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


Channel log from 27 February 2008   (all times are UTC)

00:01
<warren>
vagrantc, nope, I think my initscript is still broken
00:01
vagrantc, and I don't have a lts.conf installed in the chroot anymore
00:01
hmm
00:11
jammcq, you sent me a TK-3550 right?
00:23achandrashekar has joined #ltsp
00:24
<vagrantc>
warren: there anyone in the fedora crowd who could help with i18n for ldm ?
00:25* vagrantc is kind of surprised both the debian and ubuntu communities haven't turned anyone up for this
00:25
<warren>
vagrantc, not without a few months warning
00:25
vagrantc, everyone is too busy with F9's current defined goals
00:25
<vagrantc>
warren: sure.
00:26
<warren>
vagrantc, ldm needs some serious rewriting in order to properly do l10
00:26
l10n
00:26
<vagrantc>
really.
00:26
<warren>
yes
00:26
the ldminfod needs to feed more/different info
00:26
the greeter must display localized and native language translations
00:26
<vagrantc>
what about just getting the UI localized?
00:27
<warren>
there's already code to do the latter in gdm
00:27
<vagrantc>
the greeter is mainly what i'm concerned with, at least for starters
00:27
<warren>
and both are gtk
00:27
vagrantc, outside of language picking?
00:27
<vagrantc>
there's already code to do it, but it's broken.
00:27
or at least, not enabled.
00:27
<warren>
I wont be able to get anyone to help anytime soon
00:27
and ldm still doesn't work in fedora at all
00:28
due to the locale -a issue
00:28mccann has quit IRC
00:28
<vagrantc>
ah well. i'll try and prod some debian channels and see if i get any bites. it's been a while since i've asked.
00:29
<warren>
darn, the latest amd driver still doesn't work
00:29
in fedora
00:32
<daduke>
vagrantc: howdy! updated ldm yesterday, and autologin works!! great stuff!
00:32
<vagrantc>
daduke: great! :)
00:32
daduke: you running sid or a backport?
00:32
<daduke>
vagrantc: alioth backport on etch.
00:32
<vagrantc>
nice.
00:33
there was a really weird bug with etch RyanRyan52 figured out a workaround for... though why it's an issue is perplexing to us all.
00:34
<daduke>
vagrantc: one thing left tho: local USB devices.... the first one inserted on any thin client gets mounted, but all the others don't. I saw many ppl having this issue, but no real fix?
00:34
<vagrantc>
daduke: i haven't heard much of this.
00:35
<daduke>
vagrantc: all right, I'll tell you if I figure it out...
00:35RyanRyan62 has joined #ltsp
00:35
<vagrantc>
daduke: i.e. the first usb stick inserted in client1 gets mounted, but when you put a second stick in it doesn't ? or when you put a usb stick on client1 it gets mounted, but when you put a usb stick in client2 it doesn't get mounted?
00:36
the first case should be fixed in ltspfs. haven't heard reports of the second case.
00:36
<daduke>
vagrantc: I haven't tried 2 sticks on one client, but stick1 on client1 -> mounted, leave it in, stick2 on client2 -> see stick1, but no stick2
00:37* vagrantc wishes people would report bugs.
00:37
<daduke>
I'm trying to.... ;)
00:37
<vagrantc>
good :)
00:38
daduke: that's enough to start. i'd recommend using reportbug to submit a bug on ltspfs
00:39
the bug tracking system is my todo list and i won't forget about it and others can contribute to the debugging process
00:39
<daduke>
vagrantc: will do. I'm using chmod 777 /dev/fuse to grant permissions since we have LDAP users and the fuse group is local, but 2 weeks ago you said this should be fine.
00:39
<vagrantc>
whereas some random conversation in irc might be forgotten :)
00:40
daduke: should be... i think i tried that when i told you about it.
00:40RyanRyan52 has quit IRC
00:40
<vagrantc>
daduke: do the systems have floppy or cdrom drives ? or is it all USB devices?
00:41
daduke: i could easily test that multiple clients with multiple CDs and floppies work.
00:41
<daduke>
vagrantc: no other local devices, I just tried USB sticks
00:41RyanRyan62 is now known as RyanRyan52
00:41
<vagrantc>
actually, with the latest backports, i think i can finally get freegeek to the point of being able to support localdevs
00:42
<daduke>
vagrantc: I believe you, others got it up and running too, but I've seen quite some ppl with this first-mount-only thingy
00:42
<vagrantc>
so i'll be able to test it
00:42
daduke: where do you see them? :)
00:43
<daduke>
vagrantc: out there on the internet ;) lemme check if I still got the URLs...
00:44plamengr has joined #ltsp
00:45
<daduke>
vagrantc: it's from Nov 07, but seems to be the same... http://www.mail-archive.com/ltsp-discuss@lists.sourceforge.net/msg32902.html
00:45
vagrantc: but this still seems to be the old :10 issue...
00:45
<vagrantc>
daduke: ah!
00:46
daduke: that bug should be fixed in debian.
00:46oh207 has quit IRC
00:46
<daduke>
vagrantc: I know. But you have to believe me, I still got the same symptoms! I swear!
00:47
<vagrantc>
both the backports for etch, and debian unstable ... and technically lenny (although it's broken for other reasons)
00:47oh207 has joined #ltsp
00:47
<daduke>
vagrantc: before submitting the bug, I will set up a 2-client test bed on my desk so that I can easily debug n stuff
00:48
<vagrantc>
daduke: i can do a test with CD/floppy ... see if that works ok.
00:48
<daduke>
vagrantc: funny thing is, no user has asked for local USB so far... But I want to have it when they do.
01:07
<warren>
oh great
01:07
cd_pinger segfaults for some reason =)
01:09oh207 has quit IRC
01:09oh207 has joined #ltsp
01:13
<ogra>
warren, and only sbalneav knows the code
01:13
<warren>
cd_pinger segfaulted immediately upon startup
01:13
<ogra>
(nobody else ever touched the C version)
01:13
<warren>
so it is probably something simple triggered by our ultra paranoid C flags
01:13
hopefully...
01:16* ogra wonders why he never saw the -c option in nbd-server before ....
01:16
<ogra>
we wouldnt need unionfs at all ...
01:16
or squashfs
01:17
<vagrantc>
ogra: well, i get the impression it's not very stable.
01:17
<ogra>
(-c enables a builtin copy on write function in nbd-server)
01:17
vagrantc, i tried it last night
01:17
<vagrantc>
ogra: and gets increasingly slow the more it's used
01:17
<ogra>
yeah, that i can imagine
01:18
for a standalone client its been reasonably fast though
01:18
<vagrantc>
if it works decently though, that would be pretty slick.
01:18
i keep meaning to look at the nbdroot code in the newer versions of nbd
01:18
<ogra>
well, i'd love to combine it with squashfs ... but the kernel blocks writes as soon as it seen squashfs involved
01:19
<vagrantc>
oh, right.
01:19
so it would need to be ext2 or something like that?
01:19
<warren>
yeah
01:19
<ogra>
well, i'd take ext3
01:19
<warren>
which you really don't want =)
01:19
<vagrantc>
ogra: why bother with ext3 ?
01:20
<warren>
Fedora's livecd is ext3 compressed in a squashfs image
01:20
we get read-write root with devicemapper's dm-snap to do block-level damage
01:20
<ogra>
vagrantc, journalling helps keeping the corruption rate low ... less bribes :)
01:20
<warren>
another type of copy on write
01:21
(there is no reason we do squashfs in this, except it makes it smaller to fit on a CD)
01:21
<vagrantc>
warren: the whole squashfs + ext3 seems kind of bizzare ... i mean, i see why.
01:21
<warren>
bizarre yes
01:22
<vagrantc>
i still don't get ext3, though ... ?
01:22
<warren>
vagrantc, you can't do dm-snap on squashfs
01:22
vagrantc, dm-snap tracks block changes, not file changes
01:22
<vagrantc>
warren: yes, but why ext3 vs. ext2 ... with all the additional overhead of journaling ...
01:22
<warren>
ext3 is on a "block device" loop mounted from the squashfs image
01:23
vagrantc, good question
01:23
I dunno
01:23
<vagrantc>
on something that is basically throw-away
01:23
<warren>
ah
01:23
but it isn't throwaway
01:23
<ogra>
vagrantc, depends where you use it
01:23
<vagrantc>
journaling is not likely to help much ... or is it?
01:23
<warren>
you can optionally persist on a disk or network share
01:23
<vagrantc>
ah. then it becomes very important.
01:23
<ogra>
the partition the squashfs resides on is likely to no be unmounted properly
01:24
<vagrantc>
crazy magic.
01:24
<ogra>
(if squashfs is your rootfs its hard to unmount something that sits even below)
01:25
<warren>
vagrantc, with this crazy hack you can even reboot and fsck it
01:25
<vagrantc>
ogra: do you have your debian dirs for ltsp and ldm in bzr, and if so, could you push them somewhere visible? :)
01:25mikkel has joined #ltsp
01:25
<ogra>
vagrantc, yeah, sorry, longstanding thing on my list
01:26
<vagrantc>
ogra: i'm curious how we've diverged, and could put some time into syncing what's syncable
01:26
<ogra>
i'll try to do it today ... classmate and cd reorganization keep me quite busy
01:26
<vagrantc>
ogra: cool :)
01:26
<ogra>
(if i only could solve that grub mess)
01:27
<vagrantc>
you did some changes to the udeb, no?
01:27
<ogra>
lots
01:28
<vagrantc>
ogra: last time we sync'ed debian dirs may have been 5.0.39 in gutsy ... october
01:28
<ogra>
i reordered a lot
01:28
but i dont think you use most of that code
01:29
<warren>
are you folks going to move things we discussed on the list?
01:29
in ltsp-trunk
01:29
<vagrantc>
debian-edu is the primary user of the udeb, as far as i know
01:29
<warren>
I created some base directories where you can put both common and distro specific things below it
01:29
<ogra>
well, i default to image building etc
01:29
<warren>
for example, client/initramfs is Debian specific. Could you put all that under client/mkinitramfs/debian/ ?
01:30
<vagrantc>
ogra: sure ... so i suspect i'll need to patch up the code to only use that stuff if it's enabled. :)
01:30
<warren>
client/mkinitramfs could be renamed to client/initramfs if you want thereafter
01:30
<ogra>
surely not under mkinitramfs ... but surely we can put it in a dedicated dir :)
01:30* vagrantc has always thought it should be initramfs-tools
01:30
<ogra>
vagrantc, i think i added something
01:30
vagrantc, ++
01:30
<vagrantc>
named after the package itself
01:31
<ogra>
fully agreed
01:31
<vagrantc>
and then it doesn't matter what distro
01:31
unless there's namespace conflict with package names
01:31
<warren>
client/foo where foo is a generic name
01:31
initramfs-tools isn't a generic name
01:31
we don't name anything that way
01:31RyanRyan62 has joined #ltsp
01:31
<warren>
I'm OK with client/initramfs
01:31
<ogra>
initramfs-tools indicates that iots distro agnostic and can be used with any distro using initramfs-tools
01:31
<vagrantc>
"initramfs-tools" is a specific piece of software ... is there any other project named "initramfs-tools" ?
01:32
<warren>
is "initramfs-tools" a package name?
01:32
<vagrantc>
yes.
01:32
<ogra>
yes
01:32
<warren>
I'm putting mkinitrd related stuff under /client/foo/k12linux/
01:33
<ogra>
the thing is that mkinitramfs cant be that generic
01:33
<warren>
I suppose foo can be initramfs-tools even though we don't have a package named that.
01:33
I'm not tied to mkinitramfs. I only named it that because you have initramfs currently containing debian specific stuff
01:33
<ogra>
the initramfs-tools we use in ubuntu/debian will work on any other distro using initramfs-tools
01:33
<warren>
I think you're not understanding what I'm suggesting...
01:33
<vagrantc>
or so we hope.
01:34
<warren>
I can't use initramfs-tools.
01:34
I need to adapt mkinitrd to do the equivalent
01:34
<vagrantc>
so don't use initramfs-tools.
01:34
<ogra>
warren, what i mean is that mkinitramfs will be distro specific, while initramfs-tools isnt
01:34
<warren>
I want to share a common library between the two
01:34
<vagrantc>
there are definitely incompatible projects all named mkinitrd
01:34
<warren>
thus I want a generic name
01:34
<vagrantc>
hmmm...
01:35
<warren>
client/genericname/k12linux (containing mkinitrd related stuff)
01:35
client/genericname/common
01:35
(containing a shared library of stuff
01:35
client/genericname/initramfs-tools if you want it
01:36
I suggest initramfs as genericname
01:36
<vagrantc>
warren: i'm not sure what will be common between your mkinitrd code and any distro using initramfs-tools
01:36
<cyberorg>
due to unavailability of aufs/unionfs, "split" image support was recently added to kiwi http://lists.berlios.de/pipermail/kiwi-devel/2008-February/000324.html
01:36
<warren>
no idea what split image is
01:36
<cyberorg>
it allows temporary/persistent rw to any file or folder
01:36
<ogra>
yeah, its pretty unlikely there is anything in common
01:37
<vagrantc>
seems kind of like fedora's readwrite stuff
01:37
<warren>
Is the function in /home/warren/work/k12linux/ltsp-warren/client/mkinitramfs/initramfs-common possibly useful?
01:37* ogra watchs dmraid being refused to enter hardys default install ...
01:38* vagrantc doesn't have access to warren's home dir.
01:38
<warren>
oops
01:38* vagrantc chuckles
01:38
<warren>
vagrantc, ltsp-trunk has it
01:38
<ogra>
warren, how should we know if you dont give us access to your machine :P
01:38
heh
01:38
snap
01:38
<warren>
vagrantc, from the description it sounds like it is different. all directories are writable.
01:39
<vagrantc>
ah, yes.
01:39
<warren>
maybe I have no clue how they implemented this
01:39
what is kiwi?
01:40* vagrantc wonders when we're going to see opensuse code in ltsp-trunk
01:40
<warren>
is kiwi novell's livecd?
01:40
<ogra>
yes
01:40
<warren>
oh
01:40
<vagrantc>
warren: it's opensuse's something or other that generate's ltsp5 type stuff amoung other things
01:41
<cyberorg>
http://kiwi.berlios.de/
01:42
warren, live cd/dvd/qemu, ltsp, vmware,xen - system imaging thing
01:42RyanRyan52 has quit IRC
01:42
<cyberorg>
i am using kiwi for ltsp5 implementation on suse
01:42
<warren>
cyberorg, except for ltsp, we have a crazy project called revisor that does much of that.
01:42
and a competing implementation in cobbler/koan
01:43
cyberorg, how does split image work?
01:43
<cyberorg>
let me get the actual part in linuxrc that does it
01:48
http://svn.berlios.de/wsvn/kiwi/kiwi-head/modules/KIWILinuxRC.sh see mountSystemCombined function
01:49
rootfs.tar.gz contains all the directories/files that require rw
01:51
<Pascal_1>
hello
01:52
<cyberorg>
it would be great if all the ltsp code is documented like this :)
01:52
<Pascal_1>
i asked this question yesterday someone gave me a link but i can make works ldap authentication with thin client
01:52
anybody could help me ??
01:52
the problem i think is about pam but i dont know lot of things about it
01:55
it works from an other client with ssh but not with graphical interface on a thion client
01:56
when i try to connect from the graphical interface no error in auth.log but the login interface come back every time
01:56
<ogra>
warren, i just found time to look at your initramfs-common code ... i dont thik thats needed, we can just uname -a :) if we are in initramfs already we have a running kernel
01:56
<warren>
ogra, uname -a shows the server's running kernel
01:57
<ogra>
??
01:57
oh
01:57
<warren>
ogra, which may not necessarily match the exact version of the kernel you're trying to manipulate in the chroot
01:57
<ogra>
your script is used during creation ?
01:57
<warren>
yes
01:57
<ogra>
ah
01:57
thats something we dont have to care about at all
01:57
<warren>
that particular function could be useful to other things
01:57
<ogra>
the kernel packages care for it
01:57
<warren>
hmm
01:57
<ogra>
update-initramfs is called with the right version if you install a kernel package
01:58
<warren>
I might be able to get rid of mkinitrd-ltsp-wrapper entirely now
01:58
<ogra>
our initramfs is created on the fly durin package install
01:58
<warren>
oh here's a question
01:58
ogra, yes in my case as well
01:58
<ogra>
so why that code than ?
01:58
*then
01:58
<warren>
client/update-kernels
01:59
well
01:59
hmm
01:59
<ogra>
the initramfs is tied to the kernel anyway
01:59
<warren>
let's talk about client/update-kernels
01:59
<ogra>
oki
01:59
thats vagrants code :)
01:59
<warren>
# this script is run either chrooted on the server, or by a client with write
01:59
# access to the NFS mount point. (much of this code was originally in
01:59
# server/ltsp-update-kernels). --vagrant 20060801
01:59
<ogra>
right, he did the split
02:00
<warren>
In my case, I have no interest at all in a script writing to any boot loader or pxe configuration files
02:00
I only want scripts to generate the different image types
02:00
mknbi, mkelf, etc.
02:00
<ogra>
and how do you tell your clients where to boot from ?
02:01
i mean
02:01
<warren>
ogra, whatever is the latest kernel is symlinked from vmlinuz.ltsp
02:01
<ogra>
how do you generate your pxe defult file then
02:01
<warren>
initrd.ltsp symlink to latest initrd
02:01
<ogra>
right thast what we do as well
02:01
<warren>
pxe default file is shipped as a static file
02:01
most users don't need to ever edit the pxe config
02:02
<ogra>
but for multirach setups and multiple chroots/images you cant rely on /vmlinuz
02:02
we regenrate it every kernel install
02:02
<warren>
well, PXE config is only for i386 and x86_64
02:02
<ogra>
since it holds info like the nbd port
02:03
and tehr is no other easy or clean way to hand that out to the kernel otherwise
02:03
<warren>
and if you have both i386 and x86_64 clients on the same network, you are going to need dhcpd.conf tricks to point them to a different arch directory for tftp boot
02:03
ah, but there is =)
02:03
the dhcp server can tell it the nbd address and port
02:03
which is what I'm working on soon
02:03
<ogra>
warren, i have about ten images here i jump around between ... thats 9 customized pxe files
02:03
<warren>
why ten images?
02:04
<ogra>
for different features i work on
02:04
or tests etc
02:04
<cyberorg>
any reason for creating different image for x86_64 TC?
02:04
<warren>
I think we're talking of a very different design philosophy.
02:04
<ogra>
in any case i know ther are many eduuntu users using multiple chroots
02:04
yeah
02:04
<warren>
cyberorg, you need a 64bit kernel and to point it at a 64bit NFS or NBD
02:05* ogra doesnt
02:05
<ogra>
but we in turn need the amd64 compiled packages :)
02:05
<cyberorg>
warren, 32bit image would work as well on x86_64 cients
02:05
<ogra>
our generic kernel covers both arches
02:05
<warren>
huh?
02:05
that makes no sense
02:05
<ogra>
linux-image-generic is identical on both arches
02:06
<warren>
What if you really want a 64bit kernel in order to run 64bit VM's on a diskless client?
02:06
<ogra>
i wuld ask my vm team :P
02:06
no idea
02:06
<cyberorg>
after all they are thin clients, unless there is serious number crunching happening on the client isnt 64bit overkill?
02:06
<ogra>
i gess we have a pckage for that
02:06
<warren>
Anyway it sounds like I'm heading down a different design direction with 64bit parallel in mind
02:07
cyberorg, I am leaving open the possibility of that being wanted one day
02:07
<cyberorg>
ah :)
02:07
<ogra>
cyberorg, i actually know people runing amd64 clients
02:07
<warren>
cyberorg, since I need to redo the stuff for Fedora anyway, I might as well design with those future needs in mind.
02:07
<ogra>
but simply ecause they had spare the hardware
02:07
<warren>
I have a pretty good idea how a dhcpd.conf could drive different clients to a i386 or x86-64 kernel/userspace.
02:08
<Nubae>
wow, ogra u really seem to live here :-)
02:08
<ogra>
Nubae, well, my office is only one channel away ... so i hang around here too :)
02:08
<warren>
Nubae, unfortunately I seem to be moving in too... =)
02:08* ogra hands warren a coffee ....
02:09
<Nubae>
I mean, times... early morning to late night
02:09
<ogra>
but dont forget the dishes once a wekk at least :P
02:09
Nubae, i had platform team meeting at 7:00 UTC ...
02:09
<Nubae>
ah
02:09
<ogra>
so i had to get up early .... for my colleagues in US and asia
02:10
<Nubae>
such dedication and discipline ;-)
02:10
<ogra>
well, once a week i can bear that
02:10
<warren>
wash the dishes?
02:10
<rcy`>
hi
02:10
<ogra>
and we're rotating the times
02:11
warren, well, if you want to move in you have to participate in housekeeping a bit ... :)
02:11* warren already has to wash dishes daily or the girlfriend is mad.
02:11
<warren>
She makes surprise visits sometimes.
02:11
<rcy`>
im having a problem when netbooting clients off of one of my ltsp servers. dhcp works fine, tftp fetches pxelinux.0, but it fails to find pxelinux.cfg/MACADDR, pxelinux.cfg/IPADDR..., and finally .../default
02:11* ogra has a dishwasher :)
02:11
<rcy`>
switching nics sometimes solves the problem
02:12
er, rather some nics work, some dont.
02:12
<warren>
ogra, btw, what is the purpose of mknbi, mkelf, mkelfimage?
02:12
how are they different?
02:12
what kind of clients are supported?
02:12
<Nubae>
damn... and the list continues: It was unacceptable to suggest to a user, on a user's list, that paying
02:12
Canonical for support was the appropriate solution for this and the
02:12
other out-of-the-box gotchas from 7.1.
02:12
<ogra>
warren, one of them supports LinuxBIOS
02:12
<warren>
which?
02:12
<ogra>
i think mkelfimage
02:12
<Nubae>
jeez...
02:12
<ogra>
the one we default to in ubuntu and debian atm :)
02:12
<johnny_>
Nubae, wtf are you reading?
02:12
and why are you posting it here?
02:13
<Nubae>
that was the edubuntu-users mailing list
02:13
<johnny_>
this isn't #edubuntu
02:13
keep the depressing news out of here :)
02:13
<warren>
Yes, non-Ubuntu people like warren might get all upset.
02:13
<ogra>
warren, its mkelfimage
02:13
<Nubae>
ok, but it was for ogra's ears.... I'll move it to edubuntu
02:14
<johnny_>
i'm sure ogra already knows
02:14
<Nubae>
i live on both channels too...
02:14
<warren>
hmm, I thought we packaged that
02:14
<Nubae>
anyway, it was a response to me suggesting getting paid support
02:14
<johnny_>
this channel is one of the most harmonious channels i'm on
02:14
i'd like to keep it that way :)
02:14
<Nubae>
I mean, I suggested once could get support, and that is the mail that came back
02:15
its not totally edubuntu related
02:15
its an 'open source' matter
02:15
<johnny_>
my advice, ignore it
02:15
<ogra>
Nubae, LaserJock cared for it already... no need to freak out :)
02:16
<Nubae>
well, it was in response directly to me
02:16
all I did was recommend possibly looking at support
02:16
I dont see why that is unnacaptable
02:16
<ogra>
it isnt .. but wont get the bug fixed
02:17* warren wonders if he gets Fedora's code into Ubuntu, and can reproduce Fedora problems on Ubuntu, he can pay for Ubuntu support.
02:17
<ogra>
paid support will tell you the workaround (and probably nag me to do an update of the package)
02:17
or a backport
02:17
warren, sure, as long as your questions will refer to ubuntu you can
02:17
<Nubae>
well, whatever, I tried to point out that there are solutions to the problem if one doesnt want to be a developer themselves
02:18
<warren>
ogra, what is mkelf-linux for?
02:18
<ogra>
warren, its a replacement for mknbi ... supposed to be more free and to support everything mknbi does
02:18
sbalneav did very extensive testing wth it during gutsy development
02:19
he has a lot different etherboot HW
02:19
<warren>
hmm, mkelf-linux on my system is provided by the mknbi package
02:19
<ogra>
i trusted his judgement with pulling it in as default
02:19
<warren>
mkelf-linux has a different upstream?
02:19
<ogra>
mkelfimage
02:19
not mkelf-linux
02:20
<Pascal_1>
any idea for my problem ?
02:20
<ogra>
mkelf-linux is on the mknbi package here as well
02:20
<warren>
I should pull in mkelfimage (which has mkelf-linux) and I have no reason for mknbi anymore?
02:20
<ogra>
according to sblaneav, yes
02:20
<warren>
ah, that explains things
02:20
<ogra>
i only have two etherboot clients here
02:20
and one thin can
02:21
(well, two thin cans, but one is EOL)
02:21
<warren>
older etherboot right?
02:21
Etherboot-5.4 can do PXE
02:21
well... sort of...
02:21
<ogra>
well
02:22
the rom-o-matic images can do PXE emu since three years now
02:22
<warren>
but?
02:22
<ogra>
i think they added it in 5.2 already
02:22
but people using etherboot are usually to lazy to burn new EPROMs
02:22
and might still use 4.x
02:22
<warren>
I've seen vendor class identifier "Etherboot-5.4" do PXE
02:23
<ogra>
or older crap
02:23
<warren>
will some clients reporting "Etherboot-5.4" not PXE?
02:23
<ogra>
depends
02:23
in rom-o-matic its a checkbox you dont *need* to set
02:23
<warren>
ogra, Jim has a recipe that lets you chain-load a newer etherboot from an old etherboot, which sounds cool, but he says if the BIOS (like LinuxBIOS) doesn't support PXE, then even the latest Etherboot wont PXE.
02:23
<ogra>
not sure ther is any PXE compatibility without it being checked
02:24
right
02:24
you need some special binary blobs in the bios
02:24
thats what LinuxBIOS doesnt have
02:24
<warren>
grr
02:25
<ogra>
its all about freedom :)
02:25
<warren>
this means I can't rely on any "Etherboot-5.4" vendor class identifier
02:25
and I instead need to just point all Etherboots at mkelf-linux wrapped images?
02:25
<ogra>
you can likely for the majority ... but surely not for everyone
02:25
<warren>
oh shit, that doesn't work either
02:25
Etherboot-5.4 in PXE mode can't load anything that large
02:25
<ogra>
you have to pick one :)
02:26
<warren>
stupid manufacturers
02:26
<ogra>
our code just respects what you install in the chroot and default to one of the options
02:26
i.e. we install mkelfimage ... but if you install mknbi it overrides
02:27
so the minority for whom it doesnt work can just install mknbi in the chroot and is done
02:27
<warren>
but you don't know for sure if anybody really needs mknbi
02:27
only theoretical
02:27
?
02:28
<ogra>
i know its needed with very old etherboot images
02:28
so i wont drop the option ...
02:28
<warren>
gah!
02:29* ogra goes to make some more coffee ... waaay to early ... *yawn*
02:29
<warren>
ogra, I wonder if all etherboot images can chainload another etherboot, and THAT is a known version, so it can be pointed to a single known working image.
02:32
<ogra>
find soemone like scottie to prove that :)
02:32
i dot have the HW
02:33
*dont
02:33
<warren>
In theory it will work
02:33
scottie has been MIA though
02:33
<ogra>
in theroy you can travel at lightspeed (nearly at least)
02:33
well, he'll be back
02:34
<warren>
vacation?
02:34
<ogra>
the question is when :)
02:34
no, work
02:34
<warren>
does he respond to e-mail?
02:34
<ogra>
he spent 805 of his worktime here the last year i guesss
02:34
80%
02:34
so i think his boss demands that he does some of the work he's actually hired for atm
02:35
:)
02:35
i didnt mail him since some time ... try it
02:35
but i'd assume he does
02:35
and i know jim is on the phone with him from time to time
02:35
<warren>
3:35am
02:35
I should sleep
02:35
<ogra>
yeah
02:36
9:39 here ...
02:36
now wher do the 4 mins come from
02:36
<warren>
i'm going to not worry about mknbi for now
02:36
mkelf-image sounds like it supports the most users
02:36
<ogra>
yeah, worst case just steal our code
02:36
if people actually have probs
02:37
<warren>
I might ship etherboot for chain loading as well
02:37
=)
02:37
<ogra>
its a trivial check for the executable iirc
02:38
woah, suses gfxboot is the worst code i've seen *in my whole life* !!
02:38
<warren>
gfxboot?
02:39
graphical boot?
02:39
<ogra>
an assembler interpreter that crates an undocumented script language
02:39
yeah
02:39
my sole in the company changed, i have to take care of such crap now
02:39
s/sole/role/
02:39
we use it on the cd images as bootmenu
02:40
if i'm familiar enugh with it i'll proably go for a graphical bootmenu in tsp :)
02:40
*ltss
02:40
GAH
02:41
*$CANNELNAME
02:41
<warren>
Is xfs really needed anymore?
02:41achandrashekar has quit IRC
02:41
<ogra>
xfs ?
02:41
<warren>
We haven't installed it at all since like Fedora 6
02:41
<ogra>
you mean the font stuff or the FS ?
02:41
<warren>
ogra, x font server
02:41
<ogra>
ah
02:41
its not
02:41
but you need the fonts installed in the chroot
02:42
<warren>
ohhh
02:42
<ogra>
i discussed with scottie to drop the XFS code from the initscripts as well
02:43
if people run stuff like xfig the fonts need to be available in the x server
02:44
gnome handles that inside the session and renders them server side i think kde as well ... its just the corner case of Type1 requiring apps
02:44
<warren>
[root@newcaprica ltsp]# yum list '*font*' |wc -l
02:44
173
02:44
yikes
02:44
<ogra>
now *that* slows down your boot :)
02:44
or X startup rather
02:44
<warren>
well I have only 62 font packages installed on my laptop
02:44
<ogra>
by default we install 6-10 in ubuntu
02:45
<warren>
so every font that is installed on the server, you need the same font on the client?
02:45
<ogra>
no
02:45
<warren>
if you want it to display
02:45
?
02:45
<ogra>
but every Type1 font you will need
02:45
<warren>
true type no
02:45
<ogra>
ttf comes fien from the session side ...
02:45
*fine
02:45
<warren>
oh
02:45
k
02:45
<ogra>
but type1 cant be renederd that way, needs to happen in the x server
02:46
<warren>
Do I need something like hpijs installed on the client?
02:46
If I want the client to support local printers
02:46
<ogra>
for prinitng i suppose
02:46
we have it in the standard install
02:47
so its always been there for me ....
02:47
no idea what happens without it
02:47
<warren>
hpijs pulls in net-snmp (2.8MB) and lm_sensors (2.0MB) into my client chroot
02:47
the latter two are pretty useless =)
02:47
<ogra>
and ...
02:47
as long as they are not loaded they will do no harm
02:48
<warren>
whoa
02:48
I'm getting dizzy
02:48
time to sleep
02:48
<ogra>
even though snmp sounds like something i wouldnt want ...
02:54
<warren>
vagrantc, oh, interesting thing I learned today. apparently dnsmasq supports pretty much all crazy dhcpd.conf options that we rely upon, and its codebase is WAY simpler.
03:00johnny_ has quit IRC
03:04mcfloppy has quit IRC
03:08
<warren>
attempting a ldm login
03:08
tcpd is using 100% CPU
03:09
not sure what's going on
03:10
oh
03:11
tcpd /usr/sbin/ldminfod
03:11
this is using 100% CPU
03:12toscalix has joined #ltsp
03:17
<warren>
Feb 27 04:15:30 newcaprica tcpd[19434]: connect from ::ffff:172.31.100.219 (::ffff:172.31.100.219)
03:17
Feb 27 04:16:31 newcaprica tcpd[19434]:last message repeated 98485 times
03:17
well, this is interesting
03:17
ldm is going nuts
03:18
trying to connect to ldminfod
03:20
<johnny>
warren, i just had a dnsmasq problem
03:20
am currently using dnsmasq for my ltsp setup
03:20
and it works great, tftp, local caching dns, dhcp, etc
03:21
but when i try to move it to another machine.. i can't point it the another root server via option 17
03:22
<warren>
oh
03:23
<johnny>
so as long as it is on the same machine.. all is fine
03:36
<vagrantc>
regarding mkelfimage, mkelf-linux and mknbi-linux ...
03:37
mkelfimage and mkelf-linux both create ELF boot images, which are supported by recent etherboot versions typically, but not by older versions. the really old stuff needs NBI images, which are created by mknbi-linux.
03:38alekibango has joined #ltsp
03:38
<vagrantc>
and i don't know if there's any way to detect which modes a given etherboot supports ... i.e. it can have support for PXE, ELF and NBI, and a custom image may support any or all of those.
03:39
but i think we shouldn't focus all our development efforts on customizations ... we only should ship a default file that works most of the time ...
03:39
<daduke>
vagrantc: still around? I got two clients up and running now... things look funny tho.
03:39
<vagrantc>
for example, you *can* configure etherboot to use NFS *instead* of tftp
03:39
daduke: define "look funny"
03:40
<johnny>
or even http..
03:41
<daduke>
vagrantc: trying to... local USB device only works for *one* user, for others it doesn't get mounted, regardless of the working user being logged in or not. He must've been the first one to ever try it or something strange...
03:42
<johnny>
sounds like a known bug..
03:42
i know it is at least registered for ubuntu in launchpad
03:43
<daduke>
vagrantc: another thing is that df on the terminal server now shows df: `/tmp/.horeizo-ltspfs/usbdisk-sda1': No such file or directory, I think that's related to some coreutils update you mentioned last time...
03:43
johnny: yeah that's what I thought too. I'm with alioth backports on etch tho
03:44
<johnny>
daduke, that's the same bug
03:44
<daduke>
johnny: great! I'm not alone ;) any known fix/workaround?
03:45
<johnny>
i thought vagrantc said he posted a new version that included the fix
03:45
but i could be wrong
03:45
i was only half here at the time
03:45
<daduke>
johnny: new version of? ltspfs? ldm? I updated from alioth 10 min ago..
03:45
<johnny>
i don't know if it's in etch or what
03:45
<Nubae>
Failed to run users-admin as user root.
03:46
Unable to copy the user's Xauthorization file.
03:46
hmmm... suddenly get that...
03:46
<johnny>
i just did what was in the launchpad bug report
03:46
search for local devices
03:46
<vagrantc>
daduke: yes, the df issue is mentioned on http://wiki.debian.org/LTSP/Howto in the known issues for backports
03:47* vagrantc suspects daduke's issue is some new issue
03:48
<daduke>
vagrantc: does it have any impact other than a nonworking output?
03:48
<johnny>
i got them both at the same time i know that for sure
03:48
<vagrantc>
daduke: dpkg -l ltsp* | egrep ^ii ; dpkg --root=/opt/ltsp/i386 -l ltsp* ldm | egrep ^ii
03:48
daduke: i don't know if it has other impacts.
03:48
<daduke>
!past
03:48
<ltspbot>
daduke: Error: "past" is not a valid command.
03:49
<daduke>
where's this pastebot?
03:49
<vagrantc>
!pastebot
03:49
<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.
03:49
<ltsppbot>
"daduke" pasted "dpkg output" (6 lines) at http://pastebot.ltsp.org/453
03:50
<vagrantc>
so, while virtualbox works fine on my laptop with a server and thin-client ... adding a second thin-client is brutal....
03:50
but i'm trying to test this also, at least as best i can at the moment
03:54
daduke: well, for floppy and CD, both mounts happen.
03:54
<ogra>
warren, still awake ? i'd try: telnet localhost 9571 and see what happens
03:54
<vagrantc>
daduke: hmmmm... are the directories present in /media/USERNAME/* ?
03:54
daduke: and maybe the icons just aren't showing up?
03:54
<daduke>
vagrantc: sorry, can't test with fd/cdrom due to missing hardware. Only usb sticks.
03:55
vagrantc: no, it's not an icon prob (I'm on e17 anyway), I'm checking the mounts in /media directly
03:55
<vagrantc>
daduke: well, i'm just telling you that ltspfs works with multiple devices at least on some level. i don't know if there's anything special with usb sticks that causes it to break.
03:55
daduke: as the user for each directory?
03:55
<daduke>
vagrantc: well one thing might be the 'need to be partitioned' issue. But they are.
03:55
<vagrantc>
even root can't read other user's directories with fuse ...
03:56
<daduke>
vagrantc: thing is, only the first user gets a directory in /media, for the others there's none
03:57
<vagrantc>
daduke: any messages in ~USER/.xsession-errors ?
03:58
<daduke>
vagrantc: stand by, lots to read...
04:01
vagrantc: I tail -f'ed them, nothing upon removal/insertion of sticks...
04:05
<vagrantc>
daduke: so ... is it the first user who plugs in a USB stick ... or the first user who ever plugged in a USB stick?
04:05
daduke: have you tried it with: ln -s /proc/mounts /etc/mtab
04:06
daduke: i can't think why that would make a difference ... but it might
04:06
<daduke>
vagrantc: I guess he's the first who plugged one in since terminal server reboot, uptime 5 days (after vmsplice reboot)
04:07
<vagrantc>
well, on the server-side .... there's nothing really different between CD/floppy/usb sticks
04:07
so there must be something different in your configuration from mine.
04:08
<daduke>
vagrantc: I guess there is. But how to find it? The ln didn't change anything...
04:09
<vagrantc>
daduke: didn't even change the df output issues?
04:09
<daduke>
vagrantc: oh yes it does. but no /media dir for the 2nd user
04:09
<vagrantc>
right
04:10
daduke: this is an ldap login environment with /dev/fuse set to world writeable ?
04:11
<daduke>
vagrantc: exactly. one thing might be that I got 3 lines of ltspfs on /tmp/.horeizo-ltspfs/usbdisk-sda1 type fuse (rw,nosuid,nodev,user=horeizo) in my mount output...
04:11
vagrantc: and another one /tmp/.horeizo-ltspfs/usbdisk-sda1 on /media/horeizo/usbdisk-sda1 type none (rw)
04:11
<vagrantc>
does that directory exist?
04:12
or /media/horeizo/ ?
04:12
<daduke>
vagrantc: 4 lines for one stick? the /tmp doesn't exist, but /media/horeizo does, that's where the stick is in
04:12
<vagrantc>
daduke: is it currently plugged in?
04:14
<daduke>
vagrantc: yep. another thing: I could create a local account for you on our terminal server, this might make things a little easier...
04:15
<vagrantc>
i tested the /dev/fuse as world-writeable ... but i probably didn't test with multiple users ... doing so now.
04:15
daduke: that would be useful ... if you can trust some random nick on irc :)
04:15
daduke: i could post an ssh key ...
04:16
<daduke>
vagrantc: you're not random any more... exactly, I was just about to ask for the key.
04:16
<vagrantc>
well, someone could easily impersonate me on irc :)
04:17
<daduke>
vagrantc: how's the weather in Portland? (trick question)
04:19johnny is now known as vagrantc_
04:19
<vagrantc_>
rainy
04:19vagrantc_ is now known as johnny
04:20
<daduke>
johnny: yeah but what if I trust you too?
04:20
<johnny>
i was just jokin around :)
04:20* daduke checking google weather.... that's correct! Hi vagrantc!
04:20
<vagrantc>
heh
04:20
<johnny>
daduke, for portanld it's easy
04:20
pacific northwest == rainy :)
04:21
<vagrantc>
actually, it's been really sunny lately ... for like two or three weeks
04:21
<daduke>
johnny: I know, been to Seattle last spring...
04:21
<vagrantc>
well, i can post a public ssh key that's gpg-signed by the same key that signs my ltsp uploads and such
04:21
<johnny>
i'll come up to portland and then pretend to be him :)
04:22
to his face
04:22
hehe
04:22
<daduke>
vagrantc: do it, I'm just waiting for the key
04:22
<vagrantc>
daduke: http://llama.freegeek.org/~vagrant/vagrant.pub
04:24
<daduke>
vagrantc: all right, try vagrantc@plimpy.ethz.ch
04:24
<ogra_cmpc>
oh, that remonds me of a little piece of paper i have lying around
04:25
*reminds too
04:26alekibango has quit IRC
04:26
<vagrantc>
ogra_cmpc: heh.
04:26
ogra_cmpc: that was ... back in ... june or july? :)
04:27
<ogra_cmpc>
vagrantc, i'll manage to sign it within the year :P
04:27
btw, do you plan to come to europe soon ?
04:27
<vagrantc>
daduke: password prompt
04:27
daduke: did you strip the gpg-signature ?
04:27
<ogra_cmpc>
you seem to work hard on getting into our timezone :)
04:28
<vagrantc>
ogra_cmpc: i pulled a 34-hour stretch sunday morning to monday night
04:29
<ogra_cmpc>
i'm always tempted to do that on thursdays now that i have that horrible 7:00am meeting on wed.
04:29
<daduke>
vagrantc: hmm.. I copied the ssh-rsa part into id_rsa.pub
04:30
<ogra_cmpc>
it needs to be in authorized_keys as well iirc
04:30
<vagrantc>
s,as well,instead,
04:31
<ogra_cmpc>
dont you need the pub part locally ?
04:31
<vagrantc>
nah
04:31
<ogra_cmpc>
ah, i thought you do
04:31
doesnt do harm though :)
04:31
<vagrantc>
i never have
04:31
<daduke>
vagrantc: ok try now
04:31
vagrantc: welcome to plimpy!
04:32
<vagrantc>
a quote from kurt vonnegut
04:33* ogra_cmpc always thought vagrantc was a vagrant ... and now he has accounts in switzerland
04:33
<ogra_cmpc>
tsk
04:34
<vagrantc>
daduke: so ... horiezo is a member of the fuse group
04:35
daduke: and my very recent experiments suggest that ... /dev/fuse being writeable ... isn't enough on etch.
04:35
<daduke>
vagrantc: yep, I think I included him manually, but for the others I'd like to use the 777 of /dev/fuse
04:35
vagrantc: ok, bad ;(
04:35
<ogra_cmpc>
is the udev rule right ? is it group owned ?
04:35
<vagrantc>
no, this is server-side
04:35
it's failing to mount
04:35
oh, you mean the udev rule for /dev/fuse ...
04:36
<ogra_cmpc>
yeah
04:36
crw-rw---- 1 root fuse 10, 229 2008-02-24 14:47 /dev/fuse
04:36
thats how it should be
04:36
i know debian goes a completely different way wrt creating the devices though
04:37
the ubuntu package has a different magic here
04:37
<vagrantc>
plimpy's got /dev/fuse world readable, writeable, executable ...
04:38
ogra_cmpc: that looks about the same as debian
04:38
<ogra_cmpc>
is fusectl mounted ?
04:38
fusectl /sys/fs/fuse/connections fusectl rw 0 0
04:38
<vagrantc>
not on plimpy
04:39
<ogra_cmpc>
i *think* debian does that with an initscript
04:39
(we have modprobe.d rules for it)
04:39* vagrantc checks on vagrantc's sid install
04:39
<daduke>
ogra_cmpc: *we* as in ubuntu?
04:39
<ogra_cmpc>
yep
04:40
<daduke>
ogra_cmpc: are debian and ubuntu finally slowly drifting apart?
04:40
<ogra_cmpc>
only with some packages
04:41
debian didnt like the approach of automating the fusectl fs from modprobe.d
04:41
<daduke>
ogra_cmpc: too modern? :)
04:41
<ogra_cmpc>
heh
04:41
well
04:42
we're sometimes just to fast in ubuntu ... often things just need some time to be recognized as good by debian
04:43
i can wait and go on maintaining the fuse patches ... its not that much :)
04:44
<vagrantc>
daduke: ok so ... the bad news ... chmod o+rw /dev/fuse ... doesn't work on etch or sid, as far as i can tell.
04:44
why it doesn't work ... i don't know.
04:45
<daduke>
*sniff*
04:45
<vagrantc>
how you could get it to work ... i don't know
04:45
maybe there's some setuid/setgid binaries ?
04:45
<ogra_cmpc>
vagrantc, fusermount is sgid fuse
04:45
<daduke>
vagrantc: but if we find a way of temporarily including users into the fuse group, we should be fine?
04:46
<vagrantc>
daduke: /usr/bin/fusermount
04:46
<ogra_cmpc>
oh, not /bin/fusermount ?
04:46
<vagrantc>
daduke: don't know how safe it is ... but chmod o+x that might work ...
04:47
<ogra_cmpc>
ah, likely because we use it in d-i now
04:47
<daduke>
vagrantc: I'll try...
04:48
*drumroll*
04:48
Yeah!!! we have a winner!
04:49
<vagrantc>
well cool.
04:49
<daduke>
vagrantc: indeedy!
04:49
vagrantc, ogra_cmpc: thanks so much guys!
04:49
<vagrantc>
that's a trick worth documenting.
04:49
<ogra_cmpc>
:)
04:50
vagrantc, but with a huuuuge security warning
04:50
<vagrantc>
ogra_cmpc: well, yeah.
04:50mcfloppy has joined #ltsp
04:50
<vagrantc>
just how evil is it?
04:50
<daduke>
vagrantc: is there any login script I could use to add $USER to fuse?
04:51
<vagrantc>
well, i think i'd best sleep on this.
04:51
<ogra_cmpc>
if you have ntfs3g installed (as we do in ubuntu by default) your users can wipe the windows disk for example
04:51plamengr has quit IRC
04:51
<ogra_cmpc>
every user can mount fuse filesystems
04:51
<vagrantc>
but i was asking about evil.
04:51
<ogra_cmpc>
rw
04:51
<vagrantc>
can mount fuse filesystems that the have access to the devices ...
04:52
<ogra_cmpc>
well, gnomes new gvfs defaults to fuse in all its backends
04:52
so you could do a lot damage with nautilus through fuse that way
04:52
<vagrantc>
daduke: i don't rightly know where to hook in to automatically add groups. but that would be a little better.
04:52
<ogra_cmpc>
its an open invitation to play with your fs
04:53
<daduke>
vagrantc: that's why I'm asking ;)
04:53
<ogra_cmpc>
probably pam
04:53
<vagrantc>
well, on a massive multi-user system where all the users need ltspfs anyways .... is that really much different from setting it o+rw and o+rx ?
04:54
<ogra_cmpc>
there is that pam-scripts module
04:54
<daduke>
vagrantc: IIRC you once said all login scripts are owned by $USER, right?
04:54
<ogra_cmpc>
probably not, but i wouldnt take that risk *especially because* its massive multiuser
04:54
<vagrantc>
daduke: yes ... so you'd need to figure out some criteria and give them sudo access or something.
04:55
daduke: the debian-edu folks might have some answers to this ... they're in the same boat with ldap auth and ltspfs and such
04:55
<daduke>
vagrantc: I'll see what I can do. If I figure it out, I'll let you know so that you can present an alternative to debian users.
04:56
<vagrantc>
daduke: i'd drop in #debian-edu on irc.oftc.net and ask about it ...
04:57alekibango has joined #ltsp
04:57
<daduke>
vagrantc: will do.
04:57
<vagrantc>
it's high time i get some sleep ...
04:57* vagrantc waves
04:58
<daduke>
vagrantc: thanks again and sweet dreams!
04:58* daduke needs some food...
04:58
<vagrantc>
daduke: of course
04:58vagrantc has quit IRC
05:07Pascal_1 has quit IRC
05:13spectra has joined #ltsp
05:29alekibango has quit IRC
05:33Nubae has left #ltsp
05:38alekibango has joined #ltsp
05:51spectra has quit IRC
05:51Guaraldo has joined #ltsp
06:29alekibango has quit IRC
06:35alekibango has joined #ltsp
06:45jammcq has quit IRC
06:57alekibango has quit IRC
07:02cyberorg has quit IRC
07:06alekibango has joined #ltsp
07:13ogra_cmpc has quit IRC
07:16cyberorg has joined #ltsp
07:20cyberorg has quit IRC
07:24cliebow_ has joined #ltsp
07:31cyberorg has joined #ltsp
07:33spectra has joined #ltsp
07:37EXP_ has joined #ltsp
07:37ogra_cmpc has joined #ltsp
07:37
<EXP_>
i have ubuntu 7.10 and ltsp5 how i can get nvidia drivers work?
07:37K_O-Gnom has joined #ltsp
07:38
<EXP_>
i installed nvidia-glx to chroot
07:38
and lts.conf XSERVER=nvidia
07:38
<ogra_cmpc>
you need linux-restricted-modules as well
07:38
<EXP_>
but i get a blank screen
07:38
<ogra_cmpc>
they are not installed by default
07:40Guaraldo has quit IRC
07:40
<EXP_>
ok. in some terminals have nvidia graphics and after logout mouse cursor becomes invisible.
07:40Guaraldo has joined #ltsp
07:41ogra_cmpc_ has joined #ltsp
07:44slidesinger has joined #ltsp
07:45
<EXP_>
this should fix it
07:45
anyone know where i can find instructions how to connect scanner to terminal?
07:46
it was quite easy in 4.2 but how about ltsp5?
07:49ogra_cmpc_ has quit IRC
07:59mikkel has quit IRC
08:04Dominik has joined #ltsp
08:05
<Dominik>
Hi all, anybody there? Jemand da, der vielleicht auch deutsch spricht?
08:05
<laga>
maybe / vielleicht
08:05
<Dominik>
08:06
ich hoffe du kannst mir helfen
08:06
hab hier einen relativ alten Thinclient (Compad ipaq 733Mhz)
08:06
ltsp 4.2 auf debian
08:06
pxe klappt alles wunderbar
08:07
08:08
08:08* laga kennt sich mit LTSP kaum und mit LTSP 4.2 im besonderen nicht aus
08:08
<laga>
schonmal den traffic mitgelesen mit wireshark z.b.?
08:09
also, PXE tut, und welches image will er nicht laden?
08:10
<Dominik>
naja, den kernel eben
08:10
08:10
08:10
<ogra_cmpc>
am gleichen switch ?
08:10
<Dominik>
aber ich schau mal grad im log
08:11
08:11
noch ein switch
08:11
<ogra_cmpc>
aha
08:11
probier erstmal direkt am ersten mit einem der clients
08:12
klingt als waer irdendwas geblockt
08:12
<Dominik>
ok
08:16
<EXP_>
now i get this "busybox v1.1.3 built in shell (ash) enter "help" for all built in commands
08:16
how i can pass this
08:16
<daduke>
hi all! Does anyone know about LDM_REMOTECMD? I set it to /usr/bin/xterm just to test, but it doesn't seem to have any effect...
08:17
<Dominik>
so, habs mal direkt am switch
08:18
mal schauen was er jetzt bringt
08:18
weder im syslog noch in der dmesg steht was drinne
08:22
hmhm, jetzt bekommt er keine ip mehr vom dhcp...
08:22
wie nervig
08:22
:)
08:22
ok, gleiches TFP.... Problem wie vorher
08:23
08:24
achso, derselbe client is mit ubuntu und ltsp5 hochgekommen....
08:25
hab auch schon verschiende configs verglichen und getestet, bringt aber keinen unterschied
08:27Gadi has joined #ltsp
08:28
<ogra_cmpc>
daduke, try LDM_SESSION instead ;)
08:29
<Dominik>
kein idee?
08:29
<ogra_cmpc>
Dominik, wenn er mit ubuntu bootet liegts wharscheinlich am alten kernel in 4.2
08:29
<Dominik>
deswegen findet er denn tftp nicht?
08:30
08:30
<daduke>
ogra_cmpc: can I give it any executable? or just xsessions? I'm still looking for a way to do this user group adding.. you remember
08:30
<ogra_cmpc>
thats whats executed in the ssh command ... you should be able to give it any executable on the server
08:31
<daduke>
ogra_cmpc: cool, lemme try...
08:33
<ogra_cmpc>
Dominik, rennt der dhcp server auf dem ltsp server ? oder separat ?
08:33
<Dominik>
08:33
nis, dhcp, dns
08:33
<ogra_cmpc>
hast du ne next-server zeile mit der ip des tftp servers drin ?
08:34
<Dominik>
steht drinne
08:34
<ogra_cmpc>
wo stoppts denn genau ?
08:35
<daduke>
ogra_cmpc: working! thanks.
08:35
<ogra_cmpc>
siehst du den kernel auspacken (ne reihe von punkten)
08:35
oder stoppts vorher
08:37_dominik_ has joined #ltsp
08:37
<_dominik_>
so
08:37
hat sich aufgehangen
08:37
stoppt vorher schon
08:37
TFTP.....
08:37
dann tftp timeout
08:38
<ogra_cmpc>
timeount ... hmm
08:38
<_dominik_>
PXE-E32
08:38
<ogra_cmpc>
siehst wahrscheinlich auch nix in den server logs
08:38
<Gadi>
http://www.bootix.com/support/problems_solutions/pxe_e32_tftp_open_timeout.html
08:38
<_dominik_>
ja, steht gar nichts drinne
08:38
<Gadi>
and gutentag
08:38
:)
08:38
<_dominik_>
hi :)
08:39
gadi, alles in frage kommt, also dieser seite ist leider auszuschliessen
08:39
<Gadi>
thank goodness Google translate knows what you guys are saying :)
08:39
<_dominik_>
next server steht
08:39
<ogra_cmpc>
Gadi, well, the tftp server is running (other clients run fine on the same machine)
08:40
next-server is in the dhcp config as well
08:40
so thats slightly weird
08:41
<Gadi>
old version of PXE?
08:41
on the client
08:41
<_dominik_>
build 071
08:41
<Gadi>
is this an identical client (hardware-wise) to a working one?
08:41
<_dominik_>
PXE-2.0
08:41
it is another one
08:41
<ogra_cmpc>
no
08:42
<_dominik_>
Compaq ipaq 733
08:42
<ogra_cmpc>
and its the 4.2 kernel ... apparently the same machine runs with ubuntu ltsp5
08:44
<Gadi>
1 sec -ph call
08:45
<ogra_cmpc>
_dominik_, ich wuerd echt ueberlegen ltsp5 zu nehmen .... 4.x wird seit mehr als 2 nicht mehr entwickelt oder gepflegt
08:45
*2 jahren
08:45
so ganz generell :)
08:46
debians ltsp5 ist fast identisch mit ubuntus ... sollte also fein funktioniern
08:47
<_dominik_>
ja, ltsp5 lief bisher nicht auf debian
08:48
war auch schon ein versuch, da hat fast nichts funktioniert
08:48
:/
08:48
uund finden tut man zu ltsp5 mit debian auch nichts
08:48
08:49
deshalb haben wir auch wieder den 4.2 aufgesetzt
08:49
<ogra_cmpc>
!debian
08:49
<_dominik_>
08:49
<ltspbot>
ogra_cmpc: "debian" is is a GNU/Linux based operating system that makes an excellent LTSP server. You can find it at http://www.debian.org. for information about LTSP on debian see http://wiki.debian.org/LTSP
08:49mccann has joined #ltsp
08:49
<ogra_cmpc>
jede menge :)
08:50
das edubuntu handbuch im edubuntu-docs paket von ubuntu hat ziemlich ausfuehrliche doku
08:50
sollte auch in debian installierbar sein
08:51
(ca 50 seiten, etwa 30 davon ltsp)
08:51
<_dominik_>
08:51
<ogra_cmpc>
sarge oder etch ?
08:51
<_dominik_>
etch
08:51
<ogra_cmpc>
etch mit den backport paketen sollte sehr gut laufen
08:52
komisch
08:52
<_dominik_>
ja, allerdings
08:52
<ogra_cmpc>
wobei selbst dioe relativ alt sind
08:52
etwa ubuntu feisty
08:53
<_dominik_>
08:55vaisarger has joined #ltsp
08:55
<ogra_cmpc>
tja
08:56
jeder wie ers braucht :)
08:56
<_dominik_>
jo :)
08:56
ich brauchs leider so :)
08:56
<vaisarger>
Hello people....
08:56
<ogra_cmpc>
waer mir zuviel gefummel ... :)
08:56
hi vaisarger
08:56
<_dominik_>
08:57
08:57* ogra_cmpc is veroehnt nach 4 jahren ubuntu
08:57
<ogra_cmpc>
*verwoehnt
08:57
<_dominik_>
hehe :)
08:57
<vaisarger>
ogra, I have to say you a little sad thing...
08:57
<_dominik_>
ja, in dem fall sollten wir auch mal umsteigen
08:57tarzeau has joined #ltsp
08:57mccann has quit IRC
08:57
<tarzeau>
HELP!
08:58
<_dominik_>
ich werde mich einfach mal an den ubuntu setzen
08:58
08:58
08:58
<ogra_cmpc>
tu das :) wenn fragen sind ... bin hier
08:58
<cliebow_>
Gadi:you really translating all that?
08:58
<ogra_cmpc>
hehe
08:58slap_ has joined #ltsp
08:59
<_dominik_>
joa, gut zu wissen, das einer da ist, der helfen kann :)
08:59
<ogra_cmpc>
cliebow_, i bet he"s working on a bot that parses it in realtime through google
08:59
<_dominik_>
wenn du immer da bist, trifft sich das noch besser :)
08:59
<cliebow_>
it figures..i am back online 14
09:00
<ogra_cmpc>
hab das meiste vom neuen ltsp geschrieben und heang hier viel rum :)
09:00
<vaisarger>
I asked some time ago how to help you all in lts project, but I really don't have time... sorry
09:00
<ogra_cmpc>
fuer den debian port is vagrantc zustaendig
09:00
vaisarger, thats ok
09:00
vaisarger, once you have time, come back :)
09:00
we wont run away
09:01
<_dominik_>
ok, also kann ich hier immer auf hilfe hoffen :)
09:01
<cliebow_>
i get..Long-term care can be every meet
09:01
<ogra_cmpc>
meistens, jup
09:01
<cliebow_>
If it is foreseeable that the weather. Not improving
09:01
<_dominik_>
ok, also bis dann , heut abend oder die tage ;)
09:01
<ogra_cmpc>
ciao
09:01
<_dominik_>
cu
09:01_dominik_ has quit IRC
09:01Dominik has quit IRC
09:01
<slap_>
Hello there! If someone can help me. I would like to install LTSP 5 on a fresh install of hardy alpha 4. First I installed ltsp-server-standalone and openssh-server. Then I checked for my router's configuration (dhcp off, range adress, ...). Next I openned dhcp.conf and everything seems alright. Build client. Update sshkeys.
09:02
<vaisarger>
Thank you.. see you all !
09:02
<slap_>
What's next?
09:02
<laga>
slap_: boot client?#
09:02
<ogra_cmpc>
boot a client ?
09:02
snap :)
09:02
<slap_>
;)
09:03
Well, the sesion stop at (i don't remember) init rmfs or something like that
09:03
IT's like a prompt
09:03mhterres has joined #ltsp
09:03
<ogra_cmpc>
is there an .img file in /opt/ltsp/images ?
09:04
and does it show up in /etc/inetd.conf as well ?
09:04
<slap_>
And now the PXE-E11 gives me a init timeout
09:05jobdrb has quit IRC
09:05
<slap_>
Yes, there is a image in /opt/ltsp/images...
09:05
<cliebow_>
check inetd.conf for nbdroot
09:06
<slap_>
I have this:"2000 stream tcp nowait nobody /usr/sbin/tcpd /usr/sbin/nbdrootd /opt/ltsp/images/i386.img"
09:06
<ogra_cmpc>
looks good
09:06
inetd is running ?
09:07
ps ax|grep inet
09:07
<vaisarger>
Greetings from Italy, bye!
09:07
<ogra_cmpc>
ciao vaisarger
09:07
<cliebow_>
netstat -anp|grep ":2000"
09:07
?
09:07vaisarger has quit IRC
09:07
<slap_>
ogra_cmpc: 5129 ? Ss 0:00 /usr/sbin/inetd
09:07
<ogra_cmpc>
or that
09:07
yeah, looks fine as well
09:07
hmm
09:08
<slap_>
There's something strange on the session...
09:08
<cliebow_>
slap_:what network card?
09:08
<ogra_cmpc>
i havent done any testbuild the last week ...
09:08
oh, wait
09:08
the kernel package just changed
09:09
<slap_>
After the the dhcp loockup...
09:09
<ogra_cmpc>
it can well be that the linux-ubuntu-modules package wasnt built yet
09:09
that leaves you without unionfs and squashfs support
09:09
<cliebow_>
ahh!
09:09
<slap_>
It shows me networking information and cannot connect to/with tftp
09:09
<ogra_cmpc>
which indeed wont boot then
09:09
<slap_>
yep
09:10
any suggestions?
09:10
<ogra_cmpc>
wait for a proper kernel build :)
09:10
<slap_>
lol
09:10
<ogra_cmpc>
without the unionfs and squashfs modules it wont work
09:11
the l-u-m package was uploaded yesterday i think ... should be ready tomorrow latest
09:11
<slap_>
That's splendid!!
09:11
thanks
09:11
<ogra_cmpc>
hardy is in flux, so it breaks sometimes
09:11
<slap_>
flux?
09:12
<ogra_cmpc>
moving forward all the time
09:12
changing
09:12
:)
09:12
<slap_>
We hope so
09:12
thanks again
09:12
<cliebow_>
8~)
09:13
<slap_>
Ubuntu or UbuntuStudio won't make any difference ( because of the RT kernel) ?
09:13
<ogra_cmpc>
i dont think so
09:14cliebow_ has quit IRC
09:14
<ogra_cmpc>
but note that you need to wipe /opt/ltsp/i386 and re-run ltsp-build-client if the new kernel is up
09:15
(or at least thats the easiest way :) )
09:16
<rjune>
probably a good plan to move /opt/ltsp/i386/ out of the way, then do it, yes?
09:16
<ogra_cmpc>
right
09:17
<daduke>
ogra_cmpc: unfortunately, whatever one executes via LDM_SESSION replaces the original session. Is there a way to daisychain it? The login scripts in /opt/ltsp/i386/usr/share/ldm/rc.d don't seem to know $USER ;(
09:17
<ogra_cmpc>
well in that case you can actually wipe it ... it has no working kernel anyway
09:17
daduke, its debian, right ?
09:18
<daduke>
ogra_cmpc: aye!
09:18
<ogra_cmpc>
you should be able to do something like: "/usr/bin/mycommand ; /etc/X11/Xsession"
09:18
<daduke>
ogra_cmpc: ahhh, good idea!
09:18
<ogra_cmpc>
(untested)
09:19cliebow_ has joined #ltsp
09:24
<daduke>
ogra_cmpc: working again!
09:26
<ogra_cmpc>
:)
09:26EXP_ has quit IRC
09:34mikkel has joined #ltsp
09:40alumno10 has joined #ltsp
09:40
<alumno10>
Hi all
09:40
i tried a coreduo as a client and youtube still see slow
09:41
the server is a amd 1,2ghz but now i was using just core duo client, so . whats the problem?
09:51
:/
09:51
all sleeping
09:53
<laga>
ogra_cmpc: sorry rfor the bug spam last night. just kept thinking of more stuff that needed to be changed..
09:53
ogra_cmpc: what's the convention for intending those scripts? spaces, tabs..?
09:55
<alumno10>
laga, do you think where is the lag?
09:55
<laga>
alumno10: probably your server or the network.
09:55
uncompressed movie playback over a fast ethernet network is gonna be.. slow
09:56
<ogra_cmpc>
right
09:56
i explained that yesterday already
09:56sepski has joined #ltsp
09:56
<ogra_cmpc>
laga, four spaces is a tab
09:56
<laga>
maybe you can use a local firefox instance.
09:56
ogra_cmpc: thanks. i'll clean my plugin then
09:57
<alumno10>
ogra_cmpc, the server is an amd 1.2ghz but the core duo is the only client in a base100 network
09:58
<Gadi>
alumno10: using LDM_DIRECTX?
09:58
<ogra_cmpc>
alumno10, yes, and still it needs to refresh the complete screen 24 times in a second to get fluid pics
09:58
<alumno10>
yes i am using it
09:59
<ogra_cmpc>
the only thing you can speed it up with would be smaller resolution or less colors
09:59
(if diectx is already set)
09:59
<alumno10>
yes, but, which kind of server or clients, can i use to get ok youtube vids?
09:59
<Gadi>
or better vid card
10:00
<ogra_cmpc>
gigE cards in the clients :)
10:00
<alumno10>
in coreduo vidcard is an intel 945, i catted xorg.conf in tty client console and is using right drivers
10:01
<ogra_cmpc>
and trunking together the server<-> switch connection to multiple gigE cards
10:01
<alumno10>
and what about if i run locally firefox?
10:01
<ogra_cmpc>
that will work as on a local desktop
10:01
but there ios no general implementaion of that, you will need to set up somethihng yourself manually
10:01
<alumno10>
how can i do to get locally running just firefox?
10:02
in an specific machine, cause now people are using another clients, cybrcafe are open now
10:03* Gadi thought this was one client
10:03
<Gadi>
:)
10:03
<alumno10>
early morning i did one client test
10:03
its the only time i can do that
10:03
<Gadi>
ah
10:04indradg has quit IRC
10:04
<Gadi>
gigabit switch?
10:04
or 100Mb?
10:04
<alumno10>
wait, ill see...
10:04indradg has joined #ltsp
10:04
<ogra_cmpc>
alumno10, internet cafe ? swo i assume you use autologin etc ?
10:05
<alumno10>
100
10:05
no i didnt use autologin
10:05
<rjune>
howdy Gadi
10:05
<Gadi>
alumno10: if it is a cheap switch, that may be a bottleneck
10:05
rjune: !!
10:05
not all switches are created equal
10:06
<alumno10>
but using a single client still video runs slow
10:06
<rjune>
what's new?
10:06
my company is looking at selling switches someday
10:06
<Gadi>
alumno10: maybe test tomorrow with a croosover cable between client and server
10:06
<alumno10>
how bandwith do i need to watch video fine?
10:06
<rjune>
lots
10:06
<alumno10>
lol
10:06
<Gadi>
its not bandwidth that you are up against, necessarily
10:06
<alumno10>
Gadi, that may be the solution
10:06
<ogra_cmpc>
alumno10, so how do you run that cafe ? do users get accounts ?
10:07
<rjune>
alumno10: xres * yres * colordepth * fps
10:07
do that math, then cry
10:07
<alumno10>
itis not just a internet cafe, its a basic computing learning center
10:07
<Gadi>
could be: latency, could be the flashplayer not syncing audio and video
10:07
<rjune>
if xdamage is setup, it'll be less
10:07
<alumno10>
thats cause we havent autologin
10:08
<ogra_cmpc>
alumno10, so you have different accounts one for every user ? or is it logged in in a generic way (i.e. account per machine)
10:08
<Gadi>
bbiab
10:08
<alumno10>
all user have different acounts
10:08
<ogra_cmpc>
ok
10:08
i thought you could get away with a kiosk system
10:08
but that wont work easily if you use real accounts
10:09
<alumno10>
so, if i run locally firefox i could make youtube run faster
10:09
<ogra_cmpc>
it will be as fast as on any normal desktop system, yes
10:10
<alumno10>
tomorrow i'll try with a crossover cable... that might show me what's the problem here
10:10
<Gadi>
alumno10: do your users need to print or save files from the web?
10:10
<alumno10>
yes, server have scanner/printer and users print from clients, uses usbkeys, ipods, etc...
10:11
<Gadi>
then, setting up local firefox will be a lot of manual work atm
10:11
<alumno10>
clients are crappy since they originally were bought to basic use (there are ebox systems)
10:11
<Gadi>
heh - and forget about running firefox locally on those
10:11
:)
10:11
<ogra_cmpc>
well, you can mount /home via nfs
10:12
<alumno10>
thats what i though :/
10:12
xD
10:12
<ogra_cmpc>
and have to copy the groups and password files regulary
10:12
that and an sshd on the client should suffice
10:12
its kjust a lot of maintenence work
10:13Egyptian[Home] has quit IRC
10:13
<alumno10>
humm, i though there was an lts.conf value like localrun or some
10:13
<ogra_cmpc>
but yeah, ebox might be a major blocker here
10:13
<alumno10>
localrun=firefox
10:14
yes, i know ebox ==shit
10:14Egyptian[Home] has joined #ltsp
10:14
<ogra_cmpc>
no,not shit
10:14
<alumno10>
thats why i used an coreduo client, to know whats the real problem videos run slow
10:14
<ogra_cmpc>
just for "specific use"
10:14
:)
10:15
<alumno10>
well.. a little shitty compared to other clients
10:15
<ogra_cmpc>
as i said, they're for sure great presentation display drivers etc .... database kiosk frontends ..
10:15
for single app usage with focused purpose :)
10:16
<alumno10>
yes, i know
10:16
that was the original use
10:16
a couple of apps
10:16
lightweight text processor, browser (no media)
10:17
<ogra_cmpc>
yeah
10:17
<alumno10>
and now nowadays kernel isnt supporting sis7019 sound card :/
10:17
<ogra_cmpc>
thats what they are good for
10:17
<alumno10>
so, i think the only way i have is to buy new faster clients
10:18
<ogra_cmpc>
you wouldnt be happy ifit would work, belive me
10:18
stuttering sound while moving the mouse isnt fun
10:20
<alumno10>
before, server had 2.6.8 kernel and sound works fine, but 2.6.22-14 current kernel doesnt support
10:20
ive tried to compile driver but i got bad module format
10:20
..."bad module format" message
10:21
<ogra_cmpc>
given that upstream will drop oss at some point in the near future thats pointless anyway
10:21
<alumno10>
ok, then the very only way i have is change clients?
10:22
<ogra_cmpc>
and yes, its knwn that the code doesnt compile proper modules with anything past 2.6.8
10:22
<alumno10>
i tried to run 2.6.8 kernel in clients but i got no boot lol
10:23
ok, client booted but boot stops
10:23
<ogra_cmpc>
i dont think there was ever a properly supported initramfs-tools kernel of 2.6.8 in ubuntu
10:24
<alumno10>
hummm
10:24
<ogra_cmpc>
so that cant work
10:24
<alumno10>
damn :/
10:24
<ogra_cmpc>
2.6.8 is four years old or so
10:24
ltsp5 only 2.5
10:24
<alumno10>
yes, its the kernel working on the other server
10:25
but the system is a debian
10:25
<laga>
2.6.8 was in stable version of debian, they have released a _new_ stable version since then. so it's really old. :)
10:25
<ogra_cmpc>
laga, 2.6.8 was in ubuntu arty (4.10)
10:25
so it was recent in mid 2005
10:25
err
10:25
3004
10:25
err
10:26
grmbl
10:26
2 0 0 4
10:26
<alumno10>
i copied 2.6.8 kernel from the debian server to this server and i got messed up boot as i said :P
10:26
(i copied to chroot system)
10:27
i though it worked since debian server have ltsp boot too, and the same clients
10:28
but older ltsp version
10:28
<ogra_cmpc>
the datasheet says that sound isnt supported btw http://www.disklessworkstations.com/cgi-bin/web/200110.html?id=zTWayLq3
10:30
<alumno10>
how old are ebox systems?
10:30
this info i can use to make a inform, so bosses buy new ones
10:30
:P
10:30
<ogra_cmpc>
a bit over a year
10:31
<alumno10>
the server and clients are from same date
10:31
so i think 3 or 4 years old
10:31
<ogra_cmpc>
the ebox was released about a year ago
10:32rjune has quit IRC
10:32
<ogra_cmpc>
i saw the first ones last year in april
10:32
<alumno10>
this isnt my sistem
10:32
<ogra_cmpc>
i think ebox released them in march
10:32
<alumno10>
in the website
10:32
the clients are uglier
10:32
xd
10:33
<ogra_cmpc>
i thought you said ebox 1000
10:33
but there might be former models with different cases
10:33
<alumno10>
http://sistemas.hiller.com.bo/imagenes/ebox-hand.jpg
10:34
thats the clients
10:37
did you see it?
10:38
<ogra_cmpc>
yep
10:38staffencasa has joined #ltsp
10:38
<alumno10>
Vortex86 200 MHz, 128 MByte SDRAM
10:38
<ogra_cmpc>
likely an older model
10:38
yeah, the newer one is 233 or 300 MHz
10:39
it has two models
10:40
<alumno10>
are you an ltsp developer?
10:41
<ogra_cmpc>
yes
10:41
one of a handfull
10:42
<alumno10>
i work in this basic computing learning proyect, who is financed by a huge fundation
10:42
and thats why i have to do things right here
10:43
ok, the point is , which client do you recommend me
10:45
<warren>
sigh, nothing really good out there
10:45
<alumno10>
the most compatible, the most long time support
10:45
not necesarily the most powerful, since in that case we would buy pcs
10:46
<warren>
alumno10, a huge problem is that pretty much all of the thin clients available on the markets have video chips poorly supported by X
10:46cdealer has joined #ltsp
10:46
<ogra_cmpc>
koolo or koolu makes good clients, but the amd driver is broken in gutsy and they need that
10:47
if you can wait til april ... it will definately be fixed in the april release
10:47
<warren>
ogra_cmpc, without an xorg.conf?
10:47
<alumno10>
i can install hardy alpha5
10:47
<ogra_cmpc>
warren, with or without ...
10:47
<cdealer>
ogra_cmpc, so far the wraper in firefox is working, we havent had no more problem with connection drop in our application ...
10:48
<ogra_cmpc>
they offer the systems as miminal desktops as well with a disk built in ...
10:48
no matter how ... that soecific chip is not working
10:48
<alumno10>
ill check it out
10:48
<ogra_cmpc>
cdealer, cool !
10:48
great to hear
10:49
<cdealer>
yeah ...
10:49
I was a bit incredulous about this was going to work ... but until now no problem
10:49
<ogra_cmpc>
just note that /usr/bin/firefox gets overwritten on package upgrades
10:49
ff is a typical candidate for regular security upgrades
10:49
<cdealer>
I made a $USER_ltsp variable to be the temp folder and everything is beeing keeped inside
10:49
Hmmm okay, good to know
10:50
<alumno10>
ogra_cmpc, i wanna be a developer too, im computing engineer, how can id o?
10:50
<ogra_cmpc>
alumno10, get familiar with bzr ... thats what we use to maintain our code
10:51
<alumno10>
ok
10:51
which are most compatible clients? koolu?
10:51
<ogra_cmpc>
the bigger ones from disklessworkstations are all fine
10:51
koolu as well (apart from gutsy)
10:52
<alumno10>
ok, so you think the disklessworkstations are best ones
10:52
<ogra_cmpc>
https://code.launchpad.net/~ltsp-upstream/ has the bzr branches with the code for ltsp
10:52
disklessworkstations is the company that funds ltsp developer meetings and owns the ltsp trademark ...
10:53
they are as good as all the other ones out there with similar specs
10:53
but if you buy them you fund the next ltspo developer meeting :)
10:53
<alumno10>
ok, i read about that
10:53
when its next meeting?
10:54
<ogra_cmpc>
none planned yet
10:54
we had one in sevilla in spain in april last year
10:54
<cliebow_>
ogra, ubuntu meeting today?
10:54
<ogra_cmpc>
and there was a gentoo porting spriint recently in portland afaik
10:54
cliebow_, edubuntu you mean ?
10:55
was at 12:00 UTC
10:55
<cliebow_>
heh..just thought oif it
10:55
<alumno10>
ok guys, i gotta lunch now, see you 1 or 2 hours later
10:56
i really like to cooperate with ltsp project
10:56
see you, thanks for all
10:56
bye
10:56
<ogra_cmpc>
ciao
10:57
<cliebow_>
i or 32 hours for kunch aye?
10:57alumno10 has quit IRC
10:58
<cliebow_>
keyboard is acting up again..8~)
11:06rjune has joined #ltsp
11:07likuidkewl has joined #ltsp
11:09likuidkewl has quit IRC
11:31mccann has joined #ltsp
11:40vagrantc has joined #ltsp
11:42
<Gadi>
anyone know in udev how to say "if ACTION!="add" or ACTION!="remove"
11:45
eh, nm
12:05jQWGfmOos has joined #ltsp
12:06
<jQWGfmOos>
Hi! I'm an experienced system administrator, but I'm absolutely a newbie towards LTSP: I was thinking about using it for a project: may I ask you a simple question? Just to know it ti's the right tool to use...
12:08
<vagrantc>
!question
12:08
<jQWGfmOos>
My customer is an hotel with 75 rooms
12:08
<ltspbot>
vagrantc: "question" is is if you have a question about ltsp, please go ahead and ask it, and people will respond if they can. :)
12:08
<jQWGfmOos>
!question
12:08
<ltspbot>
jQWGfmOos: "question" is is if you have a question about ltsp, please go ahead and ask it, and people will respond if they can. :)
12:08
<jQWGfmOos>
ok
12:08
My customer is an hotel with 75 rooms
12:09
they want to let their guests access internet (and pay for it!) from their rooms
12:09
i would like to put a thin client in every room and connect it to the LCD TV
12:09
now, the problem is:
12:10
i sould use their current authentication and payment system, that tracks every user by its IP... it's a so called "captive portal" system
12:10
so
12:10
how it will be?
12:11
every room will present the same ip to the "captive portal system" because firefox runs on the server?
12:12rcy` has quit IRC
12:12
<jQWGfmOos>
Is there any way to let every room navigate the internet using a different IP?
12:13
<vagrantc>
jQWGfmOos: yes, although if the thin-clients are powerful enough, you could run them as "diskless workstations" and be running the applications on the thin-client itself.
12:13ogra_cmpc has quit IRC
12:13
<vagrantc>
jQWGfmOos: then you'd be able to distinguish between the users.
12:14
jQWGfmOos: well, you'd get one IP per user, anyways.
12:14
<jQWGfmOos>
I was thinking about assigning 75 IPs to the server and using iptables "owner" match to SNAT the various firefox instances
12:14
is it feasible?
12:14
<vagrantc>
no idea ... would love to hear about sucess with it :)
12:15ogra_cmpc has joined #ltsp
12:16ogra_cmpc_ has joined #ltsp
12:16
<jQWGfmOos>
But tell me: how is it on the terminal server? (because I've never seen or used one) are there several firefox instances running? every instance run by a different user ID?
12:16
<vagrantc>
jQWGfmOos: every instance is run by a different user, yes.
12:16
<ogra_cmpc>
yes
12:16
<jQWGfmOos>
oh, great!
12:16
so i think it should work! :-)
12:17* johnny comes to the hotel and hacks the ltsp server
12:17
<vagrantc>
so if you can match iptables rules against user processes, you're set.
12:17
<johnny>
hehe
12:18
<vagrantc>
jQWGfmOos: will they need any apps other than firefox ?
12:18
<jQWGfmOos>
can I autologin a thin client?
12:18
just a PDF viewer i think, and of course flash plugin
12:18
<vagrantc>
jQWGfmOos: what linux distro and release?
12:18
<jQWGfmOos>
I think I will use Gentoo
12:19
<johnny>
jQWGfmOos, don't do that yet..
12:19
it's not ready with ltsp5 yet
12:19
<jQWGfmOos>
you mean autologin?
12:19
<johnny>
gentoo
12:19
<jQWGfmOos>
ah!
12:19
<johnny>
it's still WIP
12:20
<vagrantc>
the current distro's with good LTSP5 support are basically debian and ubuntu
12:20
<jQWGfmOos>
so is there a way so that the guest enters the room, switchs on power and the thin client is ready to be used without any login?
12:20
<johnny>
jQWGfmOos, unless you want to help ltsp5 on gentoo, upon which you would be more than welcome :)
12:20
yes
12:21
that's all good , it does require a seperate ebuild on ubuntu, but easy enough to get
12:21
to get the fixed version
12:21
lol.. err seperate deb
12:21
<vagrantc>
fedora seems to be making progress, and gentoo started ... and there's a few distros which have implemented ltsp5, but haven't merged with upstream
12:22
<jQWGfmOos>
so ok, if you tell me with ubuntu or debian I will have easyer life, I think I will go with ubuntu
12:23
<vagrantc>
although, autologin on ubuntu probably requires hardy
12:23
<ogra_cmpc>
yeah
12:23
<vagrantc>
i *think* it's broken on gutsy
12:23
<jQWGfmOos>
And of course if any of you comes to Toscany you will have special internet fees in the hotel! ;-)
12:23
<ogra_cmpc>
vagrantc, btw, we need a plan for autologin -> logout
12:23
<vagrantc>
ogra_cmpc: yeah, probably.
12:23
<ogra_cmpc>
i think a 10 second delay or so
12:24
<vagrantc>
ogra_cmpc: just put a sleep after the autologin stuff ?
12:24
<ogra_cmpc>
yeah, something like that
12:25
\so that the user has a chance to reach the shutdown button
12:25
<vagrantc>
ogra_cmpc: physical, or GUI ?
12:25
<ogra_cmpc>
gui
12:26
<jQWGfmOos>
Supposing I want to use the diskless workstation solution: does LTSP do it? Or it's another story?
12:26
<ogra_cmpc>
gdm has a countdown telling you "will log in user $USER in $time seconds"
12:26
<vagrantc>
this sounds like it would require Gadi's patches
12:26
because currently, it just restarts X after autologin session ends
12:26
<ogra_cmpc>
yeah
12:26
that needs to change
12:27
at least it shouldnt immediately log in again
12:27
<vagrantc>
jQWGfmOos: it's possible with LTSP5 ... but it requires some manual configuration
12:28
<jQWGfmOos>
last question, just to be sure I understood well: autologin is possible with LTSP5 but I need some extra patches, right? patches for what? LTSP or GDM?
12:29
<vagrantc>
jQWGfmOos: well, it works in debian unstable (not yet released), debian etch with backports, and ubuntu hardy (not yet released)
12:30
<jQWGfmOos>
vagrantc: so no need for patches?
12:30
<vagrantc>
jQWGfmOos: the patches are for LDM (LTSP Display Manager) ... you could use GDM but you'd have to configure sound and local device support somehow ...
12:31
<johnny>
just stick with dm...
12:31
err ldm
12:31
<vagrantc>
jQWGfmOos: you could probably add the patches to ldm in ubuntu gutsy, if you wanted to go that way.
12:31
<ogra_cmpc>
and the latter might get *very* tricky without ldm
12:31
<vagrantc>
jQWGfmOos: what time-frame are you looking at for deployment?
12:31
<jQWGfmOos>
Can you please point me at those LDM patches?
12:32
vagrantc: 1 month from now
12:32
<vagrantc>
ogra_cmpc: you think jQWGfmOos could just rebuild the ldm packages on hardy for gutsy?
12:32
<ogra_cmpc>
hardy will release around april 20th ...
12:32
<johnny>
they are on launchpad attached to the autologin bug
12:32
<ogra_cmpc>
vagrantc, i think francis has some in his ppa
12:33
dont have an url handy, sorry
12:33
<vagrantc>
ppa?
12:33
<jQWGfmOos>
Ok guys, you have been super-helpful! :-) Thank you all!
12:34
<vagrantc>
jQWGfmOos: so you've got some options. first thing i would do is make sure you can get your firewalling magic working ... and if you do, let us know how you do it! :)
12:34
<jQWGfmOos>
I will do some "experiments" and let you know, expecially about the "owner" iptables match, just in case someone else needs it in future... maybe if it works we could also add an entry in the wiki...
12:34jammcq has joined #ltsp
12:35
<vagrantc>
sure
12:35
<jQWGfmOos>
vagrantc: ok, you can bet it! I've got to go now, dinner time here! bye!
12:36jQWGfmOos has left #ltsp
12:39
<ogra_cmpc>
vagrantc, https://help.launchpad.net/PPAQuickStart
12:40* vagrantc does the m68k build of ldm for debian dance
12:40
<ogra_cmpc>
heh
12:41
<vagrantc>
as soon as that hits the archive, i should be able to do uploads of ltsp, ldm and ltspfs without requiring a sponsor :)
12:41
<ogra_cmpc>
how many weeks of build time did you plan?
12:41
<vagrantc>
as long as we don't introduce any new packages in the mix
12:42
ogra_cmpc: well, i'm a little confused ... seemed to start the build for ltsp and it took 24+ hours to upload, but i'm pretty sure that was mostly waiting for a human to sign and upload the package.
12:44
hppa, mips and mipsel are still blocking migration into lenny, though :(
12:44* vagrantc is tempted to upload from vagrantc's qemu-system-mips(el) machines
12:46
<vagrantc>
ogra_cmpc: so ppa builds packages on ubuntu's buildd's ?
12:46ogra_cmpc has quit IRC
12:52ogra_cmpc has joined #ltsp
12:52
<vagrantc>
ogra_cmpc: so ppa builds packages on ubuntu's buildd's ?
12:52
<ogra_cmpc>
right
12:52
<vagrantc>
pretty cool
12:53
would seem fairly easy to try and use that for backports
12:53
<ogra_cmpc>
wait until you only have to do one click to build a package from a bzr tree ;)
12:53
thats what francis does
12:54
<vagrantc>
seems like francis just patched versions from gutsy, as far as i can tell
12:54
<ogra_cmpc>
well, he put the patches from hardy in
12:55
or the ones he proposed for it at least
12:55
i think you didnt include them 1:1
12:56
<vagrantc>
no, i only included a slightly modified version of his autologin patch
12:56
<Guaraldo>
ltspbot: seen sbalneaves
12:56
<ltspbot>
Guaraldo: I have not seen sbalneaves.
12:56
<Guaraldo>
ltspbot: seen sbalneave
12:56
<ltspbot>
Guaraldo: I have not seen sbalneave.
12:57
<ogra_cmpc>
Guaraldo, about two weeks
12:57
and its sbalneav :)
12:57
!seen sbalneav
12:57
<ltspbot>
ogra_cmpc: sbalneav was last seen in #ltsp 3 weeks, 0 days, 3 hours, 58 minutes, and 11 seconds ago: <sbalneav> What are we looking at?
12:57
<Guaraldo>
ogra_cmpc: Ops...
12:57
<ogra_cmpc>
oh, three even
12:57
<vagrantc>
warren: i'm thinking your patch to support /usr/share/ltsp/screen.d/ should actually take priority over /usr/lib/ltsp/screen.d
12:57
<Guaraldo>
ogra_cmpc: Thanks...
13:00
<warren>
vagrantc, ok, I'll change the order.
13:00
<ogra_cmpc>
is there a reason we want both ?
13:00
<warren>
vagrantc, I'd like to eliminate the old location one day in order to further speed up boot
13:00
<vagrantc>
ogra_cmpc: backwards compatibility
13:00
<warren>
ogra_cmpc, I am trying to be careful to not break anything you depend on right now
13:01
<ogra_cmpc>
ah, right, that thing
13:01
<vagrantc>
ogra_cmpc: also means we can more easily drop Conflicts fields
13:01
<warren>
If you think we should fully eliminate it now, I wouldn't mind that.
13:01
<ogra_cmpc>
warren, i'm all for using share/
13:01
<warren>
Conflicts?
13:01
<ogra_cmpc>
packaging
13:01
<vagrantc>
warren: versioned package conflicts
13:01
<warren>
right now /usr/lib/ltsp is hard coded in various places in client/
13:02
<vagrantc>
warren: i.e. this version of FOO conflicts with this version of BAR
13:02
<warren>
I'm using it in my client package only because it would be difficult to code in the other location
13:02* vagrantc prefers /usr/share too
13:02
<warren>
well, /usr/share works for the arch-neutral scripts and stuff
13:02
<ogra_cmpc>
right
13:02
<vagrantc>
which is most of the scripts
13:03
<warren>
but if you build any binaries they belong in libexecdir, which would be /usr/libexec/ltsp on Fedora, or /usr/lib/ltsp in Debian/Ubuntu.
13:03
which means we need to add a make install that does substitution to ltsp-trunk
13:03
which means extra work
13:03
this is the reason why I didn't ask for moving away from /usr/lib/ltsp hard coded in client yet.
13:05
<vagrantc>
sure, but the screen.d scripts was easy ... so i'm all for moving that as soon as possible.
13:05
<ogra_cmpc>
++
13:07* vagrantc suggested moving most of the stuff in /usr/lib to /usr/share ... sometime around june of 2005
13:08
<ogra_cmpc>
i think mdz chimed in back then
13:08
cant remember why
13:08
<vagrantc>
i think it was simply "no" ... or maybe "no, not now"
13:09
<ogra_cmpc>
heh
13:09
<vagrantc>
that was before i really proved my mettle :)
13:09
<ogra_cmpc>
the typical detailed explanation you mean :)
13:17
<warren>
OK, how about this
13:17
1) we move everything from /usr/lib/ltsp to /usr/share/ltsp NOW
13:17
2) old location is unsupported, remove from all scripts
13:17
3) if we have any binaries they go into /usr/bin or /usr/sbin
13:17
#3 is already true
13:17
right?
13:18
<vagrantc>
warren: breaking backwards compatibility will be a major pain.
13:18
<warren>
to what?
13:18
what outside of ltsp depends on that location?
13:18
<vagrantc>
warren: it will require versioned dependencies and conflicts on all of our packages.
13:18
<warren>
<vagrantc> sure, but the screen.d scripts was easy ... so i'm all for moving that as soon as possible.
13:19
As soon as possible is when?
13:19
<vagrantc>
as long as the old location doesn't break, i would be happy to do it today. so i'm contesting point 2
13:19
anything that continues to work with /usr/lib i am happy to move immediately.
13:20
<warren>
screen.d is one case
13:20
in other cases it is less easy
13:20
<vagrantc>
agreed.
13:20* warren searches for other cases
13:24
<vagrantc>
actually, everything is part of a single package ...
13:24
at least on debian.
13:25slap_ has quit IRC
13:25
<vagrantc>
so, other than screen.d, there is no need to maintain any backwards compatibility, as far as i'm concerned
13:25
yay!
13:26
and screen.d is easy to maintain backwards compatibility
13:32alumno10 has joined #ltsp
13:32
<alumno10>
hi all
13:32
is there a problem with intel gigabit integrated adapter?
13:32
..and edubuntu.
13:34
<laga>
what problem would that be?
13:36
<alumno10>
i got clients closing session
13:36
and in logs i get some nic error
13:36
related with ssh i think
13:37
<vagrantc>
ogra_cmpc: any objections ? all the stuff in /usr/lib/ltsp is either in screen.d (which is easy to fix) or part of ltsp-client-core
13:40
<alumno10>
vagrantc, how can i do to boot clients from another kernel?
13:40
2.6.22-14 doesnt support ebox audio
13:40
but 2.6.8 does
13:46ariane has joined #ltsp
13:48
<vagrantc>
alumno10: become a kernel hacker.
13:48
<alumno10>
but if i put 2.6.8 kernel into chroot
13:48
and ltsp-update-kernels
13:49
do you thing does it work?
13:50mikkel has quit IRC
13:51
<alumno10>
vagrantc, what do you think?
13:52toscalix has quit IRC
13:53
<vagrantc>
alumno10: the short answer is "no, it will not work"
13:53
<alumno10>
could you pls tellme why?
13:54
<johnny>
too old
13:54
too much has changed
13:54
since 2.6.8
13:55
<alumno10>
johnny, ok, then what can i do if this is the last kernel that supports my client's sound hardware?
13:55
theres a source, but i compile it and when i try to modprobe it i get "bad module format"
13:55
i have like a week in this problem
13:56
spended
13:56
my clients are ebox
13:56
sound sis7019
13:56
<warren>
vagrantc, ok, should I change all occurrances of /usr/lib/ltsp then?
13:57
<vagrantc>
warren: i'm ok with it.
13:57
<warren>
vagrantc, except I'll keep the screen.d in screen_sessions
13:57
ogra_cmpc, are you ok with this?
14:10cdealer has quit IRC
14:11
<alumno10>
vagrantc, so i just can change chroot kernel by a newer?
14:11
<vagrantc>
alumno10: no.
14:12
<alumno10>
and how can i update it?
14:12
<vagrantc>
alumno10: you have to backport all of the depending software as well.
14:12
alumno10: it is extremely complicated.
14:12
alumno10: and well over my head.
14:13
<alumno10>
inwhich cases should i use ltsp-update-kernels ?
14:13
<vagrantc>
alumno10: any time you update the kernels in the chroot
14:13
<alumno10>
i mean, if its so difficult to use another kernel, why i should want to?
14:14
in which cases could i change kernels in chroot?
14:14plamengr has joined #ltsp
14:15
<alumno10>
in which cases could I need change kernels in chroot?
14:15
<vagrantc>
alumno10: if you don't know why supporting an ancient kernel version is difficult with current software, then i don't know how to respond.
14:16
<alumno10>
vagrantc, i understand that, i just ask in which cases could i need to change kernel i chroot? when an official ltsp kernel releases only?
14:16
<ogra_cmpc>
warren, vagrantc fine with me, i'm not sure i will make another upstream checkout before release, i'll likely cherrypick patches from now on
14:17
so its likely to only affect interpid ibex
14:17
<cliebow_>
intrepid...nice..
14:18
jammcq:still using vmware?
14:25
<rjune>
ARG!
14:26
<ogra_cmpc>
??
14:27alumno10 has quit IRC
14:29bobby_C has joined #ltsp
14:31
<laga>
rjune: ditto
14:32
cheap usb ethernet adapter just crashed my box :/
14:32sepski has quit IRC
14:32
<rjune>
laga: you want to beat your co-workers?
14:32
<laga>
no
14:40Gadi has left #ltsp
14:41
<slidesinger>
Has anyone ever tried to run x across a vpn?
14:42
<jammcq>
cliebow_: yeah, I use vmware all the time
14:42
<slidesinger>
Hey jammcq
14:42
<jammcq>
hey slidesinger
14:42
how goes it?
14:42
<slidesinger>
Quite well, thanks.
14:42
And with you?
14:43
<jammcq>
very well
14:43
<warren>
vagrantc, ok, i'll do it
14:43
jammcq, hey, did you send me a TK-3550?
14:43
<cliebow_>
jammcq. my bridge seems afu..looks fine with vmware-config but get no ip
14:43
<jammcq>
warren: ummm
14:43
<cliebow_>
in either win or gutsy as a vm
14:43
<warren>
(I have a TK-3550 with no label of where it came from. I suspect it was you.)
14:43
<jammcq>
cliebow_: is that on a laptop?
14:47
<cliebow_>
no..desktop
14:47
<jammcq>
cliebow_: which ethX did you bridge to?
14:48
<cliebow_>
vmnet 0 and 8 show fine on host..eth0
14:48
thi in hardy
14:48
any wat to tell how that is bridged?
14:49
<jammcq>
is there a cable plugged into eth0 ?
14:49
<cliebow_>
yessiree
14:49
<jammcq>
hmm
14:49
<cliebow_>
works finee using winblows as a native os
14:49* ogra_cmpc pats his virtualbox
14:50
<cliebow_>
.me cliebow pats J45p3r
14:52
jammcq:just thought something obvious...
15:04plamengr has quit IRC
15:05plamengr has joined #ltsp
15:07oh207 has quit IRC
15:08mhterres has quit IRC
15:14cliebow_ has quit IRC
15:43
<vagrantc>
warren: ok ... so the scripts in ldm ... y'all moved those from /usr/share to @libexecdir@ ...
15:43plamengr has left #ltsp
15:43
<warren>
vagrantc, yes, didn't everyone agree on that at the time?
15:43
<vagrantc>
warren: i don't see the difference between /usr/share for ltsp and /usr/share for ldm ...
15:43* warren looks at it again
15:44
<vagrantc>
warren: why is it ok for scripts of similar functionality to go in /usr/share in ltsp and for it to need to go in @libexec@ for ldm ?
15:44
<warren>
isn't ldmgtkgreet a binary?
15:44
<vagrantc>
warren: yes, but the rc.d scripts aren't
15:45
<warren>
ok yeah, the rc scripts don't need to be in libexec
15:45
although I vaguely recall ogra suggesting it should all go in there
15:45
<vagrantc>
i mean, i've made the transition already, so i'm just wondering why y'all insisted that they be moved.
15:46
<warren>
vagrantc, compliance reviewers in Fedora pointed out that it violated our packaging standards
15:46
/usr/lib is a no-no
15:46
<vagrantc>
warren: that i understand ... but /usr/share ?
15:46
<warren>
so we went along with the GNU standard of libexecdir
15:46
vagrantc, /usr/share isn't good for non-arch things?
15:46
<vagrantc>
warren: this was moving something *from* /usr/share *to* @libexecdir@
15:46
warren: and most of the stuff is arch-independent.
15:47
warren: ldmgtkgreet, i agree with moving to libexec ... but the rc.d scripts for ldm are pretty much in the same boat as the screen.d scripts for ltsp.
15:48
so it just seems like there's some inconsistant changes going on, and i'd like to try and avoid unnecessary changes as much as possible.
15:50
but again, i've already made the transition... but i don't want people to suddenly leap up and shot "now ldm rc.d scripts belong in /usr/share" and have to transition back.
15:51
<laga>
wow. i ran ltsp-build-client with --arch i386 on amd64 and it created an amd64 chroot. file name is /opt/ltsp/i386, though
15:51
<vagrantc>
cool.
15:52
<warren>
laga I fixed that particualr thing in fedora =)
15:52
<laga>
warren: cool.
15:52
<warren>
vagrantc, I agree it is inconsistent, and it was a bad idea to move rc.d there
15:52
<laga>
could be caused by my patches or by the fact that i'm running the hardy packages on gutsy but i doubt it
15:53gonzaloaf_work has quit IRC
15:58
<warren>
vagrantc, what is that rc.d for anyway?
15:58
vagrantc, my rc.d is empty
16:00
<vagrantc>
warren: ltspfs puts hooks in there
16:00
warren: it's for login/logout scripts on the client
16:00
<warren>
OK, I'm all for moving that to /usr/share/ltsp
16:00
is that the only thing to move in ldm?
16:01
<vagrantc>
i think so ... that's a pretty easy move, i think.
16:01Guaraldo has left #ltsp
16:01
<warren>
you want to do that part?
16:01
I'm working on ltsp-trunk now
16:01
vagrantc, does ltspfs need to be changed for that move too?
16:02
<vagrantc>
warren: this can support all locations
16:02
<warren>
vagrantc, what can support all locations?
16:02
<vagrantc>
the ldm rc.d scripts
16:02
<warren>
vagrantc, we don't need to support libexecdir for rc.d scripts
16:02
it wasn't that way for long
16:02
right?
16:03
<vagrantc>
it's been that way for nearly 3 months in debian, and ubuntu's about to release a long-term-support release with it
16:03
and it's trivial to keep support for
16:04
<warren>
huh? we didn't move libexecdir until january
16:04
during fudcon raleigh
16:04
right?
16:04
too late to change it back in Ubuntu?
16:04
It was a bad idea.
16:04
<vagrantc>
ah, perhaps you're right.
16:04
<warren>
I recall it was ogra's idea
16:04
but I agreed at the time
16:05
<vagrantc>
i dragged my feet with it but couldn't stop the forces of change and chaos
16:09
warren: well ... i'll work on switching ldm back, but i'll keep the libexecdir support in there since we've already done the work to support multiple locations.
16:10
<warren>
sigh
16:11
I will accept partial responsibility for this.
16:15otavio has joined #ltsp
16:15
<vagrantc>
otavio: hi! :)
16:15
pwd
16:15
ls
16:15
<otavio>
hi
16:24
<vagrantc>
warren: committed changes to ldm rc.d stuff
16:26
warren: i guess those changes were done in mid-january ... seems like ages ago :)
16:27
heh... and actually just introduced in my most recent upload ... on feb 19th
16:28
so it hasn't been in debian very long at all ...
16:29
so i guess if ogra doesn't have any objections, we could actually drop support for it ...
16:29
ogra_cmpc, ogra ?
16:31
warren: you still working on the switch to /usr/share for ltsp, or would you mind if i did it?
16:33ariane has quit IRC
16:37
<warren>
sorry back
16:37
vagrantc, I'll do it
16:41
vagrantc, so confirmed that screen.d is the only place we might support the old location?
16:42
<vagrantc>
warren: as best i can tell, it's the only place that should need to.
16:42
<warren>
vagrantc, k
16:44K_O-Gnom has quit IRC
16:48elisboa has quit IRC
16:52elisboa has joined #ltsp
16:55slidesinger has quit IRC
16:55
<warren>
vagrantc, where is bzr shelve from?
16:55* warren doesn't see it
16:55otavio has quit IRC
16:55
<warren>
anyway, I'm working more on this on the busb
16:55
bus
16:55
bbl
16:57
vagrantc: hmm, does bzr have anything like git's local checkin?
16:57
<vagrantc>
warren: not knowing what local checkin is ...
16:57
warren: bzr shelve is from bzrtools
16:58otavio has joined #ltsp
16:59mistik1 has quit IRC
17:03mistik1 has joined #ltsp
17:04
<laga>
ogra, ogra_cmpc: any ETA on a new LTSP upload?
17:25jammcq has quit IRC
17:33cliebow has joined #ltsp
17:34oh207 has joined #ltsp
17:39bobby_C has quit IRC
17:40supreme has joined #ltsp
17:40
<supreme>
hi all
17:40
im alumnoXX
17:46gonzaloaf has quit IRC
17:50gonzaloaf has joined #ltsp
17:51andarilho has joined #ltsp
17:53
<andarilho>
Hello, how disable sound in the thin clients?
17:54
<supreme>
andarilho, edit lts.conf
17:55
<andarilho>
Yes
17:55
<supreme>
and change sound value into false
17:55
SOUND=false i think
17:55
<andarilho>
Change, and nothing changed
17:56
Carry on, playing on the thin cliente
17:56
clients
17:56
<supreme>
i assume that you reboot clients
17:56
cliente? do you speak spanish?
17:56
portuguese?
17:56
<andarilho>
Português
17:57
<supreme>
i see... eu nao falo portugues :/
17:57
<andarilho>
Estou me esforçando no meu inglês
17:57
<supreme>
but english , my main language is spanish
17:57
<andarilho>
all right
17:58
<supreme>
ok, if lts.conf modification doesn't works you always can disable sound module in chroot filesystem
17:58
but lts.conf might work
17:58
<andarilho>
Supreme, i reboot the clients and continue playing
17:58
<supreme>
which ltsp version do you have?
17:58
<andarilho>
Yes, but want the sound in the server
17:58
This is problems
17:59
Version 5.0
17:59
Ubuntu 7.10
17:59
<vagrantc>
andarilho: what linux distro and release?
17:59
andarilho: where did you edit lts.conf ?
17:59
<andarilho>
/opt/ltsp/.....
18:00
<vagrantc>
andarilho: you need to edit /var/lib/tftpboot/ltsp/i386/lts.conf
18:00
<supreme>
in older versions you had to modify /opt/ltsp/i386/etc/lts.conf, but..... vagrantic continued my sentence
18:00
xD
18:01
<andarilho>
Ok,
18:02
<vagrantc>
it should have a big warning in /opt/ltsp/i386/etc/lts.conf about that
18:02
<andarilho>
vagrantc: no find this file a /var/lib/.....
18:03
<vagrantc>
andarilho: create it
18:03
<supreme>
andarilho, you have to create
18:03
<andarilho>
Ok
18:03
<vagrantc>
andarilho: be sure to include [Default] at the top
18:04
<andarilho>
I did create
18:05
Ok, [Default]
18:06
And now?
18:06
<vagrantc>
SOUND=false
18:06
<andarilho>
add ::
18:06
Ok
18:06
Just: Sound=false
18:07
<vagrantc>
andarilho: did you change any other options in /opt/ltsp/i386/etc/lts.conf ?
18:07
<andarilho>
I change, i know
18:07
I did have read about LTSP
18:08
<vagrantc>
you shouldn't have to change anything by default
18:08
only add options for things you want to change.
18:08
<andarilho>
Ok, thanks vagrantc
18:08
Now, is time of test
18:21twinprism has quit IRC
18:22
<supreme>
andarilho, how is it going?
18:23
<andarilho>
I'm waiting a user finish your work
18:23
One no, various
18:24twinprism has joined #ltsp
18:28
<vagrantc>
andarilho: you shouldn't need to reboot the server ... just a thin-client ...
18:29
<supreme>
damn, vagrantic always spokes what im just going to say,,, if my main language were english ...
18:33supreme has quit IRC
18:34mccann has quit IRC
18:35
<andarilho>
I did know vagrantc, i be waiting the user in the terminal
18:35
<vagrantc>
only one terminal?
18:36
<andarilho>
no, 15,
18:37
Laboratory of school
18:37
<vagrantc>
all users are using all 15 ?
18:37
<andarilho>
yes,
18:37
<vagrantc>
ah.
18:38
<andarilho>
vagrantc: but this resolve my problem?
18:39
<vagrantc>
andarilho: hope so
18:40
<andarilho>
ok
18:40RyanRyan62 is now known as RyanRyan52
18:56
<andarilho>
vagrantc: the sound continue playing
18:56
I reboot one client
19:00
<vagrantc>
a mystery to us all.
19:04staffencasa has quit IRC
19:10
<andarilho>
Verdade
19:10
true
19:11
vagrantc, tomorrow i come back to resolve this problema
19:11
problem*
19:11andarilho has quit IRC
19:24vagrantc has quit IRC
19:55otavio has quit IRC
19:56otavio has joined #ltsp
20:18jammcq has joined #ltsp
20:19cliebow has quit IRC
20:30spectra has quit IRC
20:39steph_ has joined #ltsp
20:41
<steph_>
Can someone tell me why the new distribution upgrade tries to remove linux-rt package ? (hardy alpha 4)
20:53RyanRyan52 has quit IRC
21:00RyanRyan52 has joined #ltsp
21:11mccann has joined #ltsp
21:19elisboa has quit IRC
21:19elisboa has joined #ltsp
21:20elisboa has joined #ltsp
21:26steph_ has quit IRC
21:44ogra_cmpc has quit IRC
21:45ogra_cmpc has joined #ltsp
22:04alekibango has quit IRC
22:08RyanRyan62 has joined #ltsp
22:15RyanRyan62 has quit IRC
22:18RyanRyan52 has quit IRC
22:21yyeago has joined #ltsp
22:24
<yyeago>
I am having trouble understanding the general setup described in step 2 in this doc https://help.ubuntu.com/community/UbuntuLTSP/LTSPFatClients
22:24
My goal is to have a generic desktop setup on several machines that have perfectly fine hardware specs
22:25
this document seems to cover everything but I am not quite decrypting how this chroot jail installation works, where it runs from, and how the clients use it.
22:25
I imagine that it is installed on opt/ltsp of the server but that doc doesn't indicate that
22:27
<Ryan52>
yyeago: actually, what part confuses you
22:27
<yyeago>
it says to install it to /opt/ltsp but ... on the client? the server?
22:27
if its the server.... where does this doc indicate how to get the client to load from it
22:28
<Ryan52>
okay so how it works is the client gets its filesystem from the servers /opt/ltsp/something
22:28
<yyeago>
i see.
22:28
super cool.
22:28
I don't see where this doc indicates how to configure the client to the server.
22:29
<Ryan52>
the client gets the kernel location from dhcp and downloads it from tftp. Then it mounts the /opt/ltsp/something as its root using nfs
22:29
<yyeago>
NFS. aha.
22:29
<Ryan52>
yyeago: sudo chroot /opt/ltps/something
22:29
<yyeago>
right.
22:29
cool.
22:29
that clears a bunch up.
22:29
<Ryan52>
and you will be in the filesystem that the client will get
22:29
<yyeago>
I like that.
22:29
super maintainable.
22:30
<Ryan52>
ya
22:30
<yyeago>
I've already got LDAP working (although I will have to redo it in this chroot jail)
22:30
<Ryan52>
yyeago: well good luck
22:31
<yyeago>
Ryan52, thanks. I really apprecaite it
22:31
<Ryan52>
np
22:35basanta has joined #ltsp
22:37
<yyeago>
Ryan52 -- sudo chroot /opt/ltsp/fati386 mount /proc -t proc /proc
22:37
can you explain how chroot and mount are interacting here?
22:39
<Ryan52>
yyeago: what said to run that command?
22:40
ohh
22:40
<yyeago>
that doc, Ryan42
22:40
<Ryan52>
do you know what proc is?
22:41
<yyeago>
processor?
22:41
or ram...?
22:41
<Ryan52>
proc is a filesystem that has lots of info
22:41
so /proc/cpuinfo has info about your cpu
22:41
and other stuff like that
22:42
<yyeago>
ok
22:42
<Ryan52>
so what that command does is mount the proc filesystem in the chroot
22:42
<yyeago>
the proc filesystem of the local machine.... right?
22:42
<Ryan52>
because one of the commands that you are about to run needs the proc filesystem to get certain information
22:42
its just temporary
22:43
It mounts your servers proc filesystem in the chroot
22:43
<yyeago>
weird.
22:43
<Ryan52>
so you can run commands that use info in proc
22:43
<yyeago>
I thought it was there so that processor / ram wasn't going thru network but thru the local machine
22:43
<Ryan52>
no...
22:43
proc is called proc because it was originally designed to get information about different running processes
22:44
<yyeago>
oh.... that was the whole point of 'fat' client vs thin.
22:44
<Ryan52>
but they added other stuff and made a big mess
22:44
no...
22:44
a fat client is where the programs run on the client and a thin client is where the programs run on the server
22:44
*I think*
22:44
ya..thats the only diffeerence
22:47* Ryan52 hopes he's not being too confusing...
22:47
<yyeago>
right.
22:47
no, I gotcha.
22:48
<Ryan52>
good :)
22:48
<yyeago>
I simply misunderstood the whole reason for mounting proc
22:48
=)
23:05RyanRyan52 has joined #ltsp
23:05BGomes has joined #ltsp
23:07yyeago has quit IRC
23:12mccann has quit IRC
23:20Jester45 has joined #ltsp
23:20
<Jester45>
with ltsp does the server do all the processing or the clients?
23:22
<RyanRyan52>
Jester45: The server
23:22
Jester45: Once you log in using ldm, you are on the server
23:23
<Jester45>
humm i guess i should be looking at a diffrent project
23:23
<RyanRyan52>
Jester45: You can use a fat client setup where the processing is done on the clients though
23:23
<Jester45>
thats what im looking for
23:23
<RyanRyan52>
debian or ubuntu?
23:23
<Jester45>
doesnt matter i love both but i think xubuntu would be the easier
23:24
server would be debian
23:24
<RyanRyan52>
ubuntu has a nice fat client guide
23:24
<Jester45>
they have ltsp install option
23:25
<RyanRyan52>
it should be similar with debian
23:25
ya
23:25
<Jester45>
and im sure i can but can all the clients use the same image?
23:26
<RyanRyan52>
yes
23:26
they use the chroot in thats on the server
23:26
<Jester45>
they have diffrent hardware so i guess they might need the live cd or somthing that will auto config them
23:26
<RyanRyan52>
it will autoconfigure them
23:26
<Jester45>
cool
23:26
<RyanRyan52>
ltsp is smart...
23:26
<Jester45>
lol
23:27
and can i keep users from making changes to the image but login with a admin account or somthing to change things
23:27
like disply settings or settings for firefox
23:27
<RyanRyan52>
so to do fat clients, you do the same thing with ltsp-build-client. But to install new stuff (you'll want to install gdm) you do chroot /opt/ltsp/some_chroot and then do admin stuff
23:28
<Jester45>
k
23:28
so ill just have to shroot and start a 2nd x server
23:28
its for my school
23:28
<RyanRyan52>
no...you'll just want to chroot and run some commands
23:29
https://help.ubuntu.com/community/UbuntuLTSP/LTSPFatClients
23:29
<Jester45>
we have a lab that im going to setup so the other non linux students can learn and play with and not mess up the system
23:29
<RyanRyan52>
Thats for ubuntu...it might help with debian
23:29
cool
23:29
<Jester45>
i could run the server with ubuntu also
23:30
<RyanRyan52>
ya...that would be the easiest
23:30* RyanRyan52 usually says to use debian...but in this case ubuntu has better docs
23:32yyeago has joined #ltsp
23:33alekibango has joined #ltsp
23:33
<Jester45>
well im sure i will be back in here later
23:34
<RyanRyan52>
Jester45: good luck
23:51BGomes has quit IRC
23:52yyeago has quit IRC