[04:59] Good morning. I can't get evolution to work with fat clients, it needs to create a lock in ~/.evolution and sshfs doesn't support that. [04:59] I wonder if an NFS _option_ would be accepted upstream... [10:19] Hi [10:20] I've got a problem with LTSP and disabling ssh [10:20] I've add this line in lts.conf without success : LDM_DIRECTX=True [10:35] g4tsu: what do you mean "disabling ssh" ? [10:36] when you connect the thinclient to the ltsp server [10:36] you make a ssh tunnel [10:36] (apologies for my bad english) [10:36] I search how to disable ssh [10:37] you cant [10:37] I find that I have to add a line in lts.conf [10:37] the tunnel is always there [10:37] really ? [10:37] LDM_DIRECTX just disabled the heavy transport of graphics through the tunnel [10:37] *disables [10:38] g4tsu: how are the users going to login if you disable ssh? [10:39] ok but when I add this line, It seems to be that the graphics support continue to through the tunnel [10:40] ssh -Y -t -S /var/run/ldm_socket_5030_192.168.1.68 -l adminsu 192.168.1.68 XDG_DATA_DIRS=/tmp/ltsp-localapps-adminsu-diDRKu/:/usr/local/share/:/usr/share/ LTSP_CLIENT=192.168.1.102 LTSP_CLIENT_HOSTNAME=ltsp102 LC_ALL=fr_FR.UTF-8 LANGUAGE=fr_FR.UTF-8 LANG=fr_FR.UTF-8 DISPLAY=192.168.1.102:7 PULSE_SERVER=tcp:192.168.1.102:4713 ESPEAKER=192.168.1.102:16001 /etc/X11/Xsession default < /dev/null > /dev/null ; /usr/sbin/ltspfsmounter all cleanup [10:41] I've got this line [10:41] So it's seems that Xsession use the ssh tunnel [10:41] DISPLAY=192.168.1.102:7 [10:41] it doesnt [10:41] ah ok [10:41] else DISPLAY would point to localhost [10:42] ok great [10:42] sshd sets up a local proxy for X on the server [10:42] if you use X forwarding through the tunnel [10:42] thanks for help me see things clearly [10:42] :) [13:37] Good Morning Edubunteros :) [13:48] Edubuntistas? [13:55] hi dgroos [13:56] Ahmuck: how are you? [14:02] heh, rebuilding my ltsp server [14:03] It's always so satisfying to do that, feels like one is getting so much done, using so much computing power, and one only need type in a little command... and wait. [14:10] heh [14:11] ubuntu has a problem with the forcedeth driver, as a result i usually have to search for a couple of hours how to fix it every time [14:11] the real problem is this problem has existed since hardy or prior, and keeps getting marked "won't fix" [14:12] in part because nobody can decide who should fix it [14:12] Well, I'd have to search for a couple hours to find out what a forcedeth driver is... [14:13] heh [14:13] it's the driver for nvidia chipsets on the mobo, the driver that drives the nic [14:14] the software behind it reads the nic wrong or the nvidia has implented it incorrectly, and upon every reboot it increases the number of the eth by 1 [14:15] so you end up with 300+ eth in 70-persistant-net-rules [14:15] ie, eth300 [14:15] making it impossible to set a static ip address for that interface [14:15] that sounds like one would want to rip out ones (that is ones computer's) innards. [14:15] and get a new one. [14:16] or put a new nic in a pci slot [14:16] :) [14:16] you can get the nic set, by using some hacks, but the problem is that the mac address still changes upon every reboot making outside access via ports impossible [14:17] er, some port implementation ... if your tunneling for vpn, etc. [14:17] as i understand it [14:17] anywho, it's a mess, and i would have expected it to have been fixed already by karamic considering how long it's been in the system [14:17] since gutsy [14:17] at least [14:17] I remember someone commenting here one day that they had a really high number for there eth port and when alkisg asked, I think this person's comment was something like, 'long story'. Now I know... [14:18] real long story. [14:18] heh [14:18] it happens on boards where the bios allows you to set your own nic [14:18] which is a really neat idea, cause it makes irc banning near impossible :) [14:23] irc banning? disallowing someone from participating in an irc channel? [14:23] yes [14:28] This page http://www.wikihow.com/wikiHow:IRC-Bans seems to make it sound like banning is a valuable thing? Not allowing someone to virtually sabotage a community? I'm missing something... [14:29] dictorial ops [14:29] who get up on the wrong side of the bed that day [14:30] Ahhh... [14:32] True I've never seen the need for it with my limited experience here on #edubuntu, but I can theoretically imagine a situation where it might be needed. There are some cats that are even stranger than I 'out there'... [14:33] democracy and free speech is a double edge sword. kings and plebs. even kings can get out of order, look at history and see the revolutions that have happened [14:34] you're telling me. [14:34] I see it often enough in educational leadership... [14:34] not evil, just misguided. [14:35] i agree [14:35] do you get dilbert cartoons where you are? [14:35] nope :( [14:36] i don't recall the last time i read a newspaper ... [14:37] I understand that! When you get a chance... http://www.dilbert.com/ [15:36] Morning all [15:41] Hi Scotty [15:43] sbalneav: have you ever used pam environment variables? I can't get pam to read my env vars.. [15:50] Hmmmm, No, but.... [15:51] http://freshmeat.net/projects/pam_envfeed [15:53] That'd be perfect, but I don't think we want it as an ltsp dependency... :-/ [15:54] there is a "pam_env" module.... I think lemme look [15:54] Yes, I was trying to use that [15:55] What's it doing/not doing? [15:55] But I could only make it read static variables from /etc/security/pam_env.conf [15:55] You want something dynamic [15:55] So I couldn't make it set CKCON_X11_DISPLAY=$DISPLAY [15:55] (i.e. read DISPLAY or some other environment variable) [15:56] yeah, pam_env won't do that, that's why you'd want pam_envfeed. [15:56] I think that "pam environment" is a special, system-wide environment... I'd like a per-session environment [15:56] Want me to package it? [15:56] Nah... [15:56] if we need it, then we need it. [15:56] If it comes to that, I'll just use a python function instead [15:57] i.e. call OpenSessionWithParameters to properly notify consolekit [15:57] I've just committed a workaround that ...works, but not a proper one [15:58] If we won't find any better way in the next weeks, then I'll just replace `su user` with a python function [15:58] (i was trying to avoid it because not all distros are using consolekit, so I preferred environment variables + su) [16:01] Well, pam_envfeed could be useful to us for other things. [16:02] If we package it, and make it useful for debian, and potentially other distros, it might become a standard we could rely on. [16:02] That's true... /me shrugs [16:02] Well, tell you what. I'll package it, and we can look at it post-lucid. [16:02] highvoltage, mgariepy, you might want to try the latest commit in ldm-trunk, it makes the fat client sessions CK/PK aware (so device mounting works) [16:02] sbalneav: sounds good to me [16:03] Ah, we probably need to disable LOCALDEV for fat clients, too [16:03] nice alkisg thanks, i won't be able to test this today though. [16:04] np [16:06] alkisg: cool! [16:11] There's also http://manpages.ubuntu.com/pam_exec - maybe that could be used instead... [16:18] Not sure if it will preserve the enviornment... We could try.. [16:20] I'm only interested in modifying the pam environment temporarily [16:20] Not the process environment... [16:20] I.e. pam must "see" CKCON_X11_DISPLAY in its "pam environment" in order for `su` to work correctly, but we don't need that variable in the session environment [16:21] (I hope I'm making sense...) [16:22] yeah, makes sense. [17:47] highvoltage: can you send me your newusers script, to try to make it a little faster? [17:48] e.g. we could use pam_mkhomedir to make the home dirs for only the users that actually log in... [17:53] alkisg: yep [17:54] alkisg: ok this is sttill one with a bashism: [17:54] for user in {1..255}; do echo "ltsp$user:ltsp$user:200$user:1000:LTSP Guest:/home/ltsp$user:/bin/sh" >> /tmp/userlist newusers /tmp/userlist [17:55] done [17:55] ok, ty [17:55] ok that doesn't paste well [17:55] but you get it [18:08] highvoltage: hi there, is there a list of uncompleted taskss before release? [18:08] some place... [18:09] public preferably :-) [18:10] nubae: our meeting notes usually contain to-do items, otherwise it's in tickets in LP [18:11] nubae: the website work is listed out there quite well, besides that we basically have some artwork issues left and getting the liveltsp and ltsp-installer in [18:11] nubae: oh and the squashfs required for that [18:14] ok, can I get started on something? whats most urgent? [18:15] I've been away for quite a while and want to show that I still care :-) heh [18:17] can I see the to do items? [18:17] I'd also like to run something by you, get your ok... [18:17] I would like to link linux-for-education.org to edubuntu.org [18:17] just a link [18:17] for extra help [18:18] the lforE site has a ubuntu specific part, so we could redirect users to there [18:19] especially the howto jazz up you ubuntu instalation, etc. [19:03] nubae: I think the best would be to get an edubuntu iso, boot from it (even if in a virtualmachine) and find/fix bugs [19:04] nubae: I think at this stage in the release cycle that's quitee critical and we can definitely do more of that right now [19:04] nubae: not sure how good your packaging is, when we get the new logos from canonical there will be quite some artwork work to be done [21:18] new edubuntu logos? [21:29] Would anyone care (or mind) for also allowing a local partition (identified by LOCAL_HOME=volume-name) to be mounted as /home for fat clients? It shouldn't need more than 10 lines... [21:32] ...anyway, i'll leave that post-lucid === joerg_ is now known as joerg [22:44] hi folks