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


Channel log from 20 March 2009   (all times are UTC)

00:00
<vagrantc>
cyberorg: either "icewm" or "/usr/bin/icewm"
00:00
<cyberorg>
-s does not like anything other than names of the session, no path
00:00
<vagrantc>
ok, so we could get the basename ...
00:01
<cyberorg>
we'll try and work on that and send another patch :)
00:03
about sound, would something like this be acceptable?
00:03
mkdir ~/.ltspvar && mount -t tmpfs tmpfs ~/.ltspvar/ && echo
00:03
"PULSE_SERVER=xx.xx.xx.xx" > /.ltspvar/ltspsound
00:04Nubae1 has joined #ltsp
00:04
<vagrantc>
cyberorg: the problem is mucking with the user's home directory...
00:04
<cyberorg>
we can delete and unmount tmpfs immediately after the variable is read
00:04Nubae has quit IRC
00:05
<vagrantc>
cyberorg: putting it in tmpfs doesn't really gain anything
00:06
<cyberorg>
hmm, i am out of ideas then, any other suggestion?
00:07
<vagrantc>
cyberorg: "-s shell" doesn't give any way to pass arguments?
00:08
<cyberorg>
vagrantc, no
00:08
just name of the session
00:09
<vagrantc>
so, the options are either writing files in the user's home directory, or patching xrdp...
00:10
and xrdp developers weren't really amenable to much in the way of patches?
00:15
cyberorg: removing it immediately after login helps ...
00:15
then you'd only have the unusual case of a partially sucessful login ...
00:16
with the auto-resume feature, if multiple thin clients log in as the same user, do they get a connection to the same session on all desktops?
00:19
<cyberorg>
never tested multiple login by same user
00:20
there is a xrdp-sesman, will test what happens
00:22
<vagrantc>
the resumed sessions will also be really weird with sound, if you set the PULSE_SERVER from one session, and then log into a resumed session from a different thin client, i don't think you can change the value of PULSE_SERVER to point to the new thin client
00:22
<cyberorg>
it allows multiple login so resume would have wrong PA variable
00:24
multiple login behaves like vnc behaves on windows, display is same on all screens
00:25
<vagrantc>
that's both really cool, and really frustrating...
00:25
<cyberorg>
i guess best would be to let user select the TC PA via padevchooser?
00:26
that would be manual process and they will have to remember IP of the client when they log in
00:26
<vagrantc>
ouch
00:27
cyberorg: i'd like for this nomad thing to somehow be selectable by the user ... but that would require a lot more patching...
00:28
<cyberorg>
vagrantc, we can have three checkboxes on ldm screen one for LDM_NOMAD, LDM_DIRECTX and normal with encryption
00:29
or somewhere in preference menu
00:29
<vagrantc>
it's been a while, but i remember it being tricky to do for LDM_DIRECTX, possibly having to do with Xauthority files or something
00:30* cyberorg is lost with sound handling :(
00:34
<vagrantc>
the sound handling seems pretty tricky.
00:34
<cyberorg>
it would break on session resume from another client no matter what we come up with
00:35
<vagrantc>
pretty much.
00:37alkisg has quit IRC
00:50epsas_ has joined #ltsp
01:01epsas has quit IRC
01:08
<vagrantc>
gah. looks like the only thing that's preventing debian's ltsp/ldm packages from migrating from unstable to testing is an alpha build of ldm ...
02:10alkisg_work has joined #ltsp
02:24try2free has joined #ltsp
02:36vagrantc has quit IRC
02:45try2free has left #ltsp
03:33alkisg_work has quit IRC
03:39alkisg_work has joined #ltsp
03:48try2free has joined #ltsp
04:00alkisg_work has quit IRC
04:16try2free has left #ltsp
04:28mikkel has joined #ltsp
04:34alkisg_work has joined #ltsp
05:12ogra has quit IRC
05:13ogra has joined #ltsp
05:17alkisg_work has quit IRC
05:21valent1 has joined #ltsp
05:21
<valent1>
hello
05:28
how to setup LTSP for old clients? is there a distribution setup for old clients - P1 with 64MB of ram... ?
05:41
what is RAM used in thin clients?
05:42
by LTSP documentation it says that clients need 128-258MB of ram
05:43
I thought that some much lover amount of memory was needed because server is actually starting applications and server's memory gets used...
06:27hanthana| has joined #ltsp
06:31makghosh has joined #ltsp
06:33makghosh has quit IRC
06:34makghosh has joined #ltsp
06:42hanthana_ has quit IRC
07:03pmatulis has joined #ltsp
07:06
<valent1>
anybody here?
07:09din_os has joined #ltsp
07:09
<din_os>
The new chromium browser should rock on ltsp, even on alpha... if there was only a way to add flash support
07:10
<Appiah>
there is no flash for chromium yet?
07:10
what
07:12alkisg has joined #ltsp
07:15
<din_os>
well it is still on pre-alpha so surely not
07:15bobby_C has joined #ltsp
07:17ogra has quit IRC
07:17ogra has joined #ltsp
07:19din_os has left #ltsp
07:21alekibango has joined #ltsp
07:22lsuse has joined #ltsp
07:22lsuse has quit IRC
07:25lucy- has joined #ltsp
07:27
<lucy->
someone has to compile the LBE errors, inappropriate ioctl for device, you can help me?
07:36makghosh has quit IRC
07:39valent1 has quit IRC
07:39Selveste1 has joined #ltsp
08:05Nubae has joined #ltsp
08:05Nubae1 has quit IRC
08:27Gadi has joined #ltsp
08:36wima has joined #ltsp
08:45bobby_C has quit IRC
09:01hanthana| has quit IRC
09:05cliebow has joined #ltsp
09:16OvaKill has joined #ltsp
09:17F-GT has quit IRC
09:17OvaKill has quit IRC
09:17F-GT has joined #ltsp
09:30vagrantc has joined #ltsp
09:36twinprism has joined #ltsp
09:39
<wigwam>
hi
09:39
i use ltsp with rdesktop to connect to windows terminal server
09:41makghosh has joined #ltsp
09:41
<wigwam>
but at random times, as user switches the keyboard using alt+shift shortcut, alt act as still pressed (every key pressed after is processed as alt+<key> shortcut)
09:41
any ideas please?
09:42
<Gadi>
what keymap?
09:46
@help
09:46
@locate
09:46
hmm
09:46
that bot is not up
09:46
:)
09:46
<wigwam>
Gadi: XKBMAP="us"
09:47
<Gadi>
wigwam: are you using us keyboard and us rdesktop keymap, as well?
09:47
<wigwam>
no
09:47
<Gadi>
what keyboard layout are you using?
09:47
<wigwam>
i have to use czech keymap in windows
09:48
<Gadi>
do you use the -k flag in rdesktop?
09:48
<wigwam>
i tried to run rdesktop with -k none switch and this problem was fixed
09:49
but then win key and some shortcuts didn't work
09:49
<vagrantc>
Gadi: i thought the bot used ! .. or are you trying to talk to a different bot?
09:49
<Gadi>
which -k flag do you use normally?
09:49
vagrantc: nubae had another one
09:49
<wigwam>
Gadi: tried to use -k gb (win key works)
09:50
or -k cs2 (where cs2 is our company built keymap for rdesktop)
09:50
with both this alt problem appears
09:51
<Gadi>
do they both have "include common" ?
09:51
<wigwam>
yes
09:52
<Gadi>
yeah, I had found an rdesktop keymap bug with respect to caps lock
09:53
which I fixed by editing the common keymap
09:53
perhaps it is similar?
09:53
have you searched for rdesktop bug?
09:55
<wigwam>
no
09:55
now i tried and i found some interestingly similar problem, thx
09:59makghosh has quit IRC
10:01
<Gadi>
wigwam: great! let us know what you find
10:08
<wigwam>
http://osdir.com/ml/network.rdesktop.user/2004-08/msg00058.html
10:08
not proved yet
10:12_UsUrPeR_ has quit IRC
10:12twinprism has quit IRC
10:13topslakr_ has quit IRC
10:13_UsUrPeR_ has joined #ltsp
10:16
<wigwam>
Gadi: yeah, xmodmap -e "keysym Alt_L = Alt_L" works
10:17
<Gadi>
weird
10:17
<wigwam>
but i don't know where to write this command to make the change permanent, don't you know?
10:17
<Gadi>
I wonder if you could fix it in an rdesktop keymap, as well
10:17
<wigwam>
/etc/X11/xinit/xinitrc doesn't work
10:17
<Gadi>
are you running rdesktop from the server after you login through ldm?
10:18
<wigwam>
no, i use rdesktop screen script
10:18
<Gadi>
ah, so add it to the screen script
10:18
<wigwam>
k, thx
10:19* Gadi wishes he knew if this was an rdesktop bug or an X11 bug
10:20
<Gadi>
also, wigwam, what version of rdesktop are you using?
10:20
<wigwam>
btw. i wrote very simple script called useroption which shows ascii menu and reads option from keyboard... and call screen script selected by user
10:20
should i send it somewhere? it's quite stupid, but maybe useful for somebody
10:20
<Gadi>
we love stupid scripts
10:20
:)
10:21
I guess you can send it to the ltsp-devel mailing list
10:21
<wigwam>
Gadi: 1.5.0 .. tried to upgrade to 1.6.0 but problem doesn't vanished, so i downgraded back to ubuntu latest
10:22
<Gadi>
ubuntu's latest isnt 1.6?
10:36
<wigwam>
no
10:37
or i don't have latest ubuntu version :) i'm a little bit lost in ubuntu versions + debian repositories
10:37
<Gadi>
wigwam: from /usr/share/rdesktop/modifiers:
10:37
# Translate Meta, Super and Hyper to Windows keys.
10:37
# This is hardcoded. See documentation for details.
10:37
it seems that upon release, it sends Meta rather than Alt
10:38
and the Windows scancode for meta seems to be hardcoded
10:38
so, that must be the source of the problem
10:38johnny has joined #ltsp
10:38
<Gadi>
your fix fixes it by having X send Alt instead of Meta upon release
10:39
which must be something that is done in the ubuntu keymap by default
10:46CAN-o-SPAM has joined #ltsp
10:47* CAN-o-SPAM has recieved all the votes for the LTSP Logo Contest ...
11:05RobertLaptop has quit IRC
11:10epsas_ is now known as epsas
11:10alkisg has quit IRC
11:11
<epsas>
vagrantc - thanks
11:13twinprism has joined #ltsp
11:14* SDuensin cheers wildly! He fixed FreeNX on Ubuntu!
11:14
<epsas>
Sduensin - congr4atulations
11:15
<SDuensin>
Now if I could get Gmail to work, I'd post to the list. :-)
11:15
<CAN-o-SPAM>
Ubuntu 8.10? What was the problem?
11:16
<SDuensin>
"netcat -z 127.0.0.1 7000" waits damn near forever.
11:16
And it does it 5 times on different ports.
11:17
The intent it to ensure that the port isn't already in use. If it *is* in use, netcat returns immediately.
11:17
So I tossed a "-w 1" in there.
11:22faustino3331 has joined #ltsp
11:22faustino333 has quit IRC
11:28kaeles has joined #ltsp
11:29epsas is now known as JesusChrist
11:29JesusChrist is now known as Jesus_
11:29Jesus_ is now known as epsas
11:30twinprism has quit IRC
11:31
<SDuensin>
Is IDS-Server part of iTALC?
11:36mib_cgprfw has joined #ltsp
11:40mib_cgprfw has left #ltsp
11:40mib_cgprfw has joined #ltsp
11:41Nubae1 has joined #ltsp
11:41Nubae has quit IRC
11:43mib_cgprfw has quit IRC
11:44alkisg has joined #ltsp
11:44japerry has quit IRC
11:53lucy- has quit IRC
11:55Lns has joined #ltsp
11:56daduke has quit IRC
11:59daduke has joined #ltsp
12:17
<Lns>
Hey all, CAN-o-SPAM - any results for the logo contest? =)
12:18alkisg has quit IRC
12:25elisboa has quit IRC
12:26elisboa has joined #ltsp
12:26
<vagrantc>
Lns: CAN-o-SPAM said they were all in ... surely they're compiling the info as we speak!
12:26elisboa_ has joined #ltsp
12:27
<Lns>
vagrantc: woohoo!
12:27
<cliebow>
cant believe he waited fro mine to tabulate the results..just like in tribal council
12:35elisboa has quit IRC
12:36elisboa has joined #ltsp
12:39_UsUrPeR_ has quit IRC
12:40
<rjune>
cliebow, ping
12:40
Do you know how long the maple season is up your way?
12:41
<CAN-o-SPAM>
yes the votes are being tallied as we speak ;)
12:44
<SDuensin>
Anybody ever try to use iTALC with FreeNX?
12:45
<NeonLicht>
I've tried with just LTSP and never got it working
12:45
<SDuensin>
hehe
12:45
I got it to work with LTSP clients. Wish I remember how.
12:46
I wrote it down somewhere. :-)
12:46
<NeonLicht>
SDuensin, if you happen to find it, gould you care to share that info? :-)
12:46
<SDuensin>
I think I'm going to have to roll some code for the FreeNX part.
12:46
Lemme dig real fast.
12:46
I'm on Ubuntu, so you may have to translate.
12:47
<NeonLicht>
hehe
12:47maginot has joined #ltsp
12:47
<maginot>
hi
12:47alkisg has joined #ltsp
12:49
<maginot>
If Im able to boot my client using a floppy with etherboot r8169 image, and even the client bios having pxe support Im not able to boot without the floppy (But shows the message that his trying to get a dhcp offer and the server shows a dhcp offer but seen the client dont accept/responde) so where is the problem? Is the Client NIC Card/BIOS is the server r8169 module or is the dhcp server?
12:51
<SDuensin>
NeonLicht - Here's the voodoo: http://jaeger.pastebin.com/m4fef60b6
12:51
<vagrantc>
maginot: most likely the PXE implementation on the client
12:51
maginot: possibly issues with the switch
12:51
<SDuensin>
NeonLicht - Do that as root or sudo each line.
12:52
<NeonLicht>
athanks a lot, SDuensin, let see if it works on Debian :-)
12:52
<SDuensin>
NeonLicht - If I remember right, the problem was the avahi-daemon not running on the clients.
12:52
<NeonLicht>
I see
12:52
SDuensin, what do you install joe for? XDDDD
12:53
<SDuensin>
Was that in my paste? :-)
12:53
I install joe everywhere. :-D
12:53
<NeonLicht>
yup!
12:53
haha
12:53
I do the same with Vim XDD
12:53
<SDuensin>
Surprised you also didn't get a paste of "mc" and "screen" as well. :-)
12:53
<maginot>
vagrantc, well, Im testing in a closed enviorement so cant be the switch. The client has a onboard RTL8111/8168B and seens to use the r8169 driver. And it ask for a dhcp offer, I can see the request and the server answer at /var/log/daemon.log but the client stays in the waiting dhcp offers screen and dont boot.
12:53
<NeonLicht>
there was a time I used to install joe everywhere, until I started using Vim, haha
12:54leio has quit IRC
12:54
<SDuensin>
I'm old school from the Borland "Turbo" days. Still got all the WordStar keys in my head.
12:54
<NeonLicht>
the first thing I do after installation is aptitude install vim screen unclutter ion3, myself, hahaha
12:54
<vagrantc>
maginot: i don't see how you can rule out the switch just because it's a closed network.
12:55
maginot: when troubleshooting, always think of ways to reduce your variables.
12:55
<maginot>
vagrantc, using a etherboot .dsk image for the r8169 nic I can boot normaly, so I started to think if could not be some problem/configuration in the server
12:55leio has joined #ltsp
12:55
<maginot>
vagrantc, well, I can tell that is not the switch because we have used 4 different so far
12:55
<vagrantc>
maginot: in many cases etherboot is more reliable than on-board PXE implementations. that's no surprise.
12:56
<maginot>
vagrantc, but with etherboot im stuck in the floppy/disk incovinient
12:56
<vagrantc>
maginot: i just recently found an issue with spanningtree enabled on two chained switches. caused all sorts of grief.
12:56
maginot: i understand...
12:56
<SDuensin>
Hey, that ion3 is cool. Never seen that one before. :-)
12:57
<NeonLicht>
SDuensin, sure, it is, that's why I use it :-)
12:57
<vagrantc>
maginot: i spent months trying to figure out configuration differences and it turned out to be numerous switches all with the same options enabled.
12:58
<maginot>
vagrantc, well, the switch we're using are 10/100 and the server nic and client nic (r8169) are both gigabit, you think this can be a switch problem?
12:58
<NeonLicht>
SDuensin, I remove the frame bars, and have a black background with a few frames running xterm with ssh sessions to different servers, and then run screen on them... I diferenciate among frams because I use different font color for different servers XD
12:58
<vagrantc>
maginot: could ... try plugging a cable between the client and server directly ... see if it makes a difference.
12:58
<SDuensin>
Oh, NeonLicht - forgot a step. In lts.conf, add: START_ITALC = True
12:58
<maginot>
if I use a crosscable and connect then directly so we can see if is realy a switch problem
12:59
*they
12:59
<vagrantc>
maginot: shouldn't need a crossover if either is gigabit ... gigabit does auto-sense
12:59
<NeonLicht>
SDuensin, that I already had done :-)
13:00
<maginot>
vagrantc, didnt know that ... well, I will test right now
13:00
<vagrantc>
maginot: also, check what version of the PXE client is ... might be some way to upgrade or downgrade it to a working version
13:00
<NeonLicht>
SDuensin, well, the first problem comes: there is no /etc/ital/keys subdir :/
13:01
<maginot>
vagrantc, with a bios/firmware update I suppose
13:01
<SDuensin>
On the server or the chroot?
13:01
<alkisg>
maginot: also, see what dhcp options the client asks for - and if the server doesn't provide them, add them to dhcpd.conf
13:01
<NeonLicht>
SDuensin, on neither of them
13:02
<SDuensin>
And you installed the italc-master package on the server?
13:02
<NeonLicht>
sure, your instructions told me so :-)
13:03
in fact, SDuensin, the /etc/italk/ directory does exists, but it is empty
13:03
<SDuensin>
I reread my script/notes. That's all I got. :-/
13:04
Maybe you have to run it first on Debian or something?
13:04
<NeonLicht>
I see, thanks anyway
13:04wima has quit IRC
13:04
<maginot>
vagrantc, PXE-2.1 (build 082) - For realtek RTL8111B/8111C Gigabit Ethernet Controller v 2.09 (071017)
13:04
<NeonLicht>
I'm rebuilding the image anyways to see if it works even w/o the keys step
13:04warren has joined #ltsp
13:05
<warren>
CAN-o-SPAM: got my e-mail?
13:05
<CAN-o-SPAM>
warren: yup ... we've been working on disabling that before and were aware of the issue, thanks for your analysis.
13:06
warren: same problem with TC5, or no?
13:06
<alkisg>
NeonLicht: If I remember correctly, the command to create the keys is: ica -createkeypair -role teacher
13:06
<warren>
CAN-o-SPAM: it will become available in a bios update?
13:06
CAN-o-SPAM: no, TC5 is fine in that regard
13:07
<NeonLicht>
alkisg: as root, I suppose...
13:07
<alkisg>
Yes
13:07
<NeonLicht>
creating new key-pair ... ...done, saved key-pair in
13:07
<SDuensin>
alkisg - guess the Ubuntu package does that for you. I never had to.
13:08
<maginot>
when I receive the TFTP message what means?
13:08
<NeonLicht>
creating new key-pair ... ...done, saved key-pair in -role and teacher (seems strange, doesn't it?)
13:08
<alkisg>
SDuensin: yes - and also the windows installer - never tried on another OS/distro :)
13:08
NeonLicht: try without the -role teacher
13:09
<NeonLicht>
alkisg: and then? am I supposed to placce those files on //etc/italcc/keys?
13:09
<alkisg>
NeonLicht: where did it put the keys? It should report the directory...
13:10
<NeonLicht>
alkisg: . (mmy home dir)
13:10
alkisg, I know where they are, I need to know where they are supposed to go :-)
13:10
<alkisg>
Bah... then try specifying the directory as part of the ica command line
13:10
<NeonLicht>
ok, man ica, then XDDD
13:11
<alkisg>
Yup :)
13:11
/etc/italc/keys/private/<role>/key
13:11
<SDuensin>
iTALC is cool, but man, what a pain.
13:12
<NeonLicht>
is it cool, really?? never saw it workinng yet :-)
13:12
<SDuensin>
hehee
13:12
Yea, it's nice. :-)
13:13
<NeonLicht>
updating the image again (after having moved the key files into /etc/italc/keys/ ---man page did not help at all)
13:15
now, in case it works, I'd need to know if there is a way to run iTalc other than a Gnome applet (I won't run Gnome just for that :-)
13:15
<SDuensin>
So close. Dumb ica-launcher keeps handing out the same port numbers for non-LTSP connections.
13:16
NeonLicht - Go crazy! Run the Windows one in Wine. :-P
13:16
<NeonLicht>
hahahaha
13:17
the only wine I use comes on glass bottles, most of the time either from Rioja or from Rivera del Duero XDD
13:19
<kaeles>
silly question, I was just wondering if when using ltsp, the xserver is actually run on the server itself and the X "data" is forwarded to the clients, or do the clients run their own copy of the xserver, wondering cause I'm using mythbuntu, and having some issues with ati drivers, but I'm using the diskless version which uses ltsp, and IIRC I had issues trying to pxe boot machiens and use custom xorg settings before.
13:20
<NeonLicht>
Xserver runs on the thin clients, kaeles, the X server *always* run where you work at
13:21
<SDuensin>
X11 uses the terms "server" and "client" backwards from what most people think.
13:21
<NeonLicht>
13:21
<kaeles>
I thought thats how it was , I was just double checking, must be something else thats funky then :|
13:23
thats what I get for being cheap and having old machines with old ati cards :P
13:23
<NeonLicht>
ok, I run italc from the command line now and I get an error msg telling me that ica doesn't seem to be running on the server :-(
13:23
man, I will never be able to understand the iTalc thing XDD
13:24
<ltsppbot>
"alkisg" pasted "italc ubuntu postinst script" (41 lines) at http://pastebot.ltsp.org/288
13:24
<alkisg>
NeonLicht: see the commands from the ubuntu installation script ^^^
13:25
What are you trying to do? Run italc normally, or in the chroot?
13:25
<SDuensin>
What I gave him was for chroot.
13:25
<NeonLicht>
alkisg: "normally"
13:26
<alkisg>
Then why did you say you were updating the image?
13:26
<NeonLicht>
am I supposed to run it from the chroot?
13:26
<alkisg>
Both methods are valid; each has its advantages
13:26
<NeonLicht>
alkisg: because SDuensin instruction told to update the chroot,