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


Channel log from 5 July 2009   (all times are UTC)

00:09Faithful has joined #ltsp
00:45hanthana has joined #ltsp
00:46hanthana has joined #ltsp
00:47alkisg has quit IRC
00:50alkisg has joined #ltsp
01:44Egyptian[Home] has quit IRC
01:51Lns has quit IRC
02:00Egyptian[Home] has joined #ltsp
02:06cyberorg has quit IRC
02:12cyberorg has joined #ltsp
03:04alkisg has quit IRC
03:05alkisg has joined #ltsp
03:57sunru has joined #ltsp
04:11Vonor has joined #ltsp
04:12
<Vonor>
hi again
04:13
ok, i managed to install the vbox additions within the chroot on the ltsp server, the client has all modules loaded also, but still it doesn't work
04:13
I even provided it a valid xorg.conf for the use of virtualbox additions
04:13
but mouse integration still jails the mouse and display resize is also not possible
04:15
<alkisg>
Vonor: are the vbox drivers loaded? (lsmod from a client terminal)
04:16
<Vonor>
alkisg, yeah, i checked that and even the xorg.1.log says it's loaded
04:17ogra has quit IRC
04:17
<alkisg>
Vonor: so what does xvinfo say?
04:18ogra has joined #ltsp
04:19
<Vonor>
alkisg, from the ssh-tunneled x-session? no adaptors present
04:25sunru has quit IRC
04:39hanthana has quit IRC
05:01
<Vonor>
alkisg, i take it you have ltsp running in virtualbox too?
05:02
<alkisg>
Vonor: yeah, but not with guest additions. I've only put a custom vesa resolution...
05:02
<Vonor>
ah, too bad
05:02
<alkisg>
Vonor: no, that's how I want it, to minimize chroot changes
05:03
Otherwise I'd use vmware, there's built in support for that
05:04
<Vonor>
vmware isn't free, or has that changed?
05:05
<alkisg>
The player is. I don't suppose the workstation package is needed for ltsp clients...
05:06
Vonor, out of curiosity, you spend a lot of time using the vbox ltsp clients? I only use them to see if something works or not, so I don't mind the mouse grabbing...
05:08
I change chroots so often that updating vbox drivers for every change would be more trouble than its worth
05:08
<Vonor>
alkisg, well, so far it's only experimenting and fiddling around, but i might use it in production environments too. for the experimenting i don't have any hardware resources, thats why i use virtualbox, both for the ltsp server and for the client
05:19pmatulis has joined #ltsp
05:54
<Faithful>
Hey I need some net cafe software to divvy out my pcs between my kids over the holidays
06:42
<knipwim>
why is https://bugs.launchpad.net/ltsp/+bug/372135 not being solved?
06:49Pulga has quit IRC
06:57
<ogra>
knipwim, because nobody committed it yet
06:58hanthana has joined #ltsp
07:07
<knipwim>
is there a reason for that?
07:08
<ogra>
time :)
07:08
<alkisg>
(speaking of time) ogra hi, do you have some time to have a look at http://users.sch.gr/alkisg/temp/udhcp and maybe come up with something better for the bring_up_interfaces() function?
07:08
<ogra>
i committed it upstream now, thanks for the patch
07:08
<alkisg>
It runs in init-premount, I don't know exactly the problem, aren't the devices set up yet?
07:09
<knipwim>
thanks
07:09
<ogra>
the devices should be up in init-top ...
07:09
udev too
07:09* alkisg tries to put it in init-top
07:11
<ogra>
no, i mean you should have them if your script runs in init-premount
07:11
premount runs after top
07:11
does sleep accept numbeds below 1 ?
07:11
*numbers
07:11
<alkisg>
Yes, e.g. 0.1
07:11
Should I put something lower there?
07:12
<ogra>
well, you are waiting that an interface is up
07:12
does the sleeptime matter ?
07:12* alkisg has no idea about timings :)
07:12
<alkisg>
I'm not really sure. Let me try with 0.1...
07:13
<ogra>
why the other sleeps ?
07:13
<alkisg>
while ! ip link show > /dev/null 2>&1; do ==> waits for the devices to become available (they're not! afaik)
07:13
<ogra>
if "while ! ip link show > /dev/null 2>&1; do" is exiting the loop you can be sure to get output from the ip command
07:14
so why the sleep 1 below
07:14
<alkisg>
Because there may be more than 1 interfaces
07:14
And ip could be only returning the first one... I"ll try without it.. :)
07:14
<ogra>
ip link show will return all of them
07:14
which makes your while loop exit
07:15
<alkisg>
So all devices are created in the same time? ok
07:15
And the final sleep if to enable the interface to come up
07:15
<ogra>
no, not really
07:15
<alkisg>
*is
07:15
<ogra>
i'm pretty sure there is a more proper way using one of the udev userspace tools
07:16
<alkisg>
Agh, I'm wondering if calling ipconfig with dummy parameters would solve all that :)
07:16
<ogra>
also all that piping to cut is very expensive
07:17
you should work out a proper sed line rather
07:17
<alkisg>
Is awk more efficient? I could replace both grep and cuts with a single awk
07:17
Or I could use var=${var%something*} and stuff...
07:17
<ogra>
sed is the most efficient
07:18
the point is that every pipe spawns a new shell
07:18
cut -d' ' -f2 | cut -d':' -f1 uses two ... one sed here is more efficient
07:18
yeah, or even shell vars ... but be careful to stay POSIX compatible with that
07:19
<alkisg>
I'm using `man sh`, I hope that's POSIX compliant?
07:20
So replacing 1 grep and 2 cut's with 1 awk would be good thing, right?
07:20
(it's all busybox, anyway...)
07:21
<ogra>
for the first while loop looking at /sys/class/net might be appropriate
07:21* ogra isnt a fan of awk
07:21
<ogra>
why not sed ?
07:22
<alkisg>
ok, sed it is :)
07:22
ogra, which other constructs cause spawning a shell? Do << or $() also spawn a shell?
07:22
<ogra>
yes
07:23
<alkisg>
OK, I'll try to minimize all those, thanks
07:23
<ogra>
$() is the most expensive one afaik
07:23
<alkisg>
Ouch, and I thought it would be the lighter..
07:23
<ogra>
and if you have pipes in there you might have subshells for each one plus the one for $()
07:24
<alkisg>
got it.
07:24
<ogra>
using vars properly is the least expensive ...
07:24
since that hapens in the same process
07:24
<alkisg>
Is eval is about the same (expensive-wise) as spawning a shell?
07:25
<ogra>
i think eval is a shell builtin so if you use it in a script it should be fine
07:25
<alkisg>
Hmm... that's a lot of power there :)
07:36
woah sed is powerful!
07:39litlebuda has joined #ltsp
08:37cabrey has joined #ltsp
08:37
<knipwim>
if i am correct on the following
08:37cabrey has left #ltsp
08:37
<knipwim>
the selected plugin dir for ltsp-build-client is controlled by VENDOR (from the ltsp-build-client bin)
08:38
what is the purpose of 001-set-dist?
08:38Catman has joined #ltsp
08:38
<Catman>
Hello:)
08:38
Anyone feel like helping me out?
08:39
<knipwim>
depends on the problem
08:39
<Catman>
Ok
08:39
DHCP starts
08:39
but my client doesnt pick up a lease
08:39
any other services I need to kick start?:)
08:40
<knipwim>
do you have network drivers compiled in the client kernel?
08:41
<Catman>
Network drivers are all sorted.
08:41
the client is a thin-term
08:41
PXE bootable
08:42
<knipwim>
and dhcp can provide an ip to the incoming client mac?
08:42
<Catman>
no it doesnt pick it up.
08:42
It just scans for the DHCP:(
08:43
<knipwim>
your dhcp server configuration, you set it up to actually provide an ip for the client mac?
08:44
<Catman>
not the mac in perticular.
08:44
Just broadcast the IP across the subnet
08:46
<knipwim>
can you get an ip normally, with a disk booted client?
08:46
<Catman>
Didnt try that.;
08:46
It should just get a Lease though.
08:46
feel free too SSH in and take a look if you have time/
08:47
<alkisg>
Catman: could you upload your dhcpd.conf to pastebin?
08:47
http://pastebot.ltsp.org
08:49
<ltsppbot>
"Catman" pasted "DHCP" (17 lines) at http://pastebot.ltsp.org/401
08:49
<alkisg>
Catman: you don't provide a boot filename
08:50
<Catman>
sugar .
08:50
where do I add that?:S
08:50
<ltsppbot>
"alkisg" pasted "Default LTSP dhcpd.conf config file (Jaunty)" (22 lines) at http://pastebot.ltsp.org/402
08:50
<alkisg>
See the example above ^^
08:52
<Catman>
thank you very much
08:57fotanus has quit IRC
08:57agike_ has joined #ltsp
08:57
<ogra>
knipwim, you once could build chroots for older releases in debian and ubuntu, thas what set-dist was for (i dont think anyone used/tested it since severeal years though)
08:58hanthana_ has joined #ltsp
08:58
<ogra>
knipwim, i.e. to be able to run "--dist hardy" on jaunty or "--dist lenny" on sid
08:58jhaig|home has joined #ltsp
08:59
<alkisg>
One sed instead of 1 grep and 2 cuts coming up: ip -oneline link show | sed -e '/ether/!d;s/[0-9]*: //;s/:.*//' :)
09:00
<jhaig|home>
I have just installed an ltsp server on Ubuntu (installing from the 'alternative' CD). I have successfully booted a client but I cannot log in. Currently, there is only one user, which is the user set up by Ubuntu on install. Where can I find out why it will not let me log in. I can log into the server. If it makes any difference (it shouldn't) I am running the server on a VMware guest.
09:01
<knipwim>
ogra: okay
09:01
i'm looking for a way to create a debian chroot on gentoo
09:01
<ogra>
knipwim, but you are right, VENDOR rules it ...
09:01
then lsb_release inside the actual distro
09:02
--dist only overrides the latter
09:02
<knipwim>
changing the VENDOR tries to install debian now
09:02
<ogra>
right
09:02
<jhaig|home>
I actually want to authenticate with LDAP, but I think that is the next step.
09:02
<alkisg>
jhaig|home: select the recovery entry on the boot loader, get a root shell and do: passwd user
09:02
<ogra>
but i'm doubtfull that will work on gentoo unless your server has the necessary debian bits
09:02
like debootstrap at least
09:03
<Catman>
knipwim, thank you so much:) saved my ass there.
09:03
<ogra>
jhaig|home, check the users .xsession-errors first
09:03
jhaig|home, and auth log of the server
09:03
<knipwim>
ogra: yes, :) line 3: dpkg: command not found
09:04
Catman: rather thank alkisg
09:04
<ogra>
knipwim, yeah, i expected something like that
09:04
<jhaig|home>
There is no .xession-errors file in the user's home directory.
09:04
<Catman>
oops
09:04
I mistyped there.
09:04
<ogra>
jhaig|home, that suggests something with ssh is wrong
09:04
<Catman>
and yeah
09:04
Kernel panic on first boot:()
09:04
<ogra>
jhaig|home, check auth.log
09:05
<Catman>
how do I update the bootable image?
09:05
<ogra>
jhaig|home, also check if the user can ssh into the server from another system
09:05
<jhaig|home>
ogra: auth.log on the server? Nothing appears there when I try to log into the client.
09:05
<alkisg>
Catman: does it say anything about a rootserver ip? Is that the ltsp server?
09:05
<jhaig|home>
I have already ssh'ed into the server.
09:05
<Catman>
No
09:05
<ogra>
jhaig|home, so ssh is definately running ?
09:05
<Catman>
the client kernel panicked
09:05
<jhaig|home>
Yes.
09:05
<Catman>
someting about not syncing?
09:05
<alkisg>
Catman: what does it say before the kernel panic?
09:06
<ogra>
jhaig|home, weird, did you fiddle with lts.conf ?
09:06
<alkisg>
Catman: do you have another dhcp server in your network?
09:06
<Catman>
65.454943 Kernel panic - not syncing attempted to kill iit
09:06
no:)
09:06
<jhaig|home>
ogra: No. I did move the user's home directory (as I want LDAP authentication and nfs home directories) but I have moved it back again (and rebooted the server).
09:06
<alkisg>
Catman: Sure? No router with dhcp enabled nearby?
09:07
<Catman>
Yeah but the client isnt connected too itS:
09:07
<ogra>
jhaig|home, before doing anything you should confirm the default setup works to exclude possible points of failure ;)
09:07
<alkisg>
OK, a little before the not-syncing it should be reporting the IP it gets, take a look at the root server, is that the ltsp server ip?
09:08
<jhaig|home>
ogra: Yes, you are right. I have changed it all back again, though.
09:08
<ogra>
jhaig|home, so just to get what you are seeing, your client properly boots to the login manager screen ?
09:08
<Catman>
Nope.
09:08
ill check the DHCP
09:08
<jhaig|home>
I have the graphical login screen.
09:08
<ogra>
if you then log in there, do yu see any error message, does the login manager restart immediately or what exactly happens ?
09:09
<alkisg>
Catman: ok, to update the kernels you do: chroot /opt/ltsp/i386 update-initramfs -u and then ltsp-update-kernels
09:09
But if you've just installed I don't expect that would change much...
09:09
<jhaig|home>
ogra: It waits for a while on "Verifying password. Please wait." and then goes back to the login screen (appears to reload X).
09:09
<ogra>
alkisg, if he rund debian or ubuntu ;)
09:09
*runs
09:10
<alkisg>
ogra: sure, but that's the only ones I know about :)
09:10
<Catman>
Ubuntu:)
09:10
<ogra>
other distros dont have update-initramfs :)
09:10
<Catman>
I lost my debian LTSP in a SATA failure:(
09:10
<alkisg>
So if he runs anything else he'd have to tell me to shut up :)
09:10
<ogra>
alkisg, right, bt he didnt mention his distro above if i'm not blind, just wanted to make sure
09:10
<Catman>
I didnt.
09:10
but im glad he assumed :)
09:11
<ogra>
jhaig|home, and during that please wait there is nothing from ssh in auth log ?
09:11
<alkisg>
Sure I was just lazy to type all the "what distro etc" and I assumed he would correct me if I guessed wrong :P OK no harm done
09:11
<ogra>
jhaig|home, you should see login attemts from the user there
09:11
!distro
09:11
<ltspbot>
ogra: Error: "distro" is not a valid command.
09:11
<ogra>
!dist
09:11
<ltspbot>
ogra: "release" is please mention the linux distro and release you're using :)
09:12
<ogra>
alkisg, ^^^
09:12
<alkisg>
See? it took 2 lines :D
09:12
<ogra>
no need for excessive typing :)
09:12
<jhaig|home>
ogra: Just trying it again now, but I don't think so. This is in the server while trying to log in to the client. Is there an auth.log file for the client?
09:13
<ogra>
no, the ssh server runs on the server machine so only the auth.log there is intresting
09:14* alkisg has misread jhaig|home's problem, thought he had problems logging in while sitting on the server...
09:14
<Catman>
hmm
09:14
<ogra>
no, from the client only apparently
09:14
<Catman>
should I try restarting the server?
09:15
<alkisg>
Catman: you could try removing the splash screen to better see what the error is
09:15
<ogra>
jhaig|home, did you try sshing in with the same user from another machine already ?
09:15hanthana has quit IRC
09:15
<Catman>
tried
09:15
<alkisg>
I.e. modify /var/lib/tftpboot/<ARCH>/ltsp/i386/pxelinux.cfg/default
09:15
Err
09:15
<ogra>
heh
09:15
<jhaig|home>
ogra: I can ssh into the server as another user, but not the client because (I think) sshd isn't running.
09:16
<ogra>
if you can ssh in sshd is running
09:16
<jhaig|home>
s/as another user/from another machine/
09:16
<ogra>
but try it as the same user
09:16
<alkisg>
...modify /var/lib/tftpboot/ltsp/<ARCH>/pxelinux.cfg/default :)
09:17
<jhaig|home>
I get connection refused trying to ssh into the client.
09:17
<ogra>
how did yo ssh in fro the other machine ?
09:17
thats funny ...
09:18
so ... sudo /etc/init.d/ssh restart on the server and see if you can log in then
09:18
<jhaig|home>
And log into the client? Still connection refused.
09:19
<ogra>
(it should also tell you if it was already running)
09:19
<Catman>
it doesnt get that far
09:19
<ogra>
no log in to the server via ssh with the same user you try out on the client
09:19
<alkisg>
Catman: it doesn't display the boot splash screen?
09:20
<jhaig|home>
I'm already logged into the server via ssh with the only user on the system.
09:20
<Catman>
No it does.
09:20
but fails before anything loads.
09:20
<ogra>
jhaig|home, hmm, so your sshd is running then
09:20
<alkisg>
Catman: ok, but if you remove the splash screen (and the "quiet" param there), you'll be able to better see the error...
09:20fotanus has joined #ltsp
09:20
<ogra>
jhaig|home, where exactly do you get "connection refused" ????
09:21* ogra is confused
09:21
<jhaig|home>
Trying to ssh onto the client.
09:21
<ogra>
heh
09:21
that wont work
09:21
<jhaig|home>
Well, it doesn't, so that is working OK. :-)
09:21
<ogra>
you cant log in to the client in any way by default, its completely secured
09:21
right
09:22
<Catman>
i resent Kernels.
09:22
<ogra>
there are not even users on a client and root is locked ... so even if someone would manage to sneak in an sshd he would have a prob ;)
09:23
<jhaig|home>
Oh, I thought it inherited users from the server.
09:23
<Catman>
Same:)
09:23
alkisg
09:23
I dont understand the command.
09:23
<ogra>
no, the login manager largely does: ssh -Y user@server /etc/X11/Xsession
09:23
<alkisg>
Catman: it's not a command, it's a file.
09:23
<ogra>
so all you need is users on the server side
09:24
<alkisg>
Catman: OK, let me try that again. What's your server arch? i386 or amd64?
09:24
<Catman>
i386
09:24
sorry should off realised:(
09:24
<alkisg>
ok, give: gksu gedit /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default
09:24
And remove the "quiet splash" from there and save and reboot the client
09:25
<ogra>
jhaig|home, did you change the ssh config on the server in any way ?
09:25
<Catman>
restart it?
09:25
<ogra>
or do you have a firewall between client and server
09:26
<alkisg>
Catman: yes, the client (not the server)
09:26
<Catman>
:)
09:26
I know
09:28
its a broken interface.
09:28
eth0.
09:28
hmph:(
09:28
<jhaig|home>
ogra: No change of the ssh config, no firewall - both server and client on the same network.
09:29
<alkisg>
Catman: You mean not supported? What interface model and what's the exact message?
09:29
jhaig|home: try putting SCREEN_02=shell SCREEN_07=ldm in lts.conf, reboot the client, press Alt+Ctrl+F2 and try `ssh server` from there, to see the exact ssh error message
09:29
<ogra>
jhaig|home, very weird
09:29
<Catman>
eth0
09:29
link is not ready
09:29
<ogra>
jhaig|home, what alkisg said ...
09:29
<jhaig|home>
But you said there should be no users set up by default, so how is the first user added?
09:30
<Catman>
fixed it
09:30
It was a broken interface card
09:30
<alkisg>
ok
09:30
<ogra>
jhaig|home, during installation of the server you added a user
09:30
<Catman>
comes from being a recycled hardware group
09:30
all our stuff is donated.
09:30
and we only work in linux;)
09:30
<alkisg>
Heh... same thing in my school :)
09:30
<Catman>
ha.
09:30
In your school!
09:31
<alkisg>
(teacher, not student)
09:31
<Catman>
was about to say.
09:31
I'd hate too think children were that smart:)
09:31
I would be up shit creek
09:31
<alkisg>
Heh... nah, ltsp is easy enough, it doesn't need smart guys :)
09:32
<Catman>
you use it as a school network?
09:32
<alkisg>
Yes, for the past 2 years we've been using ltsp only in my school(s)
09:33
<Catman>
brilliant
09:34
my thin client manager doesnt work:S
09:34
I cant see the client attached
09:34Pulga has joined #ltsp
09:35
<ogra>
jhaig|home, there are no users *on the client*, there is a user on the server... the client boots a login manager that does ssh user@server to ssh into the server, for that you dont need any user on the client
09:35
<alkisg>
Catman: You mean the program? Switch to italc for the time being (until Lns picks tcm up)
09:37
<Catman>
ok
09:38
how do I add users?:)
09:39
<jhaig|home>
alkisg: Do those lines go in /var/lib/tftpboot/ltsp/i386/lts.conf ? I have done that but I am not getting a text console.
09:39
<alkisg>
jhaig|home: distro/version?
09:39
.../arch?
09:39
<ogra>
[default]
09:39
SCREEN_02=shell
09:39
SCREEN_07=ldm
09:39
thats how your lts.conf needs o look like
09:39
<alkisg>
^^^
09:40
<jhaig|home>
OK, I didn't have the [default] - the file didn't exists.
09:40
<ogra>
ctrl-alt-f2 should get you to a running shell
09:40
no, it doesnt exist by default
09:41* alkisg thinks task switching from reading sed manuals to chatting here is more difficult than it sounds :)
09:41
<Catman>
ok
09:41
Im off out to deploy
09:41
cya in a bit.
09:41
<alkisg>
bye
09:41Catman has quit IRC
09:41
<ogra>
heh, yeah, sed twists your brain :)
09:41
<alkisg>
ogra: does it support backreferences?
09:41
<stgraber>
hey ogra
09:42
<ogra>
stgraber, hey ...
09:42
<alkisg>
I.e. match <something><here's the match><something else> and return the match?
09:42
Hi stgraber
09:42
<stgraber>
ogra: do you happen to know if the current 2.6.31 kernel has VFS union in it ?
09:42
I'd love to get rid of unionfs-fuse :)
09:42
<ltsppbot>
"ogra" pasted "devpts mount plugin" (8 lines) at http://pastebot.ltsp.org/403
09:42
<jhaig|home>
OK, from the console on the client I can do "ssh administrator@server" and log into the server
09:43
<ogra>
stgraber, no idea
09:44
<alkisg>
jhaig|home: did you see any warning when ssh'ing?
09:44
<stgraber>
ogra: there's a certain lack of documentation on how it's supposed to work so not really easy to test :)
09:44
<ogra>
stgraber, can you take a look at that plugin ^^^^ i'd like to commit it, bt i'm not sure about the options
09:44
stgraber, yeah, lest wait until the week starts and nag cjwatson about it, he should know more
09:45
<jhaig|home>
"The authenticity of the host 'server (192.168.0.156)' can't be established.
09:45
<ogra>
did you change the ip of the server after installation ?
09:45
<jhaig|home>
Ah, yes I did.
09:45
I went from a dynamic ip to a fixed ip.
09:45
<stgraber>
ogra: looks good, we have a plugin to unmount everything at the end right ?
09:46
<ogra>
sudo ltsp-update-sshkeys && sudo ltsp-update-image
09:46
<jhaig|home>
(Ubuntu doesn't give that option on install).
09:46
<ogra>
run that line
09:46
stgraber, chroot_mount does that automatically if you use it
09:46
<stgraber>
ogra: regular ubuntu also has the "rw" option, not sure it matters though
09:47
(I'd suspect rw to be the default)
09:47
<ogra>
oh, thats new
09:47
<jhaig|home>
ogra: Thanks. It will take a bit of time to update the image, but this is almost certainly the solution.
09:47
<jammcq>
ogra: hey
09:47
<ogra>
i pulled the line out of an intrepid script i had lying around :)
09:47
jammcq, hi :)
09:47
<jammcq>
ogra: scotty and I have been discussing BTS-2009
09:48
we're proposing that it be Oct 21-25 in Maine
09:48
<ogra>
jammcq, cool, i have a lot of commitments in oct, did you plan o do it in oct or nov ?
09:48
<jammcq>
but that's just a proposal
09:49
<stgraber>
that's the week before Karmic's release
09:49
still better than release week itself though :)
09:50
<jammcq>
hmm
09:50
<ogra>
i'm likely travelling that exact weekend
09:50
<jammcq>
we want to avoid the next week, cuz of halloween
09:50
<ogra>
its maemo summit in amsterdam
09:51
<stgraber>
would the week before be possible ?
09:51
<ogra>
no way for me to get around going there i suspect
09:51
<jammcq>
how does Nov 4-8 look?
09:51
<stgraber>
there'd be less Ubuntu stress and maybe it's better for ogra too ?
09:51
<ogra>
oh, wait
09:51
wrong !!
09:51
<jammcq>
huh?
09:51
<ogra>
its otc9 to oct 11
09:51
<jammcq>
oh
09:52
so maybe oct 21-25 would work?
09:52
<ogra>
though there was something else the 21th-25th ...
09:52
but i dont remember what
09:52
<jammcq>
well... think about it for a few days. we haven't cast it in concrete yet
09:52
<ogra>
ok
09:53
i'll try to find out what it was and how important it was
09:53
<jammcq>
sue
09:53
sure
09:53
we'd really like to have you out there this year
09:53
<ogra>
though oct is quite stuffed
09:53
<stgraber>
I'll be asking at the office too, I plan on bringing at least one other guy from Revolution Linux (not sure Francis will come this year though)
09:53
<jammcq>
and hopefully stgraber and all the usual suspects can make it again
09:53
<ogra>
yeah, i dont want to miss the 10th anniversary drinking
09:53
<jammcq>
right. I'd better alert Ragnar then :)
09:53
and Gadi
09:53
<ogra>
yeah
09:54
i wonder if warren would come
09:54
<jammcq>
we'd definately invite him
09:54
it's only like a 5 hr car ride for him
09:54
an adventure
09:55
<ogra>
heh
09:57
stgraber, submitted ...
09:57* ogra dances ... i fixed two ltsp bugs today :)
09:57* jammcq has forgotten the feeling of fixing ltsp bugs :(
09:57
<stgraber>
ogra: yeahh
09:59
<ogra>
jammcq, well, my boss hunted me down about my assigned bugs and was upset that i refused to unsubscribe from a 2 year old bug i didnt solve yet :)
09:59
though it was cosmetical anyway
10:00
which didnt put it high on my prio list
10:01Patina has quit IRC
10:02
<stgraber>
ogra: we still have that "Generated chroot don't have -updates and -security in sources.list" bug if you're in a bug fixing day ;)
10:03
IIRC it was an issue to simply generate a sources.list as it'd break with d-i generating a chroot without internet or something like this
10:03
<ogra>
well, do you know how that is if you have a very important task to do and start to clean your house instead ? ... its somehow what i'm doing here, i should be writing a spec that has to be done by tomorrow :)
10:04
no, its not that easy sadly
10:05
it need a rewrite of the setup-mirrors plugin ... since that uses a std. debian setup for security which is completely incompatible to the ubuntu one
10:05
<jammcq>
ogra: I know exactly what you are talking about. my desk is VERY clean today :)
10:05
<ogra>
haha
10:10sene has joined #ltsp
10:11
<ogra>
wow, someone made an action movie based on pong ?
10:12
http://staffweb.cms.gre.ac.uk/~cs39/playpong.html
10:12
awesome
10:20Vonor has quit IRC
10:20Patina has joined #ltsp
10:28Patina has joined #ltsp
10:29hanthana_ is now known as hanthana
10:30bobby_C has joined #ltsp
10:33lucascoala has quit IRC
11:07fotanus has quit IRC
11:07alkisg has quit IRC
11:09sene has quit IRC
11:09fotanus has joined #ltsp
11:13alkisg has joined #ltsp
11:21Padhu has joined #ltsp
11:31jhaig|home has quit IRC
11:37alkisg has quit IRC
11:37Padhu has quit IRC
11:48alkisg has joined #ltsp
12:01Faithful has quit IRC
12:30lucascoala has joined #ltsp
12:35alkisg has quit IRC
12:41lucascoala_ has joined #ltsp
12:42alkisg has joined #ltsp
12:56vagrantc has joined #ltsp
12:56lucascoala has quit IRC
13:07lucascoala has joined #ltsp
13:09
<vagrantc>
ogra: shouldn't commandline options take precedence over configuration files?
13:10
ogra: regarding the ARCH settings
13:12
<ogra>
vagrantc, hmm, feel free to flip it
13:12
<vagrantc>
i didn't look closely at the code...
13:16
ogra: ah, nevermind... it looks like it's doing the right thing
13:16mikkel has joined #ltsp
13:16
<vagrantc>
first checks commandline, then if it's already set, and then falls back to it's default
13:21lucascoala_ has quit IRC
13:41GGD has quit IRC
13:42artista_frustrad has joined #ltsp
13:44GGD has joined #Ltsp
13:54Faithful has joined #ltsp
14:00lucascoala has joined #ltsp
14:01
<alkisg>
vagrantc, stgraber: much optimized version, only using echo and sed now: http://users.sch.gr/alkisg/temp/udhcp
14:01
comments welcome :)
14:16
<vagrantc>
alkisg: i committed some stuff to make pxelinux configuration more flexible
14:17
<alkisg>
vagrantc: this also does lease renewal...
14:17
<vagrantc>
still room for improvement on that, but at least it makes ipappend possible, and also makes it easy to set up a menu-based setup.
14:18
alkisg: what does it do when the lease expires?
14:18
<alkisg>
It sends a dhcp request to renew it
14:18
(udhcpc does that automatically...)
14:18
<vagrantc>
and if the server gives back a different ip?
14:19
<alkisg>
It shouldn't. Only the gateway could change...
14:22
<vagrantc>
leases typically have a default renewal time, and an expiration time, after the expiration time, the dhcp server *may* hand out a different ip address.
14:23
<alkisg>
that's what renewal is for; to prevent IPs from expiring
14:23
Also, this works automatically with 2 or more NICs with no need for ipappend or bootif...
14:29
<mikkel>
"XF86CONFIG_FILE = xorg.via" is not working, I have "/opt/ltsp/i386/etc/xorg.via" but the client says "Error! - xorg.via - File not found" I'm on fedora 11, what is the problem ?
14:29nubae has quit IRC
14:32
<mikkel>
Is there a way to push an image to a local harddisk on a thin client, so everything is booting from the client and only X is using the netnork (no NFS, NDB and swap over network) ?
14:33
<vagrantc>
mikkel: probably need X_CONF=/etc/xorg.via
14:34
alkisg: the crazy thing is i wrote a whole bunch of code to handle this stuff a long time ago called initrd-netboot-tools ... using udhcpc ...
14:34
<mikkel>
is X_CONF an option, not mentioned on the Wiki, will try
14:35
<vagrantc>
alkisg: but gladly abandoned it for the much simpler support built in to initramfs-tools
14:35
mikkel: XF86CONFIG_FILE is deprecated
14:35
should still work, but it just sets X_CONF
14:36
though i don't really know how things work on fedora, that's just what upstream ltsp code does.
14:36
<alkisg>
vagrantc: well... ipconfig is pretty broken... and the ip=xxxx command line parameter handling is also bad... and there's no way to autoboot with multiple nics
14:36
So simpler, yes, but also broken :(
14:42
<vagrantc>
actually, we could easily write code to re-run the configure_networking function with each detected NIC ...
14:58
<alkisg>
vagrantc: ipconfig hangs and upstreams isn't very interested in patches
15:03
<vagrantc>
i'll go with the tide on this one, but i haven't really had any problems with ipconfig
15:03
<alkisg>
ok
15:04
<vagrantc>
udchpc seems like a lot more code to maintain, unfortunately.
15:04
<alkisg>
?
15:05
<vagrantc>
alkisg: more code to maintain handling network booting
15:05
<alkisg>
udhcpc has a lot less bugs than ipconfig
15:05
vagrantc: I put a lot of things to do a lot of stuff - I could easily do it in only a few lines if I wanted to do what configure_networking does now..
15:06
<vagrantc>
and then the dracut folks are using isc's dhclient, which doesn't support next-server ...
15:06
<alkisg>
dhclient doesn't support next-server?
15:06
<stgraber>
nope
15:07
<vagrantc>
they's why i switched to udhcpc years and years ago.
15:07
<stgraber>
that's why I didn't use it and went for udhcpc instead
15:07
<vagrantc>
and apparently nothing's changed when i mentioned the support in dracut
15:07
<alkisg>
So what do they use instead?
15:07
<vagrantc>
they don't.
15:08* alkisg doesn't get where's the problem then...
15:09
<vagrantc>
bsically, they require putting the next-server and root-path on both, even if the ip address is the same for both, whereas we've defaulted to using next-server if ip isn't specified in root-path for 4-6 years, depending how you count it...
15:10
overall, the dracut syntax for root-path looks a lot nicer and more flexible, thankfully
15:15
alkisg: the big downside i see with udhcpc for me is that it's going to be hard to get into default initramfs-tools in debian...
15:16
<alkisg>
vagrantc: well, it's transparent; if it's there, it gets used, otherwise (1) if a static ip was requested, the script handles it because it handles it better than configure-networking, and (2) if dhcp was requested, ipconfig is used...
15:17
Well, basically, it would be nice if we could replace configure-networking with this one :(
15:18
I've heard a lot of people complaining about the busybox problem because other dhcp servers are around; opensuse deals with it with the kiwiserver variable, but ip=x:<server:x... is defined in the kernel docs and configure_networking doesn't handle it correctly...
15:19
<vagrantc>
alkisg: what do you mean it doesn't handle it correctly?
15:20
<alkisg>
I mean that if you pass it an ip=:server:::hostname:dhcp command line, it doesn't care about the server nor the hostname
15:21
<vagrantc>
it just uses the values it gets from dhcp in that case?
15:22
<alkisg>
Yes, but http://www.kernel.org/doc/Documentation/filesystems/nfsroot.txt states that the manually defined one should take precedence
15:22
<vagrantc>
right
15:22
i'll test and file a bug...
15:22
<alkisg>
That's not the only bug :)
15:23
Most of them are not handled correctly
15:23
<vagrantc>
well, i'll file a bug for each separate issue that i'm made aware of :)
15:24
i can't file a bug and say "most everything is broken".
15:24
<alkisg>
Err... want me to send you a dozen ones?
15:24
<vagrantc>
alkisg: sure, send a mail with all the cases that it breaks, and i'll try and reproduce them.
15:24
if i can reproduce them, i'll file bugs in the debian bts.
15:26
<alkisg>
OK... I guess I'm thinking like a programmer and not like a maintainer; if something is pretty broken I'd rather rewrite it than handle it as bug fixes...
15:26
I shows that I'm new to this :)
15:29
<vagrantc>
i like to have other people fix problems in code that they maintain, with me occasionally nudging them... :)
15:30
i also don't like to change out something that's been mostly working for me to fix corner cases that sound like bugs.
15:32
<alkisg>
There's absolutely no code to handle multiple interfaces... Not statically defining DEVICE=eth0 doesn't sound like ...a corner case to me :) But the main problem is ipconfig itself; the current maintainer is not the one that wrote it, and it has some serious bugs in it (i.e. dropping packets or waiting for non-existent packets)
15:33
<vagrantc>
yeah, i've long been meaning to address the DEVICE issue- though that's totally workable in the code rather than ipconfig itself.
15:35
<alkisg>
Sure... a proper code would do what I tried to do; i.e. call ipconfig for all available interfaces. But then ipconfig would also need to be changed because of its bugs :)
15:35
Anyway, /me 'll stop whining now :D
15:40GGD_ has joined #Ltsp
15:42dmc has quit IRC
15:46* vagrantc tries the NBD support in debian for the first time in ages
15:52alkisg has quit IRC
15:53GGD has quit IRC
15:54GGD_ is now known as GGD
16:07sene has joined #ltsp
16:27fotanus has quit IRC
16:28alkisg has joined #ltsp
16:56
<vagrantc>
ah, good. found some bugs... wonder how long those have been there.
17:04bobby_C has quit IRC
17:08mikkel has quit IRC
17:19
<vagrantc>
aw hell. it's borked real good...
17:40* vagrantc wonders when it actually worked...
17:50alkisg has quit IRC
18:07
<vagrantc>
using --create-ext3-image works with NBD, but the squashfs image refuses to mount.
18:13
aha! it may just be an incompatibility with the squashfs-tools package and the squashfs-modules versions...
18:23
did squashfs get into the mainline kernel?
18:23
<panthera>
jup
18:24
you need squashfs 4.0 and linux >= 2.6.29; or squashfs 3.3 and linux <= 2.6.28 (with squashfs oot module)
18:24
.oO(..and now bedtime.)
18:24
<vagrantc>
panthera: heh. thanks :)
18:25* vagrantc tries again with a downgraded squashfs-tools
18:31
<vagrantc>
gah. how many hours i spent debugging that non-issue...
18:32GGD has quit IRC
18:32GGD has joined #Ltsp
19:03
<vagrantc>
well, that works better.
19:04
of course, it still didn't resolve the issue i had with ltspfs /media/root/* mounts...
19:09hanthana has quit IRC
19:14fotanus has joined #ltsp
19:16fotanus has quit IRC
19:16fotanus has joined #ltsp
19:40japerry has joined #ltsp
19:59Egyptian[Home] has quit IRC
20:00Egyptian[Home] has joined #ltsp
20:08pmatulis has quit IRC
20:13litlebuda has quit IRC
20:29F-GT has quit IRC
20:37Faithful has quit IRC
20:40Faithful has joined #ltsp
20:43vagrantc has quit IRC
20:43F-GT has joined #ltsp
20:48try2free has joined #ltsp
21:00akSeya has joined #ltsp
21:01
<akSeya>
anyone here helping develop ltsp5 for gentoo?
21:14akSeya has quit IRC
21:17try2free has left #ltsp
21:37hanthana has joined #ltsp
21:53Pulga has quit IRC
22:14
<sbalneav>
Evening all
22:15
<stgraber>
evening sbalneav
22:21sene has quit IRC
22:59Sarten-X has quit IRC
23:07Sarten-X has joined #ltsp
23:09Egyptian[Home] has quit IRC
23:10alkisg has joined #ltsp
23:14Egyptian[Home] has joined #ltsp
23:35
<jammcq>
sbalneav: ping
23:44
<sbalneav>
jammcq: Pong
23:44
Happy belated 4th of july
23:44
<jammcq>
hey buddy
23:59agike_ has quit IRC