantc>
jhutchins_lt: have you uninstalled kde ?
11:11
<jhutchins_lt>
No.
11:11
<vagrantc>
jhutchins_lt: then please run the update-alternatives commands i mentioned above.
11:12
<gooddoogles>
Thanks for your conversation and ideas. I have a gmail account. You can reach me at SteveSings. I have to get back to grading projects.
11:12
<jhutchins_lt>
Ok, ltsp-build-client just completed, let me see what we get.
11:12gooddoogles has quit IRC
11:13
<jhutchins_lt>
Looks like this one will work.
11:13
<ltsppbot>
"jhutchins" pasted "Output" (9 lines) at http://pastebot.ltsp.org/52
11:14
<jhutchins_lt>
Oh how nice, it posted it for me.
11:14
<vagrantc>
!vers
11:14
<ltspbot>
vagrantc: Error: "vers" is not a valid command.
11:14
<vagrantc>
!ver
11:14
<ltspbot>
vagrantc: "ver" is to get version info on debian/ubuntu (please use the pastebot): COLUMNS=200 dpkg -l 'ltsp*' | awk '/^ii/{print $2}/^ii/{print $3}' ; COLUMNS=200 dpkg --root=/opt/ltsp/i386 -l 'ltsp*' ldm | awk '/^ii/{print $2}/^ii/{print $3}'
11:15
<vagrantc>
jhutchins_lt: please paste that ^^^ too
11:17
<ltsppbot>
"jhutchins" pasted "ltsp-server 5.1.10-1~40.etch.0" (10 lines) at http://pastebot.ltsp.org/53
11:17
<jhutchins_lt>
Client is starting up now.
11:17
<vagrantc>
looks like you only partially followed the howto.
11:17
<jhutchins_lt>
Or not, no tftp...
11:17
Eh?
11:17
What did I miss.
11:18
<vagrantc>
you've installed the backported packages on the server, but the original debian etch packages in the chroot.
11:19
<jhutchins_lt>
Heh. I followed http://wiki.debian.org/LTSP/Howto exactly so far.
11:19
<vagrantc>
ltsp-build-client \
11:19
--extra-mirror "http://pkg-ltsp.alioth.debian.org/debian etch-ltsp-backports main" \
11:19
--apt-key /etc/apt/trusted.gpg
11:20
jhutchins_lt: you ran that command?
11:20
<jhutchins_lt>
Ok, so to re-do it, do I remove /opt/ltsp again?
11:20
<vagrantc>
yup.
11:20
<jhutchins_lt>
No, I just ran ltsp-build-client
11:21
<vagrantc>
jhutchins_lt: yeah, step 5 in "backported packages for etch"
11:21
jhutchins_lt: did you follow those instructions, or the ones above?
11:21
<jhutchins_lt>
Actually, I ran that the first time, but of course it errored out because I had already run ltsp-build-client.
11:21
Strangely, I followed the instructions sequentially.
11:22* jhutchins_lt waits impatiently for his third build.
11:22
<jhutchins_lt>
It would seem quite useful to have a rebuild option.
11:22
Really appreciate the help, by the way.
11:22
<johnny>
rebuild is the same as do over :)
11:23
<vagrantc>
rebuild would have to just delete the directory ... otherwise you run the risk of having misconfigured stuff in the chroot
11:26
<jhutchins_lt>
Woo!
11:26
Build complete
11:27
<vagrantc>
there's also a few options to facilitate re-building so you don't have to re-download all the packages ... some of them are mentioned in --extra-help
11:28
<jhutchins_lt>
Hm.... We're not getting net boot.
11:29
dhcpd's running...
11:29
<vagrantc>
configured?
11:29
jhutchins_lt: could you paste the dhcpd.conf ?
11:30
<jhutchins_lt>
Um, yeah, but it's tftp that's not running.
11:30
Hang on a sec.
11:30
<vagrantc>
well, tftp should normally be run from inetd
11:30
unless you configured it otherwise
11:30
<jhutchins_lt>
Sorry, I'm a redhat/mandriva guy, little clumsy on debian.
11:31
<vagrantc>
no problem
11:31
<jhutchins_lt>
there's no /etc/xinetd.d/
11:32
<vagrantc>
probably /etc/inetd.conf
11:33
<jhutchins_lt>
Yeah, there's no line for tftp for some reason.
11:33
<vagrantc>
dpkg -l 'tftp*' | egrep ^ii
11:34
i should really re-write wiki.debian.org/LTSP/Howto into separate sub-pages for each release.
11:35
as the instructions are a little different for etch, etch backports, and lenny
11:35
not hugely so, but enough thta i always get some of the same problems...
11:35
<jhutchins_lt>
Yeah, it worked fine first time...
11:36
Could you paste the tftp line for inetd.confi?
11:36
.conf I mean.
11:36
<vagrantc>
jhutchins_lt: it depends on which tftp you have installed....
11:37
<jhutchins_lt>
hpa, from the ltsp install.
11:37
<vagrantc>
jhutchins_lt: by default, it should pull in tftpd-hpa
11:37
jhutchins_lt: dpkg-reconfigure tftpd-hpa
11:37
<jhutchins_lt>
Which it did, it's installed, but doesn't appear to be running. K.
11:38
Ok, do I need to restart inetd?
11:38
<vagrantc>
grep tftp /etc/inetd.conf
11:38nubae has quit IRC
11:39
<jhutchins_lt>
Still nothing there.
11:39
<vagrantc>
jhutchins_lt: make sure it's in there ... then restart it: invoke-rc.d openbsd-inetd restart
11:39nubae has joined #ltsp
11:40
<jhutchins_lt>
Crud. I've run dpkg-reconfigure four times now - toggled it on and off, still nothing in the inetd.conf file. Could you just paste the line and I'll add it manually?
11:41
<vagrantc>
jhutchins_lt: it seems like you have numerous things broken which, in years of doing this stuff, have always "just worked" ... so i
11:41
'm hesitant to work around the problem.
11:42
<jhutchins_lt>
Seriously? "which" is not part of default debian?
11:42
Years of doing it, minutes on this project.
11:43
<vagrantc>
"which" is part of debianutils, which is a required package on debian. you'd have to go through a lot of hacking to get it to not be installed.
11:44
<jhutchins_lt>
It's not in the chroot though, especially after the chroot the shell's "in" has been deleted twice.
11:45
tftp stream tcp nowait root /usr/sbin/tcpd /usr/sbin/in.tftpd -- that should do it.
11:45
<vagrantc>
jhutchins_lt: sounds like you have a very broken system.
11:46
jhutchins_lt: 'shell's "in"' ?
11:46
<jhutchins_lt>
vagrantc: No, not really. Only slightly broken. I'm running screen on the server, I had a window where the shell was in the chroot, which has since been deleted twice, once when I reinstalled ltsp, and once when I rebuilt it.
11:47
I was looking for the tftpd daemon for the inetd.conf, and which tftpd said no such command from there.
11:47
<vagrantc>
jhutchins_lt: ah, that shell is likely to not work very well ...
11:47
<jhutchins_lt>
Once I exited that chroot, it was fine.
11:47
<vagrantc>
right
11:47
<jhutchins_lt>
Yeah, no kidding. Rebooting the client agian after adding the line and restarting inetd, let's see if it works...
11:47
<vagrantc>
so where are we?
11:48
<jhutchins_lt>
Blasst.. Still stuck at tftp.
11:48
<vagrantc>
that line might work, although it's not the same as the default ...
11:49
jhutchins_lt: am i mistaken, or did you actually have it sucessfully booting before?
11:49
<jhutchins_lt>
Yes, it was.
11:49
<vagrantc>
jhutchins_lt: and, if so, what changed since then?
11:49
<jhutchins_lt>
At least to the GUI login.
11:50shogunx has quit IRC
11:50
<jhutchins_lt>
1) aptitude remove ltsp-server-standalone, rm -f /opt/ltsp/ ; 2) aptitude install ltsp-server-standalone; ltsp-build-client; rm -f /opt/ltsp/; <more complex build client command to do updates>... That's pretty much it, other than this messing with trying to get tftp running.
11:52
<vagrantc>
none of that should have broken tftp
11:52
... unless aptitude auto-removes stuff ...
11:52
but even then ... dpkg-reconfigure tftpd-hpa should take care of it.
11:54
jhutchins_lt: grep TFTP /etc/inetd.conf
11:54shogunx has joined #ltsp
11:54
<jhutchins_lt>
#:BOOT: TFTP service is provided primarily for booting. Most sites
11:55
<vagrantc>
right ...
11:55artista_frustrad has joined #ltsp
11:55
<jhutchins_lt>
I ran it from the command line and it appears to have succeeded.
11:55
Woo! White screen over there...
11:56
It's the login screen. Guess it would help if I chroot and add some users...
11:56
<vagrantc>
update-inetd --group BOOT --add "tftp dgram udp wait root /usr/sbin/in.tftpd /usr/sbin/in.tftpd -s /var/lib/tftpboot"
11:57
running that might reveal why dpkg-reconfigure wasn't working ...
11:57
and/or add it properly...
11:57
jhutchins_lt: no need to add users to the chroot
11:57
jhutchins_lt: everything basically happens on the server.
11:59shogunx has quit IRC
12:01feinbein has joined #ltsp
12:02
<jhutchins_lt>
so add the users to the server, not the chroot?
12:02
<johnny>
yes
12:02shogunx has joined #ltsp
12:02
<johnny>
just don't try to login as a user you're already logged in as
12:02
via gui
12:02
<jhutchins_lt>
Um, no, I get a login screen.
12:03
Oh, sorry, mis-parsed.
12:03
<vagrantc>
jhutchins_lt: if you're logged into the server directly as a particular user, don't log in form a thin-client as the same user.
12:03
<johnny>
gnome doesn't like it.. and firefox doesn't like it .. and probably a bunch of other apps
12:04Subhodip has quit IRC
12:04
<jhutchins_lt>
Yeah, multiple gui logins to the same home directory have "unexpected consequences".
12:04
I've tried that over X forwarding and using multiple X sessions on the same box.
12:06
WOOT! Viola, KDE.
12:06
Thanks much for your help.
12:07
Now I need auto-login of the default user, and to do configuration for the screen and apps and such.
12:07
How do I add apps? Just install 'em on the server?
12:07
<johnny>
default user?
12:07
yes
12:08
<jhutchins_lt>
There will be two users, a default and an alternative (or two).
12:08
<johnny>
a default user per computer?
12:08
if you name the computers by hostname and give them also that password
12:08
then you can set autologin for each automatically
12:08
<jhutchins_lt>
With a fullblown workstation installed to local HD, I jused kcontrol to set auto-login for the one I wanted as defualt.
12:08
Yes, I saw that you can set it by mac ID.
12:08
<johnny>
and hostname
12:09
<jhutchins_lt>
Set it by hostname, which is set where?
12:09
<johnny>
same notation
12:09
[hostname]
12:09
<jhutchins_lt>
Ah.
12:09feinbein has quit IRC
12:09
<jhutchins_lt>
I think that would assume some other method than ltsp to get the hostname.
12:09
<johnny>
yeah.. dns and dhcp
12:10
<jhutchins_lt>
Weh have here a single, diskless workstation.
12:13
Users can do things like set up their desktop resolution, screensaver, etc., right?
12:17
<stgraber>
nubae: hi, I just saw you pasted the Testsetup wiki page some hours ago, it's been renamed to: https://wiki.stgraber.org/LTSP-Cluster/Documentation/OpenVZSetup
12:17
<jhutchins_lt>
vagrantc: In the "known issues" section, as well as a couple of others, it's not clear what environment you're supposed to be in, ie server command line, chroot, client command line.
12:19
I had formerly shared directories on the server with clients via samba. How can I add these directories to the ltsp client? Seems like a symlink somwhere would do it nicely.
12:22
<nubae>
stgraber: I'll update my records... thanks :-)
12:24
<vagrantc>
jhutchins_lt: i write the documentation to be explicit such that you are never in the chroot unless it says so.
12:24
jhutchins_lt: so all the paths are full paths (assuming default install on i386)
12:29
jhutchins_lt: why do you want to share directories with the clients?
12:30
jhutchins_lt: the whole user's session is on the server. so when the user logs in, whatever's mounted on the server will be available to a user logged in via a thin-client.
12:30X0d_of_N0d has quit IRC
12:31X0d_of_N0d has joined #ltsp
12:40rezon8 has joined #ltsp
12:49r3zon8 has quit IRC
13:17
<jhutchins_lt>
Symlinked the shares into the users home directories.
13:17
Mkay.
13:18
So I can make changes from the cleint just as if I were in a normal session on the server.
13:18
Need to get sound working, alsa and pulseaudio weren't installed on the server.
13:18
<jammcq>
that's the thing, you ARE in a normal session on the server
13:19
the client is just a terminal that gets you a session on the server
13:19
<nubae>
does anyone else thing sound is just way too low compared to a normal machine on a thing client?
13:19
think
13:20
<johnny>
nubae, on which distro?
13:20
intrepid should have the fix for it, as well as newer releases of fedora and opensuse i imagine
13:22
<vagrantc>
jhutchins_lt: you sure you have ltsp-server-standalone installed?
13:23
<nubae>
nope intrepid is still way too low
13:24
I've tested it on various machines and the only way it sounds ok is through external speakers... laptop speakers just won't do it, even though they are fine on a normal or fatclient
13:24
<vagrantc>
jhutchins_lt: you'll also probably want to read /usr/share/doc/ltsp-server/README.Debian.sound
13:24
<jhutchins_lt>
vagrantc: Yup.
13:25
Heh. As you said, it's just like you're running on the server... so shutdown -r now doesn't reboot the client...
13:25
<vagrantc>
correct.
13:25
<jhutchins_lt>
vagrantc: I don't think that sound instruction is current, alsaconf didn't create the conf file there.
13:26
<vagrantc>
jhutchins_lt: think of a thin-client as an extra monitor, mouse and keyboard for the server.
13:26
<johnny>
nubae, there's an lts.conf var to set the baselink i think
13:26
err baseline
13:26
<jhutchins_lt>
My main problem is that the documentation assumes that you have a full, working GUI on the server.
13:26
<vagrantc>
jhutchins_lt: there's no reason why it would.
13:27
jhutchins_lt: i guess the missing information is "add the following lines to /etc/asound.conf, or create /etc/asound.conf if it is not already present"
13:27
<nubae>
yeah I'm just saying the defaults are too lpow
13:27
low
13:27
<jhutchins_lt>
Yeah, that's more specific.
13:28
<vagrantc>
jhutchins_lt: though it's been a while since i've tested sound on etch.
13:28
<jhutchins_lt>
Hm. Server was "entering maintenance mode". Off to the basement I guess...
13:30nubae has quit IRC
13:30
<jhutchins_lt>
Yup, dropped to single user for some reason. Perhaps that's what Ctrl-Alt-Del does on debian.
13:31
So how does one properly reboot just the client?
13:31
<vagrantc>
no, ctrl-alt-del defaults to a simple reboot
13:31nubae has joined #ltsp
13:31
<vagrantc>
jhutchins_lt: log out and select reboot from the preferences menu
13:33bobby_C has joined #ltsp
13:33
<jhutchins_lt>
Not sure how I managed single user mode then, but oh well, careening on.
13:34
<vagrantc>
jhutchins_lt: you've had just enough quirks with this system i'd almost suspect some filesystem corruption or something like that
13:35
<jhutchins_lt>
No, really, it's all the result of installing ltsp directly on a headless server that had harly even had a local console login before.
13:35
99% of what I've done with that server was over ssh.
13:36
The other factor here is that as I said, I'm more familiar with RedHat and related distros than with Debian, although I've dabbled in kubuntu and managed a couple of Gentoo systems for a few years.
13:37
I would suggest a note that a working GUI desktop should be set up on the server prior to installing LTSP.
13:39
The other thing would be just a little more wordiness about the updated pacakges; that you need to rebuild the client by first removing the chroot, that setting the root password in the chroot is obsolete, etc.
13:39
Also, something about when the client needs to be rebuilt/updated, ie for new kernels, or modules, or whatever.
13:41
I would change the line "adduser fuse USER" to "for each ltsp user, adduser fuse <username>". Also explicitly say that ltsp users need to be local users on the server.
13:43ogra has quit IRC
13:43
<vagrantc>
jhutchins_lt: i don't see how installing over ssh would cause any of these problems.
13:43
jhutchins_lt: i do it all the time. almost exclusively.
13:43
<jhutchins_lt>
Hm. "NFS over TCP not available"... Something didn't restart. What's the equivalent of "chkconfig" in debian?
13:43
vagrantc: No, installing over ssh isn't the problem, the problem is that there was no GUI environment on the server, no sound, etc.
13:46
<vagrantc>
yes, ltsp-server-standalone should ensure that at the very least, a window manager was installed.
13:46
i've never quite figured out how to make a politically neutral choice on the default, though. :)
13:47
<jhutchins_lt>
vagrantc: I dunno, maybe a choice menu?
13:47
<vagrantc>
jhutchins_lt: what do you mean about "setting the root password in the chroot is obsolete" ?
13:48
jhutchins_lt: where would you put the menu?
13:48
jhutchins_lt: when have you ever needed a root password set?
13:48
<jhutchins_lt>
Ok, that's in a different howto.
13:49
I think that was probably for the non-updated packages.
13:49
<vagrantc>
jhutchins_lt: and you might note, it is a wiki.
13:49
jhutchins_lt: so you could fix it, and i'm subscribed to them, so i'll correct any factually incorrect changes.
13:49
<jhutchins_lt>
What's supposed to start NBD.
13:49
<vagrantc>
on debian, nothing.
13:49
<jhutchins_lt>
vagrantc: It's a deal!
13:49
NFS?
13:50
<vagrantc>
yes
13:50* vagrantc makes sure vagrantc is still subscribed to the updated pages.
13:50
<jhutchins_lt>
Ok, and does it run from inetd or as a daemon.
13:50
?
13:50
<vagrantc>
it runs as a daemon
13:51
<jhutchins_lt>
So what's the equivalent of chkconfig nfs-kernel-server on?
13:52
<vagrantc>
i honestly don't know the equivalent.
13:52gate_keeper_ has joined #ltsp
13:52
<jhutchins_lt>
It would set nfs to start at boot by default.
13:52
<vagrantc>
unless you've changed the default runlevel, you can look in /etc/rc2.d/ if the S*-nfs-kernel-server symlink is there.
13:54
<jhutchins_lt>
Nope. Must've turned it off at some point, wasn't using it before.
13:54
<vagrantc>
jhutchins_lt: you have anything in /etc/exports ?
13:56
<jhutchins_lt>
Yep, the ltsp stuff, per the howto.
13:56
<vagrantc>
jhutchins_lt: but you had configured it to not start?
13:56
<jhutchins_lt>
Daemon wasn't running at update for some reason. Like I said, I think I turned it off.
13:58ogra has joined #ltsp
13:59
<vagrantc>
ogra, stgraber: just uploaded ltspfs 0.5.5-1 to debian, should set the udev priority right based on which distro it's built on.
13:59
so should be syncable
14:02
<jhutchins_lt>
Yeah, def. something strange on the server. sysv-rc-conf crashes out if I try to toggle bind or tftp.
14:07
<vagrantc>
ah. sysv-rc-conf is supposed to be something sort of like chkconfig ?
14:07* vagrantc hasn't found a replacement since switching away from the rpm family 7-9 years ago
14:09
<vagrantc>
of course, i usually just remove stuff to have it not run.
14:09
haven't really made use of runlevels.
14:09johnny has left #ltsp
14:10johnny has joined #ltsp
14:10johnny is now known as Guest26626
14:13
<stgraber>
vagrantc: cool
14:14
<jhutchins_lt>
vagrantc: Actually, update-rc.d is the straight replacement for chkconfig, sysv-rc-conf gives you a text menu, but it's crashing when I try to turn certain things on & off.
14:15
<vagrantc>
jhutchins_lt: well, update-rc.d is a bit ... klunkly.
14:16
<jhutchins_lt>
Yeah.
14:16
Would apparently work better if the headers were more standard.
14:16
... just like chkconfig.
14:24Guest26626 is now known as johnny
14:25johnny is now known as Guest36191
14:26Guest36191 is now known as johnny
14:27
<johnny>
i like gentoo's init system much better
14:27
debian's is a total mess
14:32
<rjune_>
isn't debian's init system a sysV variant?
14:34
<johnny>
i'm sure it is
14:34
yes
15:00K_O-Gnom has quit IRC
15:01Egyptian[Home] has quit IRC
15:02marilyn has joined #ltsp
15:03marilyn is now known as jhutchins_ltsp
15:03Egyptian[Home] has joined #ltsp
15:03
<jhutchins_ltsp>
Still no joy on the sound. KDE Sound System just keeps trying to start it.
15:03
What other tools do I have to tweak it in ltsp?
15:04
<rjune_>
what are you trying to do ?
15:04
KDE audio needs to be piped over something else.
15:05
<jhutchins_ltsp>
Hi rjune_! Trying to get the sound on my diskless micro-atx ltsp client running.
15:05
<rjune_>
NAS or ESD both make good transports
15:05
<jhutchins_ltsp>
rjune_: Is there a howto somewhere?
15:05
<rjune_>
there is for ltsp4
15:05
<jhutchins_ltsp>
... which of course is totally useless for 5 I'd guess.
15:05
<rjune_>
if you are operating as a terminal rather then running everything remotely it should still apply
15:06
<jhutchins_ltsp>
This is so far just a stock installation from the debian howto, using the updated files.
15:06
<rjune_>
http://www.bravegnuworld.com/~rjune/ltsp/audio/HOWTO-arts.html
15:06
I know nothing about ltsp5
15:06
but I know a bit about ltsp4
15:06
if you have X providing the display, and all the work happening on the server
15:06
then that document should help
15:07
I would suggest using esound for the base transport
15:07
it's likely already in your distro and arts likely has support for it
15:07
I have to take a quick shower. I'll be back in 30 min or so
15:08
<jhutchins_ltsp>
I'll see how far I get, thanks.
15:08
<rjune_>
yup
15:13
<jhutchins_ltsp>
Where's my lts.conf file on 5?
15:16
<vagrantc>
what the .... ?!?! ... my ltspfs upload, for some bizarre reason, didn't include the orig.tar.gz
15:19radoeka has joined #ltsp
15:20radoeka has left #ltsp
15:21mccann has joined #ltsp
15:21
<jhutchins_ltsp>
wow. Must've been in the chroot.
15:25
<vagrantc>
jhutchins_lt: what's KDE configured to use for sound?
15:25
jhutchins_lt: should configure it to use alsa
15:26
<jhutchins_ltsp>
yes, it is. lts.confi had #SOUND=FALSE. I've replaced that with SOUND=Y
15:26
<vagrantc>
defaults to True.
15:27
<jhutchins_ltsp>
Well crud. What else then?
15:27
<vagrantc>
i haven't messed with KDE sound much, but i know that's what the debian-edu folks use, and it "just works" as far as i've heard.
15:27
<jhutchins_ltsp>
Does the new pulseaudio setup need to have ESD set up?
15:28
<vagrantc>
it emulates ESD ... so if you can redirect to esound, then it should work.
15:36
<rjune_>
jhutchins_ltsp: see if you get pulseaudio together manually
15:36
aka, first make sure it works
15:37
then make it automagic
15:39
<jhutchins_ltsp>
How?
15:40
<rjune_>
don't know how to do pulseaudio
15:40
IIRC, esd was ESPEAKER="terminal"
15:40
then run esd on terminal
15:40
and esdplay on the server
15:41
get pulse working, then arts should be easy
15:41
KDE control panel should have option for esd
15:42
<vagrantc>
jhutchins_lt:
15:42
<jhutchins_ltsp>
I have no idea what packages might install esd on debian. I've tried a few things, I'm going to restart the client and see what happens.
15:42jhutchins_ltsp has quit IRC
15:42
<vagrantc>
don't mess with esd that way.
15:42johnny has left #ltsp
15:43
<vagrantc>
jhutchins_lt: can you check if it's registering the pulse device with alsamixer ?
15:43
jhutchins_lt: in the package alsa-utils
15:47Vince-0 has joined #ltsp
15:48marilyn has joined #ltsp
15:48marilyn is now known as jhutchins_ltsp
15:48
<jhutchins_ltsp>
Still no sound.
15:48
<vagrantc>
jhutchins_lt: does alsamixer register the pulseaudio device?
15:49
<jhutchins_ltsp>
I wonder how I'd get the MAC of the workstation instead of the server?
15:49
<vagrantc>
look in your dhcp logs
15:50
<jhutchins_ltsp>
/var/log/messages for that matter, I 'spose.
15:50
<vagrantc>
or /var/log/syslog ...
15:50ogra has quit IRC
15:50
<jhutchins_ltsp>
Yes, alsamixer runs and shows pulseaudio as the device.
15:51
Oops, no it doesn't since reboot.
15:51
*** PULSEAUDIO: Unable to connect: Connection refused
15:51
<vagrantc>
jhutchins_lt: logging in from a thin client?
15:51
jhutchins_lt: and in an xterm?
15:51
<jhutchins_ltsp>
Yes.
15:51
<vagrantc>
jhutchins_lt: echo $PULSE_SERVER
15:52
<jhutchins_ltsp>
Nothing.
15:52
<vagrantc>
maybe it didn't detect your sound hardware ...
15:52
but even then, PULSE_SERVER should get set ...
15:52
!ver
15:52
<ltspbot>
vagrantc: "ver" is to get version info on debian/ubuntu (please use the pastebot): COLUMNS=200 dpkg -l 'ltsp*' | awk '/^ii/{print $2}/^ii/{print $3}' ; COLUMNS=200 dpkg --root=/opt/ltsp/i386 -l 'ltsp*' ldm | awk '/^ii/{print $2}/^ii/{print $3}'
15:52
<vagrantc>
jhutchins_lt: can you paste that again?
15:53
<jhutchins_ltsp>
!paste
15:53
<ltspbot>
jhutchins_ltsp: Error: "paste" is not a valid command.
15:53
<jhutchins_ltsp>
!pastebin
15:53
<ltspbot>
jhutchins_ltsp: Error: "pastebin" is not a valid command.
15:53
<vagrantc>
!pastebot
15:53
<ltspbot>
vagrantc: "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.
15:53
<vagrantc>
ltspbot: learn paste as try !pastebot
15:53
<ltspbot>
vagrantc: The operation succeeded.
15:54
<vagrantc>
ltspbot: learn pastebin as try !pastebot
15:54
<ltspbot>
vagrantc: The operation succeeded.
15:54
<vagrantc>
!paste
15:54
<ltspbot>
vagrantc: "paste" is try !pastebot
15:54
<vagrantc>
!pastebin
15:54
<ltspbot>
vagrantc: "pastebin" is try !pastebot
15:54
<vagrantc>
:)
15:54
<ltsppbot>
"jhutchins_ltsp" pasted "ltsp-server 5.1.10-1~40.etch.0" (14 lines) at http://pastebot.ltsp.org/54
15:55
<jhutchins_ltsp>
I'm going to take the audio stuff from the wiki back out of lts.conf and restart.
15:55
<vagrantc>
audio stuff from the wiki?
15:57
<ltsppbot>
"jhutchins_ltsp" pasted "# SOUND=Y # SOUND_DAEMON=esd #" (5 lines) at http://pastebot.ltsp.org/55
15:57
<jhutchins_ltsp>
Without the #'s.
15:57
brb
15:57jhutchins_ltsp has quit IRC
16:00jhutchins_ltsp has joined #ltsp
16:00
<jhutchins_ltsp>
Ok.
16:00
Now I get alsamixer,
16:01
kmix doesn't find any mixer though.
16:01
<vagrantc>
echo $PULSE_SERVER
16:01
jhutchins_lt: could you paste your whole lts.conf?
16:01
<jhutchins_ltsp>
tcp:192.168.76.6:4713
16:02
vagrantc: Dead stock default, everything commented except X_COLOR_DEPTH=16
16:03
The IP above is the clent, btw.
16:03
<vagrantc>
jhutchins_ltsp: yes, that sounds correct.
16:04mccann has quit IRC
16:05
<vagrantc>
jhutchins_ltsp: ok, so there's something amiss in the KDE configuration -> alsa layer, i'm guessing.
16:06
jhutchins_ltsp: i try to support backports as much as possible, but would you consider using debian lenny? it's currently in freeze and due to release "real soon now(tm)"
16:07* jhutchins_ltsp cackles maniacly
16:07
<jhutchins_ltsp>
I barely know what to do with etch! Half the documentation is barely up to etch, let alone etchandahalf!
16:08
Would a lenny client be good enough?
16:08
<vagrantc>
well, it's mostly the server-side stuff that's probably not working so well.
16:08
<jhutchins_ltsp>
Hm.
16:09
I think that may be a bigger job than I have time for this weekend, I'm running out of time as it is.
16:09
<vagrantc>
i don't test KDE much at all, so it's hard for me to know how well each and every feature works.
16:09
<jhutchins_ltsp>
vagrantc: Understood.
16:09
<vagrantc>
jhutchins_ltsp: might consider installing vorbis-tools and seeing if ogg123 works for sound...
16:09
<jhutchins_ltsp>
vagrantc: What do you run on your clients?
16:10
Hm - sure.
16:10
<vagrantc>
just try simple, individual applications
16:10
<jhutchins_ltsp>
Install from a shell on the server, not the client, or does it matter?
16:10
<vagrantc>
jhutchins_ltsp: mostly, icewm or gnome. though i don't personally use either of those :)
16:11
<jhutchins_ltsp>
appears to be here already.
16:11
<vagrantc>
jhutchins_ltsp: i'll be explicit if you have to do anything anywhere but on the server.
16:11
<jhutchins_ltsp>
Now you probably want me to find an ogg.
16:11
<vagrantc>
oh, yeah, i just assumed that'd be handy ... sorry.
16:12
jhutchins_ltsp: here's a smallish one from a recent debian-edu gathering: http://ftp.skolelinux.org/skolelinux/press/20081013-conference/20081013-conf-0of9.ogg
16:12
<jhutchins_ltsp>
cannot open device alsa009
16:12
er, 09.
16:13
<vagrantc>
i remember the remote sound on etch was a little finicky.
16:13
<jhutchins_ltsp>
But will it let go of my terminal? Nooo.
16:14
<Vince-0>
ltsp whoot
16:14Vince-0 has quit IRC
16:15
<jhutchins_ltsp>
Not that it matters, but the speaker jack is usually green, right?
16:15
vagrantc: On another note, on 5 do I need to explicitly enable local USB devices?
16:16
<vagrantc>
jhutchins_ltsp: the instructions on the backports page should be sufficient.
16:16
jhutchins_ltsp: namely, add the users to the fuse group.
16:16
make sure the fuse module is loaded.
16:17ogra has joined #ltsp
16:17
<jhutchins_ltsp>
K. 'cause next on the list is getting the ipod nano working.
16:18
<vagrantc>
make sure it shows up on a "normal" system first.
16:21
<jhutchins_ltsp>
It'll be interesting. I don't think they show up as normal USB Storage devices. (I should check one of those first.)
16:21
At least it's charging.
16:21
<vagrantc>
that'll be a challenge, then.
16:21
<ogra>
grmbl ...
16:22
getting the size of gtk widgets before they are mapped is a pain in several not to mention body areas
16:25johnny has joined #ltsp
16:26bobby_C has quit IRC
16:26
<stgraber>
ogra: btw, do you have that ldm code uploaded somewhere ?
16:27
<ogra>
not yet
16:27
as soon as i have something thats clean enough
16:27
i'm still experimenting
16:28
<stgraber>
ok, I received a couple of suggestions from our test users that I'd like to implement in LDM
16:28
<ogra>
its non functional anyway, i'm only playing with layout possibilities
16:28
if thats doable in the normal ldm, do it there, i will port all gui items that exist
16:29
m current cde only has the clock (which might optionally get a cairo analog clock btw :) ) the logo and an input field
16:30
<stgraber>
ogra: analog cairo clock, that'd rock :)
16:30
<ogra>
i'm not really happy with the size detection code i got yet, i want a proper half transparent bounding box around the input field but thats very hard to do in a clean way
16:31
well, the items are freely positionable and each sits in its own function
16:31
<stgraber>
I have one easy change: make <tab> to do the same as <enter> (similar to gdm's behaviour) the other will be a bit more tricky
16:31
<ogra>
so you can easily drop the digital clock core in favor of an analog one :)
16:31
<stgraber>
as it's to basically have a way to make LDM try the password only once, then get back to the login prompt (so you don't have to enter 2 more buggy password because you can fix the typing mistake you did in the login)
16:32
<ogra>
that sounds like helpful for current ldm
16:32
stgraber, ++
16:32
<jhutchins_ltsp>
rebooting to test auto-login, brb
16:32jhutchins_ltsp has quit IRC
16:32
<ogra>
i'D love that
16:32
i hacked ldm before to do that but with sbalneav's changes thats not as easy anymore
16:32
<stgraber>
second will need an extra ssh parameter and probably some changes in the expect code
16:32
<ogra>
yeah
16:34jhutchins_ltsp has joined #ltsp
16:34
<jhutchins_ltsp>
Oh for...
16:34
It's a _PERMISSIONS_ issue on /dev/dsp.
16:36
brb
16:36jhutchins_ltsp has quit IRC
16:38jhutchins_ltsp has joined #ltsp
16:38
<stgraber>
ogra: btw, if you don't know what to do one of the days: https://wiki.stgraber.org/LTSP-Cluster/Documentation/OpenVZSetup
16:38
<jhutchins_ltsp>
Houston we have audio!
16:38
<ogra>
stgraber, after release
16:42
<jhutchins_ltsp>
vagrantc: Ok, note to add ltsp users to the audio group.
16:42
Next, on to flash!
16:44
<ogra>
*shudder*
16:53alkisg has left #ltsp
16:53
<jhutchins_ltsp>
Well, no youtube this weekend, but we're getting there. Thanks so much for all the help!
16:53jhutchins_ltsp has quit IRC
16:57
<stgraber>
ogra: btw, I just noticed there's a TB meeting on Tuesday, maybe this time they'll have time to look at my upload rights for LTSP+iTalc :)
16:58
<ogra>
heh, just in time for the "nothing goes" freeze
16:58
<stgraber>
yeah ..
16:59
<ogra>
though i think we're mostly fine for intrepid
16:59
<stgraber>
yep, I just need to do some more QA on iTalc this week, just to be sure everything works as expected but other than that it looks like a really good release for LTSP
17:00
(iTalc SHOULD be good too but the last patch went in a rush so I'd prefer to make sure all supported usecase are fine now)
17:00
<ogra>
would you mind triggering a discussion about a safe way to get user lists for ldm ? i would be willing to put something like that into the UI, but i cant imagine a safe way to get a list witut exposing user names anywhere
17:00
<jammcq>
ogra: is release this week or next?
17:01
<ogra>
30th or 31th
17:01
<jammcq>
ah
17:01
k
17:01* ogra alwas forgets the exact date
17:01
<stgraber>
should be on the 30th
17:01
<jammcq>
it's usually a thursday, eh?
17:01
<stgraber>
we tend to have all deadlines on the Thursdays
17:01
<ogra>
yeah
17:02
meh, gtk is evil ... if you want non standard things
17:03
but i think i found a way for the size detection of widgets
17:03warren has quit IRC
17:10Loto has joined #ltsp
17:10
<vagrantc>
what's all this talk about user lists?
17:10
like, a drop-down menu to select your login?
17:13Loto has quit IRC
17:14
<vagrantc>
jhutchins_lt: ah yes. audio group. who would've thought. :)
17:16
<ogra>
vagrantc, either that or even something with icons to click on
17:16
<vagrantc>
icons? eeeyk.
17:16
<ogra>
redesiging UI code gets me weird ideas, sorry
17:16* vagrantc remembers the sppoky head icon problem from using XDMCP with KDM
17:16
<vagrantc>
there were hundreds of head icons, and it would eventually crash the thin client
17:17
<ogra>
heh
17:17
well
17:17
i would program that better then
17:17
<vagrantc>
on a server with lots of logins, that seems very likely to be hard to implement in a useful way...
17:17
<ogra>
ldm-ng already has about 50% of debug code already
17:18* ogra is doing a massive code cleanup of each function he ports over to the new UI
17:19
<ogra>
and adding lots of ifdefs to switch on different debugging modes
17:22ogra has quit IRC
17:25
<johnny>
you could do it like google suggest i guess.. type in the first bits of letters
17:25
ie: the parts you remember
17:29ogra has joined #ltsp
17:34warren has joined #ltsp
17:36* vagrantc attempts to tackle the Xsession mess again
17:36mikkel has quit IRC
17:36
<vagrantc>
one piece at a time.
17:37
<ogra>
i'm only taking care of the ui atm :)
17:38
ooh, the new ubuntu system cleaner is sweet
17:40cliebow has joined #ltsp
17:40
<vagrantc>
hrm. trickier than i thought. we have to selectively define the Xsession based on which server is selected.
17:40Egyptian[Home] has quit IRC
17:41
<ogra>
thats why it should come from ldminfo
17:41
<vagrantc>
i was hoping we could just handle it in the screen script.
17:41
right, i'm looking at implementing it in ldminfod...
17:41Egyptian[Home] has joined #ltsp
17:41
<vagrantc>
implementing the ldminfod side of it is pretty simple.
17:42
<ogra>
yeah
17:42
<vagrantc>
may as well start there.
17:42
<ogra>
we could even hardcode it at compile time based on lsd-release
17:42
err
17:42
s/d/b/
17:42
:)
17:42
<vagrantc>
indeed.
17:43
though i'll start with the simplest implementation first, and we can improve it later.
17:59gate_keeper_ has quit IRC
18:03warren has quit IRC
18:06warren has joined #ltsp
18:19X0d_of_N0d has quit IRC
18:20warren has quit IRC
18:20
<vagrantc>
hrm. ldminfo stuff is currently only handled in the greeter...
18:22X0d_of_N0d has joined #ltsp
18:22
<vagrantc>
would have to somehow include that in ldm itself...
18:27ccherret1 is now known as ccherrett
18:45
<vagrantc>
ok, how to actually make use of ldminfo.c is over my head.
18:49X0d_of_N2d_ has joined #ltsp
18:50
<Ryan52>
vagrantc: just link it too...see Makefile.am in the src dir.
18:51
if it and the greeter share a header file, then copy the functions defined in the header file that are relavant to ldminfo.c into a separate .h file
18:51
then: #include "file.h"
18:51
in the ldm's and the greeter's header file.
18:53
<vagrantc>
Ryan52: that part i can sort of stumble through ... it's more just following how ldminfo.c is used at all
18:55
<Ryan52>
oh. wow. the word over complicated comes to mind...
18:55
<vagrantc>
and just rudiments of C
18:55
ah, well, at least it's not just my incompetence :)
18:57
i just need a way to parse output generated by ldminfod and set a variable based on the xsession: value
18:58
i was first just trying to implement it in the greeter itself, just to wrap my head around it ...
18:58
even though that wouldn't work for autologin
19:01* Ryan52 goes back to parsing purposes
19:18cliebow has quit IRC
19:37johnny has left #ltsp
19:41nubae1 has joined #ltsp
19:41nubae has quit IRC
19:48siki has quit IRC
20:05X0d_of_N0d has quit IRC
20:05X0d_of_N2d_ has quit IRC
20:13warren has joined #ltsp
20:20X0d_of_N2d_ has joined #ltsp
20:48warren has quit IRC
20:56* vagrantc does cargo cult dances and wonders why the gods have shown no interest in C
20:59
<jammcq>
huh?
21:01
<vagrantc>
jammcq: i don't know what i'm doing, but i'm fiddling with things in the greeter to try and make it work :)
21:01
<jammcq>
ah, in C ?
21:01
<vagrantc>
http://en.wikipedia.org/wiki/Cargo_cult_programming
21:02
<jammcq>
hahahahha
21:02
that's hilarious
21:02
<vagrantc>
maybe i'm not even qualified to be cargo cult programming
21:03
i always think of it when i'm way over my head in something, and trying to fumble my way through it :)
21:03
<jammcq>
I know a few of those Cargo Cult programmers
21:03
<vagrantc>
well, looks like i finally got it to compile again...
21:04
<jammcq>
well, if it compiles, it must be correct, eh?
21:04* vagrantc isn't *that* naive
21:13
<vagrantc>
ldminfo.c: In function ‘_ldminfo_parse_string’:
21:13
ldminfo.c:155: warning: assignment from incompatible pointer type
21:13
hmmm... doesn't bode so well, does it.
21:13
<jammcq>
heh
21:13
what does line 155 look like?
21:15
<vagrantc>
ldm_host_info->xsession = val[1];
21:15
what are the -> do-hickeys?
21:15
<jammcq>
de-reference a pointer
21:15
val is an array of something
21:16
ldm_info_info.xession is a pointer to something
21:16
just make sure the 2 "somethings" are the same
21:16
same type
21:17
err
21:17
-> isn't exactly dereferencing a pointer
21:18
well, it sort of is. ldm_host_info is a pointer to a struct
21:18* vagrantc wonders how to ensure the somethings are of the same type
21:18
<rjune_>
in C?
21:19
<vagrantc>
rjune_: yes...
21:19
<jammcq>
ldm_host_info->xession = val[1] says "take the value from val[1] and stuff it into the xession member of the struct pointed to by ldm_host_info
21:19
<rjune_>
hard to do unless the structure includes a parameter that indicates the type
21:19
<jammcq>
vagrantc: need to see the definitons of the struct pointed to by ldm_hist_info, and the array called 'val'
21:20* jammcq hasn't programmed in C in a while
21:20
<vagrantc>
jammcq: i can't even figure out where val comes from ...
21:21
<jammcq>
there's prolly a header file containing the definition of the struct
21:21
<rjune_>
val is an array
21:21
<vagrantc>
er ... gchar **val;
21:21
<jammcq>
ah
21:21
<rjune_>
a gtk char isn't it?
21:21
<vagrantc>
val = g_strsplit(lines[i], ":", 256);
21:21
<rjune_>
vagrantc: pointer to a string
21:21
<vagrantc>
followed by that ...
21:21
<jammcq>
val is a pointer to an array of pointers
21:21
<vagrantc>
which looks like a split on :
21:22
the example line i'm trying to deconstruct ...
21:22
<jammcq>
so the type of value stored in the array is of type 'gchar'
21:22
<vagrantc>
xsession:/etc/X11/Xsession
21:22
<rjune_>
jammcq: isn't it a pointer to an array of gchar's? not array of pointers
21:23
<jammcq>
umm
21:23
<rjune_>
or a pointer to a gchar *
21:23
<jammcq>
I think it's a pointer to an array of pointers to gchar's
21:23
<rjune_>
which normally would be called a string
21:23
<jammcq>
just like argv, is usally defined as **argv
21:23
which is an array of pointers to strings
21:24
<rjune_>
which is a pointer to an array of char* ...
21:24
<jammcq>
or pointer to an array of strings
21:24
<rjune_>
ok. I can see that.
21:24
<jammcq>
and a string is nothing more than an array of char (or gchar in this case)
21:24
<rjune_>
right
21:24
<jammcq>
so, how is the struct declared with the xsession member
21:24
<ltsppbot>
"vagrantc" pasted "more cargo cult patches" (36 lines) at http://pastebot.ltsp.org/56
21:24
<rjune_>
so it can be an array of gchar * or a pointer to an array of gchar
21:25
<jammcq>
prolly as a pointer to an array of 'char' strings
21:25
<rjune_>
vagrantc: which bit are we looking at?
21:25
<jammcq>
in which case, possibly: ldm_host_info->xession = (char)val[1];
21:25
might do the trick
21:25* vagrantc digs up a URL for the whole code...
21:25
<jammcq>
or: (gchar*)(ldm_host_inf0->xession) = val[1];
21:26
i'm a bit rusty tho
21:26
<rjune_>
no, what is xsession?
21:26
<vagrantc>
i blindly made it a GList in the .h file ...
21:26
<rjune_>
I'm not sure what a GList is.
21:26
<vagrantc>
gchar didn't work out for me...
21:26* vagrantc isn't sure what any of this is :)
21:26
<rjune_>
vagrantc: should probably be a gchar *
21:26
though that means you have to do memory handling
21:27
<Ryan52>
rjune_: no, glib does all that.
21:27
<rjune_>
if you make it a gchar[256] you avoid dealing with memory
21:27
<vagrantc>
http://bazaar.launchpad.net/~ltsp-upstream/ltsp/ldm-trunk/files/919?file_id=ldmgtkgreet0.1-20070612162729-e6zhalymbgx16ejx-1
21:27
<jammcq>
ugh
21:27
<rjune_>
glib has automagic garbage collection?
21:27
<vagrantc>
that's a browseable link for this source
21:27
<rjune_>
jammcq: ugh which?
21:27
<jammcq>
[256]
21:27
<rjune_>
yeah
21:27
<Ryan52>
rjune_: it makes things easier, lets put it that way.
21:27
really easy.
21:28
<rjune_>
but ifn he's not a coder, that might be simplest for him.
21:28
<jammcq>
it's either wasteful, cuz your strings might alwasy be small, or it's dangerous cuz your strings could be longer than that
21:28
<rjune_>
jammcq: not per that snippet
21:28
+ val = g_strsplit(lines[i], ":", 256); <-- I would have to check that function, but I bet it breaks at 256 bytes
21:28
<jammcq>
ah
21:28
<rjune_>
so odds are it's wasteful, but not a memory leak
21:29
<jammcq>
right
21:29
<rjune_>
Ryan52: malloc and free aren't difficult.
21:29
it's the remembering to malloc and free that suck
21:29
<Ryan52>
okay, ya.
21:30
<rjune_>
vagrantc: where else are you usnig xsession?
21:30
<vagrantc>
rjune_: nowhere yet
21:31
wheee: ldminfo.c: In function ‘ldminfo_init’:
21:31
ldminfo.c:59: warning: assignment makes integer from pointer without a cast
21:31
ldminfo.c: In function ‘_ldminfo_parse_string’:
21:31
ldminfo.c:155: warning: assignment makes integer from pointer without a cast
21:31
that's after switching it to gchar ... in greeter.h
21:32
<rjune_>
vagrantc: do you know the difference between char and char[] ?
21:32
<Ryan52>
vagrantc: "gchar *" or gchar var[size];
21:32
<vagrantc>
rjune_: nope.
21:32
<rjune_>
gchar refers to a single character
21:32
so 'a'
21:32
<vagrantc>
Ryan52: gchar xsession;
21:33
Ryan52: so i used neither
21:33
<Ryan52>
vagrantc: you need to add [256] to the end, or a * in the middle :)
21:33
<rjune_>
gchar[9] can be up to 8 characters + NULL
21:33
so 'local\0'
21:34
<vagrantc>
Ryan52: with [256]; ...ldminfo.c: In function ‘ldminfo_init’:
21:34
ldminfo.c:59: error: incompatible types in assignment
21:34
ldminfo.c: In function ‘_ldminfo_parse_string’:
21:34
ldminfo.c:155: error: incompatible types in assignment
21:34
<rjune_>
\0 represents NULL
21:34
<Ryan52>
vagrantc: how are you assigning?
21:34
<rjune_>
vagrantc: when using [256] you have to do typecasting
21:34
(gchar *)...->xsession = val[1];
21:34
<vagrantc>
gchar xsession[256];
21:35
<Ryan52>
vagrantc: xsession = blah;
21:35
(that part)
21:35
<rjune_>
val is a gchar *
21:35
so you have to assign it to a gchar *
21:35
<vagrantc>
Ryan52: no idea!
21:35
<Ryan52>
vagrantc: so your not using this variable anywhere?
21:35
<rjune_>
he's trying to add it.
21:35
<Ryan52>
he must be using it somewhere. hence the "incompatible types in assignment"
21:35
<vagrantc>
Ryan52: this is the closest i have to trying to use it: ldm_host_info->xsession = val[1];
21:36
<rjune_>
vagrantc: if you have some free time, I suggest you grab "A Book on C" or poke google for "Thinking in C"
21:36
<Ryan52>
vagrantc: val is an array of what?
21:36
<rjune_>
vagrantc: (gchar *)(ldm_host_info->xsession) = val[1];
21:36
Ryan52: gchar I think.
21:36
<vagrantc>
rjune_: mostly, i just try and do the occasional small patch and if i get in over my head i rope in others :)
21:36
<rjune_>
val = g_strsplit(lines[i], ":", 256); <-- there's the assignment
21:36
vagrantc: still, good to know the basic concepts.
21:37
<vagrantc>
Ryan52: i think it ends up being an array of xession and /etc/X11/Xsession ...
21:37
<rjune_>
C is the stepping stone to many other languages
21:37johnny has joined #ltsp
21:37
<rjune_>
vagrantc: that's the data stored in the array, not the type of the array
21:37
<Ryan52>
vagrantc: where and how is the variable called val declared?
21:38
vagrantc: like the "gchar xsession[256];"
21:38
<vagrantc>
Ryan52: http://pastebot.ltsp.org/56
21:38
Ryan52: gchar **val;
21:38
<Ryan52>
+ GList *xsession;
21:39
whereas the other is a gchar
21:39
those are not compatible with each other :)
21:39* vagrantc wonders how bad of a performance hit it might be to re-write the ldminfo parsing in python...
21:39
<rjune_>
LOL
21:39
<vagrantc>
Ryan52: well, that's the one i switched to gchar foo[256];
21:39
<rjune_>
Ryan52: hence the suggestion he use a gchar[256]
21:40
though it should probably be a 257
21:40
then when you set gchar do xsession[256] = '\0'; that ensures a terminated string
21:40
<vagrantc>
Ryan52: pastebot is behind the times already... i'm a fast-moving clueless developer :)
21:40
<jammcq>
C is perfectly well suited for this task
21:40
<Ryan52>
rjune_: i think the 256 in that call is part includes the null character...
21:40* vagrantc is not perfectly well suited for C
21:40
<rjune_>
jammcq: but is he? ;-)
21:41
<jammcq>
Python makes it easy, but there's way too much overhead for something as simple as this
21:41
<Ryan52>
heh
21:41
<rjune_>
Ryan52: I don't know that.
21:41
<jammcq>
he can learn, this is a perfect opportunity
21:41
<rjune_>
I haven't done much GTK stuff with C, so I'm taking the "safe" route
21:41
<Ryan52>
rjune_: way off!
21:41
rjune_: that's the number to split it into, not the size of the string.
21:42
<rjune_>
damn, split into up to 256 strings?
21:42
that's impressive.
21:42
vagrantc: ignore me and read the docs. *always*
21:42
<vagrantc>
heh
21:42
<Ryan52>
vagrantc: same goes for me :)
21:43
<rjune_>
unless jammcq agrees with me. then it's cool.
21:43
<jammcq>
is that saying "split this string into no more than 256 pieces" ?
21:43
<Ryan52>
so it's probably dynamically allocated to whatever it needs to be.
21:43
jammcq, yup :)
21:43
<jammcq>
makes sense
21:43
<vagrantc>
ok ... so: gchar *xsession; gives no compile warnings...
21:44
<rjune_>
nope. it won't
21:44
you're assigning a gchar * to a gchar *
21:44
<vagrantc>
but val was: gchar **val;
21:44
<rjune_>
that's ok.
21:44
<Ryan52>
vagrantc: but your saying to use the first.
21:44
<rjune_>
val is a gchar **, val[1] is a gchar *
21:44
<Ryan52>
vagrantc: see ** means it's an array of arrays of gchars.
21:44
<jammcq>
that's an array of gchar pointers
21:44Tshital has joined #ltsp
21:45
<jammcq>
and val[1] will return one of those pointers
21:45
<vagrantc>
*** would be 3 items?
21:45
<rjune_>
by indexing gchar, you dereference one of the *
21:45
<jammcq>
and stuffing it into xession works, cuz it's declared to hold a gchar pointer
21:45
<vagrantc>
that much makes sense.
21:45
<rjune_>
vagrantc: do you understand the concept of an array?
21:45
<vagrantc>
rjune_: difference between a hash or dictionary?
21:45
<rjune_>
gchar ** is a two dimensional array
21:45
vagrantc: nope, mathematic array
21:45
[ 0 1 2 3 ]
21:46
*** is a three dimensional array
21:46
<vagrantc>
each part contains something, and you can reference them with foo[0], foo[1], etc ... ?
21:46
<rjune_>
right.
21:46
<jammcq>
carefull, ** isn't necessarily a 2 dimensional array
21:46
it's an array of pointers
21:46
<rjune_>
jammcq: true, but let's get the basic concept first.
21:46
<jammcq>
or, points to an array of pointers
21:46
val[][] is a 2 dimensional array
21:46* vagrantc points the finger at The Man
21:47
<rjune_>
LOL
21:48
<jammcq>
I should actually stay out of this conversation. I'm right at the edge of my memory of C programming
21:48
ask me a Perl question tho :)
21:48
<rjune_>
heh
21:48
I need to learn perl.
21:49
<vagrantc>
ok, so it compiles and doesn't explode.... now how do i get information back out of ldminfo ?
21:49
<Ryan52>
vagrantc: same as everywhere else. ldminfo->xsession
21:50* vagrantc doesn't see that used anywhere.
21:50
<Ryan52>
err, ldm_host_info
21:50
<vagrantc>
aha.
21:51
i can simply treat ldminfo->xsession just like any other variable?
21:51
<rjune_>
it is any other variable.
21:51
<jammcq>
yeah, zackly
21:51
<vagrantc>
ok.
21:52
<rjune_>
ldminfo is a structure(variable, kinda)
21:52
if you declare a variable of type ldminfo, it will have a property called xsession
21:52
<vagrantc>
of course, where the heck in this code it's actually available... is a bit confusing
21:53
<rjune_>
ldm_host_info <--- that's the variable name you're working with in ldminfo.c
21:55
<vagrantc>
now if i can just figure out all these g_something_with_a_bazillion_underscores do
21:56
<Ryan52>
http://library.gnome.org
21:56
:)
21:56
<vagrantc>
heh
21:57
it seems like the ldminfo related stuff should be split out from greeter.h ...
21:57* Ryan52 nods
22:02* vagrantc eyeballs g_hash_table_lookup
22:02
<johnny>
vagrantc, glib is quite the library
22:02
saves you from having to deal with alot of nonsense
22:03
<vagrantc>
looks like a lot of nonsense from this angle
22:03
<johnny>
if you were to look at gtk itself.. you'd disagree :)
22:03
<vagrantc>
but i recognize my ignorance
22:03
<johnny>
saves you from having to deal with memory management and alot of cross platform issues
22:04petre has quit IRC
22:04
<johnny>
plus.. gio is awesome :)
22:04* vagrantc doesn't have the background to find any of this very interesting
22:04
<johnny>
vagrantc, somebody even took the concept of gobjects that were built into glib.. and built another programming language on top called vala :)
22:04
vagrantc, i'm not a C programming btw
22:05
i just follow gnome devel
22:05
and of course.. i'm a gentoo user.. so that helps
22:05* vagrantc just tries to prevent massive amounts of toxic equipment from hitting landfills and stupidly overpowered desktops from running screensavers
22:06
<Ryan52>
"stupidly overpowered desktops from running screensavers"?
22:06
<vagrantc>
Ryan52: well, screensavers are about the only thing most people run these days that actually test the limits of the hardware
22:07
<Ryan52>
oh
22:08* vagrantc gives up
22:10jhutchins has joined #ltsp
22:10
<vagrantc>
i took the low-hanging fruit and committed it 4 hours ago and have an infinitessimally small increased understanding of C.
22:11
<jhutchins>
gotta love it when it works, even if you're not sure why.
22:11
<vagrantc>
well, the low-hanging fruit was in python :)
22:12
i could re-write all this ldminfo parsing in 5 minutes in shell or python... but 4 hours and i'm just tired.
22:12
<jhutchins>
I appear to have sound, youtube, and printing working. These were the primary complaints from running MCNLive, which is essentially a Live CD on a USB key.
22:13
<Ryan52>
vagrantc: if you wanna tell me what it is you were trying to do, I'll do it :)
22:13
<jhutchins>
The thing _I_ appreciate about this, over MCNLive, is that I can add apps, change settings, customize it.
22:14
vagrantc: Seriously, from my limited C exposure (after six mainframe languages), that sounds about right.
22:14
What I can do in a shell script would take days to do in C.
22:15shogunx has quit IRC
22:16
<ltsppbot>
"vagrantc" pasted "patial patch to get Xsession from ldminfod" (68 lines) at http://pastebot.ltsp.org/57
22:16
<johnny>
jhutchins, thus glib/gobject/gtk+.. which is supposed to make things easier.. especially if you can use anjuta or some other IDE that offers code completion and help.. instead of searching docs all day
22:16
then you can just get directly to your work
22:16
<vagrantc>
Ryan52: all i need is the "magic" referenced in that patch :)
22:17zoredache_ has joined #ltsp
22:17
<johnny>
jhutchins, i'm a web developer/scripter by trade.. it's just the gospel i hear :)
22:17
from all the gnomers
22:18
writing raw C is no fun for anybody who is trying to get something done
22:18
that's why glib == awesome :)
22:18
<jhutchins>
heh.
22:18
<Ryan52>
johnny: but it's really fun when your just bored and don't have a project to do :)
22:18
<johnny>
Ryan52, when is that???
22:18
<jhutchins>
Well, I'm a sysadmin now, and devoutly not a coder.
22:18
<johnny>
lemme know when i get some of that?
22:18
i can*
22:18
<Ryan52>
heh
22:18
<jhutchins>
I can script your sox off, but I leave code to those who do it well
22:19
<johnny>
jhutchins, it's cold in here.. please don't do it
22:19
:)
22:19shogunx has joined #ltsp
22:19
<jhutchins>
I _won't_ code.
22:19
<ltsppbot>
"vagrantc" pasted "more Xsession patch" (104 lines) at http://pastebot.ltsp.org/58
22:19
<johnny>
the scripting
22:19
if you were to code.. at least it would warm my machine :)
22:19
lol
22:19
<vagrantc>
Ryan52: that's actually everything i've figured out in the last 4 hours ^^^
22:20
<jhutchins>
I will refer the job to a coder, and I will translate his utterances to the user. that's what I do.
22:21
<Ryan52>
vagrantc: k. first question is how do I get ldminfod to give me whatever it is supposed to give me? and second is, do mind if I do it tomorrow? :)
22:21
<jhutchins>
Thanks for the work, and thanks for the help. I have a significant milestone done this weekend, and it may blossom into a major industrial project.
22:22
<vagrantc>
Ryan52: ltsp-trunk/server/ldminfod
22:22
Ryan52: i committed the changes to it so it should spit out xsession:
22:22
<Ryan52>
oh, I must've missed that one.
22:22
<vagrantc>
Ryan52: and you can just run it manually ...
22:25X0d_of_N3d_ has joined #ltsp
22:28zoredache has quit IRC
22:38
<vagrantc>
maybe i should rip out the setxkbmap call and stick it in an rc.d/I* script just to feel like i accomplished something today.
23:29rcy has quit IRC