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


Channel log from 22 March 2008   (all times are UTC)

00:01vagrantc has quit IRC
00:04Faithful has quit IRC
00:05Faithful has joined #ltsp
00:08PMantis has left #ltsp
01:55subsume has quit IRC
02:02captain_magnus has quit IRC
02:07captain_magnus has joined #ltsp
03:48harry_ has joined #ltsp
03:49
<harry_>
hi, I'm migrating an office to LTSP 5 / Ubuntu Gutsy (7.10) but having a bit of problem, could someone please help. The LTSP server is setup, I tried connecting from my laptop and it's ok. But when it's on my friend's office, none of their computer will show the GUI. The GUI seemed to be about to come up at first, but then all went back to the console. No /var/log/ldm.log, so I'm at loss. Any idea where else I can look for clues ?
03:53
<johnny>
give it a lil while
03:53
it's hard to find good answers here on the weekend
03:54
<harry_>
thanks johnny :)
03:55
I've had this problem before actually
03:55
but back then I had some time
03:55
<johnny>
you said you had s3 card... aren't they old?
03:55
<harry_>
so I just reinstalled everything into ubuntu dapper (6.06) + LTSP 4.2
03:55
yes
03:55
<johnny>
i only know ltsp5
03:55
<harry_>
but X detected them just fine
03:55
<johnny>
ltsp4 is not recommended
03:55
<harry_>
I know
03:55
<johnny>
you could try to lower the color depth
03:55
<harry_>
but it seems to be more friendly to these old computers
03:56
<johnny>
well you can also turn on LDM_DIRECTX in ltsp5
03:56
in lts.conf
03:56
<harry_>
yup, I've tried X_COLOR_DEPTH = 8
03:56
read ubuntu wiki entry saying that it should be 24, so tried that
03:56
<johnny>
which should bring you a bit closer
03:56
<harry_>
but still no joy
03:56
oh ok
03:56
<johnny>
performance wise
03:56
in ltsp5
03:56
<harry_>
wonder what's LDM_DIRECTX is for
03:56
<johnny>
you could try dropping to a console on the thin client and running
03:57
<harry_>
I'm reading the docs about it now
03:57
<johnny>
it acts similiar to ltsp4
03:57
try running ldm from the console in ltsp5
03:57
from a thin client
03:57
ldm vt7 :0
03:58
<harry_>
ok, let's try
03:58
gimme 5 mins
04:03
alright !
04:03
one workstation is showing the gui
04:04
I typed :
04:04
ldm vt7 :5
04:04
<johnny>
but it didn't on boot?
04:05
<harry_>
no
04:05
wonder why
04:05
but this is a good start
04:05
ok, I'm trying it on other workstations
04:05
gimme 10 mins
04:17
ok, it shows GUI, but after I entered user/pass, it stuck with message "Verifying password"
04:17
after a while, it bombed out to console again
04:17
it's having problems authenticating the user?
04:18
noticed something funny, the client have hostnames of ltsp
04:19
when I "ping ltsp", ltsp resolved to 127.0.0.2
04:22
<johnny>
hmm... that's why i used dnsmasq, to hand out hostnames with the dhcp
04:23
oh.. silly.. but they all use the server too
04:23
<harry_>
alright, never stumbled on that on the docs before
04:23
<johnny>
since all the accounts on the server
04:23
sorry. i woke up after not enough sleep
04:24
and thinking about going back to catch the last 4 hours
04:27
<harry_>
can you share me your dnsmasq config
04:27
found an article http://www.quixotic.org.uk/Articles/LTSP.shtml
04:27
but is it really that simple ? :)
04:28
<johnny>
you don't need to use dnsmasq.. that's just what i chose
04:28
ubuntu chooses to use dhcpd
04:29
but they don't do dns.. i wanted local dns anyways
04:29
you can set it up with bind.. but no other ltsp devs are going to support the usage of dnsmasq until a distro picks it up as the one they choose
04:29
altho we're going to choose it for gentoo..
04:30
<harry_>
dang
04:30
<johnny>
check your logs on the server when you get "Verifying Password"
04:30
on that client
04:30
that account does exist on the server right?
04:30
<harry_>
suddenly LTSP 4.2's manual hostname assignment looks much simpler, he he
04:30
yup
04:30
here it is
04:30
<johnny>
manual hostname assignment ?
04:30
<harry_>
saw sentinel logged in succesfully
04:30
established ssh session
04:30
process returned no status
04:30
executing rc files
04:30
beginning x session
04:30
xsession ended
04:30
executing rc files
04:30
killing x server
04:30
ending ssh session
04:30
expect saw : exit
04:30
that's from ldm.log
04:31
<johnny>
check the /var/log/auth or /var/log/messages
04:32
also.. can you ssh to your server as that user?
04:32
<harry_>
absolutely nothing in those files
04:32
yup
04:32
<johnny>
there should be something in /var/log/messages.. perhaps not with your problem..
04:32
<harry_>
got home dir as well
04:32
<johnny>
but it should be there
04:32
<harry_>
that's what baffles me
04:32
so little info in this case
04:32
<johnny>
what does ~/.xsession-errors of that user say
04:33
if it exists
04:33
oh.. and only in the case that you didn't manually log into the server account as that user via gui
04:33
<harry_>
it does, and it looks interesting actually
04:33
hold on
04:33
<johnny>
don't paste it here
04:33
use the pastebot
04:33
!pastebot
04:33
<ltspbot>
johnny: "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.
04:34
<harry_>
oh sorry, yes the log was from my last GUI session from my laptop
04:34
<johnny>
delete it
04:34
and try again
04:34
<harry_>
will do
04:34
hold on
04:35
!pastebot
04:35
<ltspbot>
harry_: "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.
04:36
<harry_>
hmm, nothing on server:/home/user/.xsession-errors
04:38
btw I have set LDM_DIRECTX=true
04:38
still the same
04:40* harry_ scratching his head
04:47
<johnny>
no.. that won't fix your issue
04:48
it was seperate suggestion
04:48
for giving you something closer to the performance of ltsp4
04:48
it doesn't tunnel X, which is the default for ltsp5 implmentations so far
04:48
hmm.. here's an idea
04:49
do you have a cd drive in that machine?
04:49
can it boot an ubuntu live cd?
04:49
if so.. see if it gives you graphics, then copy the /etc/X11/xorg.conf file to your server machine
04:49
and then in your lts.conf
04:49
do
04:50
[Your Machine]
04:50
X_CONF = /path/to/config/on/server
04:50
if that works, then you might want to try setting some specifics
04:50
like forcing the x server
04:50
i know you already tried X_COLOR_DEPTH, but there are other useful ones ib et
04:53
<harry_>
you know, it might work
04:53
let's see if my friend got a live cd around
04:53
brb
04:58
good news, 1 workstation is logged in, no problem. But I had to set SCREEN_07=startx
04:58
no localdev :)
04:58
I'm trying sdm now
04:58
hold on
04:59
<johnny>
you have to use ldm to get localdev
05:00
sdm will be dropped
05:00
<harry_>
alright
05:00
yeah, it's not installed as well
05:00
so it's either startx or ldm for now
05:00
thanks
05:00
thank you johnny, this is good progress
05:00
<johnny>
ldm should be installed?
05:00
<harry_>
sorry, sdm is not currently installed by default by gutsy
05:02
my friend is very happy now :)
05:02
OK, I've got to go now, got another appointment
05:02
I'll try ldm again later
05:19Pascal_1 has joined #ltsp
05:26
<harry_>
all workstations are booting up :)
05:27
thanks you
05:27
thanks johnny
05:27
I'm going now
05:27
cheers
05:28harry_ has quit IRC
05:38indradg_ has joined #ltsp
05:43mikkel has joined #ltsp
05:56indradg__ has quit IRC
05:57Pascal_1 has quit IRC
07:08J45p3r has joined #ltsp
07:56K_O-Gnom has joined #ltsp
08:05jhp has quit IRC
08:05jhp has joined #ltsp
08:13gvy_ has joined #ltsp
08:13
<gvy_>
morning
08:49vagrantc has joined #ltsp
08:54joebake1 has quit IRC
08:58indradg__ has joined #ltsp
09:17indradg_ has quit IRC
09:18subsume_ has joined #ltsp
09:18subsume_ is now known as subsume
09:19cliebow has joined #ltsp
09:19subsume has left #ltsp
09:21
<cliebow>
any around can answer an xterm question..in redhat ltsp terminal i have a student who claims he has a 'method" to gain an xterm even though /usr/bin/xterm is 700..besides gnome-terminal what other xterm would be available?
09:23
<mnemoc>
xterm binary downloaded from the net?
09:24
<cliebow>
could be i supposee//nothing in his homedir....
09:24
<laga>
just ask him how he did it? :)
09:25
<cliebow>
i supose ill have to...nautilus is locked down so he cant go up directories in nautilus
09:25
<mnemoc>
there are _many_ terminal apps, rxvt rxvt-unicode, Terminal, mrxvt, ....
09:26
<cliebow>
this is icewm on top of nautilus..
09:27
my group has alienated a lot of high tech kids with security so they are all trying to give us grief
09:27J45p3r has left #ltsp
09:27
<cliebow>
i shut off gnome-terminal..mayb that is what he was using
09:28
cause durned if i can find a way...
09:31
<mnemoc>
imo it's pretty silly to "secure" the computer by trying to block xterm
10:02
<warren>
cliebow, likely he downloaded a terminal app using a web browser or some other network client, then used the web browser to launch it.
10:02
<cliebow>
no..looks like he wrote his own in c
10:02
<warren>
cliebow, and yes, it is silly to secure it by blocking the terminal
10:02
unlikely
10:03
cliebow, you can mount /home and /tmp in such a way that you cannot execute anything from it
10:03
<cliebow>
maybe that is the way to fly...
10:04
<vagrantc>
cliebow: debian or ubuntu system?
10:04
<cliebow>
no rhel4
10:04
nfs mounted homedir
10:04
<vagrantc>
ah.
10:04
i'd know how to find all the likely xterm-like programs on debian ... using update-alternatives
10:05
<warren>
cliebow, anyway, preventing them from using a terminal is not "security"
10:05
<cliebow>
it does look like his own code...
10:05
well no
10:05
<vagrantc>
well, giving them access to the terminal certainly makes it considerably easier to compromise the system ...
10:06
<laga>
heh
10:06
<cliebow>
i was not contemplating anything but the most spurious dos stuff
10:06
<laga>
if they want to compromise your system, then the lack of a terminal won't stop them ;)
10:07
<cliebow>
true...
10:07
<vagrantc>
it won't stop a dedicated attacker, but it can make it harder for a script-kiddie.
10:07
<cliebow>
scriptkiddie stuff
10:07
<warren>
If you want to DOS a system, just hold down the print screen button in GNOME
10:07
<cliebow>
on a terminal?
10:08
i suppose..
10:09mistik1 has quit IRC
10:09
<cliebow>
wish J45p3r were here to throw in his 2 cents worth..
10:10
gotts read this kids code..see if i can figure out what it is really doing
10:13xcasex has quit IRC
10:13xcasex has joined #ltsp
10:17
<warren>
cliebow, most likely he downloaded code from the internet
10:18
<cliebow>
yeah...smart kid..trying only to put the eth in ethical for him
10:18makghosh has joined #ltsp
10:18petre has joined #ltsp
10:19
<petre>
morning all
10:19
<cliebow>
hey petre
10:19
<petre>
hey cliebow! long time no chat
10:19indradg__ has quit IRC
10:19
<petre>
(my fault)
10:19
<cliebow>
heh!
10:20
<petre>
I was on #ltsp last night for the first time in months
10:20indradg__ has joined #ltsp
10:20
<petre>
been too busy otherwise: new jobs, including teaching linux admin
10:20
what's new in Maine?
10:21
<cliebow>
not a lot..just trying to age gracefully
10:22
<petre>
are you generally running LTSP on Debian at your schools?
10:22
<cliebow>
umm.still in rhel4 at the high school..
10:23
<petre>
ltsp 4.2?
10:23
<cliebow>
otherwise ubuntu or slack..
10:23
still 4.2..my pos terminals are too downwardly mobile
10:24* cliebow cliebowsteps out to get mail
10:24
<petre>
yeah, but that's the beauty of ltsp: clients can be really old, just use previous release of ltsp
10:30
<warren>
point of sale terminals?
10:30
=)
10:32
hah... Headline: Pony Express strike enters day 54,000
10:33
<cliebow>
piece of doodoo
10:33mistik1 has joined #ltsp
10:34
<warren>
http://www.amazon.com/exec/obidos/ASIN/0470050101/thebigpictu09-20 "True Enough: Learning to Live in a Post-Fact Society (Hardcover)"
10:34
<cliebow>
here is his .h..
10:34
<ltsppbot>
"cliebow" pasted "#include <FL/Fl.H> #include <F" (36 lines) at http://pastebot.ltsp.org/482
10:35
<warren>
the title alone makes me laugh
10:35
<cliebow>
hah
10:37
petre:ive never gotten a 2.6 kernel to run on these things..thoughi have one that loads but never actually does anything
10:39
<warren>
cliebow, how much RAM? what processor?
10:40
what kind of video?
10:43
<gvy_>
cliebow, well you might be better off by handing the kid a "real challenge"
10:43
it might be hard to come up with decent one though
10:46
cliebow, well you can try *** linux :)
10:46
or ****** linux
10:46
or open****
10:47
to at least figure out the geneicness of the problem wrt different 2.6.x kernels
10:48
<petre>
cliebow, what kind of hardware are you using for those clients?
10:50
<cliebow>
netvista..
10:50* cliebow cliebow ducks
10:50
<petre>
ah, "specialized" ;-)
10:51
<cliebow>
gvy_:yeah i was thinking that he could use a project
10:51
wierd uncompressed kernel./.and a tech lead intent on moving everything to apple
10:51
<gvy_>
cliebow, which netvista? i remember a bunch of them
10:52
petre, not specialized, a soho desktops iirc
10:52
<cliebow>
geode 8363 if memory serves
10:52
<gvy_>
i remember via/duron one :)
10:52
ouch
10:52
<petre>
from ibm?
10:53
<cliebow>
been showing off th term150..but state of budget in Maine puts a damper on all spending
10:53
yeah
10:53
<petre>
don't get me started on state budgets for schools
10:54
<cliebow>
donated clients from maine state library
10:54
<petre>
my wife just got elected to the school board, knowing that we're facing a $1m budget cut
10:54
<cliebow>
oh yeah..were w00 million short in Maine..
10:55
state share of spec ed spending down to 50 % from 85\\
10:55
<gvy_>
anyone on schoolforge-discuss@?
10:55
same topics from time to time
10:55
<cliebow>
i am sure..
10:56
<petre>
spec ed is expensive, costs $40k per student but state funding for them is same as all other kids
10:56
yeah, I follow schoolforge list
10:56
<cliebow>
id thought of puttin ibooks in for terminals..but they are dropping like flies these days
10:57
video issues..mobo issues
10:57
<petre>
hardware *can* wear out, basically bake itself to death
10:58
<cliebow>
g3 has an inherent flaw in video chip soldering..
10:58
<petre>
otoh, finding "new" donated equipment shouldn't be too hard, I'd think
10:58
the local tech colleges here in Minneapolis have a warehouse full of old Dell PCs
10:58
<cliebow>
so you're supposed to heat the chip til the solder melts again
10:59
<laga>
cliebow: sounds like something you'd do to a xbox 360
10:59
<cliebow>
he hee
11:00
Maine gave every high school teacher in the state a new ibook..then found themselves 220m in the red
11:00
<petre>
they'd love to give 'em away to schools but the schools don't want them because they still think in fat client terms
11:01
grrr! laptops in schools are an awful idea
11:01
same category as buying every student a cadillac to get to school in
11:01
waste of money
11:02
<cliebow>
i'm pretty discouraged anyway....but at 57 dont seem to be the ideal candidate for anything else
11:03
<petre>
for the cost of one--One!--desktop machine, they could outfit an entire classroom with an ltsp server and enough clients for 1:1 student to computer ratio
11:03
Put that in every class, and you don't need laptops
11:03
<cliebow>
they all want wreless
11:03
<petre>
hell of a lot cheaper short- and long-term
11:03
<cliebow>
agreed
11:03
plus management..
11:05
<petre>
didn't bristol wireless have a proposal to have someone do the code for wireless thins?
11:05
or am I confused and it was for pcmcia + wired card booting?
11:05* petre 's memory is fuzzy
11:06
<petre>
I equate laptops to giving each kid his own folding chair, so he can carry it with him from class to class
11:06
<cliebow>
not sure..
11:06
<petre>
thus saving all the expense of putting chairs in the classroom
11:07
<cliebow>
heh!
11:08
<gvy_>
petre, well in russian school pilot it seems they've grokked the idea of fitting a class for a price of ~2 systems
11:08
or a few
11:08
<laga>
(i love my laptop for university. :))
11:08
<gvy_>
plus monitors
11:08
<petre>
that way, they can have a laptop/chair at those crucial times such as going to the lunchroom, on the bus, going from one room to another, etc.
11:08
<cliebow>
i Do like my lappy..even though ti is possessed
11:08
<laga>
(of course, it wasn't given to me by my school..)
11:09
<petre>
laga, I think laptops make sense for adults (I'm typing on one) but school kids don't need them
11:09
<laga>
yeah
11:09* gvy_ . o O ( fgrep petre > ~/doc/Cookies.txt )
11:09
<petre>
if they can have access to a computer in every classroom and their files go with them everywhere
11:09
<cliebow>
you woudnt believe how many broken screens ive replae for middle school..
11:10mistik1 has quit IRC
11:10
<laga>
i wonder why school kids need computers in every class room.. call me old-fashioned, but i prefer the old pen&paper approach for most things
11:10
<petre>
in the lappys?
11:10
<cliebow>
yes
11:10
<petre>
pen & paper *are* important
11:11
<gvy_>
laga, +1
11:11
<petre>
computers in the classroom provide immediate access to info,
11:11
<gvy_>
and chalk
11:11
<laga>
of course, our computer lab at school sucked ;)
11:11
petre: sure.. how would a 1:1 ratio be useful?
11:12
<petre>
don't have to plan to take the kids down to the lab, get them ready, work it into the schedule with all the other classes who want to use the computers
11:12
writing assignments, for example; everyone can work on them
11:13
pencil & paper is fine, but at, say age 10, why force a kid to re-write by hand a whole paragraph because he wants to make 4 or 5 small edits but which won't fit onto the paper?
11:14
should they still learn to write by hand? yes
11:15
<laga>
haha, of course :)
11:16
<petre>
the point is that for the same amount of money you can give kids *access* to computers either 30 minutes a week (lab model) or hours every day
11:16
<laga>
yeah, that's true
11:16
<petre>
going the lab route is like not having books in the classroom, keeping them only in the library
11:16
they just won't read as much
11:17
and they can do a lot of reading on the computer, which can act as a carrot
11:17
<gvy_>
petre, re "by hand": http://www.paulgraham.com/nerds.html
11:18
iow depends on what's the perceived purpose of the school
11:18
<petre>
I just read someone's blog or some such, about observing kids in the class, in which the author speculated that after a bit the novelty of doing stuff on the computers would wear off for the kids
11:18
<gvy_>
as usually with any novelty... :)
11:18
<petre>
but it didn't!
11:18
The kids always wanted to go do stuff on the computers, it didn't wane
11:19
<laga>
petre: what about supervision? it's very easy to get distracted when using a computer
11:19
<petre>
I don't know the psychology or neurology behind it, but I suspect something about the bright colors and glowing screen have something to do with it
11:20
distracted by what? actually, people can get immersed quite deeply, although sometimes in the wrong things ;-)
11:20
<gvy_>
and not having to get something done probably too :(
11:21
<petre>
but that's the beauty of k12ltsp, teacher-tool, which lets the teacher monitor what the kids are doing
11:21
or edubuntu's tool--which I can't remember the name of at the moment
11:23
<cliebow>
client-manager...something like that
11:23
<petre>
Linux's two big advantages over Mac & Win is that it's free, no matter how many places you use it, and it can be deployed in such a way that it drives the cost of computing down so low it's not worth limiting it to things like labs
11:24
<cliebow>
i must say apple has some things going fro it
11:24
<petre>
it's like electricity: we take it for granted that it's always there, which allows our society to do all sorts of things that would not be possible if the power were on only, say 12 hours per day
11:25
linux allows ubiquitos computing that Mac and Win will never be able to match
11:25
yeah, technically macs are rather interesting, but so are Rolls Royces
11:25
<cliebow>
true..
11:26
the abuse the ibooks have stood up to is incredible
11:26
<petre>
right, but would it be better to have something that can stand up to a lot of abuse or something so cheap that replacing it even a few times per year is trivial
11:27
think about the change the printing press made
11:27
<cliebow>
i am all for the terminals...
11:27
<petre>
books & paper were rare, expensive, etc.
11:28
<cliebow>
but it is all about choice too.
11:28
<petre>
linux makes that scarcity go away
11:28* gvy_ actually starts logging fortunes
11:29
<petre>
right, but choices that put the state $200mil in the red don't seem like a good idea
11:29
<cliebow>
boy we need a lug around here..closest active one is 2 hours away..
11:29* petre has to go pick his car; bbiab
11:29
<cliebow>
and i hae too many irons in the fire to get one going
11:30
<gvy_>
cliebow, lugs seem falling apart -- linux is way less club-thing and more commodity these days
11:30* gvy_ maintains lug.org.ua and kiev lug site, sort of
11:30
<cliebow>
i guess..i just need more input...face to face..
11:31
ukraine?
11:31
<gvy_>
yup
11:31
<cliebow>
know where stryj is?
11:31
<gvy_>
more or less
11:31
western part :)
11:31
<cliebow>
my dad was born there..
11:31
<gvy_>
wow.
11:32
world is a small place
11:32
<cliebow>
someday id like to come visit..
11:32
<gvy_>
cliebow, well you can take my phone just in case you need another local contact
11:32
i'd be very glad to see you
11:33
<laga>
gvy_: very nice article (http://www.paulgraham.com/nerds.html)
11:33
<gvy_>
btw we're doing annual conferences in october
11:33
<cliebow>
illlookyou up if i ever see a chance..
11:33gvy has quit IRC
11:33gvy_ is now known as gvy
11:34gvy_ has joined #ltsp
11:39
<laga>
gvy: ok, it gets a bit paranoid at times, so i'll call it "insightfull" ;)
11:39
<gvy>
:)
12:03
<steph_>
Is there someone here who use ltsp with virtualbox ?
12:03vagrantc has quit IRC
12:37Jbassett has joined #ltsp
12:37
<Jbassett>
god evening
12:38
I am having an issue after reinstalling Linux/LTSP after a HD failure
12:38
Server boots and has local x windows with Kde
12:38
remote terminal boots and attempts to startx
12:39
at which point it says xserver failed
12:39
even when trying just the VESA driver
12:39
xserver = vesa
12:39
I thought vesa driver worked pretty much no matter what
12:39
any ideas?
12:42
<gvy>
could you try to e.g. set root password inside ltsp chroot (chroot /where/it/is and run passwd)
12:42
and run xinit by hand?
12:43
don't remember whether there was useful output above "xserver failed" msg...
12:43
<Jbassett>
from the remote clients 'terminal'?
12:43
<gvy>
yup
12:44
nb: i have either *old* or specific ltsp experience, might not serve you well
12:44
<Jbassett>
ltsp 4.1
12:44
no output after failed msg
12:44petre has quit IRC
12:45
<Jbassett>
on client: passwd: command not found
12:45
<gvy>
above, not after
12:45
4.1... ugh
12:45
<Jbassett>
was working fine before reinstall
12:45
<gvy>
can you get local shell at the terminal, anyways?
12:45
<Jbassett>
yes
12:46
just no x
12:46
<gvy>
what happens if you run "X" command?
12:46makghosh is now known as makghosh|afk
12:46
<Jbassett>
x: command not found
12:46
X: command not found
12:46
startx:
12:46
xserver failed
12:47
fatal server error shown above: cannot open log file "/var/log/Xorg.0.0.log"
12:47
<gvy>
Xorg
12:47
ah.
12:47
seems like any overlay that was used there to make it r/w wasn't done
12:47* gvy trots to the office, logs into VE with ltsp4 and tries to recall how things were done then
12:49
<Jbassett>
bbs, dinner is ready
12:49* Jbassett away eating
12:49
<gvy>
bon apetit
12:53
<johnny>
yhi
12:56chupacabra has quit IRC
12:56chupacabra has joined #ltsp
13:02rcy` has joined #ltsp
13:04makghosh|afk is now known as makgosh
13:04makgosh is now known as makghosh
13:09* Jbassett returns
13:40rcy` has quit IRC
13:55petre has joined #ltsp
14:09makghosh has quit IRC
14:17petre_ has joined #ltsp
14:17petre has quit IRC
14:17petre_ has quit IRC
14:29slipttees has joined #ltsp
14:30rjune has joined #ltsp
14:33slipttees_ has joined #ltsp
14:33chup has joined #ltsp
14:34chupacabra has quit IRC
14:36slipttees has quit IRC
14:36slipttees_ is now known as slipttees
14:47gvy has quit IRC
15:15mistik1 has joined #ltsp
15:17elisboa has quit IRC
15:18ballongen has joined #ltsp
15:19Shingoshi has quit IRC
15:20Shingoshi has joined #ltsp
15:32
<johnny>
who here uses the kernel level ip configuration for ltsp?
15:38TelnetManta has joined #ltsp
15:44
<ballongen>
where is the logs for ltsp?
15:44
some times i get the busybox, sometimes i can login in gdm, get a black screen, but then returns directly to the loginscreen again
15:46
<johnny>
gdm?
15:47
check your server logs first...
15:49abadger1999 has quit IRC
15:55slipttees has quit IRC
16:09mistik1_ has joined #ltsp
16:13mistik1 has quit IRC
16:13mistik1_ is now known as mistik1
16:16ballongen has quit IRC
16:21
<warren>
hah
16:21
it turns out I was really close to making xauth work on Friday afternoon
16:21
just a typo
16:24
<johnny>
hi warren
16:24
sup?
16:25
<warren>
working on proper xauth support for ldm
16:26
so we can close the security hole
16:26
<johnny>
sweetz
16:26
while still using LDM_DIRECTX?
16:26
<warren>
yes
16:26
<johnny>
go warren :)
16:26
so, in your initrd or whatever, do you use the kernel level ip configuration too?
16:27
<xcasex>
go warren! *echo*
16:27
<jammcq>
warren: what's the target release date for fedora-9 ?
16:27
<johnny>
i see IP-Config in the bootup of ubuntu
16:27
<warren>
jammcq, late April I think
16:27
<jammcq>
will there be a beta before then?
16:27
<warren>
jammcq, beta is out on Tuesday
16:27
<jammcq>
cool, including LTSP-5 ?
16:28
<warren>
yes the basics at least
16:28
a few parts are broken
16:28
because I need to adapt some parts from Ubuntu
16:28
<jammcq>
I'm heading down to Brazil on april 15th, and I want to show off what you've done
16:28
<warren>
ask me before then and I'll point out how to set it up
16:29
<jammcq>
k
16:29
that would be great
16:29
<johnny>
can you keep me in the know on the changes warren ?
16:29
<jammcq>
I'm gonna attempt to pplay with the stuff the Opensuse guys have done too
16:29
<johnny>
being that i'll run into similiar issues
16:29
<warren>
johnny, everything I do goes into ltsp-trunk
16:41alkis1 has joined #ltsp
16:43
<warren>
ok, one typo and one thinko
16:43
<johnny>
put your cap on
16:44
<warren>
battery dying, bbl
16:48
<alkis1>
Hi, has anyone managed to correctly install italc so that the ltsp clients can be shown/managed? Ubuntu hardy/Ltsp 5.0, it currently reports "no user logged in" for the clients...
16:48
<johnny>
that's beta
16:48
give it awhile
16:49
nobody here uses hardy afaik
16:49
<alkis1>
I have a presentation Monday afternoon, I'd really like to show it off!
16:49
<johnny>
ask in #edubuntu perhaps
16:49
<alkis1>
OK, I'll try that. Thanks!
16:49
<johnny>
they were talking about it there
16:49
nobody has ever talked about it here
16:51
<alkis1>
Oh, may I ask another question? In my laptop, when I change the screen brightness, a window pops up to show the brightness. The problem is that I have the laptop as ltsp server, and all the clients also see the same popup window!
16:52
But in hardy again, so it may be a beta thing...
16:54
<stgraber>
why are the clients running gnome-power-manager ?
16:55
btw, what's your problem with italc ? I'm the one who packaged it for Hardy :)
16:55
<alkis1>
Hi, I know that, forget about the power management! italc is my priority!
16:56
In hardy, I installed italc with synaptic.
16:56
I add a client, but it only reports "No user logged in"!
16:56
<johnny>
why would you ask about your laptop here ..
16:57
<alkis1>
johnny, it's ltsp related, I see the brightness window in the ltsp clients
16:57
<johnny>
hmm.. odd
16:57
<stgraber>
ok, so that's italc in a LTSP environement and release 1.0.7 ?
16:57
<alkis1>
stgraber, yes
16:57
<stgraber>
how did you add your client ?
16:58
<alkis1>
right click > add classroom and then right click > add client and I added the client ip
16:58
<stgraber>
ok, so here is your problem
16:58
ica doesn't run on the client but on the server
16:58
<alkis1>
What should I do? How to declare the clients?
16:58
<stgraber>
so all clients are listening to 127.0.0.1 but on different port
16:59
just close italc and start it with : run-italc
16:59
and it should generate a .xml file with all your clients and the right ports
16:59
hmm, italc-launcher
16:59
we changed the name recently :)
17:00
<alkis1>
should I remove any configuration files first?
17:00
<stgraber>
no
17:01
it'll make a backup of your config, then generate a new one and restore your previous one when you close italc
17:01
<alkis1>
ok, just let me reboot the client because it hanged before, I don't know why
17:01
<stgraber>
ok
17:02
<alkis1>
btw, you're a life saver! I use it every day at schools with windows installations, and installed hardy only because of italc!
17:02
(on another lab with edubuntu)
17:03
<cliebow>
alkis1, Good Man!!!
17:03
<stgraber>
hehe, we switched to iTalc in edubuntu because our previous tool wasn't maintained anymore and thought that having an upstream would be a good idea rather than developing our own :)
17:04
it seems to have been a good idea as iTalc is from my point of view better integrated than tcm (our previous tool)
17:04
and Tobias (italc's upstream) is very reactive and is working on some cool feature (LDAP and avahi support for client list being my favourite one)
17:04
<alkis1>
much better, I didn't like the old one. Too much manual configuration for simple teachers... And far lesser options (to control the students) :)
17:05
<stgraber>
so with Intrepid we should be able to have classroom generated on the fly include thin clients, fat clients, Windows computers and Linux workstations
17:05
that'll just rock
17:06
<alkis1>
the client booted and upon logging in it reports: the isd-server could not be started because port 11225 is already in use
17:06
<stgraber>
hmm, that's because when your client crashed the processus left haven't be killed
17:07
you can run that in a terminal on the client : killall -9 ica ; ica-launcher
17:07
that should start a clean instance of ica
17:08
<alkis1>
now I get an error on the server for port 5800. Will you wait for me to reset both server and client? Don't go away please!!! :)
17:08
<stgraber>
I guess ica isn't running on your server
17:09
that's probably because you didn't close and reopen you session after installing it
17:09
<alkis1>
no, i installed it in the morning and done 5-6 restarts since then
17:10
message on the server: The ISD-server could not be started because port 5800 is already in use. Please make sure that no other application is using this port and try again
17:10
I'm rebooting and I'll come back in 2'
17:10alkis1 has left #ltsp
17:12alkis1 has joined #ltsp
17:13K_O-Gnom has quit IRC
17:13
<stgraber>
alkis1: I think I know what your problem is, you are using your server as a workstation right ?
17:13
<alkis1>
(waiting for the client to boot).
17:14
Yes, i sit on the server, but I also have a pc for a workstation
17:14
<stgraber>
alkis1: the scripts have been done for an environement where the server is only a server and not a workstation, so italc doesn't work correctly if you are on the server as it doesn't know it's part of a LTSP setup ($LTSP_CLIENT isn't set)
17:15
though, I can tell you how to connect to your client if that's just for demo purpose. You'll need some different scripts to have it working with a setup like yours if you don't move to a "large scale" LTSP setup
17:15
<alkis1>
No problem there, I have (on the lab) a free pc where I can sit. But can I do something to test it here at home, where I only have the laptop and a pc?
17:16
<stgraber>
ok, so on the server, just open italc
17:16
then create a classroom and add a computer
17:16
choose the name you want for it
17:16
<alkis1>
Is this going to be the same in hardy final? If so, I'd like to sit on my laptop, so I'd like to use modified scripts...
17:17
ok, done
17:18
how should I declare its IP? 127.0.0.1 and port number?
17:18
<stgraber>
do you know the IP of your client ?
17:19
<alkis1>
192.168.0.225
17:19
<stgraber>
ok, so that'll be : 127.0.0.1:10225
17:20
<alkis1>
Yeah, it's working!!! Danke!!!
17:21
<stgraber>
np :)
17:21
<alkis1>
How stable is it? I have a presentation for other secondary school teachers, and I'd like to use edubuntu/ltsp/italc for it,
17:21
so that they become interested and use it too, but if it isn't stable enough, perhaps I should use windows xp/italc?
17:22
<stgraber>
it's stable, at least as stable as it's on Windows
17:22
<alkis1>
perfect!
17:22
And something about client discovery:
17:22
<stgraber>
the only problem is the demo mode but 1.0.7 greatly improved that point and should be a lot more fluent than it previously was
17:23
<alkis1>
I'll have my laptop as a server (with dhcp and all) and 12 clients
17:23
What is the easier way to declare the clients?
17:23
<stgraber>
to wait 30s and you will get the modified script :)
17:23
<alkis1>
I'll have to give you a donation for this!!! :)
17:24
<stgraber>
http://ubuntu.pastebin.com/f52314cb8
17:25
this script will generate a new xml file, add a LTSP classroom and put the clients in it
17:25
<alkis1>
so first I run the script and then italc, right?
17:25
<stgraber>
this script will start italc
17:26
<alkis1>
Magic!
17:26
<stgraber>
so the classroom will be regenerated every-time you start it
17:27
I would have preferred a way of having clients appear/disappear when they are turned on but that'll need to wait Intrepid Ibex :) (and the amount of free time Tobias and I will have during the next 6 months)
17:28
<alkis1>
Sorry, I don't know what intrepid ibex is... :(
17:28
<stgraber>
the release after Hardy Heron, so that'll be 8.10
17:28
<alkis1>
But it is very much usuable as it is. Users can logon/logoff without re-running the script, right?
17:29
<stgraber>
indeed, though sometimes italc has some problem detecting that a user is back but I mainly saw that problem on Windows not sure it also happens on Linux
17:30
that's one of the reason I want a good way of detecting client connection/disconnection, that'll solve 90% of our current problems :)
17:30
<alkis1>
Anyway, it's not much of a problem to close italc and start it again on each class break! So, it really works for me!
17:30
<stgraber>
indeed
17:31
<alkis1>
It took me 1 hour to install italc on windows xp to the lab where the presentation will be... Had I talked to you yesterday, I wouldn't have wasted it! Italc/LTSP rock!
17:32
Thank you very much, I'm going back to testing... Good night from Greece!
17:32
<stgraber>
np, good night from switzerland :)
17:53
<steph_>
Is there someone here who use ltsp with virtualbox ?
17:54alkis1 has left #ltsp
18:04chup has quit IRC
18:04chup has joined #ltsp
18:07alkis1 has joined #ltsp
18:07
<alkis1>
stgraber, are you still here?
18:10
stgraber, demo mode doesn't work, it says "Establishing connection to 192.168.0.1:53415" and remains there... Viewing/controlling the clients does work though!
18:14
<stgraber>
alkis1: you need to set the right IP in the config panel
18:14
in some case it defaults to an incorrect NIC
18:14
in your case selecting lo 127.0.0.1 should solve that
18:16
<alkis1>
I have an eth0:1 alias to enable my laptop to access the internet through the same physical NIC, maybe that's what caused the problem
18:16
How do I "select lo 127.0.0.1" ?
18:17
Should I modify the script?
18:18
Sorry, didn't see the first message, "set the right ip"! :(
18:21
<stgraber>
that's in italc itself
18:22
<alkis1>
Did that, it didn't work: not even the "establishing connection" message
18:23
Then I closed italc and re-run it with the script, now it says "there seems to be no italc service running or the authentication-keys aren't set up properly
18:24
And I can't even run it and go to the control panel anymore!
18:24
will rm -rf ~/.italc help?
18:26
<stgraber>
that means your local ica had a problem
18:26
<alkis1>
alkisg@alkisg-L:~$ italc
18:26
[warning] isdConnection::open(): unable to connect to server on client 127.0.0.1:5800
18:26
<stgraber>
just run : ica-launcher and then italc-launcher
18:27
<alkis1>
ica-launcher is a service? should I run ica-launcher &
18:27
<stgraber>
yes
18:27
<alkis1>
(sorry, not experienced)!
18:29
I ran it with the script you gave me, I can see/control the client, but demo mode doesn't work
18:29
On the server it says "demo mode running", on the client nothing, on the terminal no error message
18:29
On the control panel "lo 127.0.0.1"
18:29
<stgraber>
you don't even see the clients trying to connect to the server for the demo ?
18:29
<alkis1>
No
18:30
The client works as if italc wasn't even installed
18:30
But I haven't rebooted the client for some time, maybe the changes I've done need reboot or logoff?
18:31
(logging of the client...)
18:32J45p3r has joined #ltsp
18:36
<alkis1>
(rebooting the server also...)
18:36alkis1 has left #ltsp
18:39alkis1 has joined #ltsp
18:39abadger1999 has joined #ltsp
18:42alkis1 has quit IRC
18:43alkis1 has joined #ltsp
18:44
<alkis1>
stgraber, thank you very much, it's running OK now. Sorry for my inexperience!
18:47chup has quit IRC
18:47chup has joined #ltsp
18:56alkis1 has left #ltsp
18:58vagrantc has joined #ltsp
19:01
<steph_>
Is there someone here who use ltsp with virtualbox ?
19:01
<johnny>
i do
19:02
<steph_>
Can you just confirm....
19:02* vagrantc uses ltsp + virtualbox
19:02
<steph_>
Here is my setup: 2 NIC on the host. eth0 is connect to the net (DHCP) eth1 is connect to a switch for lan (192.168.0 range). I created a guest who as a dhcp services for lan (192.168.1 range). So adaptor0 -> NAT for internet, adaptor1 -> host interface networking. Is this a good starting point?
19:02
<vagrantc>
i do most development on a virtualbox ltsp server and a virtualbox ltsp client
19:02
<johnny>
nat won't work if you're using virtualbox for ltsp server
19:03
without port forwarding on your host machine that is
19:03
<steph_>
even if eth1 is connected to a switch
19:03
<johnny>
you should either go with "internal network"
19:03
or use the host based one
19:04
i chose to use briding
19:04
bridging*
19:04
<steph_>
Do I use bridging with what they call "host interface networking"
19:05
<johnny>
yes
19:05
but.. then again
19:05
if you're not using your virtualbox vm as the ltsp server
19:05
are you?
19:05
<steph_>
Yes, my guest will be the ltsp server
19:07
<johnny>
i'm running them both in vms
19:07
<steph_>
johnny: If I understand, eth0 -> Nat for www, eth1 -> internal networking, right?
19:07
<johnny>
and your client is?
19:08
<steph_>
are you asking for my distribution?
19:08
<johnny>
no
19:08
is it a vm? or a real machine?
19:09
<steph_>
It's a real machine. Didn't know you can can do that !!
19:09
<johnny>
i run both with vm .. since it's simpler for me than digging up other hardware
19:10
i run both in vm
19:10
ubuntu vm for ltsp server
19:10
and then another vm that just boots from network
19:10
ltsp server is interface vbox0 and client is vbox1
19:11
<steph_>
interface vbox0 = bridge, right?
19:11
<johnny>
br0 is the bridge
19:12
you can use internal networking if you want both the client and server to be vm
19:12
but i wanted to be able to test other machines, so i setup the host interface
19:14
<steph_>
Again, just to be sure how packets will travel. #1- My switch is connected to eth1. Client (real machine) -> switch -> host.eth1 (range 192.168.0) -> br0 -> guest.eth1 (range 192.168.1) -> ltsp/dhcp then back to the real machine. Do I understand?
19:16
euh. there is no #2 :))
19:19
johnny: am I totally wrong ? :|
19:20
* steph_ hope not, but it doesn't look good...
19:20
<johnny>
why are your using two subnets?
19:20
i dont understand why you're making it so complicated
19:23
<steph_>
I thought bridging connect two subnet together. (well, thanks. you confirm me something.)
19:27
johnny: So host.eth1 (192.168.0.1) and guest.eth1 (192.168.0.2) is a better idea, right?
19:31
If they are on the same subnet, why do I need a bridge?
19:36chup has quit IRC
19:37mikkel has quit IRC
19:37chup has joined #ltsp
19:40petre has joined #ltsp
19:41
<petre>
evening all
19:44
<johnny>
steph_, the bridge is for your computer itself
19:44
so you don't have to port forward
19:48
<steph_>
I'm not sure I understand? * well I'm very sure
19:50
I'm very lost :(
19:51
*steph_ is crying... :(
19:57
<johnny>
i'm on the phone atm
19:57vagrantc has quit IRC
20:08Jbassett has quit IRC
20:18
<johnny>
the problem is that if your server is natted
20:18
all inbound connections from your client will be answered by the host server and not the ltsp server
20:18
vm
20:19
bridging gives them a real external ip
20:20
so you don't have to setup port forwarding
20:20
that's my understanding anyways
20:20
my box is 192.168.2.4 but my ltsp server is 192.168.2.3
20:25chup has quit IRC
20:25chup has joined #ltsp
20:39
<steph_>
johnny: so just setting up port fowarding will make the job, right?
20:39
<johnny>
perhaps.. but it doesn't sound the smart way to do it
20:39
otherwise nobody would be setting up briding
20:39
bridging*
20:39
there are docs for bridging on the virtualbox site
20:40
<steph_>
Yeh, I know, I read it twice
20:40
but never done that before
20:40
(bridging)
20:40
<johnny>
neither did i
20:41
first timer here
20:41
another reason for briding vs port forwading, is that you couldn't run multiple setups
20:41
otherwise you'd be twiddling the ports for every one
20:42
<rjune>
portforwarding only exposes a single port though, not the entire interface
20:42
<johnny>
sure, you'd have to do one for each
20:43
<rjune>
and you might only have 1 IP address available to you
20:43
<johnny>
you understand that right steph_ ?
20:43
<steph_>
You confirm that I must continue with portfowarding. thanks
20:43
<johnny>
not in a local lan tho
20:43
steph_, you'll have to port forward ever possible port
20:43
<steph_>
argh !
20:43
<johnny>
it sounds like a bad idea
20:43
<rjune>
every required port.
20:43
<johnny>
well port range even
20:44
<steph_>
Sorry, tried to say bridging
20:44
<rjune>
bridging is generally preferred
20:44
<steph_>
*steph_ almost had a heart attack
20:45
That's what I wanted to say. Thanks
20:45
The step by step in the doc should work well.
20:46
Did you ever try? Is there something I have to care about?
20:46
<johnny>
uhmm.. i'm using gnetoo
20:46
gentoo* for my host
20:46
so i'm not sure
20:46
check ubuntu/debian forums if you want to be sure
20:46
or vbox forums itself
20:48
<steph_>
Thank you very much, it's clearer.
20:50
Good night! By the way, we had 71cm (+- 28 inches) of snow yesterday. It's pretty nice. Bye and thanks again.
20:51
<johnny>
no snow here :(
20:51
g'night
20:51
<steph_>
Where are you from
20:51
<johnny>
baltimore,md
20:51
<rjune>
steph_: I think he's from MD
20:51
<johnny>
truly i am
20:52
<steph_>
Canada. Ciao
20:52steph_ has left #ltsp
21:00J45p3r has quit IRC
21:13captain_magnus has quit IRC
21:14captain_magnus has joined #ltsp
21:24cliebow has quit IRC
21:50indradg__ has quit IRC
21:55petre has quit IRC
21:58twinprism has quit IRC
22:40chupa has joined #ltsp
22:42chup has quit IRC
23:05
<johnny>
hmm. so.. why is it called RC*_WHITELIST.. it looks like a blacklist to me :)
23:07
or maybe i have it backwards..
23:12
silly me..
23:22|Paradox| has joined #ltsp
23:23
<|Paradox|>
LTSP development still continuing? or has it just gotten so bug free there is no need to update it anymore?
23:23
<MacIver>
uh...yeah ;)
23:26
<johnny>
certainly is continuing?
23:26
why wouldn't it be ?
23:27
<|Paradox|>
I've been away for quite some time now, and it seems LTSP is on the same version.
23:27
<johnny>
there hasn't been a release
23:27
<jammcq>
which version?
23:27
<johnny>
it's still in 5.1.0 in bzr
23:27
<|Paradox|>
heya jammcq
23:27
<jammcq>
hey |Paradox| welcome back
23:27
<johnny>
ogra is doing snaps as pre
23:27
for hardy
23:27
<jammcq>
what version are you thinking of?
23:29
<johnny>
i wonder when there will be an "official" snap or at least a bzr tag
23:29
<jammcq>
I think there will be a tag, when Ubuntu releases 8.04
23:30* |Paradox| needs to dust off his server and get back to work on weaning the rest of his family off windows
23:31
<|Paradox|>
Won't be easy since they have discovered WOW and there is no Linux client yet. Hope it works well under WINE or something.
23:32
<johnny>
it should
23:32
i know people who are personally using it
23:32
<warren>
HAH, I figured out why it was taking forever for X to start after I fixed the xauth problem in ldm.
23:32
ldm was reading 16 bytes from /dev/random to generate the magic cookie key.
23:32
<johnny>
i don't know about 100% compat graphically, but functionally it works well
23:32
<warren>
the entropy pool was entirely empty
23:32subir has joined #ltsp
23:32
<warren>
I tried to use the keyboard to figure out what's stuck, but doing so would cause it to unstuck, because key presses generate entropy.
23:33
<jammcq>
how do you fill the entropy pool ?
23:33
<johnny>
it sucks when you try to start apache on a fresh install :(
23:33
warren, did you switch to urandom?
23:34
<warren>
jammcq, kernel drivers might be coded to shove things into entropy
23:34
jammcq, keyboard and mouse are usual examples
23:34
jammcq, some network drivers do it
23:34
<jammcq>
hmm
23:34
<warren>
jammcq, I'm warned against switching to /dev/urandom for the xauth key
23:35
on fast clients it gets to X too quickly to generate entropy
23:35
even 16 stinking bytes =)
23:35
<johnny>
understandable i guess
23:36
<MacIver>
just always use 1
23:36
<warren>
so now I have to think of something that might be unpredictable that I can shove into entropy
23:36
I'm thinking "ifconfig" might be a good one. all kinds of random network stuff can change those numbers.
23:37
<rjune>
use when I chane my underwear.
23:37
change
23:38
<MacIver>
rjune: tomorrow?
23:38
<rjune>
You can't be sure. :-)
23:38
<MacIver>
but it's always 'tomorrow'
23:39
<rjune>
on a more serious note, would memory values or video data be useful?
23:39
<MacIver>
not really random...
23:40
i would take cpu temp, fan speed, and a true random number generator
23:40
can i have fries with that?
23:40
<rjune>
fan speed, that's interesting, it varies and you won't know what it is
23:41
<warren>
can you read fan speed in a way that is identical on all machines?
23:41
<rjune>
have to look at the code for sensors
23:41
<MacIver>
warren: acpi provides a handle usually...but not every system supports it
23:43
<rjune>
on a desktop system, why would video not be any more or less random then keyboard/mouse input
23:43
<jammcq>
as a thin client boots, there is no keyboard/mouse input
23:44
and the video is more or less predictible
23:44
<rjune>
ahh, this is at boot time
23:44
<warren>
cat /proc/schedstat /proc/net/dev > /dev/random
23:44
this should be good enough
23:45
<MacIver>
in the future, people will be paid to come up with random numbers...computer will rule the world, but millions of people with be placed in jars generating the life-blood of the computer systems: random numbers.
23:45* MacIver patents it
23:46
<warren>
MacIver, measuring radioactive decay wouldn't be better (and cheaper?)
23:46
MacIver, humans are surprisingly bad at coming up with random numbers.
23:46
<MacIver>
warren: not random enough
23:47
<warren>
MacIver, dude, radioactive decay is as random as it gets
23:47
<MacIver>
warren: it wouldn't be directly...the computer would observe them
23:47
warren: but decay happens at a more or less predictable rate
23:47
<warren>
ok, not decay, but the particles themselves
23:48
<MacIver>
but if we had radioactive people!
23:48
<rjune>
measure bugs int he latest software package from MS
23:48
<MacIver>
then we could count how many buildings they ripped apart or how many walls they scaled