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


Channel log from 10 March 2007   (all times are UTC)

00:09sbalneav has quit IRC
00:27Faithful has joined #ltsp
00:40
<Lumiere>
hi Faithful
00:40
<Faithful>
Hey Lumiere how is it?
00:47vagrantc has quit IRC
02:35
<g333k_laptop>
anybody home?
02:43
<Davo_Dinkum>
What is that your catch phrase?
02:49
<g333k_laptop>
Davo_Dinkum, hi
02:49
<Davo_Dinkum>
g333k_laptop: Howdy
02:49
<g333k_laptop>
Davo_Dinkum, trying to catch some bat
02:50
<Davo_Dinkum>
bat?
02:50
<g333k_laptop>
Davo_Dinkum, well I want to download the source code of ltsp to read it an start contributing to the project
02:51
Davo_Dinkum, because of the hour... its 3.47am here, just bats are awake :p
02:51
<Davo_Dinkum>
Ah right
02:52
g333k_laptop: I can't find it on ltsp.org either.
02:52
<g333k_laptop>
Davo_Dinkum, but I dont know where start from
02:52
<Davo_Dinkum>
Here it is: http://sourceforge.net/project/showfiles.php?group_id=17723
02:54
<g333k_laptop>
This is ltsp 4, I want lstp 5
02:54
<Davo_Dinkum>
Oh right.
02:56
<g333k_laptop>
I think they have stopped development of ltsp 4.2
02:56
<Davo_Dinkum>
Possibly. I have no idea.
02:57
g333k_laptop: Would you happen to know what an ltsp thin client needs on the hard drive?
02:57
Or can it boot from PXE/etherboot?
02:58
<g333k_laptop>
Davo_Dinkum, yep, the can boot from PXE
02:58
Davo_Dinkum, actually your clients doesnt need to have hard disk drives
02:58
just a NIC card
02:58
<Davo_Dinkum>
Good.
02:59
But could I set it up to use the HDD as a buffer/cache, to speed things up?
02:59
<g333k_laptop>
Davo_Dinkum, of course the mainboard should support booting from PXE
02:59
Davo_Dinkum, would it speed things up?
02:59
<Davo_Dinkum>
Dunno.
03:00
<g333k_laptop>
Davo_Dinkum, I dont think so, RAM is faster than hard disks
03:12Egyptian[Home1 has quit IRC
03:34robbie has quit IRC
03:40Core has joined #ltsp
04:49Egyptian[Home1 has joined #ltsp
04:56Egyptian[Home] has quit IRC
05:11Egyptian[Home1 has quit IRC
05:31antgel has joined #ltsp
05:38antgel has quit IRC
06:16zald has quit IRC
06:49Davo_Dinkum has quit IRC
07:06Faithful has quit IRC
07:17codey has quit IRC
07:17codey has joined #ltsp
07:18freemind has joined #ltsp
07:24
<cliebow_>
g333k_laptop, if you are talking ltsp4.2 d/l the lbe..if ltsp5 root through your chrott
07:37zald has joined #ltsp
07:52plamengr has joined #ltsp
07:54
<plamengr>
hello! I'm on Debian sid with LTSP5 and the client boots fine. I have problem with logon on the client
07:54
I have another server with similar configuration which works fine
07:55
I tried to configure na 2nd server in similar way, but I have this problem with the client logon?
07:55
Did you have the same problem, or any suggestions?
08:38
<cliebow_>
plamengr, perhaps regnerate ssh keys?
08:42
<plamengr>
I'll try this, 10x
08:43
<cliebow_>
not sure of the command sorry to say..
08:53plamengr has quit IRC
08:53freemind has quit IRC
09:20vagrantc has joined #ltsp
09:21Egyptian[Home] has joined #ltsp
09:22
<g333k_laptop>
cliebow_, but where do I get the code from?
09:36
<cliebow_>
look in the wiki for lbe if you are doing 4.2..have you done a ltsp5 install?if so look in like usr/share/ltsp..something like that to see how the chrot is installed..can allso lok at .sbin/ltsp-build-client for the install code..5 uses native ubuntu binaries to build the chroot..
09:36
!wiki
09:36
<ltspbot>
cliebow_: "wiki" is the LTSP wiki. You can find it at http://wiki.ltsp.org
09:38
<cliebow_>
it is not triviala
09:38
8~)
09:40
<vagrantc>
ogra: wondering if you have any comments about my patches for syslog stuff ...
10:05
<g333k_laptop>
cliebow_, yeah I have installed ltsp5 in ubuntu, but I want to develop more for debian, I have downloaded a pkg from the debian repos called ltsp_0.99debian11.tar.gz, is it the source code? I feel something is missing
10:09
<vagrantc>
g333k_laptop: yeah, that's only the "upstream" parts.
10:09
g333k_laptop: there should be a line in /usr/share/ltsp-server/copyright explaining how to download the sources using bzr
10:09
er
10:10
/usr/share/doc/ltsp-server/copyright
10:11
<g333k_laptop>
vagrantc, I will find this file where ltsp5 is installed right?
10:11
<vagrantc>
g333k_laptop: if downloading tarballs from debian, you'll also want the .dsc and .diff.gz
10:11
g333k_laptop: yes.
10:12
if you have the ltsp-server package installed, that file will be present
10:12Egyptian[Home] has quit IRC
10:12
<vagrantc>
g333k_laptop: there's also the wiki page: http://wiki.debian.org/LTSPDevelopment though its not very well maintained
10:13
g333k_laptop: and: "bzr get http://bzr.debian.org/bzr/pkg-ltsp/main" should grab the sources too
10:13
presuming you have bzr installed
10:16
<g333k_laptop>
vagrantc, installing...
10:17
vagrantc, you are ltsp5's debian pkg maintainer right?
10:17
<vagrantc>
g333k_laptop: one of the crew
10:17
g333k_laptop: probably the most active :)
10:18
g333k_laptop: what sort of things are you hoping to work on?
10:19
<g333k_laptop>
vagrantc, I'm a user of ltsp 4.2, I know that things have changed since muekow... I'm very pleased with the project and I want to contribute with something, also I'm a debian user that's the reason because I want to contribute with the debian pkg
10:20
vagrantc, but first I'd have to read and understand the code to start developing something, maybe there is a TODO to follow
10:20
<vagrantc>
g333k_laptop: good. we need people more familiar with ltsp 4.x, as some of the developers have never used it :)
10:20
<g333k_laptop>
vagrantc, really?
10:21
vagrantc, so the start using ltsp 5?
10:21
they*
10:21
<vagrantc>
g333k_laptop: i came to ltsp after working on another project called lessdisks, which basically does the same thing as ltsp5, so i never used ltsp 4.x
10:22
<g333k_laptop>
vagrantc, ooh
10:23klausade has quit IRC
10:23klausade has joined #ltsp
10:24
<vagrantc>
i'd been doing lessdisks since 2001 ... and then ltsp5/muekow came along and i figured i'd join efforts :)
10:24
since there was really very little different
10:24
<g333k_laptop>
vagrantc, well I'm a noob on this kind of things, can you tell me what do I have to do? where to start and how? in order to start contributing to the project?
10:24
<vagrantc>
difference
10:25
g333k_laptop: read the wiki page, download the code, sign up for the mailing list, and ask questions :)
10:26
maybe look at the open bug reports, see if there's anything there you think you could fix
10:27
<g333k_laptop>
vagrantc, yeah this is a good start
10:27
vagrantc, can yo tell me the mailing list?
10:27
<vagrantc>
g333k_laptop: it's on the wiki page
10:28
though maybe it should be more at the top...
10:28
i should clean up all those crufy bzr branches
10:28plamengr has joined #ltsp
10:29plamengr has left #ltsp
10:29plamengr has joined #ltsp
10:34
<g333k_laptop>
vagrantc, ltsp-discuss mailing list right?
10:34
<vagrantc>
g333k_laptop: you looking at the wiki page i pointed you to?
10:35
<g333k_laptop>
http://wiki.ltsp.org/twiki/bin/view/Ltsp/Support
10:35
<vagrantc>
http://wiki.debian.org/LTSPDevelopment
10:36
should probably be a reference to that on wiki.ltsp.org ...
10:37
<g333k_laptop>
vagrantc, I didnt see any
10:38
<vagrantc>
much of the stuff on wiki.ltsp.org is for ltsp 4.x, not ltsp5 ... not sure how best to handle that long term ... as each distro will probably want to use whatever documentation infrastructure that it normally uses ...
10:38
g333k_laptop: search for "mailing list"
10:39
g333k_laptop: or, were you referring to a link to the debian.org page?
10:39
<g333k_laptop>
vagrantc, noup
10:40
<vagrantc>
g333k_laptop: noup, you didn't find a reference, or noup you didn't find a reference to the mailing list on that page?
10:41
<jammcq>
vagrantc: hey, currently, we're calling the ltsp-5 etch tarball 'ltsp_debian_etch_i386-1.tar.bz2'. Should we instead be using the name 'testing' ?
10:41* vagrantc HATES the use of stable/testing/unstable/
10:41
<g333k_laptop>
vagrantc, I was in the ltsp wiki page
10:41
<jammcq>
so, should I take that as "don't use 'testing'" ?
10:42
<vagrantc>
it makes so many things so difficult ... because once it switches, it's no longer stable/testing/unstable
10:42
<g333k_laptop>
jammcq, etch is still testing, but soon will be stable
10:42
<jammcq>
yes, I understand that
10:42
<vagrantc>
jammcq: yeah, don't use testing ... though maybe figure out some way to indicate that it's not the released etch
10:42
<jammcq>
our tarball is "etch pre-release"
10:42
<g333k_laptop>
jammcq, I think the best way to avoid mistakes is calling testing, stable or unstable
10:43
<jammcq>
so you two disagree then, eh?
10:43
<vagrantc>
g333k_laptop: but then, if you have this tarball sitting up there called "debian-testing" and etch is released, it's now "debian-once-was-testing-but-not-yet-stable"
10:43
<g333k_laptop>
vagrantc, the day etch become stable, the "etch" pkg made for testing debian release will not be suitable for the next debian testing anymore
10:44
vagrantc, correct
10:44
<vagrantc>
so i think it's bad to use testing.
10:44
<jammcq>
vagrantc: what's inside the /etc/apt/sources file?
10:44
<vagrantc>
jammcq: etch
10:44
<jammcq>
ok, then we have ETCh as the name
10:44
makes sense to me
10:44
<vagrantc>
so if you apt-get ugrade the chroot, you'll end up with whatever etch is at that time.
10:45
even if it may be an outdated tarball
10:45
<jammcq>
isn't it also possible to specify 'testing' inside the sources file? (i'm not suggesting we do this)
10:45
<vagrantc>
jammcq: how are you indicating the pre-release status?
10:45mrchicken has joined #ltsp
10:45
<jammcq>
vagrantc: i'm not (yet)
10:45
today is the day we figure all that crap out
10:45
<g333k_laptop>
jammcq, yes it is possible
10:45
<vagrantc>
jammcq: yes, you can specify testig in the sources file, but that's one of the things i really like to avoid doing ...
10:46
<mrchicken>
Hello =) Finnally I got ltsp working
10:46
however I have a problem with one client
10:46
<vagrantc>
jammcq: becuase i don't want an apt-get upgrade to essentially... switch releases...
10:46
<mrchicken>
I'm booting from rom-o-matic boot roms diskette based
10:46
and it loads everything
10:46
and then it says
10:47
eth0: Transmit timeout, status 00000000 00000259
10:47
<g333k_laptop>
vagrantc, what would happen if you have "testing" in your sources.list and the you do a apt-get dist-upgrade?
10:47
<mrchicken>
Error DHCPD failed
10:48
<jammcq>
mrchicken: what kind of NIC in the workstation?
10:48
<mrchicken>
Kernel panic - not syncing: attempted to kill init
10:48
SIS900
10:48
<vagrantc>
g333k_laptop: it would switch to testing ... but i typically don't want to follow testing, i want to follow a release, such as etch.
10:48
or a pre-release.
10:48
<jammcq>
isn't the "real name" of the release something like "3.1" ?
10:49
<vagrantc>
debian etch/4.0
10:49
but that's just the numeric name. both are "real" names :P
10:49
<g333k_laptop>
jammcq, I think no, the real name is sarge, etch, sid, woody, etc. they get the version number when they become stable
10:50
<jammcq>
ok
10:50
<vagrantc>
well, the version number is chosen on it's way to stable.
10:50
<mrchicken>
hello?
10:50
<jammcq>
see, with Ubuntu, I think once it's released they more-or-less forget about the text name, and refer to it by number
10:50
mrchicken: hello
10:50
<g333k_laptop>
jammcq, some of them will never become stable like sid, it will be always unstable
10:51
<vagrantc>
mrchicken: if it's only a single machine that's failing, i would say it's probably a bad NIC or bad network cable
10:52
<g333k_laptop>
jammcq, yeah Ubuntu has a different way to refer to they releases, I think they just work with stable versions and a beta testing right?
10:52
<mrchicken>
but t he NIC works fine under win98
10:52
<g333k_laptop>
jammcq, like edgy and feisty
10:53
<vagrantc>
that's one of the things i think ubuntu does right.
10:54
because with debian, you have some tools which use the release name (etch) some tools which use the alias (testing) and occasionally some that use the version number (4.0)
10:54
and it's insanity.
10:56
<g333k_laptop>
vagrantc, hahah yes
11:03
<jammcq>
vagrantc: one change we want to make to ltsp-update-kernels
11:03
<mrchicken>
The error says something about NETDEV watchdog
11:03
<jammcq>
mrchicken: it's possible that the driver in linux doesn't properly support that card
11:04
vagrantc: this script needs to run on any distro. so far, I think it will, but.... if the host has a /tftpboot directory, we want the kernels copied there too
11:04
<vagrantc>
jammcq: yeah, well, it's configurable.
11:04
jammcq: do any distros still keep the kernel/initrd in / ?
11:05
<jammcq>
dunno
11:05
i'm talking about installing the ltsp_debian_etch_i386 tarball on a non-debian machine
11:05* vagrantc wants to get rid of the copying altogether.
11:05
<jammcq>
good luck with that
11:05
<vagrantc>
well, i tested it, and it works with atftpd. no need to copy at all.
11:06
<jammcq>
vagrantc: it's "configurable", to put the kernels in a single place, based on TFTPBOOT
11:06
but i'm saying if /var/lib/tftp exists, drop the kernels there. AND, if /tftpboot exists, drop the kernels there (too)
11:06
<vagrantc>
eeyk.
11:07
well, that shouldn't be too hard, i guess.
11:07
<jammcq>
right.
11:07
<vagrantc>
if both exist, drop the kernels in both places?
11:07
<mrchicken>
jammcq ... I did do a small change in dhcpd.conf
11:07
<jammcq>
yep
11:07
<mrchicken>
!pastebot
11:07
<ltspbot>
mrchicken: "pastebot" is The LTSP pastebot is at http://pastebot.ltsp.org. Please paste all text longer than a line or two to the pastebot, as it helps to reduce traffic in the channel. A link to the content will be pasted in the channel.
11:07
<vagrantc>
mrchicken: your problem is hardware related. dhcp won't matter.
11:08
mrchicken: either broken hardware, or hardware that isn't properly supported (or some combination of the two)
11:09
jammcq: it sure adds another ugliness layer :(
11:09
<jammcq>
vagrantc: YEP :)
11:09FernandoMM has joined #ltsp
11:09
<jammcq>
worse ugliness would be to have the "debian" version of ltsp-update-kernels, and the "upstream" version of ltsp-build-kernels
11:10
<vagrantc>
well, i will think about how to make it work and be less ugly.
11:10
<ltsppbot>
"MrChicken" pasted "NETDEV Watchdog: eth0 transmit timed out" (74 lines) at http://pastebot.ltsp.org/39
11:10
<vagrantc>
mrchicken: try replacing the network card.
11:10
<mrchicken>
vagrantc ... already did with a rtl8139
11:11
and it wont even boot
11:11
it will just show some numbers on screen.
11:11
<vagrantc>
mrchicken: did you change your etherboot disk too?
11:11
<mrchicken>
vagrantc: yes
11:11
<vagrantc>
mrchicken: try putting the card in a different slot in the machine...
11:11
<jammcq>
mrchicken: why are you loading a pxelinux.0 bootloader with an Etherboot bootrom?
11:12
<mrchicken>
jammcq ... cuz the other way around it would give me another error
11:12
hold on let me find it for you
11:12* vagrantc likes that etherboot can work with pxelinux
11:12
<jammcq>
you basically have: if PXE, load pxelinux, otherwide load pxelinux
11:12
why bother having an 'IF' statement at all
11:12
<mrchicken>
I know
11:13
I changed that cuz I had IF pxe load pxelinux else load "/tftpboot/lts/2.6.17.8-ltsp-1"
11:13
<jammcq>
but.... your error is coming from the LINUX KERNEL, not from etherboot or pxe
11:13
so obviously, the kernel got loaded
11:13
and once the kernel is loaded, the bootrom is out of the way
11:14
<mrchicken>
So its my NIC?
11:14
<jammcq>
could be
11:14
can you try a different slot?
11:14
<mrchicken>
built in
11:15
but maybe I could change the IRQ
11:15
<jammcq>
with PCI ? not really
11:18
<mrchicken>
:(
11:18
then the nic is fucked
11:18
<jammcq>
well, can't win em all
11:19
<mrchicken>
Yeah... thanks buddy
11:19
Have a nice day
11:19mrchicken has quit IRC
11:19
<jammcq>
heh, did that sound sarcastic to you all?
11:20
<vagrantc>
yeah, well.
11:20
if they want, they could patch the kernel.
11:20
yay free software!
11:20aristidezzz has joined #ltsp
11:20* vagrantc shakes head thinking about ingrateful people
11:21
<jammcq>
he's already sucked enough of my support bandwidth, I didn't want to get into the 'patch the kernel' thing
11:21
<Core>
Hi all. I use xubuntu with its own ltsp packages. thin client boots kernel but then switches to graphics mode and thats it. the screen is black and the system doesnt respond to neither ctrl-alt-del nor ctrl-alt-bsps. what can that be ?
11:21
<jammcq>
hmm
11:21
screen completely black?
11:21
not gray
11:21
<Core>
yeap
11:21
<jammcq>
no cursor
11:21
<Core>
black
11:21
nothing
11:21
<jammcq>
what color is the LED on the monitor?
11:22
<Core>
i use vmware
11:22
<jammcq>
i'm guessing amber
11:22
oh
11:22
why?
11:22
<Core>
i see it changes resolution (screen dimensions)
11:22
<vagrantc>
maybe try forcing a lower resolution with X_MODE_0 in lts.conf
11:23* vagrantc backs out of supporting vmware
11:23
<jammcq>
vagrantc: too late, yer in it now
11:23* vagrantc knows how to type /quit
11:23
<Core>
I played with xorg.conf and let it use lower res, so it acctualy uses what i did in xorg. conf )))
11:24
<jammcq>
Core: what did you specifiy in xorg.conf for the video driver?
11:24
<Core>
sorry for asking such dumb quastions and messing with vmware )
11:25
it is vmware svga video driver
11:26
the "server" (xubuntu) works great with its Xfce and stuff....
11:26
<vagrantc>
it's not a dumb question, i'm just pretty clueless when it comes to vmware
11:27* vagrantc struggles to make supporting both /tftpboot and /var/lib/tftpboot not suck
11:27
<jammcq>
vagrantc: are you working on that right now?
11:27
cuz i'm hacking the script too, but if you wanna do it, have at it
11:28
<vagrantc>
jammcq: i've mostly got it. it just might break the "no copying" support.
11:28
<jammcq>
if you aren't doing copying, how are you doing it?
11:28
symlinks ?
11:28
<vagrantc>
just point your tftp server to /opt/ltsp and configure dhcpd
11:28
or symlinks
11:28
and a tftp server that supports symlinks
11:29
<jammcq>
we're going into installations where they've already got /tftpboot, and ltsp-4.2
11:29
changing where tftpd looks for the kernels really ins't an option
11:29
<vagrantc>
symlinks is probably simpler...
11:29
<jammcq>
but copying should always work (ugly as it is)
11:29
<vagrantc>
except i could never get tftpd-hpa to work with symlinks, but atftpd does
11:30
well, i like the idea that you shouldn't have to run ltsp-update-kernels whenever you update the kernels in a chroot ... it's nicer to just have it automatically work
11:30
<jammcq>
sure, it is a great goal
11:30
<Core>
btw guys, a little quastion not conserninng ltsp. I did also kindof thin client. used debian 3.1r5. it works great in console but when you do xstart, only like 1/5 times it works ok and all other 4/5 when desktop boots it freezes completely.any hints what could that be ?
11:30
<vagrantc>
otherwise, you risk having a different kernel in the chroot than the tftp dir
11:30
<jammcq>
but sometimes we have to live in the real world
11:31
<vagrantc>
well, that's why i made it work without copying, while leaving the default to copy files
11:31
<jammcq>
k
11:31
Core: sorry, no clue
11:32
<vagrantc>
jammcq: but this new requirement definitely is making it hard to support properly :(
11:32
<jammcq>
vagrantc: it's just a matter of programming :)
11:32
<vagrantc>
jammcq: details.
11:33
i guess i could make it stupid and do ugly things if TFTPDIR = BASE
11:33
it's not even that ugly.
11:33
<jammcq>
this is what I have so far:
11:34
TFTPDIRS=${TFTPDIR:-"/var/lib/tftpboot /tftpboot"}
11:34
for TFTPDIR in ${TFTPDIRS}; do
11:34
<vagrantc>
jammcq: we think alike :)
11:34
<jammcq>
ah, but you are trying to ALSO have no-copy as an option
11:34
I wasn't thinking that gar
11:34
<vagrantc>
yeah.
11:34
<jammcq>
actually, I stole those two lines from scotty, but I wouldda done it the same way
11:34
<vagrantc>
makes it harder ... but i think with a little work, doable.
11:35
yeah, it's a pretty basic idea
11:35
jammcq: so ... you ever gonna publish a bzr branch? :P
11:35
<jammcq>
my goal, is for ANYBODY to be able to grab a tarball, and if they have debian or ubuntu, they already have ltsp-update-kernels, so they just run it. but for those who don't use deb/ub, they can grab a copy of the scripts from us
11:36
<vagrantc>
i don't even remember how the no-copy stuff got set up... so maybe we kill it for now and add it back again as a TODO
11:36
<jammcq>
vagrantc: we had hoped to do it last weekend, but couldn't get our minds wrapped around the idea of 'master copy' of the upstream code
11:36
vagrantc: works for me
11:37
<vagrantc>
jammcq: basically, you designate one branch as "upstream" ... and either you have a single person, or group of people, decide what goes into it.
11:37
<jammcq>
ah
11:37
that's what we need
11:38
scotty and i don't understand how to get that started
11:38
and our timing wasn't good last weekend, ogra wasn't around, and I think you were pretty busy too
11:38
maybe, if scotty is around today, we can do something
11:39
<vagrantc>
i would either base it on ogra's mainline branch or debian's pkg-ltsp main branch.
11:39
<jammcq>
where would that upstream branch live?
11:39
<vagrantc>
ltsp.org/bzr ?
11:39
<jammcq>
served from our web server?
11:39
<vagrantc>
or if you prefer, launchpad...
11:39
yeah... generally good to serve it from a web server.
11:40
and you can sftp push to it
11:40
<jammcq>
it becomes a political issue, more than a technical issue
11:40
<vagrantc>
yes.
11:40
politically, i think it would be best to have ltsp.org upstream on ltsp.org ...
11:40
<jammcq>
prolly
11:41
<vagrantc>
it also means you have full control over it, which can be nice. and launchpad does appear to do a decent job of mirroring branches, too.
11:42
so you would probably still be able to use whatever launchpad features you like
11:42
<jammcq>
well, right now, i'm not using any launchpad features :)
11:44
btw, similar changes will need to be made to ltsp-update-sshkeys
11:45
although, perhaps it already works
11:46
<vagrantc>
i think i got ltsp-update-kernels working without too bad
11:46
<jammcq>
i'm happy to test it
11:46
<vagrantc>
ltsp-update-sshkeys is much simpler
11:46
<jammcq>
ltsp-update-sshkeys already works
11:46
although it would be nice if it displayed something about what it was doing
11:47
we could just open a bug on that
11:48
can I have a copy of your ltsp-update-kernels ?
11:48* vagrantc is pasting it now
11:48
<jammcq>
you need our command-line paster
11:49
<vagrantc>
yes!
11:49
<ltsppbot>
"jam" pasted "paste2ltsp" (251 lines) at http://pastebot.ltsp.org/40
11:49
<jammcq>
heh, it pasted itself :)
11:50
<ltsppbot>
"vagrantc" pasted "shiny new ltsp-update-kernels with multiple tftp dir support" (64 lines) at http://pastebot.ltsp.org/41
11:52
<vagrantc>
unfortunately, that probably has to be post-etch for debian
11:52
<jammcq>
no prob
11:52
for people using debian, they probably are only using /var/lib/tftp
11:53
<vagrantc>
/var/lib/tftpboot
11:53
<jammcq>
yeah, that too
11:53
<vagrantc>
oh... i wonder how the code handles if the directory doesn't exist ...
11:53
probably should start turning some of that code into a function
11:54
otherwise we'll go to indentation hell
11:54
<jammcq>
which dir ?
11:54
if [ ! -d $TFTPDIR ] ; then
11:54
mkdir -p $TFTPDIR
11:54
fi
11:54
oooh
11:54
<vagrantc>
ah... hrm.
11:54
<jammcq>
that's not good
11:54
<vagrantc>
yes.
11:55
may as well hard code to always make all dirs :)
11:55
when it was a single dir, that wasn't such a bad thing.
11:55
<jammcq>
who's job is it to create /var/lib/tftpboot ?
11:56
<vagrantc>
heh.
11:56
seems like its in ltsp-server
11:56
but not either of the tftp servers
11:56
<jammcq>
tftpd-hpa, ltsp-server: /var/lib/tftpboot
11:56
<vagrantc>
ah.
11:56
<jammcq>
what does that mean?
11:56
<vagrantc>
i've never installed tftpd-hpa probably.
11:57
<jammcq>
that's a feisty box
11:57
<vagrantc>
they both include it as part of their package
11:57
<jammcq>
is that possible?
11:57
<vagrantc>
with directories, yes.
11:57* jammcq is surprised it's not against debian policy
11:57
<vagrantc>
so, we probably don't really need the mkdir call
11:58
and, i wish i had some powerpc users to test powerpc stuff ... ogra seems to think the yaboot hack in there is no longer needed.
11:58
<jammcq>
right, instead it should be: [ ! -d ${TFTPDIR} ] && continue
11:59
but that test should go immediately after the 'for TFTPDIR in $TFTPDIRS; do'
11:59
<vagrantc>
ooooh. continue? does that do what i think it does? :)
11:59* vagrantc learns something new
12:01
<jammcq>
yep, assuming you are thinking right
12:01
<vagrantc>
hmm... works in dash but not posh
12:01
<jammcq>
posh ?
12:02
wth is posh
12:02
<vagrantc>
policy compliant shell ... aims to be a true POSIX sh
12:02
<jammcq>
sounds like it's 'mostly complient shell'
12:02* vagrantc is doing a little more thorough checking
12:03
<jammcq>
root@biglap:~# help continue
12:03
continue: continue [n] Resume the next iteration of the enclosing FOR, WHILE or UNTIL loop. If N is specified, resume at the N-th enclosing loop.
12:03
<vagrantc>
you running bash?
12:03
<jammcq>
yep
12:03
<vagrantc>
bash is most definitely not POSIX compatible
12:03
<jammcq>
POSIX sounds so 1990's
12:04
<vagrantc>
upon further testing, it works with posh too, so i'm very, very happy to have learned about it
12:04
this makes so many things sooooo much easier to implement simply.
12:05
<jammcq>
well, the alternative is more nested IF statements
12:05
and that's no fun
12:05
<vagrantc>
indeed.
12:05
i feel new doors opened to me.
12:05
<jammcq>
it's like spring time all over :)
12:07
<vagrantc>
oooh. this means that atftpd will work with the defaults, too!
12:07
i think atftpd uses /tftpboot by default
12:08
well, it's not looking so ugly, and it seems to bring great fruit, so i'm glad this issue came up :)
12:08
<jammcq>
yeah, I usually use atftpd, so that's why I'm wanting to add support for /tftpboot
12:08
heh
12:08
<vagrantc>
all for only 4 additional lines of code.
12:08
<jammcq>
my oldest daughter is home from college, and my wife just informed me that the whole family is going out to lunch. So.... I'll need to leave in a few minutes
12:08
hey, cool, 4 lines
12:09
<vagrantc>
enjoy!
12:09
<jammcq>
thanks
12:09
do you have an updated script I can test?
12:09
before I leave
12:09
<vagrantc>
sure.
12:10
http://llama.freegeek.org/~vagrant/ltsp-update-kernels.20070310
12:11
jammcq: oh... but.
12:11
jammcq: since ltsp-server creates /var/lib/tftpboot ... it will always copy to that directory by default.
12:11
so if you have both, it will always copy both
12:11
<jammcq>
yeah, that's fine with me
12:12
<vagrantc>
unless you set up a configuration file to change TFTPDIRS
12:12
<jammcq>
for people installing from the tarball, they won't have /var/lib/tftpboot by default
12:12
unless they've installed tftpd-hpa
12:12
<vagrantc>
in which case, they probably want it
12:12
<jammcq>
yeah, so no harm
12:13
so... why does ltsp-build-kernels care if it can chroot into the $CHROOT ?
12:14
ie, why is it trying to run /usr/lib/ltsp/update-kernels?
12:14
<vagrantc>
jammcq: so it can update the initrd images if necessary.
12:14
<jammcq>
hmm
12:14
we'll need a way to handle that on cross-arch setups
12:15
<vagrantc>
well, on cross-arch setups, you'll need to boot the appropriate arch to update it.... so no updates should happen that are needed.
12:15
<jammcq>
maybe it's just a bunch of steps in documentation
12:15
<vagrantc>
except for the corner case where you've tweaked /etc/initramfs-tools/*
12:16
<jammcq>
so, the tarball would already have an initramfs that matches the kernel, so we'd be fine there
12:16
<vagrantc>
it could probably be pulled out, actually.
12:16
jammcq: yes
12:16
<jammcq>
ok, lets roll with it, and see how things work out
12:16
<vagrantc>
i think it's actually legacy code from when we generated the initramfs server-side
12:16
but now all that's handled whenever you install a new kernel anyways, so it's probably not needed anymore.
12:17
except for the above-mentioned corner case
12:17
<jammcq>
if you chroot into /opt/ltsp/i386, and do: apt-get install kernel-image (or whatever the kernel pkg is called), does it automatically generate the new initramfs for you?
12:17
<vagrantc>
yup
12:17
<jammcq>
ah, I see your above answer
12:17
yeah, that should prolly come out
12:17
<vagrantc>
in those probably rare corner cases, we just don't need to worry about it, i suspect.
12:18
<jammcq>
i'm gonna head out. if you have any updates to the script, /msg me, or email it to jam@ltsp.org, or figure out some way to get the script to me. when I get back in a couple hours, I'll post whatever is the latest to the ltsp download page, and work on some docs
12:18
<vagrantc>
or when you upgrade the kernel in the chroot. that was probably my biggest contribution to the ltsp kernel handling
12:18
<jammcq>
thanks, one small step closer to greatness :)
12:40
<vagrantc>
you can emulate amd64 on i386 with qemu ... maybe i can explore some of this cross-architecture stuff finally.
12:40plamengr has quit IRC
12:42edgarin has joined #ltsp
12:43
<edgarin>
Hi people!
13:13CapriCoRN^80 has joined #ltsp
13:33
<edgarin>
hi jammcq
13:33* vagrantc suspects jammcq is out to lunch with the family
13:39bobby_C has joined #ltsp
13:44CapriCoRN^80 has quit IRC
13:52aristidezzz has quit IRC
14:00Core has quit IRC
14:01MaCaDe has joined #ltsp
14:10* jammcq is back
14:10* vagrantc realizes that vagrantc was to give an LTSP talk on .... march 14th
14:11
<jammcq>
oh?
14:11
"was" ?
14:11
<vagrantc>
is
14:11
<jammcq>
sooper, where?
14:11
<vagrantc>
albuquerque
14:11
at the local users group
14:11
<jammcq>
cool
14:11
unfortunately, I'll miss it
14:11
<vagrantc>
i'd best... ummm... prepare or something
14:11* jammcq loves to hear other people talk about ltsp
14:11
<jammcq>
so what's up with ltsp-update-kernels, is it ready to go?
14:12* vagrantc would not recommend watching the ltsp talk from debconf6
14:12
<jammcq>
who gave that talk?
14:12
<vagrantc>
otavio and i
14:12
<jammcq>
hmm
14:12
<vagrantc>
we both neglected to prepare or even do silly things like sleep
14:12
<jammcq>
I've grown bored with giving ltsp talks. it used to be fun, when it was more difficult to set up
14:13
<vagrantc>
right now, i feel the need to talk about the ltsp integration stuff with ltsp 5
14:13
<jammcq>
yeah, but that's a different audience
14:13
well, different from the groups that i've talked to
14:13
<vagrantc>
but it's a somewhat technical topic that most people would just go "huh?"
14:14
this user group has everything from people working on clustered supercomputing to people who just discovered openoffice and firefox
14:14
a little tricky to know where to start, exactly...
14:14
<jammcq>
so, were you thinking of hacking out the CHROOT stuff from ltsp-update-kernels ?
14:14
or leave it in?
14:14
<vagrantc>
jammcq: i haven't really worked on it since you left
14:15
jammcq: but yeah, it's probably ok to removing the CHROOT code ... welllllll
14:15
jammcq: ok on debian/ubuntu ... other distros might still need it...
14:15
or want it
14:15
<jammcq>
we'll make it required that they prepare their initramfs during install of the kernel
14:16
it's the only way we can do cross-arch
14:16
<vagrantc>
and upgrades
14:16
<jammcq>
and upgrades, sure
14:16
<vagrantc>
true enough.
14:16
<jammcq>
i'd like to not treat cross-arch differently
14:16
<vagrantc>
it's good to see a few lines of code removed :)
14:17
<jammcq>
i'm all for that
14:17
simple is good
14:18
<vagrantc>
and we can just document what you need to do for those corner cases, as it's still pretty simple.
14:18
<jammcq>
yep
14:19* vagrantc just talked to ltsppbot instead of ltspbot
14:21
<vagrantc>
that's sneaky
14:21
<jammcq>
hmm, that was probably an oversight
14:22
we prolly should come up with a different name for him
14:22
<vagrantc>
ltsppbot is apparently away saving humanity
14:22
or some such
14:22
is it actually a separate bot?
14:22
<jammcq>
no, ltsppbot is the name we gave to the pastebot
14:22
<vagrantc>
perhaps ltsppastebot
14:22
hrm
14:22
<jammcq>
or pastebot_ltsp
14:22
<vagrantc>
yeah
14:23
that would be tab-completion safe
14:23
<jammcq>
so that we don't have a tab-completion issue
14:23
:)
14:23
somehow, you and I are on the same wavelength today
14:24
<vagrantc>
142.47 MHz ?
14:24
<jammcq>
eh, .47, .48.... whatever
14:24
<vagrantc>
well, that would explain the variations
14:28
jammcq: you changed "gory details" to "glory details" on the integratingltsp wiki page ... was that intentional?
14:28
<jammcq>
I did that?
14:28
<mistik1>
Do any of you guys know how to write a setup.py for using to install a python module?
14:28
<vagrantc>
i think so...
14:30
<jammcq>
looks like RichardBos changed that
14:30
<vagrantc>
oh.
14:30
<jammcq>
dunno why, but we should change it back
14:30* vagrantc cannot really understand how to diff in twiki
14:31* vagrantc struggles with twiki in general
14:36
<vagrantc>
jammcq: /tftpboot/lts .... not /tftpboot/ltsp ?
14:39
<jammcq>
umm
14:39
I think that's right
14:39
one of the mistakes I made 8 yrs ago
14:40
<vagrantc>
hrm. we seem to have, uh, broken compatibiltiy with that one.
14:40
would it be so evil to also have /tftpboot/ltsp ?
14:40
<jammcq>
well, since we're talking about ltsp-5, and not ltsp-4.2, maybe it's not so bad to finally fix this
14:41* vagrantc would love to do that
14:41
<jammcq>
ok, consider it a bug then
14:41
an 8yr old bug
14:44edgarin has quit IRC
14:50
<jammcq>
vagrantc: so, are you hacking the CHROOT code out? or am I?
14:52
<vagrantc>
it's just yanking 4 lines, no?
14:52
hroot
14:53
i guess more like 9-10 including comments
14:53
<jammcq>
well, until we actually setup upstream, I just wanna make sure the right person takes ownership of it
14:53
I can change it and send you the copy
14:53
or you can change it and send it to me
14:53
<vagrantc>
i can change it.
14:54
<jammcq>
that's the answer I was looking for:)
14:54
<vagrantc>
i'll make a branch for this stuff
14:54
actually, i can probably just put it in my old ltsp-update-kernels branch
14:56
whoah. i forgot i had uncommitted patches to support a couple more architectures sitting there in that branch
14:57
i never had the hardware to test them, but they should work in theory.
15:00
oh, this branch is way too outdated. best make a new branch.
15:02
<mistik1>
grr
15:03
I am so sick of python developers that think that thier so holy online doc should give everyone all they need to know about python and they should have to offer no more support than that
15:03
<jammcq>
hey mistik1, how ya doin?
15:03* mistik1 pissed off
15:04* mistik1 makes note to NEVER write python EVER
15:04
<jammcq>
perl rocks
15:05
<mistik1>
yea, The module I was after I found a perlmod for anyway, Not as savvy but it works
15:05* vagrantc has apparently only encountered freindly python coders
15:05
<mistik1>
python devs suck
15:06
<vagrantc>
sure beats obfuscation contests :P
15:06
<mistik1>
any question you ask, all they can do is point to thier stupid manual which is written for elves
15:07
If emerge was not written in it i'd purge it from my system
15:08
*sigh*
15:09* mistik1 goes to sit in a corner
15:09
<jammcq>
vagrantc: shouldn't the ltsp-update-kernels and ltsp-update-sshkeys scripts contain a copyright notice at the top?
15:13
<vagrantc>
jammcq: generally a good thing, yes.
15:14
i think they probably fall under Canonical's copyright from the original code mdz uploaded.
15:15
<jammcq>
prolly
15:15
was that GPL ?
15:15
<vagrantc>
it'd be in the debian/copyright
15:15
<jammcq>
k
15:15
<vagrantc>
in the catch-all ...
15:15* Lumiere is a python user...
15:16
<jammcq>
Lumiere: go talk to mistik1, he's over there in the corner, sulking
15:16
<vagrantc>
there's some disagreement amoung folks about exactly what all is needed. i tend to like to put a short 2-liner in everything.
15:16
<jammcq>
vagrantc: yeah, i'd vote for that too
15:17
<vagrantc>
it just prevents any question in the future ... of course, copyright is insane when it comes to multiple contributors ...
15:17
<jammcq>
well, copyright holder is one thing, license is another
15:17
<vagrantc>
true enough.
15:17
<jammcq>
but, it should be obvious what the license is, and who holds the copyright
15:18
<vagrantc>
as i understand it, technically, even a small block of unique code is copyrighted and should probably be liscensed separately . but that's crazy.
15:19
so multiple contributors makes it kind of wierd ...
15:27
<Lumiere>
jammcq: I can't
15:27
he's being too emo
15:45Egyptian[Home] has joined #ltsp
15:48
<mistik1>
Lumiere: lol
15:49
I figured out my problem anyway but I still hate the way some devs treat people
15:49
<vagrantc>
that's not limited to python, i assure you.
15:50
<mistik1>
I agree
15:51
what are you guys working on there?
15:51
<vagrantc>
who working on where?
15:52
<mistik1>
I see you and Jim going at it but been busy to read all
15:53
<vagrantc>
i think jammcq is working on the "download a pre-built tarball of debian/ubuntu ltsp5.0" stuff
15:53
<mistik1>
ahh
15:53
<vagrantc>
and it requires a few changes to some of the scripts in debian/ubuntu to prevent forkage
15:53
<Lumiere>
mistik1: cando.schooltool.org
15:53
and most language devs
15:53
are rtfm jackasses
15:53
<vagrantc>
mistik1: hopefully it will nudge other distros to integrate LTSP :)
15:54
<mistik1>
I get the hint ;)
15:54
<vagrantc>
well, at least there's been some activity from y'all ... it's more some of the others we haven't heard from.
15:54
<mistik1>
I'm working on it, actually most of what i'm working on right now will just transfer neatly to ltsp Gentoo
15:54
<vagrantc>
especially the fat cats.
15:54
nice :)
15:55
<mistik1>
But I got to finish my little OS first
15:55
<vagrantc>
little OS ?
15:55
<mistik1>
yea, building yet another firewall OS
15:56
Gentoo based of course, this is why I said the experience will transfer to LTSP5
15:58
<vagrantc>
heh. :_
15:58
er...
15:58
:)
15:59* jammcq wonders if vagrantc has the de-chrooted ltsp-update-kernels script ready
15:59
<vagrantc>
jammcq: i do, i do!
16:00
<jammcq>
where might I get it?
16:00
<vagrantc>
i've been trying to figure out what branch i should put it into... if only i had an upstream to branch off of, everything would be so clear :P
16:01
<jammcq>
agreed
16:01
if only
16:02
<vagrantc>
i guess i'll just stick it in my branch off of ogra's mainline
16:02
<jammcq>
in the meantime, can I grab a copy?
16:02
<Lumiere>
anyways... I'm helping project manage a school information system
16:02
and competency tracking system
16:03
written in zope3/python
16:03
<Egyptian[Home]>
evening all
16:03
<jammcq>
hey Egyptian[Home]
16:04
<Egyptian[Home]>
jammcq: i have this app which connects to a database ..the database checks the clients ip and .. basically it works like a firewall disallowing clients ip + login if the two dont match.. u got a solution for this?
16:04
<vagrantc>
jammcq: http://llama.freegeek.org/~vagrant/ltsp-update-kernels.20070310
16:05
<jammcq>
Egyptian[Home]: hmm, nope
16:06
<Egyptian[Home]>
not even in ltsp5? i see the ltspinfo app can ask the client questions
16:07
<vagrantc>
Egyptian[Home]: you just want to match weather the ip and login match up?
16:07
<jammcq>
vagrantc: ok, the ltsp-update-kernels scripts displays tons of info to the screen, but the ltsp-update-sshkeys script displays nothing to the screen. which way is better?
16:07
<vagrantc>
Egyptian[Home]: you could do it in Xsession
16:07
<jammcq>
my philosophy is succeed quietly, fail loudly
16:08
<Egyptian[Home]>
vagrantc: could u explain a bit more pls?
16:08
<vagrantc>
Egyptian[Home]: you just want to match weather the ip and login match up?
16:08
jammcq: you and your philosophies.
16:09
jammcq: yeah, we could shut ltsp-update-kernels up a bit.
16:09
<jammcq>
yeah, it spits out so much, it's hard to tell if there were any errors
16:09
<Egyptian[Home]>
vagrantc: yes
16:09
<jammcq>
the cp's prolly don't need the -v
16:09
<Egyptian[Home]>
vagrantc: umm.. apps username and password and ip
16:09
<vagrantc>
jammcq: although, a lot of what it's spitting out could be seen as errors. but that's just more the script is kind of stupid.
16:10
Egyptian[Home]: "apps username" ?
16:10
<jammcq>
but I like the 'Updating tftp directors for chroot: /opt/ltsp/ubuntu_feisty_i386'
16:10
<Egyptian[Home]>
vagrantc: app has built in login process ..does not use pam
16:11
<Lumiere>
how about a debugging variable
16:11
to control how loud the program is
16:11
<vagrantc>
Lumiere: yeah, that's what i was thinking
16:11
<Lumiere>
if it is in python
16:11
or almost anything
16:11
<jammcq>
it's not
16:11
<vagrantc>
we started to implement it with ltsp-build-client
16:11
<Lumiere>
heh
16:11
bash/sh?
16:11
<vagrantc>
good ol sh
16:11
actually, maybe bash
16:12
though if you really need bash features, you may as well write it in python.
16:12
<Lumiere>
anyways
16:12
<jammcq>
or perl
16:12
<Egyptian[Home]>
vagrantc: sorry i meant client .. its past midnite here :(
16:12
<Lumiere>
my favorite was always using a function to call debugging
16:12
pass it the string and what level the debug statement was at
16:13
<vagrantc>
Egyptian[Home]: i don't really understand your setup enough to help.
16:13
<Lumiere>
and the function dealt with whatever prettyness you wanted (I was also using ansi colors at the time)
16:13
<vagrantc>
Egyptian[Home]: it would be fairly simple to hook into /etc/X11/Xsession.d that checks if the username is authorized to login from a certain IP and die if not...
16:14
<Egyptian[Home]>
client which is to be isntalled on ltsp client .. connects to a server which does authentication via username and ip of the client .. which in ltsp clients case would be the ltsp server
16:14
<vagrantc>
Egyptian[Home]: oh... uhhh... no idea.
16:14
<Egyptian[Home]>
the username is of course not pam related not even linux related. its within the db
16:14
vagrantc: k
16:14
vagrantc: thanks for trying to think of a solution :)
16:14
nite
16:15
sleep tight all
16:15Egyptian[Home] has left #ltsp
16:17Egyptian[Home1 has joined #ltsp
16:20MaCaDe is now known as McDutchie
16:28
<cliebow_>
so that is where /tftpboot/lts came from..
16:28
<jammcq>
heh
16:34
vagrantc: bug in new script :(
16:35
vagrantc: you took out this line:
16:35
> export CHROOT_NAME="$(basename $CHROOT)"
16:35
>but that line is needed further down, on this line:
16:35
< cp -a $CHROOT/boot/. $TFTPBOOT/$CHROOT_NAME/
16:35
-
16:36
so, the kernels end up being copied to $TFTPBOOT//
16:36
which isn't quite what we want
16:37
<cliebow_>
8~)
16:37
<jammcq>
i've fixed it, and reduced the amount of output displayed
16:41
<vagrantc>
jammcq: ok, cool.
16:42
jammcq: so, i should have left the "export CHROOT_NAME" in?
16:42
<jammcq>
yeah
16:43
and there's also a couple of problems where we end up with pathnames with double '//'
16:43
which doesn't hurt, but looks goofy
16:43
<vagrantc>
sure, wel.
16:46
i'd really like to kill that yaboot code.
16:46
<jammcq>
don't we need it for ppc ?
16:47
<vagrantc>
well, not in the server-side scripts, just client-side
16:47
the client-side stuff is hack we supposedly needed at some point but maybe no longer do, according to ogra
16:47
er ...
16:47
server-side
16:47
but i can't really confirm it.
16:48
<jammcq>
http://www.ltsp.org/ltsp-update-kernels
16:48
that's my fixes
16:49
<mistik1>
I suppose the server/client-side kernel update scripts are both needed because of where Ubuntu puts the ltsp kernel?
16:49
<vagrantc>
jammcq: do you take bribes?
16:49
<mistik1>
The whole /var/tftpboot deal
16:49
<jammcq>
sure
16:49
<vagrantc>
mistik1: no, it was a bug in yaboot that it needed to be a particular place
16:50
mistik1: or, you mean....
16:50
mistik1: the /var/lib/tftpboot stuff is for policy compliance to avoid having an ever-propegating number of directories in /
16:50
<jammcq>
i'd like cliebow to chime in on the yaboot stuff, he's got a good handle on it
16:51
http://www.ltsp.org/ltsp-update-sshkeys
16:51
<vagrantc>
mistik1: if you configure your tftp server to look in /opt/ltsp for files and configure DHCP properly, it's not really needed at all.
16:51
<jammcq>
that has my small mods, to display a little about what it's doing
16:51
<mistik1>
vagrantc: That is what I was trying to get to
16:51
<vagrantc>
jammcq: so, what sort of bribe would it take to get you to use bzr? :P
16:51
<jammcq>
mistik1: the problem is, we can't assume that LTSP is the only user of tftp
16:51
<mistik1>
I think we would default to the /opt/ltsp/? route
16:52
<jammcq>
vagrantc: I'm definately gonna use bzr, just not today
16:52
<mistik1>
jammcq: I suppose so
16:52
However we are talking about an ltsp server
16:52
<jammcq>
with laser printers and firewalls and other devices wanting to get updates via tftp, we can't hijack it as our own
16:52
in many places, the ltsp server is the ONLY server
16:52
<vagrantc>
mistik1: i host network-install images for debian-installer on the same server as my ltsp setups
16:53
<mistik1>
I see your points as valid
16:53
<vagrantc>
and debian-live network boot... and numerous other things.
16:53
but if you *can* avoid it, it's way better to not have to copy files around.
16:53
<jammcq>
and tftpd's only allow you to specify one directory to serve from
16:54
and sometimes, the tftp server is NOT the ltsp server
16:54
<vagrantc>
the other idea, which has it's own warts, is to use bind mounts from /opt/ltsp/*/boot to /var/lib/tftpboot/*
16:54
<mistik1>
It can be still avoided with a mount --bind /var/tftpboot/whatever into the /opt .....
16:54
yea that
16:54
<vagrantc>
jammcq: which we don't really handle either
16:54
though tftp servers may restrict to a single filesystem ...
16:55
atftpd with symlinks is my favorite approach
16:56
<mistik1>
vagrantc: DO you know if the symlink method works with the other tftpd's?
16:57
<vagrantc>
mistik1: i couldn't figure out how to disable tftpd-hpa's paranoia about symlinks
16:57
those are the only two i've tried
16:57
<mistik1>
fair enough
16:59
<vagrantc>
jammcq: hmmm.. there's probably more code we can move out of the for loop
17:01
getting there ...
17:02cliebow has joined #ltsp
17:03
<mistik1>
Speak and he shall appear ;)
17:04
<vagrantc>
cliebow: jammcq says you understand network booting powerpc machines
17:04
<cliebow>
kind of
17:04
what you got?
17:05
<vagrantc>
well, i have some code thats hackish and i want to get rid of it :)
17:05
<cliebow>
cme to the right place 8~)
17:05
dhcpd.conf stuff/
17:06
<vagrantc>
more ... yaboot
17:06
<cliebow>
heh..one laptop in the living room..one in the kitchen..one to take in the tub
17:06
yaboot.conf is the only user srviceable piec i know of
17:07
<vagrantc>
heh.
17:07
<cliebow>
yaboot either is..or isnt..i have seen minor size differences between versions
17:08
<vagrantc>
does yaboot need to be in the root dir of the tftp server ... i.e. /var/lib/tftpboot/yaboot ... or could it be in a sub-dir, say /var/lib/tftpboot/ltsp/powerpc/yaboot
17:08sahil has joined #ltsp
17:09
<cliebow>
i think it can be anywhere you like...i have put it in /..but something deep down tells me it wants to be in the root of the tftp chroot
17:09
ip ut in / ehrn i had no chroot
17:09
when i had no chroot
17:10
have to look it up on the wiki
17:12
i have it near set up on these two lappies..one pc and one mac..ill see if i can try it in a bit
17:16nubuntu has joined #ltsp
17:18
<jammcq>
vagrantc: hmm, just found an interesting problem. ubuntu-edgy's ltsp doesn't have the pxelinux stuff in it's /boot directory
17:19
<vagrantc>
jammcq: it probably didn't get merged till post-edgy.
17:19
<jammcq>
which means ltsp-update-kernels doesn't copy the pxelinux stuff to the tftpboot directory
17:19
<cliebow>
vagrantc, gonna try it,,brb
17:19cliebow has quit IRC
17:19
<jammcq>
I can copy the stuff to the chroot, and make a new tarball
17:20
I guess i'll have to
17:21loca|host has joined #ltsp
17:21
<loca|host>
i have a problem installing ltsp
17:21
<jammcq>
ok
17:21
<loca|host>
i've did what is written on the docs
17:21
<vagrantc>
jammcq: feisty should be out the door any day now, no?
17:22
<jammcq>
vagrantc: late april
17:22
<loca|host>
the client gets a gray screen with the mouse cursor, nothing else
17:22
<jammcq>
loca|host: ah, easy to fix
17:22
<vagrantc>
!greyscreen
17:22
<ltspbot>
vagrantc: "greyscreen" is (#1) No login, and only a grey screen? See http://wiki.ltsp.org/twiki/bin/view/Ltsp/Troubleshooting-GrayScreen#GDM, or (#2) No login, and only a grey screen? See http://wiki.ltsp.org/twiki/bin/view/Ltsp/Troubleshooting-GrayScreen
17:22
<jammcq>
loca|host: see that wiki page, it'll guide you to the happy land
17:22
<loca|host>
thats great
17:23
<jammcq>
vagrantc: what's up with ltspbot having 2 entries for that?
17:23
<vagrantc>
jammcq: looks like someone wanted GDM to be more prevalent.
17:23* mistik1 tries something
17:23
<loca|host>
can any one tell what's the difference between using ltsp-utils to get ltsp server running and apt-getting ltsp-server-standalone or ltsp-server ?
17:23
<mistik1>
!pastebot jammcq
17:23
<ltspbot>
mistik1: Error: "pastebot" is not a valid command.
17:24
<vagrantc>
loca|host: very, very different.
17:24
<jammcq>
loca|host: ltsp-utils is for LTSP-4.2. ltsp-server-standalone is LTSP-5
17:24
<mistik1>
!greyscreen jammcq
17:24
<ltspbot>
mistik1: Error: "greyscreen" is not a valid command.
17:24
<mistik1>
lame bot
17:24
<vagrantc>
mistik1: yeah, i was wishing for a syntax of what i think you're trying to do
17:24
<loca|host>
ok
17:25
<mistik1>
vagrantc: ;)
17:25
<loca|host>
i thought that standalone means something else
17:25
<vagrantc>
loca|host: ltsp 5 builds the environment out of debian/ubuntu/otherdistro packages, whereas LTSP 4.x is essentially it's own customized linux distribution
17:25
<jammcq>
ltsp-server-standalone is a metapackage that depends on ltsp-server
17:25
<vagrantc>
!ltsp5
17:25
<ltspbot>
vagrantc: "ltsp5" is (#1) The next version of LTSP, and the future of Thin Client computing on GNU/Linux. Find out more on wiki.ltsp.org, or (#2) ltsp5 is The next version of LTSP, and the future of Thin Client computing on GNU/Linux. http://wiki.ltsp.org/twiki/bin/view/Ltsp/Ltsp5
17:25
<jammcq>
it also depends on dhcpd, tftpd and a few other things
17:26
<loca|host>
i thought i can get a debian server with no display manager, just a server with console, and getting it serving ltsp for another distribution like ubuntu running gnome ... something standalone ...
17:26
<mistik1>
vagrantc: You guys should really do something about the package naming conventions in debian
17:26
<vagrantc>
looks like the ltsp5 entry also has some duplicates
17:26
mistik1: could you be more specific?
17:26
<mistik1>
why not just call the meta package ltsp5
17:26
make it human readable
17:27
<jammcq>
cuz at the time it was created, ltsp5 didn't exist yet
17:27
<vagrantc>
in what way is that more human readable than "ltsp-server" and "ltsp-client" ?
17:27
<mistik1>
ltsp<ver>-[server|client]
17:27
<jammcq>
there's not a client vs server package
17:27
<vagrantc>
and, if at all possible, not having version numbers in the package name is a good thing.
17:27
<jammcq>
it's not like tftp and tftpd
17:28
tell that to the apache2 guys :)
17:28
<mistik1>
*sigh*
17:28sahil has quit IRC
17:28sahil has joined #ltsp
17:28
<vagrantc>
if we upgrade to ltsp6 at some point, and it's a clean upgrade path, it complicates upgrades to have to represent how ltsp6 relates to ltsp5
17:29
instead of just upgrading ltsp
17:29
<mistik1>
Would that upgrade then remove the old version?
17:30
<vagrantc>
and if it's a clean upgrade path, that's a good thing.
17:30
<mistik1>
I disagree but different distros different policies
17:30
<vagrantc>
if it's not a clean upgrade path, then you start doing ugly things like putting version numbers in package names.
17:31
yeah.
17:31
<mistik1>
For gentoo I would suggest Slotting ltsp so a user may have as many versions as they want
17:31bobby_C has quit IRC
17:32
<jammcq>
vmware rocks. I just tested all 3 ltsp tarball images
17:32
<mistik1>
*shrug* I wont bash anything though as long as it "Just Works(tm)" ;-)
17:32
<jammcq>
works great
17:34jbtfrog has joined #ltsp
17:36cliebow has joined #ltsp
17:38
<cliebow>
vagrantc: i put my yaboot and yaboot.conf in /var/lib/tftpboot/wtf and specified /wtf/yaboot in dhcpd.conf and booted right up..seems you could put yaboot and yaboot.conf anywhere you like
17:38
<vagrantc>
this is good news.
17:38
cliebow: thank you for testing :)
17:38
<jammcq>
that's a great directory too :)
17:39
<cliebow>
i was thinking netvista which has special needs
17:39
<vagrantc>
cliebow: and the yaboot.conf needs to... sit right next to the yaboot binary?
17:39
<cliebow>
No Prob!!..i am pretty sure the two need to be tohether
17:39
there is no way to point i know of
17:40
<vagrantc>
hmmm... i wonder if this is feasible for etch
17:41
<cliebow>
i have most of a adhcpd.conf touse if statemnetd for etherboot pxe and apple..just havnt dealt with root-path yet
17:42
it is easy..Jm has that
17:42
<vagrantc>
cliebow: ooooh. tasty.
17:43cliebow__ has joined #ltsp
17:43
<cliebow__>
it is a no brainer really
17:44
see i wasnt fooling when i saod i had a chinefor the tub..
17:49cliebow has quit IRC
17:51
<vagrantc>
cliebow__: no brainer? i can't ever figure out the vendor-id strings ...
17:52
<cliebow__>
jim had that piece all done..
17:53
<loca|host>
howto change default gnome default keyboard layout
17:54
xorg is configured as i want XkbLayout = fr
17:54
but it seems us
17:56
<vagrantc>
loca|host: set XKBLAYOUT=fr in lts.conf
17:58
<cliebow__>
with wireshark you just count characters
17:58* vagrantc contemplates a potential speed improvement for the debian/ubuntu xorg.conf generation
17:58
<vagrantc>
cliebow__: i couldn't figure out how to use that thing.
17:59
<jammcq>
couldn't figure out wireshark? wow, one of the most incredibly useful programs in my tool box
18:01
<vagrantc>
is this one of those GUIs that needs to be run as root?
18:07
<jammcq>
no
18:08
<vagrantc>
i didn't have an interfaces to capture as non-root
18:08
<jammcq>
if you wanna capture packets from an interface, i think it needs to be root, but I usually use tcpdump to capture the packets, and use wireshirk to look at them
18:08
use tcpdump -i eth0 -s 1520 -w /tmp/tcpdump.out
18:08
that'll capture the packet stream
18:08
then, wireshark will open the file and let you look at the packets
18:09
<vagrantc>
nice
18:12jbtfrog has quit IRC
18:18ogra has quit IRC
18:19ogra has joined #ltsp
18:20
<ltsppbot>
"cliebow" pasted "authoritative; subnet 192.168." (34 lines) at http://pastebot.ltsp.org/42
18:20
<cliebow__>
jammcq: that option vendor class d-i looks wird but it works
18:20BadMagic has quit IRC
18:20BadMagic has joined #ltsp
18:22
<cliebow__>
whoops
18:22freemindx has joined #ltsp
18:24
<cliebow__>
yeah it is ok
18:36cliebow__ has quit IRC
18:55Davo_Dinkum has joined #ltsp
18:57cliebow has joined #ltsp
19:04Topslakr| has joined #ltsp
19:17Davo_Dinkum has quit IRC
19:24loca|host has quit IRC
19:39loca|host has joined #ltsp
19:39
<loca|host>
the ltspadmin can install ltsp packages, how can i remove an already installed package ?
19:44
<vagrantc>
loca|host: don't use ltspadmin if you installed using ltsp-build-client
19:45
<loca|host>
LTSP5 is not yet stable ?
19:45
<vagrantc>
ltsp5 is incompatible with the tools used to build ltsp 4.x
19:46
<loca|host>
i knew that, just because there's no docs/tutos for ltsp5 i thought it wasnt that stable one
19:46
<vagrantc>
ltsp5 still is still under development, for sure. it works, for the most part, but it needs... documentation. badly. and people complain about how slow it boots a lot.
19:47
<loca|host>
vagrantc, i have a debian server, can i serve ubuntu ltsp from it ?
19:47
<jammcq>
you sure can
19:48
<loca|host>
where can i find a good doc
19:48
<vagrantc>
loca|host: if you want ... might require some tweaking.
19:48
<cliebow>
errrr
19:48
<vagrantc>
and jammcq is working on that sort of stuff right now
19:48
<jammcq>
http://wiki.ltsp.org/twiki/bin/view/Ltsp/Ltsp5TarballInstructions
19:48
be the first one to try out my new installation instructions
19:48
<cliebow>
jammcq:great!!
19:48
<vagrantc>
jammcq: where did that come from?
19:48
<jammcq>
been working on that stuff today
19:49
but read those docs!!! especially the part about "Recommended only for people who have already used older versions of LTSP, and understand the administration of TFTP, DHCP, NFS and other things"
19:49
<loca|host>
jammcq, should i work with the tarball ? no need for binary packages on debian etch ?
19:49
<jammcq>
loca|host: you said you wanted to install Ubuntu's LTSP on Debian
19:50
<loca|host>
yes
19:50
<jammcq>
the only way to do that is with the tarballs
19:50
it's not like there's a .deb package that you can put on your debian system, to build an Ubuntu ltsp tree
19:50
<vagrantc>
though really, that shouldn't be too difficult.
19:51
<cliebow>
localhost:im serving an ibook off my sony laptop!
19:51
<vagrantc>
jammcq: it would be more proper to drop ltsp-update-* into /usr/local/sbin
19:52
<jammcq>
vagrantc: yeah, but you are putting yours in /usr/sbin, so I figured we didn't want to have 2 places for it
19:52
i'm happy to tell people anything to
19:52
s/to/tho/
19:52
<cliebow>
jammcq:great stuff
19:52
<jammcq>
vagrantc: should I change the doc?
19:52freemindx has quit IRC
19:53
<vagrantc>
jammcq: your call. i cringe at manually installing things outside of /usr/local
19:54
jammcq: but cool.
19:54
<jammcq>
well, putting in in /usr/local wouldn't cause a problem if they also installed the .deb, but then they'd have 2 of them
19:54
<vagrantc>
and /usr/local would take precedence ...
19:54
<jammcq>
is that the rule?
19:54
seems to be on my boxes
19:55
<vagrantc>
that's the theory behind putting custom stuff into /usr/local
19:55
we could implement some ugly hack or something ...
19:55
no.
19:56
<jammcq>
ok, docs are changed
19:56
<vagrantc>
like ... maybe a big warning message if it detects an existing /usr/sbin/$(basename $0)
19:57loca|host has quit IRC
19:57
<cliebow>
jammcq:want a tarball for ppc yet?
19:58
<jammcq>
cliebow: you have one?
19:58
<cliebow>
sur!
19:58
<jammcq>
which distro?
19:58
<cliebow>
sure!
19:58
edgy
19:58
so far
19:58
<jammcq>
dude, I will SOOO take it
19:58
can you tar cjf it for me?
19:59
and put it somewhere that I can come and get it?
20:00
<cliebow>
working on it
20:02
<jammcq>
ok, i'm gonna go grab some dinner. i'll be back a bit later
20:07
<cliebow>
http://169.244.3.137/powerpc.tar.gz..it will take 20 minutes or so
20:07
<vagrantc>
i could try qemu's ppc emulation :)
20:10
<cliebow>
i am challenged with emulation
20:12
http://169.244.3.137/tftpboot.tar..it will be there soon
20:12
<vagrantc>
the only emulator i've gotten to work decently is amd64 on an i386 ... which ... hurts my head.
20:12
<cliebow>
mine too 8~)
20:14Topslakr| has quit IRC
20:19
<cliebow>
shhot..dont know how to double cluck in ppd
20:19
shhot..dont know how to double cluck in ppc
20:19
Balls
20:20
<vagrantc>
as much as i dislike mice ... i can't really see them being functional with only 1 button.
20:20
<cliebow>
ohh..usb mouse
20:21
55c683043e1c0533635d76abd721c12c tftpboot.tar
20:22
Ha!
20:22
<vagrantc>
oh, you expect me to trust an md5sum delivered over irc? :P
20:22
<cliebow>
over a wireless connection hooked up to a wireless connection..why not?
20:23
a traceroute is a powerful thing for me 8~)
20:24
3.244->3.1->1.104->1.1->adelphia
20:31
jammcq:tftpboot.tar checks out
20:40cliebow__ has joined #ltsp
20:41
<cliebow__>
http://www.grammacams.com/dhcpd.conf
20:42
http://www.grammacams.com/tftpboot.tar
20:43
http://www.grammacams.com/powerpc.tar
20:43
powerpc is only half there
20:52cliebow__ has quit IRC
20:54cliebow_ has quit IRC
20:54cliebow_ has joined #ltsp
21:05
<cliebow>
9312e2731608c581803c22fee0e38983 powerpc.tar.gz
21:06cliebow has quit IRC
21:09McDutchie has quit IRC
21:16FernandoMM has quit IRC
21:22
<cliebow_>
all for now..
21:22
<jammcq>
cliebow_: hey
21:23
I got '404 not found'
21:23
on the powerpc.tar file
21:23* vagrantc notes the md5sum was a .tar.gz
21:24
<jammcq>
ah
21:24
yeah, .gz is working
21:32FernandoMM has joined #ltsp
21:38
<cliebow_>
cool!
21:38
running on fumes..c ya tomorrow
21:53FernandoMM has quit IRC
21:56Bhaskar has joined #ltsp
22:10TheProf has joined #ltsp
22:11
<TheProf>
Hello Hello. I hope everyone is doing well.
22:11
I just have a small question about yum on my k12ltsp server (FC).
22:11ccherrett has quit IRC
22:11
<TheProf>
After I yum clean all, I try yum check-update and I get the error
22:12
http://download.fedoralegacy.org/fedora/3/os/i386/repodata/repomd.xml: [Errno 4] IOError: HTTP Error 404: Not Found
22:12
I know that's because the FC legacy program is shut down
22:12
then I get the error: Cannot open/read repomd.xml file for repository: legacy-base
22:12
failure: repodata/repomd.xml from legacy-base: [Errno 256] No more mirrors to try.
22:13
So I'm not sure what to do. Does anyone have a suggestion of what to do/change? Thank you.
22:13ccherrett has joined #ltsp
22:32ccherrett has quit IRC
23:19spackle has quit IRC