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


Channel log from 17 July 2009   (all times are UTC)

00:01bobby_C has joined #ltsp
00:07F-GT has quit IRC
00:08F-GT has joined #ltsp
00:13Lns has quit IRC
00:48ogra has quit IRC
00:52ogra has joined #ltsp
00:52Ahmuck-Jr_ has joined #ltsp
01:30nubae has joined #ltsp
01:47
<Ahmuck-Jr_>
anyone using a ltsp server in a vm sucessfully?
01:50
<ogra>
yes, thats how most of us do development
01:50
vbox server with vbox client and internal network between the two
01:51* alkisg uses host networking with the host os as the ltsp server
01:51
<alkisg>
Good morning all
01:56fotanus1 has joined #ltsp
01:56fotanus has quit IRC
01:57
<nubae>
mornin... so I've install firefox as a local app, which is working, but no net connection, what should I be looking at to get it working?
01:57
this is on Jaunty btw
01:58
<Ahmuck-Jr_>
nubae: same here
01:59
firefox as a local app and then get 404
01:59
but i have to start it from a command line, "ltsp-localapps firefox"
01:59
<nubae>
yeah, I've just created a local user and am gonna ssh in and take a look
01:59
ah... well mine starts from the menu just fine
02:00
I had to fix [default] on first line of lts.conf to fix that though
02:00
<ogra>
you need NAT on the server
02:00
there is no forwarding set up by default
02:00
<nubae>
I have NAT on server
02:00
<ogra>
does the client resolve DNS properly ?
02:00
<nubae>
checking that now
02:01
<ogra>
(try brosing an IP from the local FF)
02:01
<nubae>
looks like it doesnt
02:01
well, I sshed in
02:01
<alkisg>
nubae: do you have the resolv.conf fix?
02:01
<ogra>
right ... the fix is in karmic, not in jaunty iirc
02:01
<alkisg>
It was fixed after jaunty was released
02:02
Ahmuck-Jr_: did you enable the local apps menu from lts.conf?
02:02
<nubae>
my local /etc/network/interfaces has no eth0 mentioned, is that ok?
02:03
<alkisg>
Yeah, it's managed by nm (if you only have 1 nic)
02:03
Ah, you mean the thin client?
02:03
<ogra>
thats fine
02:04
<nubae>
ah my resolv.conf permissions are screwed
02:04
i remember talk of this before, what was the fix?
02:04
<ogra>
known bug
02:05
<alkisg>
https://bugs.launchpad.net/ubuntu/+source/ltsp/+bug/348305
02:05
<Ahmuck-Jr_>
alkisg: yes
02:05
no workie for me
02:05
<ogra>
https://bugs.launchpad.net/ubuntu/+source/ltsp/+bug/347957
02:05
there is a hack at the bottom
02:05
<nubae>
k thanks
02:05
<Ahmuck-Jr_>
i'd like some tool or something to "migrate" users
02:05
i'd like to do a re-install
02:06* ogra would suggest to just add: chmod 644 /etc/resolv.conf ... to /etc/rc.local though
02:06
<ogra>
without all the other hacks
02:07
<Ahmuck-Jr_>
heh, i've had this issue as well
02:07
just one i had not tackled yet
02:08kamalp has quit IRC
02:08
<alkisg>
nubae: if you want you can add my ppa to your sources.list, it has a newer ltsp version that doesn't have this bug, and most greek ltsp schools will have it to their sources, so I'll be careful on what I put in there.
02:08
Also the libxcb fix is there
02:08
(copied by stgraber's PPA - I was just looking for something more stable because stgraber's ppa changes often)
02:09
https://launchpad.net/~ts.sch.gr/+archive/ppa/
02:11l0C0 has joined #ltsp
02:33lejo_ has joined #ltsp
02:33Briareos1 has joined #ltsp
02:33
<Briareos1>
good morning again :)
02:33
<lejo_>
Hi, is there already a (upstream) solution to enable numlock on a TC ?
02:35
<Briareos1>
i want to deploy ltsp on ubuntu 8.04 in virtual box - question: should i use 64 bit or 32 bit ubuntu for the ltsp in the vbox?
02:37
<alkisg>
lejo_: X_NUMLOCK=True
02:37
But you also have to have the numlockx package installed
02:38
<lejo_>
alkisg: ok thx!
02:38
<alkisg>
(X_NUMLOCK=True => in lts.conf...)
02:48
<nubae>
alkisg, thanks... will keep it in mind for deployments...
02:49
right now this is just a test machine, but it would be nice to have a working solution for the thousands of schools we are installing on
02:49* nubae is now working for Guadalinex
02:49
<alkisg>
Ah, you may want to make your own PPA then :)
02:49
<nubae>
maybe yeah
02:49
<ogra>
or wait for karmic :)
02:49
<alkisg>
I'm sure the LTSP released with karmic will also have some bugs :)
02:49
<nubae>
I need to convince the schools to let go of their wireless connections first :-)
02:50
<ogra>
sure it will
02:50
but different ones :)
02:50
fresh and shiny new bugs :)
02:50
<nubae>
weeee
02:50
<alkisg>
Yeah, that's why I wanted a PPA, so that I can minimize the number of ltsp bugs the greek schools have...
02:50dorimar has quit IRC
02:50
<ogra>
took 6 months to add them, so they must be great
02:50
<alkisg>
Heh :)
02:51
E.g. next time I'll update the PPA will be when the ldm logout bug is solved...
02:51
<ogra>
if that ever happens
02:51
<alkisg>
Sure :D
02:52
<ogra>
its such an old bug
02:52
<alkisg>
I have faith in sbalneav :)
02:52
<nubae>
ogra, how's it going in arm land? are we likely to see prototypes soon?
02:52
<ogra>
yeah
02:52
at some point
02:52
<nubae>
I'm trying to guess what will be available as hardware in September
02:52
<ogra>
i'm no HW manufacturer though
02:53
netbooks and nettops
02:53
<alkisg>
What are the key benefits of using an arm-based laptop instead of e.g. an atom-based one?
02:53
<nubae>
yeah, but u have a better idea than the rest of us for sure
02:53
<ogra>
alkisg, an arm consumes 10% of the power an atom takes
02:53
<alkisg>
That's CPU only, right? But overall, it lasts, what, 1-2 hours more?
02:53
<ogra>
alkisg, thats why your mobile phone has one :)
02:54
well, i'd guess twice as long as an atom with similar setup
02:54
<alkisg>
Mmm that's a big difference
02:54
<ogra>
with room for improvement
02:54
<alkisg>
What about performance?
02:55vagrantc has joined #ltsp
02:55
<ogra>
as ram gets better and SSDs get more common the peripherials will draw less
02:55
<alkisg>
Right, so CPU power will be more significant
02:55dorimar has joined #ltsp
02:55
<nubae>
so which models have u been working with at the mo?
02:55
<ogra>
performance is ok CPU wise a 1GHz ARM doesnt have to hide in front of a 1.2GHz ATOM
02:56
prob are the graphics drivers that are nonexistent for most HW yet
02:56Briareos_ has joined #ltsp
02:56
<Briareos_>
lost my conn.
02:56
<ogra>
nubae, a good bunch i'm not allowed to talk about yet ... public is the freescale/pagatron imx51
02:57
<alkisg>
Nice! And, did you hear any rumors about final pricing? Will they be about the same as other netbooks?
02:57
<ogra>
cheaper
02:57
<Briareos_>
is it wise(r) to use LTSP on a 64bit ubuntu _guestsystem_ in a virtualbox or would i rather use a 32bit?
02:57
<ogra>
i know pegatron lans for something like $200-250 for a decent netbook in the quality of a dell mini
02:58
*plans
02:58
<alkisg>
Now you got me excited, I'll be waiting for them before I buy a netbook :)
02:59
<ogra>
well, it will still be the first iteration ... excpect typical early adopter problems :)
02:59
<alkisg>
Yeah sure that's to be expected
02:59* vagrantc is curious about an arm or mips based netbook
03:00
<alkisg>
My main problem will be that it won't be able to run wine apps. Heh, I'll use it as an LTSP client for that :)
03:01
<ogra>
mips ones exist since a while
03:01
<alkisg>
I tried a mips-based one, it was dead slow
03:01
(only 64mb ram, low end)
03:01
<vagrantc>
i was eyeing up the ones from leemote.com
03:02
er, lemote.com
03:02
the teelong
03:03
gah. yeelong.
03:03
<ogra>
heh
03:03* vagrantc is using a box of cat5 for a table
03:03
<ogra>
we had one at a grocery discounter here about half a year ago ... 120€
03:04
specs similar to the first eee's
03:04
<vagrantc>
grocery?
03:04
<nubae>
so how many arm based models do u expect to hit the market then? (u mentioned a bunch)
03:04
<ogra>
yeah
03:04
in a supermarket
03:04
<vagrantc>
selling laptops?
03:04
<ogra>
yeah
03:04
for about $150 ... or 120€
03:04
<nubae>
I remember, the Hofer ones
03:04
called something else in Germany
03:05* vagrantc will have to bring up the idea to people's food co-op
03:05
<ogra>
400MHz MIPS, 128M, 2G
03:05
800x480
03:06
not really something you want to use on a day to day base
03:07* alkisg wonders how well does the flash player ARM port work...
03:08* vagrantc ponders uploading a new ltspfs today
03:08
<ogra>
according to my colleagues who do OEM work it runs smooth
03:08
<alkisg>
Nice, cause many people would complain if surfing without flash...
03:09
<ogra>
i havent had it in my own hands yet though
03:09
and there is no freely downloadable version
03:09
has to be preinstalled by OEMs
03:09
<alkisg>
So the main problem remaining now are the graphic card drivers?
03:10
<ogra>
yeah
03:13
<nubae>
root@ltsp20:~# cat /etc/resolv.conf
03:13
search virt-man || Change this for your own network
03:13
nameserver 192.168.0.254 || Change this for your own network
03:13
root@ltsp20:~# ping google.com
03:13
ping: unknown host google.com
03:13
and permissions are correct
03:14
<ogra>
pipes are valid comment signs in resolv.conf ?
03:14Briareos1 has quit IRC
03:15
<ogra>
and you have a DNS on 192.168.0.254 ?
03:15
<nubae>
hmmm, probably not
03:16
<ogra>
give it a valid DNS address then in your dhcpd.conf :)
03:16
<alkisg>
ogra, are you sure putting dns in dhcpd.conf works?
03:16
<ogra>
it should
03:17
but no, i'm not *sure* :)
03:17
<alkisg>
I think that if isn't defined in lts.conf, then resolv.conf isn't created at all
03:17
<ogra>
worst case use lts.conf
03:18
<nubae>
i'll tell u if it works in a sec
03:18
<alkisg>
(and I thought I heard someone saying that this was by design, for security... :()
03:20
<nubae>
nope, setting dns in dhcpd.conf doesnt work
03:20
<ogra>
then use lts.conf
03:23panthera_ is now known as panthera
03:24Briareos_ is now known as Briareos1
03:24Briareos1 is now known as Briareos_
03:29dorimar has quit IRC
03:38yanu has joined #ltsp
03:49
<nubae>
all working now.... thanks
03:50
I finally got round to looking at fatclient under jaunty, but getting some very strange stuff:
03:50
chmod: failed to get attributes of `Ubuntu': No such file or directory
03:50
error: LTSP client installation ended abnormally
03:50
Ubuntu is an option for the --fatclient extension
03:51
not sure why its doing this, but I imagine something has changed in the way plugins works
03:54ogra has quit IRC
04:06alkisg has quit IRC
04:13
<nubae>
hmm, so looks like plugins has changed to only allow values of true or false
04:13
this is quite unhelpful
04:16ogra has joined #ltsp
04:17flokuehn_ has quit IRC
04:23flokuehn has joined #ltsp
04:31hanthana has quit IRC
04:34
<vagrantc>
nubae: if it's only true/false, then it shouldn't accept options at all.
04:34
nubae: you sure you're reading the code properly?
04:34alkisg has joined #ltsp
04:35
<nubae>
well, no matter what I set as a value, I get only true or false back
04:35
but whats weirder is I then get chmod: failed to get attributes of `true': No such file or directory
04:35
error: LTSP client installation ended abnormally
04:36
anyway, something clearly changed from intrepid to jaunty because the script was working just fine before, and now it seems totally broken
04:39
I'm more interested in the second part, as I dont get what its trying to do with the get attributes
04:45
<vagrantc>
hrm. the fatclient plugin isn't in upstream.
04:46
<nubae>
nope, but the chmod part is coming from outside the script...
04:46
If I can fix it, I'll move it upstream... but at the moment, kinda pointless
04:46
<vagrantc>
nubae: is the fatclient plugin part of the ubuntu package?
04:47
or is it just somewhere else?
04:47
<nubae>
it is to be installed under /usr/share/ltsp/plugins/ltsp-build-client/Ubuntu
04:47
<vagrantc>
i don't think the plugins have changed much since 2006.
04:47
<nubae>
so uses the ubuntu plugins system... or at least used to
04:48
well, something has changed, I dont get why I'm getting these weird chmod errors
04:48
<vagrantc>
i mean, the plugins themselves have changed, but the argument processing is hasn't changed since the initial implementation 3 years ago.
04:48
<nubae>
well, must have
04:48
my script I havent touched since before intrepid
04:49* vagrantc can't comment on what vagrantc can't see
04:49
<nubae>
and many users have complained it doesnt work under Jaunty, which is what I am checking now
04:49
http://www.nubae.com/030-fatclient
04:49
u can see the script there
04:50
but its not something in the script for sure... something changed elsewhere
04:54alkisg has quit IRC
04:54alkisg has joined #ltsp
04:56
<vagrantc>
nubae: you've got some bashisms in there
04:56
[[ == ]] should just be [ = ]
04:56
oh wait, ltsp-build-client runs as bash, nevermind.
04:56
<nubae>
yup
04:57
where could I be getting chmod: failed to get attributes of `Ubuntu': No such file or directory
04:57
error: LTSP client installation ended abnormally from?
04:57
<vagrantc>
nubae: try adding "set -x" to the top of the plugin to see exactly where it dies and "set +x" at the bottom
04:57
<nubae>
k
04:58
<vagrantc>
perhaps this line: chmod 600 $ROOT/etc/sudoers
04:59
why use [[ ]] ?
04:59
does any of the code actually need it?
05:00
<nubae>
ok, well I can see it dying now where it tries to create /opt/ltsp/fati386/Ubuntu (why is it trying to create that, thats no where in the code)
05:01
<vagrantc>
and things like: FATCLIENT=$(echo $FATCLIENT)
05:01
seem completely pointless
05:01
<nubae>
yeah was being used before, I can take that out
05:02
doesnt answer the question though... why is fati386/Ubuntu being created?
05:02
aha... so its this that is screwing it up: + . /usr/share/ltsp/plugins/ltsp-build-client/Ubuntu/005-tmpdir
05:02
<vagrantc>
where is it trying to create that?
05:03
are you using a variable in your plugin that's used by another plugin, perhaps?
05:03
<nubae>
yep
05:04
nice spot... thanks
05:05
nice one, replacing $TMP for something else fixed it... still... uncool for another plugin to take my variables :-(
05:05
I guess thats what I get for not putting the thing upstream
05:15
<vagrantc>
TMP is a widely used variable :P
05:32
<nubae>
well, it wasnt smashing my script before
05:32
mine was there first! :-)
05:55
<Briareos_>
ty, cu
05:55Briareos_ has quit IRC
06:21cyberorg has quit IRC
06:21vvinet has quit IRC
06:23cyberorg has joined #ltsp
06:43ogra has quit IRC
06:46Shinhung has joined #Ltsp
06:51
<Shinhung>
is it possible that dumb terminals keep disconnecting from the server due to overheat?
06:52ogra has joined #ltsp
06:58hgrover has joined #ltsp
07:06
<nubae>
nice so seems fatclient works very nicely now with Jaunty
07:06
noticable speed difference on more potent machines to pure ltsp
07:06pmatulis has joined #ltsp
07:08pmatulis has quit IRC
07:10sene has quit IRC
07:12pmatulis has joined #ltsp
07:15Blinny has joined #ltsp
07:16gate_keeper_ has quit IRC
07:33gate_keeper_ has joined #ltsp
07:38sene has joined #ltsp
07:39vvinet has joined #ltsp
07:42DaZ- has joined #ltsp
07:44hgrover has quit IRC
07:46Shinhung has quit IRC
07:50
<nubae>
where do I edit the repo ltsp pulls from?
07:51
<vagrantc>
--mirror
07:51
or /etc/ltsp/ltsp-build-client.conf
07:52
or a handfull of other options
07:54
<nubae>
thanks
07:55
so /etc/ltsp/ltsp-build-client.conf one has to create
07:56Shinhung has joined #Ltsp
07:57DaZ has quit IRC
08:02* _UsUrPeR_ returns!
08:02
<_UsUrPeR_>
...hello
08:04
ok, so: wake on LAN
08:16nubae has quit IRC
08:25
<lejo_>
someone using card-readers on a TC ?
08:48Egyptian[Home] has joined #ltsp
08:49Gadi has joined #ltsp
08:49Egyptian[Home] has joined #ltsp
08:54
<_UsUrPeR_>
gadi: do you have any experience w/WOL on thinclients?
08:56anton_ has joined #ltsp
08:58
<Gadi>
_UsUrPeR_: good morning
08:58
yes
08:58
whats up?
08:59
oh, and I have a good one for you when we're done
08:59
one of those "I've never seen that before" moments this morning
08:59
:)
08:59
<ogra>
vagrantc, https://wiki.ubuntu.com/ARM/BuildEABIChroot :) with that ltsp-build-client --arch armel should be easy
09:00
<_UsUrPeR_>
gadi: I was wondering if you had any experience with wake-on-lan
09:00
<anton_>
Hello, can anyone help me on LTSP LocalDev problem. I have gone thru all steps of Troubleshooting, and on step 11 found that lbus_event_handler.sh prints an error. Actually, it is ltspfs that prints the error. And the message is from xauth: ""No matches found, authority file "-" not written"". I can supply ltrace output if necessary. I use Fedora 9-based distro (ASPLinux), and gdm 2.22
09:00
<Gadi>
_UsUrPeR_: yes
09:00
<ogra>
lbus_event_handler.sh ??
09:00
<Gadi>
shoot
09:01
anton_: sounds like you are still on LTSP 4.2
09:01
<anton_>
yes i am
09:01
<ogra>
ouch
09:02
<anton_>
is it bad? are things smoother in 5?
09:03
<Gadi>
anton_: I'll let ogra answer that -> he loves that question
09:03
ogra: btw, very cool on the cross-arch thing
09:03
<ogra>
its unmaintained and hasnt seen security updates since over 3 years
09:03
<_UsUrPeR_>
gadi: I have BIOS WOL set on a 1220, send the magic packet, which can be observed as hitting the client (broadcast hits all) it's got the right MAC, but it does not boot
09:03
I have no experience with WOL, but I imagine that it should "just work" when invoked
09:04
<Gadi>
_UsUrPeR_: is WOL enabled in the BIOS?
09:04
<ogra>
Gadi, its probably not stable, but enough for building a thin client chroot for sure :)
09:04
<_UsUrPeR_>
yes
09:04
<ogra>
... and so much faster than running a VM
09:05
<Gadi>
_UsUrPeR_: boot the thin client, log in at shell, and do: ethtool eth0
09:06
<_UsUrPeR_>
k
09:06
<alkisg>
_UsUrPeR_: also, for some motherboards, the client must be in a soft-off state, not turned off by the hardware button.
09:06
<rjune_wrk>
ogra: hey
09:06
<_UsUrPeR_>
alkisg: I.E. "halt" status?
09:06
<ogra>
rjune_wrk, seen the wikipage above ?
09:06
<alkisg>
Yeah, a simple power off from ldm.
09:06
<ogra>
rjune_wrk, now you can test your code yourself :)
09:07anton_ has quit IRC
09:11anton_ has joined #ltsp
09:12Blinny has quit IRC
09:13
<vagrantc>
ogra: so... it just dynamically figures out that it's an arm-based binary and wraps it in qemu?
09:14
<ogra>
right, binfmt-misc catches the header
09:14
<rjune_wrk>
ogra: the one you posted to vagrant?
09:14
<ogra>
and qemu acts as a syscall converter
09:14
rjune_wrk, righ
09:14
t
09:15
<vagrantc>
ogra: i was just talking with otavio last night about arm and mips support and he didn't want to bother until we have cross-architecture building available... :)
09:15
<rjune_wrk>
I know it works, I just wanted you to take a peek at the actual binary and tell me what you thought
09:15
<ogra>
vagrantc, there you got it :)
09:18
<_UsUrPeR_>
gadi: ok, so ethtool needs to be installed on the chroot then
09:18* _UsUrPeR_ does that
09:18
<Gadi>
_UsUrPeR_: yeah, then once installed, boot the client, log into shell, and do: ethtool eth0 |grep Wake
09:19
<_UsUrPeR_>
gadi: what am I looking for from ethtool?
09:19
<Gadi>
Wake-n capabilities
09:19
Wake-on
09:21
it will tell you what it supports and the current setting
09:22CAN-o-SPAM has joined #ltsp
09:22
<Gadi>
so, anyway, the weird thing I encountered this morning was a server that was turned off that was flooding the switch with network traffic
09:22
<_UsUrPeR_>
?!?
09:22
<Gadi>
I know!
09:22
it was weird
09:22
<_UsUrPeR_>
how "off" was it exactly? :)
09:22
<Gadi>
unplugged it, and all was right with the world
09:23
off as in shutdown
09:23
evidently, down but not out :P
09:23
<ogra>
Gadi, tis is da fjutscha ;) intelligent network cards with builting AI
09:23
<_UsUrPeR_>
weird. really weird. never heard of that
09:23
:O
09:23
<ogra>
*builtin
09:23
<_UsUrPeR_>
Cybernet!
09:23
<ogra>
"no mommy, i dont want to go to sleep yet!"
09:23
<_UsUrPeR_>
sentient computers working to kill the world while you are sleeping
09:24
ok, wake-on: d
09:24
<Gadi>
_UsUrPeR_: so, you can't wake them up and I can't put them to sleep
09:24
must be the end of the world
09:24
<_UsUrPeR_>
it's like owning a cat
09:24
<Gadi>
_UsUrPeR_: what does "Supports Wake-on:" say
09:24
<ogra>
Gadi, i would ckeck if there is a newer firmware for that chipset
09:25
*check even
09:25
<_UsUrPeR_>
gdi: supports wake-on: pumbg
09:25
<Gadi>
yeah - this was a Windows test server - Im afraid to check for firmware
09:25
:)
09:25
<_UsUrPeR_>
errr gadi ^^^
09:25
<Gadi>
_UsUrPeR_: excellent!
09:26
so, interpretation:
09:26
Your chip supports a whole bunch of wake-on-lan modes
09:26
including "g" (wake by Magic Packet)
09:26
currently, you are set to "d" = "disabled"
09:26
which is why you are not waking up
09:27
now, do:
09:27
<_UsUrPeR_>
hmm
09:27
<Gadi>
ethtool -s wol g
09:27
to change it to "g"
09:27
oops
09:27
forgot the eth0
09:27
ethtool eth0 -s wol g
09:28
then, have a look at ethtool eth0 again to see that it changed
09:29
<_UsUrPeR_>
gadi: it was ethtool -s eth0 wol g :)
09:29
ok, it's changed to wake-on: g no
09:29
w
09:29
<Gadi>
ah good
09:29
now press the power button on the client and shut it down
09:30
then, try waking it up from the server again
09:31
<_UsUrPeR_>
gadi: no dice :/
09:31
should I have "halted"?
09:31
<Gadi>
ok, so boot the thin client again, and check ethtool again
09:31
perhaps something else is overriding the setting
09:31
<_UsUrPeR_>
k
09:32
<Gadi>
like maybe some BIOS config or some such
09:33
<_UsUrPeR_>
gadi: wake-on: d again :/
09:33
<Gadi>
dum da dum dum
09:33
:)
09:33
back to the BIOS!
09:34
<_UsUrPeR_>
:O but it says it's enabled >:(
09:35F-GT has quit IRC
09:35
<Gadi>
what chipset is it?
09:36
<_UsUrPeR_>
dlw 1220 is a via chipset
09:36
<Gadi>
do you know which one?
09:37
lsmod|grep 8139
09:37
oh, I meant network chipset
09:37
<_UsUrPeR_>
VT8237R
09:37F-GT has joined #ltsp
09:37
<Gadi>
thats the southbridge
09:37
iirc
09:37
<_UsUrPeR_>
CN700
09:37
<Gadi>
what does it use for networking?
09:37
<_UsUrPeR_>
you're right :)
09:37
<Gadi>
realtek is my guess
09:38
<_UsUrPeR_>
10/100 Base-T Realtek 8100C Fast Ethernet
09:38
<Gadi>
do you know which driver?
09:38
lsmod|grep 8139
09:39
<_UsUrPeR_>
lemme get an lshw and all from the client. that'll help :)
09:39
<Gadi>
ok
09:39
I think ethtool supports setting wol in most newer realtek chips
09:41
you might need to set it to g on every boot with a script, and then ensure that the system halts rather than powers off completely whn the power button is pressed
09:42
<alkisg>
Ah, so that's the difference of halt vs power off... /me was wondering about that terminology :)
09:44
<_UsUrPeR_>
same here. I have read it, but am not familiar with the term
09:44
<Gadi>
alkisg: yeah, I struggle with that one, too
09:44* _UsUrPeR_ is a shutdown -h now man
09:45
<Gadi>
but, halt should equal "stop", whereas poer off should equal "off"
09:45
<_UsUrPeR_>
hmm... it seems "halt" does not fully shut down the client
09:45
It says "system halted" and sits there :)
09:45
<alkisg>
Mmmm reminds me of windows 95 days :)
09:46
<Gadi>
as opposed to their cousins "suspend" and "hibernate"
09:49Shinhung has quit IRC
09:58RobertBerger has joined #ltsp
10:23
<rjune_wrk>
!g
10:23
<ltspbot>
rjune_wrk: "g" is Gadi!!!!!!!!!!!!!!!!!!!!!!!!
10:24anton_ has quit IRC
10:29* Gadi waves to rjune_wrk
10:36
<Ahmuck-Jr_>
sbalneav: what do you do as work?
10:38l0C0 has quit IRC
10:40
<rjune_wrk>
what's new?
10:40staffencasa has joined #ltsp
10:42dan_young has quit IRC
10:50dan_young has joined #ltsp
10:51
<_UsUrPeR_>
hey all. Having a problem getting a client to start X. It's crashing :/
10:51
<ltsppbot>
"_UsUrPeR_" pasted "weird intel failure in Xorg.7.log." (332 lines) at http://pastebot.ltsp.org/446
10:57flokuehn has quit IRC
10:57bobby_C has quit IRC
10:58flokuehn has joined #ltsp
10:59
<Gadi>
_UsUrPeR_: laptop?
11:13vagrantc has quit IRC
11:21Egyptian[Home] has quit IRC
11:24RobertBerger has left #ltsp
11:25vagrantc has joined #ltsp
11:26hanthana has joined #ltsp
11:27Ahmuck-Jr has quit IRC
11:31bieb has joined #ltsp
11:35
<ogra>
vagrantc, geez, you added an error message to nbd mounting !
11:35
<vagrantc>
heh
11:36
ogra: well, i've been hanging out with Yoe, who maintains nbd for debian and does some upstream work on it as well...
11:36
<ogra>
i wonde if an exit 0 would work ther
11:37
<vagrantc>
define work? :)
11:37
<ogra>
so it doesnt try to move on with the mounting and drops into busybox without nasty "no /sbin/init"
11:37
i dont know if you have job control at that point
11:37
<vagrantc>
ah, yes ... that might be cleaner
11:37* vagrantc tests
11:41* vagrantc is still thinking of writing a generic "ltsp" initramfs-tools hook, which handles things a little more flexibly...
11:42
<ogra>
as long as you dont add to many regressions :)
11:44
<vagrantc>
heh
11:45
i'd still like a script that figures out what to do based on results from dhcp rather than the bootprompt, mainly.
11:46
well, exit 1 gives a kernel panic...
11:46
which is maybe worse than the busybox noise...
11:46
<ogra>
yeah, i thought i remembered something like that
11:47* vagrantc tries exit 0
11:47
<ogra>
no job control and you are actually exiting init
11:47
<vagrantc>
i think it's essentially exiting init...
11:47
right
11:47
<epsas>
hmm
11:47
i am trying to figure out the best place for setting up post-login scripts
11:48
there is no /etc/X11/xdm directory
11:48
<vagrantc>
ogra: well, with that last patch, i'm easily able to use NBD+aufs+ext2+tmpfs
11:48
<epsas>
and am not sure where ldm would store that
11:48
(and i don't even see any ldm processes on the server)
11:48
<vagrantc>
ogra: which seems just as fast as squashfs, but the image is more easy to tweak manually...
11:48
<ogra>
vagrantc, you could branch out to a different script that just doesnt do the mounting and ends
11:49
nice !
11:49
<vagrantc>
epsas: /opt/ltsp/i386/usr/share/ldm/rc.d/K*
11:50
epsas: or /opt/ltsp/i386/usr/share/ldm/rc.d/X99*
11:51
<epsas>
hmm - i need to perform actions on the server $HOME directory though
11:51
this may be tricky
11:52
<ogra>
no
11:52
<vagrantc>
epsas: well, if you haven't disabled localapps, it's available on the thin-client via sshfs ...
11:52
<ogra>
just look at the localapps scripts
11:52
<vagrantc>
epsas: oterhwise, just do it through the ssh tunnel
11:53otavio has joined #ltsp
11:53
<epsas>
the localapps scripts?
11:53
<vagrantc>
otavio: the link ogra mentioned: https://wiki.ubuntu.com/ARM/BuildEABIChroot :) with that ltsp-build-client --arch armel should be easy
11:53
<epsas>
okay - let me read up on the wiki
11:54
<Gadi>
vagrantc: you should be able to just add "panic"
11:54
vagrantc: instead of exit
11:55
<vagrantc>
Gadi: will try...
11:55
<Gadi>
vagrantc: "panic" is an initramfs function
11:55
<epsas>
okay - i think i need to figure out a way to trigger the server /etc/profile on login -- that would be ideal
11:56
<ogra>
Gadi, oh, you genious you :)
11:56
<otavio>
vagrantc: will take a look in few minutes :-)
11:56* Gadi crawls back into corner
11:56
<ogra>
vagrantc, yeah, panic can even take a message iirc
11:56
<Gadi>
:P
11:56
<vagrantc>
pretty soon, we'll be telling our users useful information!
11:56
<ogra>
so you just wrap your error into the panic call
11:56
OMG
11:59
<epsas>
did you just shit bricks?
12:00
<ogra>
heh, no, but we but showing useful info means the users will actually *know* what broke ... :)
12:01
we'll all lose our wizard status t some point and will only have to deal with *real* problems :)
12:01
<epsas>
hmm
12:01
<alkisg>
Heh I *knew* there was a conspiracy behind LTSP... :P :D
12:01otavio has quit IRC
12:01
<epsas>
it used to be that we could execute post-login scripts by adding files to /etc/X11/xdm/Xstartup or /usr/share/config/kdm/Xstartup
12:01otavio has joined #ltsp
12:01
<epsas>
there has to be an analog for this now
12:02
<ogra>
it took us *years* to work out all the cryptic error messages ;)
12:02
<Gadi>
try: /etc/xdg/autostart/
12:02
epsas: ^^
12:02
adding a launcher there should work for any xdg-compliant desktop
12:03
<epsas>
hmm - okay i have scripts to create user home directories if they do not exist... and scripts to write a custom ~/.config/user-dirs.dirs
12:04
<ogra>
the latter will work from xdg
12:04
<epsas>
no - it won't work from XDG
12:04
<ogra>
the former will hit you already earlier ... ssh wont let you establish the tunnel without homedir
12:04
<Gadi>
why not use pam_mkhomedir?
12:04
<ogra>
right
12:05
i was about to say that
12:05
<Gadi>
sorry
12:05* Gadi crawls faster to corner
12:05
<ogra>
for what ?
12:05
<epsas>
ssh already establishes a tunnel for the LDM session I thought
12:05
<Gadi>
i dunno - I have become more sheepish in this channel of late
12:05
:)
12:05
<ogra>
with the user credenttials
12:06
there is no "LDM session" :)
12:06
<epsas>
wait - what about the ldm login screen
12:06
<ogra>
there are two ssh tunnels
12:06
<epsas>
is that run locally?
12:06
<ogra>
both owned by the user
12:06
yes
12:06
<epsas>
ok
12:07
i need to get a new job -- programming is more fun than customizing ltsp
12:07
heh - oka
12:08
<ogra>
use pam_mkhomedir for the home creation, add a launcher for a script that creates ~/.config/user-dirs.dirs to /etc/xdg/autostart/
12:08
thats the clean way to achieve what you want
12:25
vagrantc, pfft, the tftp bug is a typical debian bug :)
12:25
<vagrantc>
ogra: we're doing it wrong :)
12:25
<ogra>
you guys use full busybox :)
12:25
we dont
12:26
<vagrantc>
well, full path shouldn't break for y'all, right?
12:29
<ogra>
no, it wont
12:29
i would have shouted earlier :)
12:30
not sure about fedora though, i have no idea if their tftp sits in bin sbin or where
12:32* ogra calls it a day
12:33
<vagrantc>
they don't use initramfs-tools anyways
12:34
and the hook script copies it from /usr/bin to the initramfs's /bin, anyways ... i checked before committing :)
12:41
<Gadi>
vagrantc: while ur picking at things, you could addd some code to grab the arch from uname -m and modify the default value for tftpath
12:41
:)
12:41* vagrantc has no plans to play the uname -m to debian architectures game
12:41
<Gadi>
hehe
12:41
fair nuff
12:42* vagrantc might be tagging and uploading soonish
12:42
<vagrantc>
ltsp-trunk, that is
12:42
it's got several bugfixes relevent for debian, at least.
12:44
<Gadi>
is ltspfs-0.5.13 in any of the debian repos yet?
12:44
<vagrantc>
Gadi: yup
12:44
<Gadi>
which one?
12:44
<vagrantc>
ftp.pl.debian.org
12:45
probably others
12:45
in sid
12:45* vagrantc hopes it didn't break anything
12:45
<Gadi>
ah, ok - because sometimes people ask about RDP+localapps on debian
12:45
when ur not here
12:45
er, RDP+localdev
12:45michael2 is now known as chalupacabra
12:46
<vagrantc>
i haven't tested it, but it should be installable on lenny/stable without pulling in extra dependencies.
12:46
the previous version worked on lenny without issues
12:46
and the diff is pretty small.
12:46* Gadi nods
12:46
<Gadi>
when u tested, you were using sid?
12:46
for everything?
12:47
<vagrantc>
squeeze/testing
12:47
<Gadi>
gotcha
12:48evilx has joined #LTSP
12:50gate_keeper_ has quit IRC
12:51
<rjune_wrk>
Gadi: how you doing?
12:51RobertBerger has joined #ltsp
12:52RobertBerger has joined #ltsp
12:52
<Gadi>
rjune_wrk: not bad, et tu?
12:52RobertBerger1 has joined #ltsp
12:53adrian15 has joined #ltsp
12:54
<adrian15>
ltspfs "authentication failed". Google does not help too much. Does anyone have updated instruction on how to debug and solve this?
12:55
<Gadi>
adrian15: are u using LTSP5?
12:55
<vagrantc>
adrian15: sounds like you don't have access to the ltspfs_token X property
12:56
adrian15: you running the manual debugging instructions?
12:57
<adrian15>
Gadi: Well, I think so it is the one that tcos offers to ubuntu 8.04 installations
12:58
vagrantc: Which ones are these manual debugging instructions? Can you please give me a link ?
12:58
<vagrantc>
adrian15: you're the one doing the debugging, you tell me :P
12:59
<adrian15>
vagrantc: I am trying to solve my problem ;)
12:59
<vagrantc>
adrian15: i.e. what were you doing when you got the error message? were you following any instructions?
13:00
<adrian15>
vagrantc: More or less the instructions are: Once the pendrive has been automatically mounted at the terminal (the physical machine not the server I mean) you can run ltspfs manually to see if something goes wrong. So I run: lstpfs 192.168.1.15:/mnt/sdb1 /home/manolo/Desktop/pendrive
13:00
vagrantc: The result is Authentication failed.
13:01
vagrantc: If I run xauth -n list it outputs two lines that are the same ip (one is ipv6 and the other one ipv4)
13:01
vagrantc: So... any clues?
13:03
<vagrantc>
adrian15: are the instructions posted somewhere?
13:04
and if so, could you tell me?
13:04
and while we're at it:
13:04
!release
13:04
<ltspbot>
vagrantc: "release" is please mention the linux distro and release you're using :)
13:07
<adrian15>
vagrantc: Ubuntu 8.04
13:07
vagrantc: Yes, but they are in Spanish
13:07
vagrantc: But they are what I have just said.
13:08
vagrantc: I want to know what else should I try if ltspfs mount fails.
13:08* vagrantc would really like to see the instructions
13:08
<vagrantc>
i can stumble through spanish, and there's some native speakers about who could help
13:09
otherwise, i don't know what you've tried, so i can't really suggest what to do next.
13:10RobertBerger has quit IRC
13:10* ogra bets TCOS simply misses setting the token
13:11
<ogra>
it doesnt use LDM
13:11fotanus has joined #ltsp
13:11
<rjune_wrk>
ogra: I thought you called it a day
13:11
<ogra>
yeah, just checked my running build while waiting for dinner
13:12
<vagrantc>
adrian15: ah, is this ltsp, or tcos ?
13:12
<ogra>
he said tcos above
13:12
<vagrantc>
i missed it ...
13:13
ogra: as ogra said, tcos probably doesn't the the ltspfs_token in X11 properties ... i'd file a bug with tcos, if that's what you're using.
13:13
<adrian15>
vagrantc: Its tcos but it uses ltspfs in the background for pendrive mounting.
13:13
<vagrantc>
pfft.
13:13
adrian15: that previous was meant for you :)
13:14
<adrian15>
vagrantc: I have tried: lstpfs 192.168.1.15:/mnt/sdb1 /home/manolo/Desktop/pendrive
13:16ltspbot has joined #ltsp
13:20
<Gadi>
vagrantc: can you check my work: https://help.ubuntu.com/community/UbuntuLTSP/RdesktopLocaldev#preview
13:20
<adrian15>
vagrantc: There is not output about it
13:21
vagrantc: So... How am I supposed to set this ltspfs_token property set manually?
13:21
<ogra>
you cant
13:22
<adrian15>
ogra: Using free as in freedom software, and I cannot ?! ;)
13:22
<ogra>
well, you can
13:22
but its a lot of hackery i suppose
13:23
<adrian15>
ogra: xprop has a -set option
13:24
so name is ltspfs_token and value is... ?
13:24
ogra: ?
13:24
<ogra>
thats your prob :)
13:24
<Gadi>
adrian15: the value is a magic key
13:24
<ogra>
right
13:24
<Gadi>
that is randomly generated
13:24
<ogra>
and its created by LDM
13:24
which TCOS doesnt use
13:24
<adrian15>
Gadi: A magic key that it is found in: /var/run/ltspfs_token maybe ?
13:25
<Gadi>
on the client
13:25
<adrian15>
in the terminal I mean
13:25
<Gadi>
right
13:25
<adrian15>
yes, on the cliente
13:25
ok
13:25
So...
13:25
Gadi: I am going to play with that, see what happens. Thank you.
13:25
<Gadi>
adrian15: does tcos use gdm?
13:30
<adrian15>
Gadi: yes, it uses gdm
13:30
<Gadi>
does it use lbus?
13:31
like in the old ltsp 4.2 days?
13:31
adrian15: if you know where the client launches ltspfsd, you may have luck adding the argument "-a"
13:31
so, it calls: ltspfsd -a
13:32
the -a *should* disable the authentication
13:32
<adrian15>
Gadi: I am not sure about lbus because main tcos developer is not available right now
13:32
<Gadi>
of course, use at ur own risk
13:40
<adrian15>
Gadi: What's the format that I should select ?
13:40
Gadi: it seems that when running xprop you need to set also its property format. Which one?
13:42
<Gadi>
xprop -root -f LTSPFS_TOKEN 8s -set LTSPFS_TOKEN $(cat $ltspfs_token)
13:42
thats a line from our code
13:42
obviously, substitute the last bit
13:44
<adrian15>
Gadi thank you very much
13:44alkisg has quit IRC
13:45
<Gadi>
con mucho gusto
13:45
:)
13:49
<adrian15>
Gadi: Ok. Ltspfs token has been set sucessfully. ltspfs manually mount gives the same problem: Authentication failed.
13:50
Gadi: Any more ideas ?
13:50
<Gadi>
hmm...
13:51
is this something that is supposed to work in tcos or are you a developer trying to make it work?
13:51
<adrian15>
Gadi: I have run it both as the user that uses the computer and as root. The same result.
13:51
Gadi: Is something supposed to work in tcos
13:52
<Gadi>
I wonder if following the LTSP 4.2 local dev troubleshooting guide might help
13:53* vagrantc notes that udev usually creates the ltspfs_token nowadays
13:54alkisg has joined #ltsp
13:54
<Gadi>
adrian15: what version of ltspfs do you have there?
13:54
<vagrantc>
Gadi: have you tried 0.5.13 yet?
13:54
<Gadi>
vagrantc: not yet
14:02
<adrian15>
Gadi: I think I know what's the problem. It might be related to dhcpd.conf setup. An stupid error. Let me check it.
14:03
<vagrantc>
Gadi: the rdesktop thing looks fine, as far as i can see.
14:05
<Gadi>
great!
14:05
in that case...
14:05spectra has joined #ltsp
14:06
<adrian15>
Gadi: Yeah. It was a wrong dhcpd.conf configuration. The clients hostname was not set ok.
14:06
<Gadi>
!learn RDP-localdev as when I want to use local devices in an LTSP rdesktop session. See: https://help.ubuntu.com/community/UbuntuLTSP/RdesktopLocaldev
14:06
<ltspbot>
Gadi: The operation succeeded.
14:06
<Gadi>
adrian15: ah
14:06
so it works now?
14:07
!RDP-localdev
14:07
<ltspbot>
Gadi: "RDP-localdev" is when I want to use local devices in an LTSP rdesktop session. See: https://help.ubuntu.com/community/UbuntuLTSP/RdesktopLocaldev
14:07
<vagrantc>
!rdp
14:07
<ltspbot>
vagrantc: Error: "rdp" is not a valid command.
14:07
<vagrantc>
Gadi: why not just link to it from rdp ? :)
14:08
<Gadi>
well, thats a bit broad
14:08
:)
14:08
<vagrantc>
sure, but you can have multiple topics under it
14:08
<Gadi>
oh -
14:08* Gadi knows not the ltspbot-foo
14:08
<vagrantc>
though once it gets long, it stops displaying all of them
14:08
!localdev
14:08
<ltspbot>
vagrantc: "localdev" is (#1) Information regarding using pluggable devices and cdroms with LTSP can be found at http://wiki.ltsp.org/twiki/bin/view/Ltsp/LTSP-42-LocalDev, or (#2) for ubuntu: https://wiki.ubuntu.com/EnableLTSP5LocalDevices, or (#3) for debian-edu: http://wiki.debian.org/DebianEdu/HowTo/LocalDeviceLtspfs, or (#4) http://wiki.ubuntu.com/DebugLocalDev
14:09
<Gadi>
gack
14:09
<vagrantc>
for example :)
14:09
<Gadi>
why is LTSP 4.2 in that list?
14:11
<vagrantc>
because it was the first
14:12artista_frustrad has quit IRC
14:12
<adrian15>
Gadi: That was it. Thank you everyone for your help.
14:12* vagrantc helped unload a pallet of debian-branded wine
14:13artista_frustrad has joined #ltsp
14:13
<Gadi>
wow - our wiki has so much old and confusing info on it
14:13* vagrantc long ago proposed renaming it to oldwiki.ltsp.org
14:13
<Gadi>
hehe
14:14
not that that helps when you get there by google
14:14
:)
14:15
wow - september of 2006 was when I wiki'd about RDP_Localdev in ltsp 4.2
14:16
time flies
14:18
<vagrantc>
hrm.
14:19
maybe the delayed_mounter was actually
14:19
need
14:19
ed
14:19highvoltage has quit IRC
14:20
<vagrantc>
hrm.
14:23adrian15 has quit IRC
14:24
<vagrantc>
well, i've got all week to sort it out.
14:26The_Code has quit IRC
14:27
<Gadi>
you took out delayed_mounter?
14:27
<vagrantc>
for xinitrc.d
14:28
the ldm delayed_mounter was still there ...
14:28
<Gadi>
hmm... you may not catch floppies then
14:28
right?
14:28
<vagrantc>
i had taken out the xinitrc.d/I10-delayedmounter manually before and it still worked... but i might have missed something.
14:28
<Gadi>
oh, but I guess udev should add that
14:28
hmm...
14:29
<vagrantc>
problem is when both delayed_mounter and stuff from udev tries to add the same thing.
14:29
<Gadi>
ah
14:29
well, delayed_mounter is only needed in ldm because the connection isn't plumbed, so I guess it is not needed in xinitrc.d
14:30
as long as udev behaves itself
14:31
<vagrantc>
it worked when i tested it, but now that i uploaded a package, it's not working :(
14:31
so i must have tested poorly
14:31
<Gadi>
of course
14:31
no, its just Murphy
14:32* vagrantc heads to dinner
14:33
<Gadi>
dinner?
14:33
isn't it about noon by you?
14:33
<vagrantc>
9:32 in the evenin
14:33* vagrantc flew over gadi's house a couple days ago
14:33* Gadi guesses ur not in kansas or portland anymore
14:34
<vagrantc>
viva extremadura!
14:43vagrantc has quit IRC
14:43vagrantc has joined #ltsp
14:44hanthana has quit IRC
14:46bieb has left #ltsp
14:47hanthana has joined #ltsp
15:31pmatulis has quit IRC
15:32RobertBerger1 has left #ltsp
15:36The_Code has joined #ltsp
15:42hanthana has quit IRC
16:01makghosh has joined #ltsp
16:03vagrantc_ has joined #ltsp
16:06CAN-o-SPAM has quit IRC
16:10vagrantc has quit IRC
16:17johnny has joined #ltsp
16:32pmatulis has joined #ltsp
16:50Gadi has left #ltsp
16:51staffencasa has quit IRC
16:53alkisg has quit IRC
16:57vagrantc_ has quit IRC
16:58tarbo_ has joined #ltsp
17:02fotanus has quit IRC
17:03makghosh has quit IRC
17:03tarbo has quit IRC
17:05makghosh has joined #ltsp
17:11vagrantc has joined #ltsp
17:16spectra has quit IRC
17:16* vagrantc will be tagging ltsp-trunk shortly
17:23makghosh has quit IRC
17:28makghosh has joined #ltsp
17:37vvinet has quit IRC
18:03dan_young1 has joined #ltsp
18:05dan_young has quit IRC
18:23
<Ahmuck-Jr_>
vagrantc: was in kansas?
18:23
<vagrantc>
no
18:24dan_young1 has quit IRC
18:45yusuf has joined #ltsp
19:29vagrantc has quit IRC
19:33Ahmuck has quit IRC
19:51sene has quit IRC
19:55yusuf has quit IRC
20:03makghosh has quit IRC
20:17knipwim_ has quit IRC
20:32pmatulis has quit IRC
21:37knipwim has joined #ltsp
21:47Sarten-X has quit IRC
21:51hanthana has joined #ltsp
23:47alkisg has joined #ltsp
23:52johnny has left #ltsp