IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 6 December 2007 (all times are UTC)
| 00:00 | <sadmin1> i install ubuntu-gutsy with ltsp 5
| |
| 00:00 | all good
| |
| 00:00 | but when thin clients boot GUI wont come up
| |
| 00:00 | what i do with lts.conf
| |
| 00:00 | i did these change in lts.conf
| |
| 00:00 | [Default]
| |
| 00:00 | SERVER = 10.1.2.2
| |
| 00:00 | XSERVER = auto
| |
| 00:00 | X_MOUSE_PROTOCOL = "IMPS/2"
| |
| 00:00 | X_MOUSE_DEVICE = "/dev/psaux"
| |
| 00:00 | X_MOUSE_RESOLUTION = 400
| |
| 00:00 | X_MOUSE_BUTTONS = 3
| |
| 00:00 | USE_XFS = N
| |
| 00:00 | X_MODE_0 =1024x768
| |
| 00:00 | twinprism_ has quit IRC | |
| 00:00 | <sadmin1> X_HORZSYNC =30-50
| |
| 00:00 | X_VERTREFRESH =50-60
| |
| 00:00 | SCREEN_01 = startx
| |
| 00:01 | SCREEN_02 = shell
| |
| 00:01 | # SCREEN_07 = ldm
| |
| 00:01 | SOUND = Y
| |
| 00:01 | LOCAL_STORAGE = Y
| |
| 00:01 | SOUND_DAEMON = nasd # or esd
| |
| 00:01 | LOCAL_APPS = Y
| |
| 00:01 | any idea
| |
| 00:02 | cyberorg has joined #ltsp | |
| 00:03 | rcy has quit IRC | |
| 00:03 | <sadmin1> helo
| |
| 00:17 | guys where i need to put my lts.conf
| |
| 00:17 | in /opt/ltsp/i386/etc/ or in /var/lib/tftpboot/i386/etc
| |
| 00:19 | Egyptian[Home] has joined #ltsp | |
| 00:51 | sadmin1 has quit IRC | |
| 00:51 | <open_lad> is anyone here????
| |
| 00:51 | why no resp[onse?? is this what IRC for??
| |
| 00:55 | sadmin1 has joined #ltsp | |
| 00:56 | sadmin has joined #ltsp | |
| 00:58 | open_lad has left #ltsp | |
| 00:58 | Egyptian[Home1 has joined #ltsp | |
| 01:04 | edgarin_ has joined #ltsp | |
| 01:13 | joebake1 has joined #ltsp | |
| 01:18 | joebaker has quit IRC | |
| 01:22 | edgarin has quit IRC | |
| 01:40 | twinprism_ has joined #ltsp | |
| 01:44 | mcfloppy_ has quit IRC | |
| 01:45 | mcfloppy_ has joined #ltsp | |
| 01:52 | MasterOne has joined #ltsp | |
| 01:54 | mathesis has joined #ltsp | |
| 02:10 | sadmin1 has joined #ltsp | |
| 02:10 | sadmin has quit IRC | |
| 02:19 | twinprism_ has quit IRC | |
| 02:23 | mcfloppy_ has quit IRC | |
| 02:48 | cyberorg has quit IRC | |
| 03:14 | sadmin1 has joined #ltsp | |
| 04:04 | Q-FUNK has joined #ltsp | |
| 05:00 | cyberorg has joined #ltsp | |
| 05:08 | Guaraldo has joined #ltsp | |
| 05:11 | tsurc has joined #ltsp | |
| 05:11 | tsurc has quit IRC | |
| 05:13 | tsurc has joined #ltsp | |
| 05:54 | MasterOne has quit IRC | |
| 06:05 | vaisarger has joined #ltsp | |
| 06:06 | <vaisarger> Hi! I joined in ltsp upstream branch... even if it is little scaring for me...
| |
| 06:08 | subir has quit IRC | |
| 06:17 | <vaisarger> I would like help in developing lts project (for my possibilities) ... any tip ?
| |
| 06:17 | <ogra> vaisarger, that taems purpose is that distro people channel contributions to the branch, it would really be better to contact your distro person on the branch (as described in the mailing list post about the team on ltsp-developer) to channel your changes in
| |
| 06:18 | we dont want it to grow to much since everybody can destroy everything for all distros if he'S in that team
| |
| 06:20 | <vaisarger> Ok... I'm new on that... You gave me a link, 2 days ago, but it was not clear for me how to collaborate, actually
| |
| 06:21 | <ogra> find out who of the four members is your distro chap and provide a branch to him with your changes
| |
| 06:22 | <vaisarger> Oh... sorry, I have to quit... see you later, if you want :)
| |
| 06:23 | vaisarger has quit IRC | |
| 06:23 | Egyptian[Home] has quit IRC | |
| 06:23 | meduxa has joined #ltsp | |
| 06:24 | Egyptian[Home] has joined #ltsp | |
| 06:28 | Egyptian[Home1 has quit IRC | |
| 06:50 | jammcq has quit IRC | |
| 07:05 | vaisarger has joined #ltsp | |
| 07:08 | <vaisarger> Hi! I came back. Who are developer distro team members? How I contact them?
| |
| 07:18 | mikkel has joined #ltsp | |
| 07:20 | <ogra> they are in the member list of the team
| |
| 07:22 | <vaisarger> Ok. Maybe my question was wrong. Who is the person that is the Ubuntu7.10 team member?
| |
| 07:22 | <ogra> me or sbalneav
| |
| 07:23 | <vaisarger> Phieww. May I call you Oliver? I'm Vittorio.
| |
| 07:27 | martoya has joined #ltsp | |
| 07:32 | <vaisarger> ogra: never mind... is there a piece of ltsp code that you suggest me I should work on ?
| |
| 07:39 | <ogra> wherever you see place for enhancements :)
| |
| 07:39 | stgraber, https://wiki.ubuntu.com/LTSPClientShutdownReboot ;D
| |
| 07:39 | you wanted that iirc :)
| |
| 07:40 | now we need a little daemon on the session side that hooks into the logout dialog (needs patches to gnome-session)
| |
| 07:44 | <vaisarger> very good... I will look at a little...
| |
| 07:46 | <ogra> it takes a while to understand the structure of the code etc, try to get familiar with how it works etc ...
| |
| 07:50 | <vaisarger> thanks for tip... If I ever manage to make that code better, how can I give you my enhancements?
| |
| 07:52 | <ogra> make a branch that you either put on launchpad or that you host on your own server ...
| |
| 07:52 | point me to it :)
| |
| 07:53 | i'll review the changes and discuss them with you then (or directly accept, depending on the change ;) )
| |
| 07:55 | sadmin1 has quit IRC | |
| 07:55 | sadmin1 has joined #ltsp | |
| 08:13 | sadmin1 has quit IRC | |
| 08:14 | Guaraldo has left #ltsp | |
| 08:14 | MasterOne has joined #ltsp | |
| 08:15 | <vaisarger> ogra: https://code.launchpad.net/~vaisarger/ltsp/direct-shutdown is it ok?
| |
| 08:16 | <ogra> vaisarger, sure, the comment above was directed to stgraber though ... but feel free to work on it
| |
| 08:18 | <vaisarger> Ohhh... touche. :) Ok. I hope will manage to be useful to you ltsp people
| |
| 08:21 | <loather-work> everyone's useful in some fashion. :)
| |
| 08:22 | <ogra> sure :)
| |
| 08:28 | Gadi has joined #ltsp | |
| 08:29 | <Gadi> ogra: ping
| |
| 08:39 | sonjag has joined #ltsp | |
| 08:52 | mathesis has quit IRC | |
| 09:12 | <ogra> Gadi, pongl
| |
| 09:13 | <Gadi> ogra: hey. I collected a lot of good juicy info wrt that problem of gnome not fully logging out
| |
| 09:13 | but, I think I need some help sorting it all out
| |
| 09:13 | <ogra> attach all of it to the bug
| |
| 09:13 | <Gadi> ok
| |
| 09:14 | <ogra> (the new ldm one)
| |
| 09:14 | <Gadi> https://bugs.launchpad.net/ubuntu/+source/ldm/+bug/173239
| |
| 09:14 | that one?
| |
| 09:18 | <ogra> yeah, ldm has only two bugs yet :)
| |
| 09:18 | thats the right one
| |
| 09:18 | <vaisarger> Bye, bye, people! :9
| |
| 09:18 | <ogra> ciao
| |
| 09:19 | vaisarger has quit IRC | |
| 09:19 | <Gadi> ogra: my issues are a bit larger than the title of that bug
| |
| 09:19 | maybe I should open a new one
| |
| 09:19 | <ogra> define larger
| |
| 09:19 | <Gadi> well, here aer some of the things I have noticed:
| |
| 09:20 | 1. After ldm makes its session connection to the server, the ssh socket becomes unresponsive to new ssh requests
| |
| 09:20 | This causes removable media not to work
| |
| 09:20 | <ogra> not true
| |
| 09:20 | <Gadi> and lots of add_fstab_entry processes hanging on the client
| |
| 09:20 | <ogra> i worked with exactly this socket for the whole day
| |
| 09:20 | <Gadi> Im not saying in *your* env
| |
| 09:20 | I am saying in *my* env
| |
| 09:21 | <ogra> that is feisty ?
| |
| 09:21 | <Gadi> gutsy
| |
| 09:21 | <ogra> or gutsy
| |
| 09:21 | well
| |
| 09:21 | mine is essentially gutsy as well
| |
| 09:21 | <Gadi> just bec *you* have no such bug, does not mean one does not exist
| |
| 09:21 | :)
| |
| 09:21 | <ogra> i didnt say that
| |
| 09:21 | <Gadi> and mine certainly exists
| |
| 09:22 | it could be some external networking/hardware thing
| |
| 09:22 | <ogra> i just said that in a default install here it doesnt expose
| |
| 09:22 | <Gadi> and not ldm
| |
| 09:22 | right
| |
| 09:22 | thats why I need help sorting it out
| |
| 09:22 | <ogra> can you try to set up a virtualbox client on the server ?
| |
| 09:22 | that would rule out any network
| |
| 09:22 | (with tun/tap interface
| |
| 09:22 | )
| |
| 09:23 | <Gadi> i have yet to set up such a thing
| |
| 09:24 | <ogra> apt-get install virtualbox-ose (i think)
| |
| 09:24 | <Gadi> but, I have been experimenting with ssh connections from other servers
| |
| 09:24 | and there is definitely an issue that is on the thin client side
| |
| 09:24 | let me give you another data point
| |
| 09:25 | when I turn *off* directx, the session hangs at "xrdb -merge"
| |
| 09:25 | <ogra> yes, we talked about that
| |
| 09:25 | <Gadi> and what I find is that in fact the ssh proxy cannot be connected to
| |
| 09:25 | and that is why it hangs
| |
| 09:25 | <ogra> weird
| |
| 09:26 | not seen here either
| |
| 09:26 | how did you install the client there ...
| |
| 09:26 | <Gadi> so, if you try to say, open a terminal and do: xrdb -display localhost:11 -query
| |
| 09:26 | it hangs
| |
| 09:26 | <ogra> is that a fresh ltsp-build-client or a kind of upgrade ?
| |
| 09:26 | <Gadi> fresh
| |
| 09:26 | the only non-standard thing is the IP of eth0 is 192.168.0.253
| |
| 09:27 | and not 254
| |
| 09:27 | <ogra> and the server ?
| |
| 09:27 | upgrade or fresh
| |
| 09:27 | <Gadi> everything was fresh install
| |
| 09:27 | <ogra> ok
| |
| 09:27 | hmm
| |
| 09:27 | <Gadi> now, one thing we did do was backed up our /etc/hosts
| |
| 09:28 | but, I have checked it and it looks right
| |
| 09:28 | but, that's one suspect I have
| |
| 09:28 | <ogra> grep 127.0.1.1 /etc/hosts ??
| |
| 09:29 | <Gadi> empty
| |
| 09:29 | <ogra> aha
| |
| 09:29 | <Gadi> what should that be?
| |
| 09:29 | <ogra> 127.0.0.1 localhost
| |
| 09:29 | 127.0.1.1 laptop
| |
| 09:29 | <Gadi> ok
| |
| 09:29 | <ogra> my system is called laptop
| |
| 09:29 | <Gadi> I had that at one point
| |
| 09:29 | I think
| |
| 09:29 | <ogra> well, its a difference
| |
| 09:29 | <Gadi> then, I noticed the etc hosts in the thin client had 127.0.0.2
| |
| 09:29 | and thought that was important
| |
| 09:30 | <ogra> hmm
| |
| 09:30 | <Gadi> personally, I don't understand the extra loopback ifaces
| |
| 09:30 | <ogra> right, it has 127.0.0.2
| |
| 09:31 | Gadi, btw seen ? https://wiki.ubuntu.com/LTSPClientShutdownReboot
| |
| 09:32 | did you check that your loopback interface is actually up on the clients ?
| |
| 09:33 | <Gadi> yes, its up
| |
| 09:33 | wrt that wiki, I would prefer we do shutdown with Xatoms
| |
| 09:33 | its cleaner
| |
| 09:33 | <ogra> well, nce thats ready indeed ...
| |
| 09:33 | *once
| |
| 09:34 | <Gadi> right
| |
| 09:34 | <ogra> which surely wont happen for gutsy :)
| |
| 09:34 | <Gadi> but, its just as easy to make your listener daemon check xprop
| |
| 09:34 | <ogra> so there its a valid alternative
| |
| 09:34 | <Gadi> rather than keep a socket open
| |
| 09:36 | <ogra> do you use the /etc/hosts file in the client chroot as well ?
| |
| 09:37 | i could imagine there is some discrepancy in the DNS info ssh has on the different sides
| |
| 09:38 | <Gadi> everything is IP-based
| |
| 09:38 | <ogra> hmm
| |
| 09:39 | <sonjag> Hi Ogra and Gadi-- I have stale processes left in Fiesty a lot. I just looked at the Launchpad bug referred to above. I am thinking of installing it, but am just wondering if this is for Fiesty and if it would conflict with Xterminator? Thanks!
| |
| 09:39 | <ogra> sonjag, thats a hardy package
| |
| 09:39 | dont use it on feisty
| |
| 09:39 | <Gadi> yeah, dont install that
| |
| 09:40 | <sonjag> Thanks... I still have stale bonobo and gnome processes causing me to reboot or clean them up manually. Xterminator isn't getting them. Any advice?
| |
| 09:41 | <ogra> Gadi, what about the routing, where does the defaultroute on the client point to ?
| |
| 09:41 | <Gadi> default server 0.0.0.0 UG 0 0 0 eth0
| |
| 09:42 | looks good
| |
| 09:42 | yeah, its weird
| |
| 09:42 | <ogra> sonjag, well, fesity still has the python ldm, you could try to edit /opt/ltsp/i386/usr/sbin/ldm yourself with the changes described in the bug (keep a copy of the original)
| |
| 09:43 | <loather-work> what
| |
| 09:43 | what's the new LDM written in?
| |
| 09:43 | <ogra> C
| |
| 09:43 | <loather-work> nice; i'm guessing it's much faster than the python bloatware.
| |
| 09:43 | <ogra> it is
| |
| 09:43 | <stgraber> ogra: cool, thanks for the wikipage
| |
| 09:43 | <ogra> but a lot less flexible
| |
| 09:43 | <Gadi> see for me, I think my issue is related because I think the ssh socket craps out
| |
| 09:44 | <ogra> i dont understand why though
| |
| 09:44 | and i've never seen something like that
| |
| 09:44 | <Gadi> which causes the sshd not to be able to die properly
| |
| 09:44 | yeah, me neither
| |
| 09:45 | <loather-work> how is it less flexible? I suppose I need to install ubuntu somewhere and create a chroot.
| |
| 09:45 | my current chroot is one that i've stolen from a 6.10 installation.
| |
| 09:46 | <ogra> loather-work, well, did you see what i just said to sonjag ? with C you need to compile the stuff etc ... with python you can quickly advise people to try a fix for example
| |
| 09:46 | its a lot less flexible ...
| |
| 09:47 | contributing to a scripted app is way easier than to C code which is complicated and compley
| |
| 09:47 | *complex
| |
| 09:47 | <loather-work> oh, right. for me i guess that wouldn't be a problem since i'm a C developer, but for most i suppose it is a bit more difficult
| |
| 09:47 | <ogra> but neither of that justifies 10x longer app startup times ... so we went for C
| |
| 09:48 | <loather-work> yeah, the startup time is killing me on some of my workstations
| |
| 09:48 | <ogra> try gutsy :)
| |
| 09:48 | its the fastest ltsp5 we have
| |
| 09:49 | <loather-work> i had planned on getting a couple of the ltsp 100 terminals from disklessworkstations.com, but they're significantly slower than the e150s or t1225s
| |
| 09:49 | <ogra> Gadi, do you have /dev/random (and whatever else ssh needs)
| |
| 09:50 | <loather-work> and it's slow on a t1225, so i didn't want it to end up taking forever to boot :p
| |
| 09:50 | i'll have to try out this gusty.
| |
| 09:50 | i still want to properly integrate it with fedora, but that'll take a lot of time, i'm afraid :\
| |
| 09:50 | <ogra> you dont want the ltsp 1000 term ...
| |
| 09:50 | its no fun
| |
| 09:51 | its great as kiosk system that only runs one app or so
| |
| 09:51 | but no fun with a desktop ...
| |
| 09:51 | (and there is no proper driver for the soundcard)
| |
| 09:52 | <loather-work> yeah, i have some shipping/receiving stations that run nothing but a firefox and a super light window manager
| |
| 09:52 | <ogra> loather-work, warren works on it, there should be code from fedora enetering upstream soon
| |
| 09:52 | <loather-work> bolting it to the back of the monitor would be nice.
| |
| 09:53 | <ogra> i use mine as router/firewall :)
| |
| 09:53 | Q-FUNK has quit IRC | |
| 09:53 | <loather-work> ah, i haven't met warren
| |
| 09:54 | Egyptian[Home] has left #ltsp | |
| 09:56 | <sonjag> ogra: I found the "kill -1" command. I will change that to -9. However, I can't find anywhere that has the cmd[i++]. I'm guessing that this is a different code structure between Python and C. If you have time, can you send what the change would look like in Python? No rush. Thanks!
| |
| 09:57 | <ogra> yeah, thats C vs .py i have to dig up the old code on my server
| |
| 09:58 | <sonjag> ogra, whenever... we're running quite well... I just reboot when I see swapping (and I can prevent if I look for stale processing, but rebooting is quicker :)
| |
| 09:58 | ogra, Thanks!
| |
| 09:59 | twinprism_ has joined #ltsp | |
| 10:00 | <ogra> sonjag, change: 'kill -1 $PPID'] to be: 'kill -9 ${PPID}']
| |
| 10:01 | its around line 195 in my code
| |
| 10:01 | not sure where that moved to if you use the version from sbalneav
| |
| 10:07 | ninkendo has joined #ltsp | |
| 10:08 | <r3zon8> is there a howto for widescreen resolutions on lts4.2?
| |
| 10:10 | <Gadi> r3zon8: widescreens will depend on the modeline (X_MODE_0 in lts.conf) and what the driver is capable of
| |
| 10:10 | many drivers are not capable of widescreen modes
| |
| 10:11 | <ogra> Gadi, did you see my last question ?
| |
| 10:11 | <Gadi> I see it now - I have all the bits that ubuntu pulls in
| |
| 10:12 | /dev/random included
| |
| 10:12 | <ogra> in the running client i mean
| |
| 10:12 | random and urandom are used by ssh
| |
| 10:12 | <Gadi> root@ltsp:~# ls -l /dev/*random
| |
| 10:12 | crw-rw-rw- 1 root root 1, 8 Dec 6 08:50 /dev/random
| |
| 10:12 | crw-rw-rw- 1 root root 1, 9 Dec 6 08:50 /dev/urandom
| |
| 10:12 | <ogra> ok
| |
| 10:14 | <Gadi> ogra: what are the args to l-b-c to build in a non-std location. Maybe I'll build another image and try it on one machine
| |
| 10:14 | that someone is not working on
| |
| 10:14 | <ogra> sudo ltsp-build-client --chroot=test
| |
| 10:15 | <Gadi> ok - thats not gonna adversely update anything that would affect normal users, right?
| |
| 10:22 | <ninkendo> Does anyone here know anything about using LTSP in conjunction with Cisco's DHCP snooping feature?
| |
| 10:22 | It tends to break LTSP completely... I was wondering if there's a patch written for it
| |
| 10:23 | I have information from a support contract with cisco that says the thin client is not behaving to spec with regards to DHCP requests
| |
| 10:23 | and the snoop feature is thus dropping its request packets
| |
| 10:34 | <Gadi> ninkendo: what spec is that?
| |
| 10:35 | <ninkendo> RFC 2131 http://www.faqs.org/rfcs/rfc2131.html
| |
| 10:36 | <loather-work> what, the pxe clients?
| |
| 10:36 | or is this post-boot?
| |
| 10:36 | <ninkendo> basically the dhcp requests/replies are sent through a cisco relay
| |
| 10:36 | this is after it gets the kernel over pxe
| |
| 10:36 | so that dhcp works
| |
| 10:36 | <loather-work> what, like a helper-address on the switch
| |
| 10:36 | ?
| |
| 10:36 | <ninkendo> but then the kernel boots, and it has to get dhcp again
| |
| 10:36 | and cisco's telling us that the thin client is supposed to reset the giaddr field
| |
| 10:36 | but instead it's leaving it set
| |
| 10:37 | so the cisco snooping feature just drops the packet
| |
| 10:37 | and then the thin clients doesn't get an address, and thus can't boot
| |
| 10:37 | <loather-work> 'cause i have an ltsp setup here using that same method (helper-address on the vlan) and it's working beautifully
| |
| 10:37 | <Gadi> thats in the 2nd request or the 2nd reply?
| |
| 10:37 | <ninkendo> the request
| |
| 10:37 | <Gadi> ah, ok
| |
| 10:37 | which ver of LTSP?
| |
| 10:37 | <ninkendo> apparently (it's news to me) the DHCP request has a giaddr field
| |
| 10:37 | I thought only the replies have it
| |
| 10:37 | whatever comes by default in gutsy
| |
| 10:38 | so pretty recent
| |
| 10:38 | <Gadi> ok. thats LTSP5
| |
| 10:38 | it uses the kernel ipconfig to generate the request
| |
| 10:38 | irc
| |
| 10:38 | <ninkendo> http://www.fathersfate.com.mx/wordpress/2005/11/02/linux-terminal-server-project-ltsp/
| |
| 10:38 | <Gadi> iirc, that is
| |
| 10:38 | <ninkendo> that page seems to have some good info, but it's in spanish
| |
| 10:39 | <loather-work> im guessing no hablas espanol, eh?
| |
| 10:39 | <ninkendo> heh
| |
| 10:39 | http://www.fathersfate.com.mx/wordpress/2007/02/15/dhcp-snooping-en-ciscoubuntu/
| |
| 10:39 | er, actually that page
| |
| 10:39 | <Gadi> i was going to say...
| |
| 10:39 | <ninkendo> http://www.fathersfate.com.mx/wordpress/2007/02/15/dhcp-snooping-en-ciscoubuntu/ here's the translated
| |
| 10:39 | but I'm not using a linux dhcp server
| |
| 10:40 | we're using an adonis appliance for that... and just setting the proper values
| |
| 10:40 | <loather-work> "hay una chingadera que se llama DHCP Snooping" --> "there's this fucking thing called DHCP snooping"
| |
| 10:40 | <ninkendo> I wonder if there's anything I can set in the client's dhclient.conf that will help it
| |
| 10:40 | hahaha
| |
| 10:40 | <loather-work> i'm guessing they've had trouble with it :)
| |
| 10:41 | <Gadi> ninkendo: thats not where it is set
| |
| 10:41 | it is in the initramfs
| |
| 10:41 | <ninkendo> oh, yeah you're right
| |
| 10:41 | since it hasn't gotten the root filesystem yet
| |
| 10:42 | staffencasa has joined #ltsp | |
| 10:42 | mcfloppy_ has joined #ltsp | |
| 10:43 | <Gadi> look in: /opt/ltsp/i386/usr/share/initramfs-tools/scripts/ltsp-nbd
| |
| 10:47 | http://www.filewatcher.com/p/udev-042-197.ppc.rpm.663517/usr/share/doc/packages/udev/README.klibc.ipconfig.html
| |
| 10:47 | martoya has quit IRC | |
| 10:47 | <Gadi> try passing the long device name as a kernel arg
| |
| 10:47 | IPOPTS
| |
| 10:48 | you may not need to modify the code at all
| |
| 10:51 | <ninkendo> man, that's a lot of useful information
| |
| 10:51 | I'm trying to figure out how IPOPTS is getting set
| |
| 10:51 | it appears it's getting it from the kernel command line
| |
| 10:51 | <Gadi> its a kernel arg
| |
| 10:51 | it can be passed
| |
| 10:51 | yeah
| |
| 10:51 | <ninkendo> so I would assume that the giaddr field is somewhere in IPOPTS
| |
| 10:51 | <Gadi> edit your pxelinux.cfg/default file
| |
| 10:52 | is that not the gateway ip addr?
| |
| 10:52 | <ninkendo> yeah
| |
| 10:52 | I think this is what it's trying to do...
| |
| 10:52 | <Gadi> look half way down that page
| |
| 10:52 | <ninkendo> my guess at least
| |
| 10:52 | <Gadi> the line with all the colons
| |
| 10:52 | <ninkendo> It thinks the gateway it received in the first dhcp lease (during the PXE phase) may be on a different subnet
| |
| 10:52 | <Gadi> thats what you need to pass: IPOPTS=<client-ip>:<server-ip>:<gw-ip>:<netmask>:<hostname>:<device>:<autoconf>
| |
| 10:53 | <ninkendo> so it's passing it as a kernel command line arg
| |
| 10:53 | hmm
| |
| 10:53 | I don't see anything of the sort in there
| |
| 10:53 | I see "ro initrd=initrd.img quiet splash"
| |
| 10:53 | <Gadi> right
| |
| 10:54 | by default, it autoconfigures, I think
| |
| 10:54 | <ninkendo> could I do something like IPOPTS=::0.0.0.0::::
| |
| 10:54 | and it'll autoconfigure everything but gwaddr?
| |
| 10:54 | <Gadi> try it
| |
| 10:54 | <ninkendo> but wait, that wouldn't make sense...
| |
| 10:54 | IPOPTS would be what it tries to configure itself with
| |
| 10:55 | <Gadi> try:
| |
| 10:55 | <ninkendo> I want it to not send the giaddr field
| |
| 10:55 | <Gadi> ::0.0.0.0::::dhcp
| |
| 10:55 | or something like that
| |
| 10:55 | or maybe no gw-ip
| |
| 10:55 | I would tinker
| |
| 10:55 | ;)
| |
| 10:55 | <ninkendo> right
| |
| 10:55 | thanks
| |
| 10:57 | <Gadi> good luck
| |
| 10:58 | twinprism_ has quit IRC | |
| 10:59 | johnny has quit IRC | |
| 11:05 | dniel has joined #ltsp | |
| 11:07 | Guaraldo has joined #ltsp | |
| 11:25 | cliebow has joined #ltsp | |
| 11:28 | mikkel has quit IRC | |
| 11:30 | daduke_ has joined #ltsp | |
| 11:33 | meduxa is now known as toscalix | |
| 11:37 | dniel has quit IRC | |
| 11:49 | exodos has joined #ltsp | |
| 11:52 | K_O-Gnom has joined #ltsp | |
| 12:01 | Grejao has joined #ltsp | |
| 12:01 | <Grejao> can i install ltsp to be my terminal server and install windows applications with wine to run over ltsp for sessions?
| |
| 12:06 | johnny has joined #ltsp | |
| 12:11 | edgarin has joined #ltsp | |
| 12:13 | <cliebow> Grejao:you can try..by and large windows apps just dont work..
| |
| 12:14 | <ninkendo> well, it looks like no matter what I provide as IPOPTS, it seems to be sending an invalid request
| |
| 12:14 | if I set a static address in IPOPTS, it works of course
| |
| 12:15 | I honestly think there's a bug in klibc (which provides ipconfig) that produces non-compliant behavior
| |
| 12:15 | <Grejao> cliebow, for example Lotus notes,,, i use Lotus notes at my station with Wine.
| |
| 12:16 | <cliebow> Grejao, i just have never had any luck with it..
| |
| 12:17 | <Gadi> ninkendo: you could try replacing it with dhclient or dhcpcd, reroll the initramfs, and ltsp-update-kernels
| |
| 12:17 | of course, make sure dhclient or dhcpcd is pulled into the initramfs
| |
| 12:18 | or email the klibc guys about it
| |
| 12:18 | maybe they know
| |
| 12:23 | dniel has joined #ltsp | |
| 12:26 | ary has joined #ltsp | |
| 12:26 | <joebake1> sadmin1: I tried to get ltsp5 working but couldn't.
| |
| 12:26 | ...on gutsy
| |
| 12:27 | The remote clients boot off the network but then they sit as an initramfs prompt.
| |
| 12:27 | The box was origionally ubuntu, then installed edubuntu-desktop.
| |
| 12:27 | <cliebow> perhaps your interface isnt configured...so dhcp doesnt start]
| |
| 12:28 | ohh..
| |
| 12:28 | <joebake1> DHCP works, pxe starts working. Busybox shows up on the desktop station
| |
| 12:28 | <cliebow> is it listewning on port 200 where nbd is?
| |
| 12:28 | 2000
| |
| 12:29 | <joebake1> I'll check with netstat.
| |
| 12:29 | Petaris has joined #ltsp | |
| 12:29 | <Petaris> Hi all
| |
| 12:29 | I have an AMD Geode LX800 board that I am trying to get working
| |
| 12:30 | <joebake1> nope. netstat -a | grep 2000 doesn't show anything. Thanks for the tip. I'll look into nbd configuration. cliebow.
| |
| 12:30 | <cliebow> look in inetd.conf??
| |
| 12:30 | <Petaris> xorg doesn't like the chipset apparently
| |
| 12:30 | I tried the nsc driver but no luck
| |
| 12:31 | Does the "geode" driver still exist or is nsc the only option?
| |
| 12:31 | <cliebow> Petaris, you may have to set the display in lts.conf..lemme look for a sec..
| |
| 12:32 | <johnny> ndb's only gonna show up when it's needed
| |
| 12:36 | ary has quit IRC | |
| 12:36 | ogra has quit IRC | |
| 12:37 | ogra has joined #ltsp | |
| 12:37 | <cliebow> johnny..but but...tftp shows listenng and that runs thru inetd
| |
| 12:37 | <johnny> hmm.. suprised.. guess i'm not used to running anything through inetd
| |
| 12:37 | nothing i've done over the past years runs through it
| |
| 12:38 | <loather-work> inetd is awesome.
| |
| 12:38 | <cliebow> johnny.. i have nbd running thru inet on port 2000 and 2001 and they both show listening....
| |
| 12:39 | Petaris..you triewd the vesa driver i suppose
| |
| 12:39 | yanu_ has joined #ltsp | |
| 12:40 | wanie has joined #ltsp | |
| 12:41 | <wanie> cyberorg, hi!
| |
| 12:42 | cyberorg, kiwi-ltsp is so cool! I installed it with just four commands starting my thinclient an it works verry fine! :)
| |
| 12:42 | dniel has quit IRC | |
| 12:42 | otavio has joined #ltsp | |
| 12:43 | <cyberorg> wanie, wow, glad to know it worked fine for you :)
| |
| 12:43 | wanie, 128mb client?
| |
| 12:45 | <wanie> cyberorg, but i haven a problem... as thin client i use an PC with enougth hardware ressourced, and sometime after the login the screen is frozen and have pink links ower the complete screen but i can move the mouse :(
| |
| 12:46 | <cyberorg> wanie, can you take a picture?
| |
| 12:46 | <wanie> no my old thinclient doesn' t work with ltsp 5... it have 160 MB but just an 266 MHz cpu
| |
| 12:46 | <cyberorg> ok, so it worked on one client but doesnt on other?
| |
| 12:46 | <wanie> yes one moment please...
| |
| 12:47 | yes on my big workstation with it boots up but then the screen is frozen
| |
| 12:48 | my diskless thinclient (an igel Win Net III) it doesn' t boot
| |
| 12:48 | Grejao has quit IRC | |
| 12:51 | zodman has joined #ltsp | |
| 12:51 | yanu has quit IRC | |
| 12:52 | <Petaris> cliebow: no, I tried the vga driver though
| |
| 12:55 | I will try the vesa driver now
| |
| 12:56 | <cyberorg> wanie, can you open bug reports for both of your clients describing hardware config(including vga) and the problem you faced
| |
| 12:57 | also attach the picture of both the client where it stops
| |
| 12:58 | wanie, link to bug tracker here http://developer.novell.com/wiki/index.php/KIWI-LTSP me got to run it is past midnight here
| |
| 13:03 | <wanie> strange, befor i rebootet my server after the installation i had always the bug, but recently i rebootet the server and now it works without problems until now...
| |
| 13:04 | can the first reboot after the lstp installation be the solution?
| |
| 13:05 | zodman has quit IRC | |
| 13:07 | <wanie> i know my enlish is bad, but I rigorous me ;)
| |
| 13:08 | scrapbunny has joined #ltsp | |
| 13:12 | <scrapbunny> i am trying to figure out what is maxing out my cpu's on my dell poweredge 1750 server. i set system monitor to show all processes but i still do not see anything that should by maxing out the cpus.
| |
| 13:14 | <cliebow> Petaris, Any joy??
| |
| 13:15 | <ninkendo> Gadi: I commented out the line in klibc that sets the giaddr on DHCPREQUEST, and rebuilt my ramdisk with the updated ipconfig binary... it works great now.
| |
| 13:15 | <johnny> scrapbunny, intall htop
| |
| 13:15 | <scrapbunny> any ideas on how i can find what is causing the problem? should i just try a fresh install?
| |
| 13:15 | <johnny> and then run it from the cli
| |
| 13:15 | <ninkendo> Gadi: it's an ugly hack all things considered... I'll have to see if I can write a more long-term fix and send a patch
| |
| 13:16 | <scrapbunny> thanks johnny - i will go try it
| |
| 13:16 | <johnny> it'll give you nice graphs, and you can sort by all sorts of things
| |
| 13:17 | mathesis has joined #ltsp | |
| 13:21 | <johnny> that's what i use to make sure firefox isn't being so mean
| |
| 13:21 | i just login remotely every once in awhile and kill runaway firefoxes
| |
| 13:22 | wanie has quit IRC | |
| 13:23 | wanie_ has joined #ltsp | |
| 13:23 | <scrapbunny> sounds great, i'll check back in after i run it
| |
| 13:27 | wanie_ has quit IRC | |
| 13:28 | staffencasa has quit IRC | |
| 13:32 | Q-FUNK has joined #ltsp | |
| 13:36 | <Petaris> cliebow: no, vesa didn't work either
| |
| 13:38 | MasterOne has quit IRC | |
| 13:40 | <cliebow> Petaris, Adam Drake was working on a geode based..and had an lts.conf which specified higher video settings..perhaps 1024x768 (cant rememebr) can t hurt to give it a whirl..you just get a black screen??
| |
| 13:40 | <Petaris> yeah
| |
| 13:41 | I have it set to 1024x768
| |
| 13:41 | only get black scrrens
| |
| 13:41 | er, screens
| |
| 13:41 | even on tty1
| |
| 13:43 | <cliebow> can you get to a shell and get a pci id for the vid card?
| |
| 13:44 | maybe googling that would turn something up..
| |
| 13:44 | <Petaris> not with this driver
| |
| 13:44 | sec
| |
| 13:44 | let me change something
| |
| 13:47 | ok
| |
| 13:47 | I have a shell
| |
| 13:48 | I did an lspci
| |
| 13:48 | Comes up as AMD Geode LX Video
| |
| 13:49 | VGA Compatable
| |
| 13:55 | daduke_ has quit IRC | |
| 13:58 | logyati has joined #ltsp | |
| 14:06 | martoya has joined #ltsp | |
| 14:06 | <scrapbunny> ok i am running htop but when i try to kill a process it does seem to close
| |
| 14:07 | <johnny> i don't use the interface to kill, because the keybindings interfere with my DE
| |
| 14:07 | i kill outside of it i guess
| |
| 14:07 | oh wait.. i forgot, that started working for me :)
| |
| 14:07 | but before that, i just exited and then killed them
| |
| 14:08 | or use another window
| |
| 14:08 | it's just a nicer display than regular top
| |
| 14:10 | so did anybody manage to make a working deb off that autologin patch?
| |
| 14:12 | <logyati> hello
| |
| 14:12 | i opened a xdmcp session on a thin client, how do i open a xdialog in the client from the server? i mean, i have an user at thiclient and i want him to see a dialog for a reason, and the server should sent it... im trying to export the display but the connection is being refused
| |
| 14:15 | logyati has quit IRC | |
| 14:17 | scrapbunny has left #ltsp | |
| 14:20 | exodos has quit IRC | |
| 14:34 | yanu_ is now known as yanu | |
| 14:42 | Guaraldo has quit IRC | |
| 14:42 | Guaraldo has joined #ltsp | |
| 14:44 | Guaraldo has quit IRC | |
| 14:44 | Guaraldo has joined #ltsp | |
| 14:47 | MasterOne has joined #ltsp | |
| 14:53 | martoya has quit IRC | |
| 14:54 | martoya has joined #ltsp | |
| 14:54 | bartolomeo80 has joined #ltsp | |
| 14:54 | <bartolomeo80> anybody knows how to enable sound in thin clients
| |
| 14:55 | <Petaris> sound=true
| |
| 14:55 | in lts.conf
| |
| 14:55 | but I think its on by default now
| |
| 14:59 | lns has joined #ltsp | |
| 15:05 | Gadi has left #ltsp | |
| 15:06 | bartolomeo80 has quit IRC | |
| 15:09 | Petaris has quit IRC | |
| 15:11 | MasterOne has quit IRC | |
| 15:12 | Egyptian[Home] has joined #ltsp | |
| 15:14 | <cliebow> Petaris lspci -v will teel you what bus id..lspci -n will tell you pci id of that bus id
| |
| 15:14 | <lns> anyone else experiencing runaway Nautilus procs consuming 100% CPU when users log in/out of LTSP?
| |
| 15:14 | <cliebow> crapolaa..
| |
| 15:32 | martoya has quit IRC | |
| 15:36 | Q-FUNK has quit IRC | |
| 15:37 | cliebow has quit IRC | |
| 15:41 | <johnny> lns, yes.. known bug
| |
| 15:41 | login no, but logout yes
| |
| 15:42 | perhaps after logout
| |
| 15:42 | <lns> johnny, is there a bug filed that I can take a look at?
| |
| 15:42 | can't find any in lp
| |
| 15:43 | <johnny> i didn't look
| |
| 15:43 | ogra would know
| |
| 15:43 | <lns> it's pretty bad...i have it on multiple ltsp servers, the whole system slows to a crawl multiple times a day
| |
| 15:45 | vagrantc has joined #ltsp | |
| 15:47 | <lns> ogra, ping?
| |
| 15:48 | I completely purged all compiz packages, just in case that could be causing any issues...
| |
| 15:49 | <johnny> perhaps, but that's not this issue
| |
| 15:50 | <lns> I recall seeing someone giving a workaround of 'killall nautilus' upon logout (from a script) but I don't remember where
| |
| 15:50 | not that this is an elegant approach to fix the issue, but at least i won't have to worry about slow systems in the meantime
| |
| 15:50 | <johnny> xterminator
| |
| 15:50 | try the 0.1-4 package
| |
| 15:51 | not the .2
| |
| 15:53 | <lns> url?
| |
| 15:54 | <johnny> uhmm.. search for it? :)
| |
| 15:54 | i don't have it anymore
| |
| 15:57 | <lns> i did...can't find it :(
| |
| 15:57 | google shows me a bunch of stuff about spam prevention
| |
| 15:58 | <johnny> moquist xterminator
| |
| 15:58 | ry that
| |
| 16:00 | Guaraldo has left #ltsp | |
| 16:01 | <lns> wow, looks nice. i remember looking at this before, with a bunch of stale user dbus procs left over (although not really causing an issue)
| |
| 16:02 | johnny, have you used xterminator, and if so, does it seem to play nice with ltsp?
| |
| 16:02 | <johnny> no i haven't yet
| |
| 16:03 | <moquist> lns: http://ppa.launchpad.net/moquist/ubuntu feisty main
| |
| 16:03 | <johnny> going to eventually
| |
| 16:03 | <moquist> lns: apt-get install xterminator=0.1-4
| |
| 16:03 | lns: the 0.1-4 implementation is messy but sufficient (for feisty)
| |
| 16:03 | <lns> moquist, I'm running Gutsy AMD64...
| |
| 16:03 | <johnny> or gutsy ?
| |
| 16:03 | <moquist> neither 0.1-4 nor 0.2 works in gutsy
| |
| 16:03 | <lns> :(
| |
| 16:03 | <moquist> what I really need to do is debug 0.2 for both feisty & gutsy
| |
| 16:04 | lns: it's python so the arch doesn't matter
| |
| 16:04 | lns: any help you can provide would be awesome :)
| |
| 16:04 | <lns> moquist, ahh. Well, IANAP unfortunately, what would you like me to do? Seeing is that I'm 30min away from my site, but I can install remotely...
| |
| 16:05 | if something goes wrong I'd hate to have to drive all the way down to fix stuff
| |
| 16:05 | <moquist> lns: you don't want to mess with xterminator on gutsy yet, then.
| |
| 16:05 | <lns> moquist, check. =)
| |
| 16:06 | You'd think that Gnome/X11...uhm...anything else would do process checking after a user logs out...
| |
| 16:06 | <moquist> One would think, wouldn't one?
| |
| 16:06 | <lns> heh
| |
| 16:06 | <moquist> Sorta. It's more complex than that.
| |
| 16:06 | <lns> I'm sure it is, more than I understand anyway
| |
| 16:06 | <moquist> How would any of those things know which processes should be left around and which ones shouldn't?
| |
| 16:07 | <lns> maybe ones that were launched via logging in, that are dependent on a window manager..?
| |
| 16:07 | <moquist> xterminator assumes 1) the world isn't perfect 2) system administrators need to be able to make the world seem perfect 3) none of your users need to let persistent processes run when they're not logged in 4) none of your users need to log in more than once at a time.
| |
| 16:08 | <lns> Seems fair
| |
| 16:08 | <moquist> the "perfect" stuff is in reference to the fact that xterminator is a huge kludge. The right thing is to fix the b0rk3n code that keeps running after it should've stopped. But we need something NOW.
| |
| 16:08 | martoya has joined #ltsp | |
| 16:09 | <lns> moquist, i agree wholeheartedly.. nothing will ever be perfect, but there needs to be 'cleanup' to make sure the stuff that isn't, is still thrown away.
| |
| 16:09 | <moquist> lns: sure; I was just covering my ax in case ogra shows up ;)
| |
| 16:09 | <lns> ;)
| |
| 16:09 | <johnny> cuz ogra wont' tawke something that will kill any perhaps screen or nohup sessions
| |
| 16:10 | you could try that moquist? skip any screen or nohup sessions?
| |
| 16:10 | <vagrantc> moquist: is there a way to per-user disable the killing spree?
| |
| 16:10 | <moquist> johnny: sure, we can make xterminator work however we want. :)
| |
| 16:10 | <vagrantc> because that sounds like a really useful concept, as imperfect as it may be.
| |
| 16:10 | <moquist> johnny: xterminator should never be installed by default; it doesn't even belong in main
| |
| 16:10 | <johnny> i really need something that will actually kill procs on power cycle , sadly i can't get acpid to do what it should
| |
| 16:11 | <moquist> vagrantc: there isn't yet, other than to check for UID=0
| |
| 16:11 | vagrantc: but that's also a great idea.
| |
| 16:11 | IIRC, I have comments in the code suggesting a conf file that adds exactly these features. :)
| |
| 16:11 | johnny: I find that the lack of electricity really does 'em in during a power cycle. :-D
| |
| 16:11 | <lns> I think I saw this in http://www.mail-archive.com/edubuntu-devel@lists.ubuntu.com/msg02182.html .. heh
| |
| 16:12 | <vagrantc> moquist: at freegeek, we have a handful of users who need their screen sessions to run persistantly, but most users would benefit greatly from something like what you described
| |
| 16:12 | <moquist> vagrantc: Yep.
| |
| 16:12 | <johnny> aha.. freegeek
| |
| 16:13 | <lns> I don't want to sound completely naive when it comes to issues like this, but the plain simple fact is that there are SO MANY apps that misbehave in X...when a user logs out, there should be some sort of default action to search & destroy all lingering processes after a user logs out of his/her last session.
| |
| 16:13 | <moquist> lns: Well, that's what we're trying to do here. The currently implementations just have bugs. :)
| |
| 16:13 | lns: Well, OK. Not default.
| |
| 16:14 | lns: But you see the validity of arguing that the right thing to do is fix the bugs rather than enable them, right?
| |
| 16:14 | <lns> moquist, well maybe even more specifically, ones that directly related to the session the user was logged into (as to not step on any other processes that might be owned by that user that aren't related to their GUI session)
| |
| 16:14 | moquist, of COURSE we want to fix the bugs
| |
| 16:14 | <vagrantc> johnny: aha?
| |
| 16:14 | <moquist> lns: INSTEAD OF enabling them. :)
| |
| 16:15 | <lns> moquist, enabling them meaning what, allowing them to still have bugs and band-aiding them with xterminator?
| |
| 16:15 | <johnny> a friend of mine might move to portland
| |
| 16:15 | and volunteer
| |
| 16:16 | he helped me with my initial ltsp setup at red emma's
| |
| 16:16 | <moquist> lns: I'm definitely no Xpert, but it seems like one could make a very strong case that if the X server dies there ought to be some specified set of related processes that are forcibly killed (if necessary).
| |
| 16:16 | <johnny> where we are looking to setup something similiar to freegeek
| |
| 16:16 | <vagrantc> johnny: ah, well, feel free to drop in #freegeek
| |
| 16:16 | <moquist> lns: yes, enabling as-in band-aiding so it takes the pressure off to fix them
| |
| 16:16 | <lns> moquist, depends on your perspective.
| |
| 16:16 | <moquist> [nod]
| |
| 16:17 | <lns> moquist, how about extending functionality beyond local syslog and e-mail alerting hung procs that were killed? maybe to a bugtracker?
| |
| 16:17 | <vagrantc> the other issue is you can't rely on the process being part of the current session- because misbehaving processes from an uncleanly logged out session need to be killed even more importantly
| |
| 16:18 | <lns> true
| |
| 16:19 | I'm very surprised that there isn't some piece of inherent code somewhere in X (or even Gnome) that deals with this kind of thing already...
| |
| 16:20 | <moquist> I tried to implement a daemon that traversed the process tree up to the highest-level process owned by the same user, and then watched for /proc/<that-process>/exe to change. It would then wait 30 seconds and kill every process still owned by that user. But it periodically killed all my users simultaneously, so I, er, commented that out. 8-o
| |
| 16:20 | xterminator 0.1-4 still contains that code.
| |
| 16:20 | FWIW.
| |
| 16:20 | I'd love to fix it.
| |
| 16:20 | <lns> hehe
| |
| 16:20 | yeah that'd be a pain
| |
| 16:20 | <vagrantc> lns: problem is that X processes can start programs that are not X code.
| |
| 16:21 | <johnny> lns, because ltsp stuff the only place this issue is happening?
| |
| 16:21 | <lns> johnny, i highly doubt that
| |
| 16:21 | <johnny> most people run stuff on their own PC
| |
| 16:21 | and when they logout, processes die
| |
| 16:21 | like on mine :)
| |
| 16:22 | <moquist> and if any processes hang around, the user reboots or something, and everything gets better.
| |
| 16:22 | this problem is just exacerbated in LTSP.
| |
| 16:22 | <lns> vagrantc, ok.. what about a watcher process launches during a login, tracks all processes related to that session and user (by time launched, perhaps?), and cleans up any lingering after logout?
| |
| 16:22 | <vagrantc> lns: what if the session never logs out?
| |
| 16:23 | i.e. an unclean logout
| |
| 16:23 | <lns> vagrantc, hmm
| |
| 16:24 | well there's still a hook in the system somewhere that realizes that a user (whether x, vt, vnc, etc) is not logged in anymore, right?
| |
| 16:24 | <vagrantc> there are many imperfect ways of getting that information
| |
| 16:24 | <lns> hmm
| |
| 16:24 | <vagrantc> if that were possible, it wouldn't be nearly so difficult to handle the issue at hand :)
| |
| 16:25 | <lns> well I'm getting a better grip on how things work, then... I hope I can be of help to fix this, even if it's only to provide the perspective of someone who doesn't have their hands in the code all the time
| |
| 16:26 | or at all
| |
| 16:26 | ;)
| |
| 16:28 | what's at the most basic level of someone being "logged in" (regardless of method) ? Is it a running process? a /proc entry, lockfile...those are the things i'd think to look at to get a handle on user presence
| |
| 16:29 | that you could hook into, to determine whether they are active on the system, or the processes are lingering after clean (or unclean) logout
| |
| 16:29 | i hope i'm not just babbling nonsensical (to a programmer) crap here
| |
| 16:31 | I'd almost think that the kernel itself has some sort of knowledge regarding who is active and what are just stale processes left behind (like zombie procs, but not so critical to the system)
| |
| 16:31 | I dunno
| |
| 16:34 | obviously parent processes would be a good place to go, but as moquist said, it doesn't work all the time - maybe it's something trivial to fix in the code for xterminator.
| |
| 16:40 | <vagrantc> lns: there really is no way to truely determine activity
| |
| 16:41 | lns: or staleness ...
| |
| 16:41 | <lns> ok, now we're getting more philosophical - more like my ballpark. ;)
| |
| 16:42 | sonjag has quit IRC | |
| 16:42 | sonjag has joined #ltsp | |
| 16:42 | <lns> so...i guess the question is, at WHATEVER level, be it kernel, or init, or gnome, etc...is there a way to create a presence for a user?
| |
| 16:42 | cliebow has joined #ltsp | |
| 16:42 | <vagrantc> *typically* a process is started by a logged in user, and you *might* be able to trace the history of that process back to a particular process ...
| |
| 16:43 | but it could also be started from cron
| |
| 16:43 | <lns> vagrantc, what about pam ?
| |
| 16:43 | does that not have sufficient evidence of someone who is logged in, or is it just plain auth and then let go?
| |
| 16:44 | <vagrantc> sounded like moquist's idea of tracking from the process tree would be the best approach ... presuming you can control which the initial login process is ...
| |
| 16:44 | lns: pam doesn't really handle logout
| |
| 16:44 | <lns> what does, then?
| |
| 16:44 | <vagrantc> lns: there might be hooks to handle it, but if those hooks don't get run, when something goes wrong, when you most likely need to do cleanup ...
| |
| 16:45 | <lns> hmm
| |
| 16:45 | <vagrantc> the issue is you need to figure out a way to know when something is wrong ... all of the proposed ideas work great when everything is going as it should.
| |
| 16:45 | but the issue is to figure out when something has gone wrong.
| |
| 16:45 | <lns> that's true...there are plenty of cases of logged-in users with runaway procs
| |
| 16:45 | that are legitimately logged in
| |
| 16:45 | <johnny> like firefox
| |
| 16:45 | evil thing
| |
| 16:46 | <lns> heh
| |
| 16:46 | seriously
| |
| 16:46 | <johnny> actually, i'm noticing the problem less on gutsy
| |
| 16:46 | <lns> this is intriguing
| |
| 16:46 | <johnny> err less since recent updates
| |
| 16:46 | that is
| |
| 16:46 | my processes have only spiked really high when it comes to java_vm
| |
| 16:47 | <lns> i wonder about any possible hooks with 'nice'
| |
| 16:47 | but there's no inherent 'nice' for the kernel if you don't specify it when running the program explicitly, is there..?
| |
| 16:48 | <vagrantc> your default nice is whatever the parent process was run as
| |
| 16:48 | <lns> ok
| |
| 16:49 | <vagrantc> but using nice values would be an ugly hack
| |
| 16:50 | <lns> agreed on the values - i wasn't sure if there was any inherit hooks in nice to actually determine whether a proc was actively running inside a users' parent (active) session or not
| |
| 16:50 | <vagrantc> all nice values are for is to limit the destructiveness of long-running processes
| |
| 16:52 | <lns> i guess it all really does come down to the ultimate 'parent process' when a user logs in
| |
| 16:54 | i wonder how m$ does it... ;)
| |
| 16:54 | <vagrantc> what you *might* consider implementing ... is a "master process" that's used for login ... and if when the master process logs in, it checks for any processes running as children of previous master processes and mercilessly kills them ...
| |
| 16:54 | <lns> hmmm
| |
| 16:55 | sounds completely logical to me FWIW
| |
| 16:55 | <vagrantc> i'm not sure if there are ways for processes to sneak outside the domain of their parent process ... if that happened regularly, this would essentially be useless
| |
| 16:56 | <lns> process numbers never change, correct? meaning, when a process is launched, it never would somehow change proc numbers (unless restarted of course)?
| |
| 16:56 | <vagrantc> it might spawn a new process and die
| |
| 16:57 | <lns> but that could be tracked through the 'master process' probably
| |
| 16:57 | <vagrantc> in theory
| |
| 16:57 | i'm not process guru
| |
| 16:57 | * lns isn't either | |
| 16:57 | <vagrantc> i feel fairly confident, if this issue were simple, it would be solved by now
| |
| 16:59 | <lns> I think I'm naive enough to still think that someone with little/no knowledge of the intricacies/details of the problem might be able to shed light on it... :)
| |
| 17:00 | it just seems logical to think that there could, in theory, be a 'master process' launched by PAM/etc. that would take over the "presence" of a user.
| |
| 17:00 | It's way beyond LTSP, I'm sure...but that's got to work at some level
| |
| 17:00 | <vagrantc> as long as you always log out properly
| |
| 17:00 | <lns> well that's the thing
| |
| 17:00 | <vagrantc> but the most important problem is when you don't log out properly.
| |
| 17:01 | as said earlier (moquist) ... LTSP makes the problem more noticeable
| |
| 17:02 | <lns> if the process could spawn another process (or whatever else seems more logical) that even if a user's session gets killed ugly-style, that proc would die along with it
| |
| 17:02 | and the master would see this
| |
| 17:02 | I dunno, i'm talking out of my *ss at this point ;)
| |
| 17:03 | <vagrantc> it would probably be worth trying to implement some sort of wrapper login that: 1) kills all remaining processes started by a child of itself, and 2) kills processes that are children of a previous login that weren't properly terminated
| |
| 17:03 | 1 would happen on logout, and 2 would happen on login
| |
| 17:03 | <lns> yeah!
| |
| 17:04 | <vagrantc> and of course, only handle processes still owned by that user
| |
| 17:04 | <lns> of course, if LTSP is a focal-point for stuff like this, it could be implemented directly into ldm..right?
| |
| 17:04 | <vagrantc> i.e. if you sudo or su to a different user, it ignores that part of the process tree
| |
| 17:04 | lns: you could write a wrapper and point LDM_REMOTECMD to it
| |
| 17:05 | <lns> ooo...spicy =p
| |
| 17:05 | i like the ideas, even if i have very little technical ability to put it into motion
| |
| 17:07 | the only thing i could see is that runaway procs during 'su/sudo' would still be an issue..but it's definitely a starting point
| |
| 17:07 | otavio has quit IRC | |
| 17:10 | K_O-Gnom has quit IRC | |
| 17:10 | <lns> heh...as i kill a nautilus process from a staff user remotely taking up 100% CPU..
| |
| 17:11 | <vagrantc> the important thing is to handle the majority of problems to give you more time to only deal with the unusual exceptions
| |
| 17:11 | (without introducing more problems)
| |
| 17:15 | <lns> for sure, any idea of a master-process we talked about would need to track (like in xterminator) what it finds and kills, and would ideally want to report it for investigation
| |
| 17:16 | * lns shivers as he thinks about the "Internet Explorer encountered a problem and must be closed. Report to Microsoft?" | |
| 17:17 | * vagrantc thinks it should be totally non-interactive with an optional log | |
| 17:17 | <lns> Good idea, but should never burden users with it. I think the fact that we're talking about open source would not send chills down the backs of users (at least not mine)
| |
| 17:17 | * lns agrees | |
| 17:18 | <lns> what an elegant band-aid! ;)
| |
| 17:18 | <vagrantc> i should probably take a look at xterminator ... since it sounds like some of the code already exists ... and maybe just adapting or some re-writing is all that's needed ...
| |
| 17:18 | <lns> a band-aid that seeks out other wounds
| |
| 17:18 | I'll help with whatever is needed
| |
| 17:19 | I'd love to see something like this take life and possibly help other projects/bugs at the same time
| |
| 17:20 | <vagrantc> main thing is you need to control what gets run at login ... which is easy to do with LDM ... it would (maybe) be impractical to make it the default shell
| |
| 17:25 | vagrantc has quit IRC | |
| 17:32 | sonjag has quit IRC | |
| 17:33 | sonjag has joined #ltsp | |
| 17:54 | otavio has joined #ltsp | |
| 17:59 | toscalix has quit IRC | |
| 18:05 | J45p3r has joined #ltsp | |
| 18:12 | zodman has joined #ltsp | |
| 18:30 | spectra has joined #ltsp | |
| 18:57 | zodman has quit IRC | |
| 19:12 | vagrantc has joined #ltsp | |
| 19:18 | * vagrantc wonders about wiki.ltsp.org | |
| 19:19 | zodman has joined #ltsp | |
| 19:21 | zodman has quit IRC | |
| 19:35 | vagrantc has quit IRC | |
| 19:42 | edgarin has quit IRC | |
| 19:43 | edgarin has joined #ltsp | |
| 19:52 | J45p3r has left #ltsp | |
| 20:21 | spectra has quit IRC | |
| 20:47 | Faithful has joined #ltsp | |
| 20:49 | <mathesis> se edgarin
| |
| 20:49 | you have installed ltsp5 or ltsp 4.2?
| |
| 20:55 | Egyptian[Home] has quit IRC | |
| 20:56 | Egyptian[Home] has joined #ltsp | |
| 21:00 | <edgarin> mathesis, hi!! :D
| |
| 21:12 | petre has joined #ltsp | |
| 21:12 | <petre> evening all
| |
| 21:21 | Faithful has quit IRC | |
| 21:23 | vernier has quit IRC | |
| 21:24 | edgarin_ has joined #ltsp | |
| 21:24 | mathesis has quit IRC | |
| 21:32 | edgarin has quit IRC | |
| 21:33 | mathesis has joined #ltsp | |
| 21:34 | cliebow has quit IRC | |
| 21:34 | cyberorg has quit IRC | |
| 21:34 | r3zon8 has quit IRC | |
| 21:34 | GiantPickle has quit IRC | |
| 21:34 | elisboa has quit IRC | |
| 21:34 | rjune has quit IRC | |
| 21:34 | jgbillings has quit IRC | |
| 21:34 | primeministerp has quit IRC | |
| 21:34 | sutula has quit IRC | |
| 21:34 | jbrett has quit IRC | |
| 21:34 | warren has quit IRC | |
| 21:34 | shogunx has quit IRC | |
| 21:34 | mistik1 has quit IRC | |
| 21:39 | jgbillings has joined #ltsp | |
| 21:41 | martoya has quit IRC | |
| 21:45 | cliebow has joined #ltsp | |
| 21:45 | cyberorg has joined #ltsp | |
| 21:45 | r3zon8 has joined #ltsp | |
| 21:45 | GiantPickle has joined #ltsp | |
| 21:45 | elisboa has joined #ltsp | |
| 21:45 | rjune has joined #ltsp | |
| 21:45 | primeministerp has joined #ltsp | |
| 21:45 | mistik1 has joined #ltsp | |
| 21:45 | shogunx has joined #ltsp | |
| 21:45 | warren has joined #ltsp | |
| 21:45 | sutula has joined #ltsp | |
| 21:45 | jbrett has joined #ltsp | |
| 21:48 | edgarin_ is now known as edgarin | |
| 21:51 | Faithful has joined #ltsp | |
| 21:54 | petre has quit IRC | |
| 21:57 | fgiraldeau_ has joined #ltsp | |
| 22:05 | <johnny> aha
| |
| 22:05 | fgiraldeau_, howdy
| |
| 22:19 | <fgiraldeau_> hi there
| |
| 22:20 | Are you the autologin guy?
| |
| 22:20 | <johnny> yes
| |
| 22:20 | ogra wanted to see if you could perhaps minimize your patch
| |
| 22:20 | so it could be accepted
| |
| 22:20 | <fgiraldeau_> It can't like this? Humm...
| |
| 22:20 | <johnny> he said you changed more
| |
| 22:20 | than necessary
| |
| 22:21 | i was going to email you, but your email is not on your page
| |
| 22:21 | i tried to jabber you, but you weren't there
| |
| 22:21 | <fgiraldeau_> I put a comment for another bug that I saw at the same time, but to be coherent, I think this is barely minimal.
| |
| 22:21 | I changed about 15 lines of code
| |
| 22:22 | https://launchpad.net/~francis-giraldeau/
| |
| 22:22 | <johnny> i also had some trouble with your deb
| |
| 22:22 | it was looking for debian versions
| |
| 22:22 | <fgiraldeau_> Look at the confirmed e-mail
| |
| 22:22 | you can recompile it from the sources if needed.
| |
| 22:23 | <johnny> oh lol
| |
| 22:23 | i didn't think to look over there :)
| |
| 22:23 | since all the other stuff was in the other section
| |
| 22:23 | <fgiraldeau_> https://launchpad.net/~francis-giraldeau/+archive
| |
| 22:23 | <johnny> ssh keys, jabber, wiki ,etc
| |
| 22:23 | GiantPickle has quit IRC | |
| 22:23 | <johnny> i had trouble making package
| |
| 22:23 | <fgiraldeau_> apt-get build-dep ltsp
| |
| 22:23 | <johnny> i've been using gentoo for awhile now
| |
| 22:23 | <fgiraldeau_> then apt-get source ltsp
| |
| 22:24 | <johnny> so this whole packaging thing is weird for me
| |
| 22:24 | <fgiraldeau_> and dpkg-buildpackage -rfakeroot
| |
| 22:24 | <johnny> i read the docs for packaging for ubuntu, but it wasn't very concise
| |
| 22:24 | yes.. i do have fakeroot :)
| |
| 22:24 | <fgiraldeau_> From my point of view, it's gentoo that is wired ;)
| |
| 22:25 | <johnny> well i'm used to patching stuff for it
| |
| 22:25 | very easy
| |
| 22:25 | just add 1 line to the ebuild
| |
| 22:25 | and that's it
| |
| 22:25 | not all these steps
| |
| 22:25 | <fgiraldeau_> Hep, nothing is perfect.
| |
| 22:25 | <johnny> truly
| |
| 22:25 | it has other disadvantages for sure
| |
| 22:26 | but i did learn alot about gnu/linux internals from it
| |
| 22:26 | which has been advantageous when helping folks with other distros
| |
| 22:26 | except for the packaging aspect :)
| |
| 22:26 | <fgiraldeau_> I will write to ogra about the patch. I could remove lines added to inform the user that he missed something, but I think that it's part of the bug.
| |
| 22:27 | <johnny> i think that should be for another bug
| |
| 22:27 | seperate patch
| |
| 22:27 | cliebow has quit IRC | |
| 22:27 | sutula has quit IRC | |
| 22:27 | r3zon8 has quit IRC | |
| 22:27 | warren has quit IRC | |
| 22:27 | jbrett has quit IRC | |
| 22:27 | rjune has quit IRC | |
| 22:27 | shogunx has quit IRC | |
| 22:27 | mistik1 has quit IRC | |
| 22:27 | primeministerp has quit IRC | |
| 22:27 | elisboa has quit IRC | |
| 22:27 | cyberorg has quit IRC | |
| 22:27 | <fgiraldeau_> what the hell is that
| |
| 22:27 | <johnny> also talk to sbalneav about it, since he could certainly integrate it upstream
| |
| 22:27 | netsplit
| |
| 22:27 | it'll resolve itself shortly
| |
| 22:28 | <fgiraldeau_> ok
| |
| 22:28 | <johnny> happens often on bigger IRC networks
| |
| 22:28 | or really.. on any network that involves more than 1 server in my experience :)
| |
| 22:28 | they are now in an #ltsp that contains only them
| |
| 22:28 | <fgiraldeau_> What about usb key? does it works for you?
| |
| 22:28 | <johnny> yes
| |
| 22:28 | <fgiraldeau_> https://bugs.launchpad.net/ubuntu/+source/ltspfs/+bug/174585
| |
| 22:28 | <johnny> after following some steps in autologin bug
| |
| 22:29 | <fgiraldeau_> I do have a strange behavior.
| |
| 22:29 | The key is never unmounted.
| |
| 22:30 | <johnny> hmm.. aha.. i do have the behaviour where there is no unmount option that does anything
| |
| 22:30 | via the mount icon on the desktop
| |
| 22:30 | but they do work
| |
| 22:30 | hmm. now if only i was near my thin clients
| |
| 22:31 | you always come in at the weird times :)
| |
| 22:31 | good times for me usually tho :)
| |
| 22:31 | <fgiraldeau_> you mean that the key is umounted? You saw that?
| |
| 22:31 | <johnny> no, keys are not unmounted
| |
| 22:31 | iirc
| |
| 22:31 | <fgiraldeau_> Well, I'm going to sleep very soon. The evening is over
| |
| 22:31 | <johnny> oh
| |
| 22:31 | that is sad
| |
| 22:31 | when will you return
| |
| 22:31 | i was gonna go over there :)
| |
| 22:32 | <fgiraldeau_> I don't like much IRC, but I will fire it up tomorrow morning. Will try to send new patch to ogra and get it's feedback fast.
| |
| 22:33 | And see if usb key has a bad bug
| |
| 22:33 | ltspfsd in fact
| |
| 22:33 | Why do I get TCP checksum error? So bad.
| |
| 22:34 | Anyway, take care, and have a nice day. See you soon.
| |
| 22:34 | <johnny> bye
| |
| 22:34 | you too
| |
| 22:34 | for sure :)
| |
| 22:34 | <fgiraldeau_> bye!
| |
| 22:34 | fgiraldeau_ has left #ltsp | |
| 22:36 | rjune has joined #ltsp | |
| 22:36 | cliebow has joined #ltsp | |
| 22:36 | cyberorg has joined #ltsp | |
| 22:36 | r3zon8 has joined #ltsp | |
| 22:36 | elisboa has joined #ltsp | |
| 22:36 | primeministerp has joined #ltsp | |
| 22:36 | mistik1 has joined #ltsp | |
| 22:36 | shogunx has joined #ltsp | |
| 22:36 | warren has joined #ltsp | |
| 22:36 | sutula has joined #ltsp | |
| 22:36 | jbrett has joined #ltsp | |
| 22:43 | <mathesis> It is necessary to install dhcp3-server in ltsp5?
| |
| 22:44 | <johnny> do you have an /etc/ltsp/dhcpd.conf ?
| |
| 22:44 | if so, then you prolly already have one
| |
| 22:44 | i use dnsmasq instead tho
| |
| 22:45 | so i skip the normal one
| |
| 22:45 | and the tftpd as well
| |
| 22:45 | works pretty well
| |
| 22:46 | <mathesis> i have not /etc/ltsp/dhcpd.conf
| |
| 22:47 | I do not have the file /etc/ltsp/dhcp.conf
| |
| 22:50 | <johnny> which one did you install?
| |
| 22:50 | ltsp-server or ltsp-server-standalone?
| |
| 22:50 | one of them includes it
| |
| 22:50 | one of them doesj't i guess
| |
| 22:53 | cliebow has quit IRC | |
| 22:53 | sutula has quit IRC | |
| 22:53 | r3zon8 has quit IRC | |
| 22:53 | warren has quit IRC | |
| 22:53 | jbrett has quit IRC | |
| 22:53 | rjune has quit IRC | |
| 22:53 | shogunx has quit IRC | |
| 22:53 | mistik1 has quit IRC | |
| 22:53 | primeministerp has quit IRC | |
| 22:53 | elisboa has quit IRC | |
| 22:53 | cyberorg has quit IRC | |
| 22:56 | <mathesis> johnny, ltsp-server
| |
| 22:57 | <johnny> i forgot the exact differences
| |
| 22:57 | i'msure you can look them up
| |
| 22:57 | GiantPickle has joined #ltsp | |
| 22:58 | rjune has joined #ltsp | |
| 22:59 | <johnny> i prolly just followed the instructions on the ubuntu wiki
| |
| 23:01 | <mathesis> johnny, http://wiki.ltsp.org/twiki/bin/view/Ltsp/MueKow#Debian_development
| |
| 23:06 | bruno has joined #ltsp | |
| 23:06 | cliebow has joined #ltsp | |
| 23:06 | cyberorg has joined #ltsp | |
| 23:06 | r3zon8 has joined #ltsp | |
| 23:06 | elisboa has joined #ltsp | |
| 23:06 | primeministerp has joined #ltsp | |
| 23:06 | mistik1 has joined #ltsp | |
| 23:06 | shogunx has joined #ltsp | |
| 23:06 | warren has joined #ltsp | |
| 23:06 | sutula has joined #ltsp | |
| 23:06 | jbrett has joined #ltsp | |
| 23:06 | subir has joined #ltsp | |
| 23:07 | brno has joined #ltsp | |
| 23:08 | bruno has quit IRC | |
| 23:08 | brno is now known as bruno | |
| 23:10 | cliebow has quit IRC | |
| 23:10 | sutula has quit IRC | |
| 23:10 | r3zon8 has quit IRC | |
| 23:10 | warren has quit IRC | |
| 23:10 | jbrett has quit IRC | |
| 23:10 | shogunx has quit IRC | |
| 23:10 | mistik1 has quit IRC | |
| 23:10 | primeministerp has quit IRC | |
| 23:10 | elisboa has quit IRC | |
| 23:10 | cyberorg has quit IRC | |
| 23:10 | Faithful has quit IRC | |
| 23:12 | mistik1 has joined #ltsp | |
| 23:25 | sutula has joined #ltsp | |
| 23:26 | shogunx has joined #ltsp | |
| 23:26 | alekibango has quit IRC | |
| 23:26 | sutula has quit IRC | |
| 23:26 | cliebow has joined #ltsp | |
| 23:26 | cyberorg has joined #ltsp | |
| 23:26 | r3zon8 has joined #ltsp | |
| 23:26 | elisboa has joined #ltsp | |
| 23:26 | primeministerp has joined #ltsp | |
| 23:26 | warren has joined #ltsp | |
| 23:26 | sutula has joined #ltsp | |
| 23:26 | jbrett has joined #ltsp | |
| 23:26 | cliebow_ has joined #ltsp | |
| 23:26 | sutula- has joined #ltsp | |
| 23:26 | bruno has joined #ltsp | |
| 23:27 | elisboa has quit IRC | |
| 23:28 | primeministerp has quit IRC | |
| 23:28 | r3zon8 has quit IRC | |
| 23:29 | primeministerp has joined #ltsp | |
| 23:30 | cyberorg has quit IRC | |
| 23:31 | lnxnubie has joined #ltsp | |
| 23:31 | cyberorg has joined #ltsp | |
| 23:33 | <lnxnubie> hi, i'm in need of a "centralized authentication" solution, so can anyone tell me pros or cons of "plain LDAP+NFS " and a LTSP based solution
| |
| 23:35 | warren has quit IRC | |
| 23:37 | jbrett has quit IRC | |
| 23:37 | sutula has quit IRC | |
| 23:38 | sutula- is now known as sutula | |
| 23:41 | cliebow has quit IRC | |
| 23:43 | Egyptian[Home] has quit IRC | |
| 23:53 | warren has joined #ltsp | |
| 23:58 | edgarin_ has joined #ltsp | |
| 23:58 | edgarin has quit IRC | |