IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 2 October 2007 (all times are UTC)
| 00:09 | Q-FUNK has joined #ltsp | |
| 00:25 | twinprism has quit IRC | |
| 00:41 | Q-FUNK has joined #ltsp | |
| 00:58 | cyberorg has joined #ltsp | |
| 01:01 | mikkel has joined #ltsp | |
| 01:26 | plamengr has joined #ltsp | |
| 01:37 | MagicStorm has quit IRC | |
| 02:01 | ace_suares has joined #ltsp | |
| 02:05 | frame_asys has joined #ltsp | |
| 02:07 | frame_asys has left #ltsp | |
| 02:10 | Daggett has joined #ltsp | |
| 02:11 | Jenna has joined #ltsp | |
| 02:34 | vagrantc has quit IRC | |
| 02:42 | ogra has joined #ltsp | |
| 02:45 | tux_440volt has joined #ltsp | |
| 03:08 | tux_440volt has quit IRC | |
| 03:24 | cyberorg has quit IRC | |
| 03:43 | Amaranth_ has joined #ltsp | |
| 03:44 | cyberorg has joined #ltsp | |
| 03:47 | Amaranth has quit IRC | |
| 03:54 | Amaranth_ is now known as Amaranth | |
| 03:55 | meduxa has joined #ltsp | |
| 04:26 | fernando110 has quit IRC | |
| 04:46 | Jenna has quit IRC | |
| 04:47 | Jenna has joined #ltsp | |
| 04:49 | Avatara has joined #ltsp | |
| 04:50 | fernando110 has joined #ltsp | |
| 05:02 | doggkatcha has joined #ltsp | |
| 05:02 | doggkatcha has left #ltsp | |
| 05:17 | meduxa has quit IRC | |
| 05:21 | subir has joined #ltsp | |
| 05:30 | Barbosa has joined #ltsp | |
| 05:39 | Jenna has quit IRC | |
| 05:42 | Clown-MG has joined #ltsp | |
| 05:55 | Q-FUNK has quit IRC | |
| 05:56 | jernst has joined #ltsp | |
| 05:58 | jernst has quit IRC | |
| 06:12 | cdealer__ has joined #ltsp | |
| 06:13 | <cdealer__> ogra: Hi
| |
| 06:13 | <ogra> hey
| |
| 06:13 | <cdealer__> ogra: after a reboot my clients cant loggin anymore... anyideia ?
| |
| 06:13 | <ogra> did your IP data change in any way ?
| |
| 06:14 | <cdealer__> ogra: maybe... but they are coming to the login screen .... but when they insert the username and password they back to the login screen and dont login
| |
| 06:15 | <ogra> so did the IP data of the server change ? if so, run the ltsp-update-sshkeys command to adjust the keys for login
| |
| 06:16 | <cdealer__> ogra: Hmmm god ideia ... =D
| |
| 06:17 | ogra: I knew that was something simple hahaha thank you =D
| |
| 06:17 | cdealer__ has quit IRC | |
| 06:22 | nf1 has quit IRC | |
| 06:55 | Guaraldo has joined #ltsp | |
| 06:58 | Guaraldo has quit IRC | |
| 06:58 | Guaraldo has joined #ltsp | |
| 07:00 | Guaraldo has quit IRC | |
| 07:00 | Guaraldo has joined #ltsp | |
| 07:04 | Guaraldo has quit IRC | |
| 07:04 | Guaraldo has joined #ltsp | |
| 07:15 | frame_asys has joined #ltsp | |
| 07:15 | <frame_asys> /msg NickServ IDENTIFY asoo
| |
| 07:15 | <jammcq> oops
| |
| 07:15 | <frame_asys> f***
| |
| 07:16 | frame_asys has left #ltsp | |
| 07:18 | nf1 has joined #ltsp | |
| 07:18 | nf1 has joined #ltsp | |
| 07:21 | frame_asys has joined #ltsp | |
| 07:22 | subir has quit IRC | |
| 07:24 | jammcq has quit IRC | |
| 07:44 | MagicStorm has joined #ltsp | |
| 07:53 | FernandoMM has joined #ltsp | |
| 08:04 | tiagovaz has joined #ltsp | |
| 08:05 | <sbalneav> Morning all
| |
| 08:08 | cliebow2 has joined #ltsp | |
| 08:18 | <frame_asys> morning
| |
| 08:19 | muh2000 has quit IRC | |
| 08:22 | joebaker has joined #ltsp | |
| 08:29 | Gadi has joined #ltsp | |
| 08:35 | ogra has quit IRC | |
| 08:38 | mikkel has quit IRC | |
| 08:39 | ogra has joined #ltsp | |
| 08:46 | Egyptian[Home] has quit IRC | |
| 08:57 | tux_440volt has joined #ltsp | |
| 08:58 | tiagovaz_ has joined #ltsp | |
| 08:59 | tiagovaz has quit IRC | |
| 09:03 | Egyptian[Home] has joined #ltsp | |
| 09:09 | cliebow2 has quit IRC | |
| 09:16 | MeBadMagic is now known as BadMagic | |
| 09:17 | diadicic has joined #ltsp | |
| 09:21 | <diadicic> Having a small problem. My clients seem to be ignoring my "/opt/ltsp/i386/etc/lts.conf" file. Any changes I make to it do nothing.
| |
| 09:22 | <sbalneav> diadicic: What version of ltsp are you running?
| |
| 09:22 | <diadicic> 4.2 I believe
| |
| 09:22 | <sbalneav> And, cand you paste your lts.conf to the pastebot?
| |
| 09:22 | !pastebot
| |
| 09:22 | <ltspbot`> sbalneav: Error: "pastebot" is not a valid command.
| |
| 09:22 | <sbalneav> hrrm
| |
| 09:23 | pastebot.ltsp.org
| |
| 09:23 | <diadicic> sbalneav: Ok I am figuring out how to do that. Hang on.
| |
| 09:25 | <ltsppbot> "diadicic" pasted "lts.conf" (28 lines) at http://pastebot.ltsp.org/324
| |
| 09:25 | tiagovaz__ has joined #ltsp | |
| 09:26 | <diadicic> Ok, I hope it worked.
| |
| 09:27 | tiagovaz_ has quit IRC | |
| 09:27 | cliebow_ has joined #ltsp | |
| 09:27 | <sbalneav> Looks ok
| |
| 09:28 | <diadicic> well I am really only concerned with [ws240]. I need that client to run at 640x480. I ve also tried to force that use to run a shell just to test it out and it did not work.
| |
| 09:29 | I guess I will change the default and see if that does anything first. Then I ll have to take it from there.
| |
| 09:30 | <sbalneav> Are you sure your workstations know their hostnames?
| |
| 09:30 | tux_440volt has quit IRC | |
| 09:30 | jammcq has joined #ltsp | |
| 09:30 | <sbalneav> Maybe try addressing via the mac address.
| |
| 09:30 | <jammcq> g'morning all
| |
| 09:30 | <diadicic> thats a though
| |
| 09:30 | <sbalneav> [00:11:22:33:44:55]
| |
| 09:30 | morning jammcq
| |
| 09:31 | <diadicic> sbalneav: If I logon to a workstation and type $DISPLAY that should give me its name. Correct?
| |
| 09:31 | <sbalneav> No
| |
| 09:31 | echo $DISPLAY
| |
| 09:32 | <diadicic> sorry thats what I ment. It would have taken me about 10 tries but I think I would have gotten it.
| |
| 09:32 | <sbalneav> but that may not be what the workstation thinks it it.
| |
| 09:32 | it is.
| |
| 09:32 | sepski has joined #ltsp | |
| 09:33 | <diadicic> How does the workstation know what it is?
| |
| 09:33 | <sbalneav> do a SCREEN_02 = shell, and then do a hostname on the thin client to find out what it really thinks it's hostname is.
| |
| 09:33 | from dhcpd
| |
| 09:33 | <diadicic> ok
| |
| 09:35 | after I edit the lts.conf. Do I have to reboot the server itself?
| |
| 09:40 | ok. It reports it self at ws240.ltsp
| |
| 09:41 | ace_suares has quit IRC | |
| 09:41 | ace_suares has joined #ltsp | |
| 09:42 | <diadicic> sbalneav: That was it. It's all better now. Thanks alot.
| |
| 09:48 | tomcats_ has quit IRC | |
| 09:49 | doggkatcha has joined #ltsp | |
| 09:49 | <diadicic> Thanks again
| |
| 09:49 | diadicic has quit IRC | |
| 10:08 | cyberorg has quit IRC | |
| 10:09 | staffencasa has joined #ltsp | |
| 10:14 | DonSilver has joined #ltsp | |
| 10:19 | cyberorg has joined #ltsp | |
| 10:25 | tiagovaz__ has quit IRC | |
| 10:32 | tiagovaz has joined #ltsp | |
| 10:32 | ltspbot has joined #ltsp | |
| 10:33 | <sbalneav> !pastebot
| |
| 10:33 | <ltspbot> sbalneav: "pastebot" is The LTSP pastebot is at http://pastebot.ltsp.org. Please paste all text longer than a line or two to the pastebot, as it helps to reduce traffic in the channel. A link to the content will be pasted in the channel.
| |
| 10:33 | <sbalneav> there we go
| |
| 10:33 | factoids are back.
| |
| 10:35 | <cliebow_> \!
| |
| 10:41 | BadMagic has quit IRC | |
| 10:41 | BadMagic has joined #ltsp | |
| 10:42 | cdealer__ has joined #ltsp | |
| 10:42 | <cdealer__> ogra: hey ...
| |
| 10:42 | cdealer__: have you noticed any problem with java jvm ?
| |
| 10:43 | ogra: in our tests we have noticed that sometimes the cpu use of jvm comes to 100% and the machine stops to respond and them we have to kill the session and start all overagain
| |
| 10:44 | Q-FUNK has joined #ltsp | |
| 10:44 | <ogra> ask sun i guess :) i only use the java briwser plugin here and even that one not more often than twice a year i guess
| |
| 10:44 | *browser
| |
| 10:44 | Avatara has quit IRC | |
| 10:45 | <cdealer__> ogra: Hmm we use a oracle ias aplication that uses a java applet... now the user are using for 2 hours without problem and no problem, but before we had some freezing problems = very strange
| |
| 10:45 | <ogra> hmm, weird
| |
| 10:46 | i havent heard about an issues yet
| |
| 10:46 | <rjune> ogra: howdy
| |
| 10:46 | <ogra> hey rjune
| |
| 10:54 | <cdealer__> ogra: is there anyway to force some app to run on the client (using the client resources) ?
| |
| 10:55 | <ogra> cdealer__, not easily, no currently you would have to set up a local ssh server, make sure the users /home is mounted on teh client (for the keys and apps) and need a network based auth mechanism that authenticates the client in parallel to the server session
| |
| 10:55 | tiagovaz_ has joined #ltsp | |
| 10:56 | <ogra> sbalneav already works on a really sweet solution for this but that wont be in before ubuntu hardy starts
| |
| 10:58 | <cdealer__> ogra: you know how to do it ? This java aplication must run smooth with no problems... i said already that we are open to consulting dont i? O.o
| |
| 10:58 | =D
| |
| 10:59 | tiagovaz has quit IRC | |
| 11:00 | K_O-Gnom has joined #ltsp | |
| 11:00 | <ogra> well, you could probably play guineapig for scott ;)
| |
| 11:01 | <sbalneav> I have a nice simple solution in it's embryonic stages.
| |
| 11:05 | tux_440volt has joined #ltsp | |
| 11:10 | <cdealer__> sbalneav: are you talking about my problem ??
| |
| 11:11 | DonSilver has quit IRC | |
| 11:12 | <ogra> cdealer__, yes
| |
| 11:13 | <cdealer__> hmm im very open to ideas... but just hope to be something I can do
| |
| 11:16 | vagrantc has joined #ltsp | |
| 11:18 | <sbalneav> I'd say I'd be at least a month or two away before I'd have something I was comfortable with releasing
| |
| 11:19 | And even then, there's no guarenteeing that the other LTSP developers will like my idea.
| |
| 11:20 | * ogra does | |
| 11:20 | <ogra> and we'll just corner vagrantc :)
| |
| 11:20 | * vagrantc wonders what idea this is | |
| 11:20 | <sbalneav> method for handling localapps.
| |
| 11:21 | In a nutshell:
| |
| 11:21 | <ogra> vagrantc, remenmer Gadi's idea for using getpwent ?
| |
| 11:21 | <sbalneav> 1) install libnss-extrafiles in the chroot
| |
| 11:21 | <ogra> scott thought that to an end :)
| |
| 11:22 | <sbalneav> 2) ldm has login script that gets user and group data from host
| |
| 11:22 | <ogra> in a really sexy way imho :)
| |
| 11:22 | (as usual indeed)
| |
| 11:22 | <sbalneav> 3) user and group data added to extra passwd and group file
| |
| 11:23 | 4) small (~200 lines so far) daemon sits and watches for an xprop on the root window to change.
| |
| 11:23 | 5) when xprop changes, reads the prop, that's a command line, and executes it.
| |
| 11:23 | no sshd, or authentication needed.
| |
| 11:24 | xprop -root -f LTSP_USERNAME 8s -set LTSP_USERNAME "sbalneav"
| |
| 11:24 | <vagrantc> is it possible for the user to fake the user/group data and get access to someone else's files?
| |
| 11:25 | <ogra> i have a vague ide about an Xsession.d scritpt that merges xdg dirs to replace the right .desktop files
| |
| 11:25 | <sbalneav> xprop -root -f LTSP_PROGRAM 8s -set LTSP_PROGRAM "firefox"
| |
| 11:25 | Rplata has joined #ltsp | |
| 11:25 | <sbalneav> vagrantc: well, here's my check for that:
| |
| 11:26 | on login, you get the user's passwd entry, and add it to the extra passwd file on the chroot, via ldm's rc scripts
| |
| 11:26 | <joebaker> The id command shows which groups the user is a member of. Your script should also bring across all the group ids the user is a member of. This sounds pretty neat.
| |
| 11:27 | <ogra> the nice thing is that it works without needing sshd
| |
| 11:27 | <sbalneav> the xrexec daemon looks at the username, and does a getpwname(LTSP_USERNAME), if that returns "null", then you can't run anything.
| |
| 11:27 | <ogra> hah ....
| |
| 11:27 | <joebaker> For instance user joebaker is a member of group nelfc. We user a technique called User Private Groups here with a umask of 002.
| |
| 11:27 | <sbalneav> so you'd only have the user you're currently running as known as the local terminal.
| |
| 11:28 | <ogra> any local desktop app we'll install will have a .desktop file installed in the clinet ...
| |
| 11:28 | lets have an ldm script that fills a LOCALAPPS env var with a list of these apps
| |
| 11:28 | then we have it available in the session ...
| |
| 11:28 | <vagrantc> so ... how is LTSP_PROGRAM actually used ?
| |
| 11:28 | <sbalneav> On logout, you simply erase those extra files.
| |
| 11:29 | Q-FUNK has quit IRC | |
| 11:29 | <sbalneav> LTSP_PROGRAM is the X property that you pass the command line of the program you want to run.
| |
| 11:29 | <vagrantc> and what runs it?
| |
| 11:29 | <joebaker> Is there a url for the specifications page for this program?
| |
| 11:29 | <sbalneav> This program I've written.
| |
| 11:29 | <ogra> vagrantc, an app on the client
| |
| 11:30 | Rplata has quit IRC | |
| 11:30 | <sbalneav> it sits and listens for X property events, and then exec's the program passed in the property.
| |
| 11:30 | <vagrantc> this app also handles home dir mounting?
| |
| 11:30 | <sbalneav> Right now, I just mount /home on the thin client.
| |
| 11:30 | <ogra> ldm can do that
| |
| 11:31 | <sbalneav> but yeah, something more fine grained can be worked up.
| |
| 11:31 | right now, it's barely proof of concept.
| |
| 11:31 | tux_440volt has quit IRC | |
| 11:31 | <ogra> even via sshfs (for apps that work with it at least)
| |
| 11:31 | <vagrantc> and so, you're basically adding the user's passwd/group info locally
| |
| 11:31 | <sbalneav> right.
| |
| 11:31 | from the server, at login time.
| |
| 11:31 | <vagrantc> and then looking for a setting in x properties ... and running it
| |
| 11:32 | <sbalneav> so you avoid having to set up network authentication on the client.
| |
| 11:32 | <vagrantc> which would be *slick*
| |
| 11:32 | ogra1 has joined #ltsp | |
| 11:32 | elisboa has quit IRC | |
| 11:33 | elisboa has joined #ltsp | |
| 11:33 | <vagrantc> well, it sounds worth exploring, for sure... though doing unusual authentication schemes always makes me a bit nervous.
| |
| 11:34 | <sbalneav> This little xrexec daemon, right now, is only around 200 lines of code.
| |
| 11:34 | Well, you're not actually authenticating. :)
| |
| 11:34 | You're just using nss to provide user and group data.
| |
| 11:34 | <vagrantc> well, "working around well-known authentication systems" always makes me a bit nervous :)
| |
| 11:34 | <ogra1> vagrantc, ubuntu wont add it without libnss-extrafiles being in main which means full security review
| |
| 11:35 | <sbalneav> the authentication's by dint of the fact that you've already GOT a plumbed X connection :)
| |
| 11:35 | Well, we could just add/remove from /etc/passwd and /etc/group on the thin client, but that's more gross, imho
| |
| 11:35 | with the extrafiles, you can just delete the files on logout.
| |
| 11:36 | * vagrantc can't find libnss-extrafiles | |
| 11:36 | * ogra1 could neither :) | |
| 11:36 | <sbalneav> p libnss-extrausers - nss module to have an additional passwd, s
| |
| 11:37 | Sorry, I always call it the wrong thing.
| |
| 11:37 | it's written by the same guy as libnss-db, which is in main, I think
| |
| 11:37 | it's small. Only a few hundred lines.
| |
| 11:38 | From my point of view, it means:
| |
| 11:38 | 1) We can take advantage of a package that's already there, for the user details
| |
| 11:39 | 2) The xrexec daemon is (so far) nice and small, and doesn't have to do much, so it's fairly easily maintained
| |
| 11:40 | 3) We avoid the whole "LDAP?! Kerberos??! Active Directory?!? How we gonna support all this in a thin client" mess.
| |
| 11:40 | Server handles authentication. if "getent passwd $USER" runs on the server, this will work(!)
| |
| 11:41 | <ogra1> !
| |
| 11:41 | <vagrantc> since ldm is running getent as the user, how do we prevent them from providing false data ?
| |
| 11:42 | <sbalneav> well, the script that gets the info runs from the croot.
| |
| 11:42 | <vagrantc> but if the user has some weird LD_PRELOAD environment or something ...
| |
| 11:42 | full paths would mostly handle the issue...
| |
| 11:43 | <sbalneav> yeah
| |
| 11:43 | It has the potential to be small and well contained.
| |
| 11:43 | <vagrantc> so, the LTSP_PROGRAM xprop setting ... can you run multiple programs locally ?
| |
| 11:44 | or rather, how do you run multiple programs locally
| |
| 11:44 | Daggett has quit IRC | |
| 11:45 | <vagrantc> or does the daemon delete the LTSP_PROGRAM xprop thingy once it executes the program ?
| |
| 11:45 | gonzaloaf has quit IRC | |
| 11:47 | <jammcq> oh jeez, not the old 'LDAP Scares the hell out of us, so lets not use it" thing
| |
| 11:48 | <vagrantc> if we can find a solid, simpler alternative, i'm all for it.
| |
| 11:48 | ogra has quit IRC | |
| 11:49 | <ogra1> jammcq, we can still use ldap ... just not twice ;)
| |
| 11:49 | whatever the server has configured will be used for authentication
| |
| 11:49 | <vagrantc> simply setting LOCAL_APPS=True in lts.conf ... should really be that simple.
| |
| 11:49 | <ogra1> ;)
| |
| 11:49 | vagrantc, not even that
| |
| 11:50 | vagrantc, have a script that checks /usr/share/applications for .desktop files
| |
| 11:50 | if there are any, you have a local app
| |
| 11:51 | <vagrantc> but the client-side needs to know about weather or not LOCAL_APPS should be enabled.
| |
| 11:51 | and the admin should be able to disable it, if they want to.
| |
| 11:51 | <jammcq> and whether or not to mount the users homedir
| |
| 11:51 | <ogra1> well, default to true if we find anything and the user hasnt set "False"
| |
| 11:52 | jammcq, indeed you always mount it by default ...
| |
| 11:52 | <jammcq> and then, there's a HUGE problem with gconfd
| |
| 11:52 | can't have a gconfd on the client AND on the server, both accessing the same gconf database
| |
| 11:52 | * vagrantc would *not* want homedirs mounted by default | |
| 11:52 | <ogra1> did you test that ?
| |
| 11:52 | <jammcq> ogra1: no, others have, like eHarrison
| |
| 11:53 | <ogra1> afaik gconfd is bound to $DISPLAY
| |
| 11:53 | * jammcq wonders what 'bound to $DISPLAY' means in this context | |
| 11:53 | <vagrantc> DISPLAY=:0.0 and DISPLAY=localhost:11.0
| |
| 11:53 | <ogra1> hmm, we'd finally need that dbus x protocol
| |
| 11:53 | <vagrantc> you've usually got two displays with LDM
| |
| 11:53 | <ogra1> jammcq, it wont start twice on one i mean
| |
| 11:53 | <jammcq> you'd still have 2 gconfd daemons running. one on the server, for the server-run applilcations. and one on the client, for the client-run applications
| |
| 11:54 | <ogra1> depends on the mode you use, not with LDM_DIRECTX for example
| |
| 11:54 | <jammcq> and BOTH would try updating the gconf database, and THERE is the problem
| |
| 11:54 | <vagrantc> so, some apps, due to bugs in those apps, can't be supported in LOCAL_APPS until those bugs are fixed
| |
| 11:54 | <jammcq> ok, ALL GNOME apps would fall under that category
| |
| 11:55 | <vagrantc> sure, and thus the impetus to get those bugs fixed.
| |
| 11:55 | <ogra1> yeah, you need a dbus channel ...
| |
| 11:55 | <vagrantc> jammcq: what do you propose other than fixing those bugs?
| |
| 11:55 | <jammcq> I agree we need a dbus channel. I'm just not sure how it fixes the gconfd problem
| |
| 11:55 | <ogra1> then the same gconfd for both can be used
| |
| 11:55 | apps talk to it via dbus
| |
| 11:56 | <jammcq> vagrantc: certainly bugs need to be fixed. I'm thinking we need a gnome person to help make gconfd network-aware
| |
| 11:56 | <ogra1> if the app finds a dbus socket and gets a proper reply from gconfd on the bus, it wont attempt to start a local one
| |
| 11:56 | * vagrantc hasn't yet started X today | |
| 11:56 | <jammcq> ogra1: that's probably the sane approach
| |
| 11:57 | <ogra1> not gconfd needs network awareness ;) dbus needs it :P
| |
| 11:57 | <jammcq> well, then gconfd needs dbus-awareness
| |
| 11:57 | chris_punches has joined #ltsp | |
| 11:57 | <ogra1> gconfd *uses* dbus :)
| |
| 11:57 | <jammcq> not for actually writing gconf updates, eh?
| |
| 11:57 | <ogra1> at least in 2.20
| |
| 11:58 | no, to talk to the apps
| |
| 11:58 | you wont get around a database
| |
| 11:59 | so ~/.gconf wont go away
| |
| 11:59 | <jammcq> but how many processes will actually touch it?
| |
| 11:59 | just one? on the server?
| |
| 11:59 | gonzaloaf has joined #ltsp | |
| 12:00 | <ogra1> yeah
| |
| 12:00 | the one in the session
| |
| 12:00 | <jammcq> and all other apps running not-on-the-server would communicate with the gconfd running on the server, via dbus?
| |
| 12:00 | <ogra1> right
| |
| 12:00 | <jammcq> make sense
| |
| 12:00 | doggkatcha has quit IRC | |
| 12:00 | <jammcq> but is any of that there now?
| |
| 12:00 | or is this yet-to-be-written?
| |
| 12:00 | <ogra1> we wanted hal anyway ... so dbus will come free :)
| |
| 12:01 | * jammcq likes-using-hyphens-when-he-talks | |
| 12:01 | <ogra1> dbus is missing the networking bits
| |
| 12:01 | <jammcq> from what I've heard, some of the networking bits are actually there, but not well tested
| |
| 12:01 | <ogra1> even IPC with netcat handles that better i get
| |
| 12:01 | probably time to take a look
| |
| 12:02 | <jammcq> and.... what does all this mean to the kubuntu/xubuntu folks?
| |
| 12:02 | <ogra1> no idea for kubuntu
| |
| 12:03 | xubuntu uses gconfd afaik
| |
| 12:03 | so for gconf apps it would be the same
| |
| 12:03 | <jammcq> i'm thinking we can't solve all the worlds problems, but it would be nice to have a direction to tell the kde guys
| |
| 12:04 | ok, so lets say that's how we handle the gconfd issue. what about mounting home dirs?
| |
| 12:04 | nfs?
| |
| 12:04 | fuse+reverse-ltspfs ?
| |
| 12:04 | samba? (can't believe I actually said that)
| |
| 12:05 | <vagrantc> it wouldn't be the first time LTSP used samba ...
| |
| 12:05 | * jammcq considers that a part of ltsp's DARK history | |
| 12:06 | <vagrantc> ltspfs is sure interesting.
| |
| 12:06 | * jammcq thinks so too | |
| 12:06 | <jammcq> well, a modified ltspfs
| |
| 12:06 | <vagrantc> modified ?
| |
| 12:06 | <jammcq> to work the other way
| |
| 12:06 | client-mounts-server, instead of server-mounts-client
| |
| 12:07 | * vagrantc doesn't see a need to modify it, just need to have both client and server bits installed on both the client and the server | |
| 12:07 | <jammcq> well, changes would be required. now, you'd have a daemon running on the server that MANY thin clients would be connecting to
| |
| 12:07 | you'd have to manage ports
| |
| 12:07 | and access control
| |
| 12:07 | <vagrantc> ah. fair enough.
| |
| 12:08 | <jammcq> dunno how involved that would be, but I seriously doubt the current implementation would work out of the box
| |
| 12:08 | I do know a guy who's an expert on ltspfs tho
| |
| 12:08 | <vagrantc> it's all black magic to me
| |
| 12:09 | <ogra1> the binaries should be usable the other way roud
| |
| 12:10 | the scripts wont
| |
| 12:11 | for the daemon side thats easily solved by a ldm rc script ... you can start it with the session
| |
| 12:11 | get the next free port and make sure the client knows about it
| |
| 12:11 | tux_440volt has joined #ltsp | |
| 12:12 | Amaranth_ has joined #ltsp | |
| 12:12 | fernando110 has quit IRC | |
| 12:13 | <ogra1> actually that should be easy with the binaries we have with some trivial scripting
| |
| 12:13 | cdealer__ has quit IRC | |
| 12:14 | Amaranth has quit IRC | |
| 12:14 | Amaranth_ is now known as Amaranth | |
| 12:23 | ogra1 is now known as ogra | |
| 12:25 | Barbosa has quit IRC | |
| 12:30 | <cliebow_> /join #melug
| |
| 12:32 | cliebow__ has joined #ltsp | |
| 12:32 | <cliebow__> m
| |
| 12:43 | cliebow__ has quit IRC | |
| 12:44 | fernando110 has joined #ltsp | |
| 12:45 | muh2000 has joined #ltsp | |
| 12:55 | FernandoMM has quit IRC | |
| 12:58 | tux_440volt has quit IRC | |
| 13:00 | Guaraldo has quit IRC | |
| 13:04 | Guaraldo has joined #ltsp | |
| 13:04 | tux_440volt has joined #ltsp | |
| 13:06 | Barbosa has joined #ltsp | |
| 13:10 | Guaraldo has quit IRC | |
| 13:12 | Guaraldo has joined #ltsp | |
| 13:32 | mikkel has joined #ltsp | |
| 13:35 | <sbalneav> Using ltspfs on the server would be bad.
| |
| 13:36 | Every. Single. 512. Byte. Block. Would. Genereate. An. Open. Seek. Read. Close. Sequence. The. Implications. For. Performance. On. The. Server. Would. Be. Staggering.
| |
| 13:36 | <vagrantc> oh well.
| |
| 13:38 | * Gadi reads up on discussion of the day, and wants to put his $0.02 and vote in for sshfs+some-other-stackable-fuse-fs-to-handle-stat | |
| 13:38 | * Gadi uses jammcq-hyphen-notation | |
| 13:38 | <sbalneav> I fail to see what's the problem with NFS.
| |
| 13:39 | <vagrantc> sshfs, or at least using the secure channel we already have... seems like a nice idea- using the secure channel we already have
| |
| 13:40 | with NFS, you basically are giving write permission to user's data across an insecure medium ...
| |
| 13:40 | <sbalneav> We're doing that now.
| |
| 13:41 | * vagrantc doesn't have any environment's where the home directories are writeable through insecure filesystems | |
| 13:41 | <sbalneav> Then maybe that's an enironment where you don't run localapps.
| |
| 13:41 | ltspfs would be LESS secure than nfs.
| |
| 13:41 | <vagrantc> sure
| |
| 13:42 | well, sure that ltspfs will be less secure than nfs...
| |
| 13:43 | i would like to be able to use local apps and have a more secure filesystem, and will continue to keep an eye out for how to make that possible.
| |
| 13:44 | <sbalneav> Well, you'd either have to modify sshfs, but that wouldn't be secure either.
| |
| 13:44 | In fact, until we do someting to store the ssh keys on a local medium on the thin client, none of the ssh stuff we've done is secure either.
| |
| 13:44 | <jammcq> sshfs ain't the answer. no freespace capability
| |
| 13:44 | <sbalneav> it's just obfuscation :)
| |
| 13:45 | <vagrantc> there are layers of security
| |
| 13:45 | <jammcq> i'd take obfuscation as an intermediate step
| |
| 13:46 | <vagrantc> security is not boolean ... there's a big difference between sending all your data via unencrypted XDMCP vs. through an ssh in which the ssh binary is coming from an insecure filesystem
| |
| 13:46 | <jammcq> tru dat
| |
| 13:46 | <sbalneav> The keys are also going across the network in cleartext.
| |
| 13:46 | like I say, it's obfuscation.
| |
| 13:47 | <jammcq> but there's a difference between a casual listener, and one who actively tries to grab the keys, and work them into a solution for monitoring all traffic
| |
| 13:47 | <vagrantc> the client-side key is generated at run-time, and the known_hosts data is cleartext
| |
| 13:47 | <jammcq> obviously, the most secure thing would be to unplug all devices from the network
| |
| 13:47 | but that's hardly a good choice either
| |
| 13:48 | Q-FUNK has joined #ltsp | |
| 13:48 | <vagrantc> to compromise it, you would have to consistantly hijack the NFS connection and insert your own known_hosts data, and grab and ensure all ip traffic goes to your stolen ip.
| |
| 13:48 | (or NBD connection)
| |
| 13:49 | <jammcq> the security issue with NFS isn't the fact that somebody can snoop and see the data. the problem is that ANYBODY with a laptop can connect up and modify YOUR files
| |
| 13:49 | student has laptop, plugs into network and modifies the grades in the teachers home directory
| |
| 13:49 | twinprism has joined #ltsp | |
| 13:49 | <jammcq> it's an authentication problem, more than an encryption problem. imho
| |
| 13:50 | * jammcq notes that's the first time he's ever used "imho" | |
| 13:51 | <sbalneav> Well, there's two problems with sshfs:
| |
| 13:52 | 1) The freespace issue. This can be handled by stacking a small, custom fuse filesystem on top that passes through most stuff to the sshfs filesystem, execpt the statfs call, which it handles out of band.
| |
| 13:52 | <Q-FUNK> all portmap-based services suck, anyhow
| |
| 13:52 | <sbalneav> and 2) Now you're going to have to keep the users password around on the workstation somewhere in cleartext so you can plumb the ssh connection.
| |
| 13:53 | <vagrantc> regarding point 2, can't we just use the ssh socket, like we do with all the other LDM stuff ?
| |
| 13:53 | <jammcq> sbalneav: is it possible to use the encrypted channel already created by ldm ?
| |
| 13:53 | <sbalneav> Not without modifying the source to sshfs, I don't think. Since it spawns off it's own ssh.
| |
| 13:54 | <jammcq> so maybe sshfs isn't the thing to do that with
| |
| 13:54 | <sbalneav> It would, one supposes, be a nice feature to use an existing ssh control socket, and Miklos of Fuse/SSHFS would probably accept something like that.
| |
| 13:55 | <jammcq> seems like it, eh
| |
| 13:55 | <vagrantc> we can hope
| |
| 13:55 | <sbalneav> Well, Miklos is pretty reasonable, I've spoken to him via email a couple of times, so that's a possibility there.
| |
| 13:55 | <vagrantc> "the freespace issue" is a fundamental design issue ?
| |
| 13:56 | <jammcq> yes
| |
| 13:56 | sshfs relies on sftp, which doesn't have a freespace call
| |
| 13:56 | <sbalneav> sftp has no call in it to find out the free space on the remote filesystm
| |
| 13:56 | <vagrantc> got it
| |
| 13:56 | <sbalneav> but writing a passtrhough filesystem to handle that's an afternoon job.
| |
| 13:56 | <vagrantc> sbalneav: for you, maybe :)
| |
| 13:57 | * jammcq notes that it's only 2pm where scotty is :) | |
| 13:57 | <sbalneav> Besides the "ick" factor of stacking two fuse filesystems on top of each other :)
| |
| 13:57 | tux_440volt has quit IRC | |
| 13:57 | <vagrantc> sbalneav, the non-programmer who writes filesystems
| |
| 13:57 | <sbalneav> I can look into the sshfs thing.
| |
| 13:57 | <jammcq> non-programmer? heh, scotty is uber-programmer
| |
| 13:57 | <sbalneav> Bull.
| |
| 13:57 | I know enough to cause trouble :)
| |
| 13:57 | <jammcq> just for the record, /me thinks sshfs is icky
| |
| 13:58 | sbalneav: it's your kind of trouble that we like
| |
| 13:58 | <vagrantc> jammcq: because ... ?
| |
| 13:58 | <sbalneav> Anywho, I got an appointment. I'm gonna dash off, I'll take a peek at sshfs tonight.
| |
| 13:58 | <jammcq> vagrantc: umm, cuz it's not the complete solution
| |
| 13:58 | it needs a bandaid on top of it to finish the job
| |
| 13:58 | <sbalneav> be back later.
| |
| 13:58 | <vagrantc> jammcq: sure.
| |
| 13:59 | but you can't deny the attractiveness of using an existing secure channel
| |
| 13:59 | <jammcq> oh, I LIKE that idea
| |
| 13:59 | you mean the channel created by LDM ?
| |
| 13:59 | <vagrantc> yes
| |
| 13:59 | <jammcq> yeah, I REALLY LIKE that idea
| |
| 13:59 | <Gadi> don't forget the quantum tunneling :)
| |
| 13:59 | <jammcq> darned quantum tunneling will get us every time
| |
| 13:59 | <Gadi> sorry - I only catch every other page of this conv :D
| |
| 14:00 | Q-FUNK has quit IRC | |
| 14:00 | * Gadi does not think sshfs is icky - just that ssh folks are anal | |
| 14:00 | * Gadi is always in favor of dirty hacks | |
| 14:01 | * vagrantc plays the "will the wind dry the clothes before it rains" game | |
| 14:01 | <Gadi> especially if we get to say 'quantum tunneling'
| |
| 14:09 | plameng1 has joined #ltsp | |
| 14:09 | plameng1 has left #ltsp | |
| 14:12 | <joebob777as7_> hey for my host ws001 in dhcpd.conf what do i need beside the mac address and the ip address i want it to have. i have a chroot called nvidia
| |
| 14:13 | trying to make a certain box boot to a different chroot...
| |
| 14:13 | <vagrantc> joebob777as7_: root-path
| |
| 14:14 | joebob777as7_: but does it use the exact same kernel ?
| |
| 14:14 | <joebob777as7_> so option root-path "/opt/ltsp/nvidia";
| |
| 14:14 | <vagrantc> joebob777as7_: otherwise you'd have to also specify a different filename
| |
| 14:14 | <joebob777as7_> not sure lol they are 64 bit boxes...
| |
| 14:14 | ?
| |
| 14:15 | <vagrantc> i can't tell you what you have :P
| |
| 14:16 | <joebob777as7_> i just did ltsp build client for nvidia dir and i installed graphics drivers
| |
| 14:16 | ...
| |
| 14:16 | plamengr has quit IRC | |
| 14:16 | <vagrantc> joebob777as7_: then you probably need to specify something like: filename /ltsp/nvidia/pxelinux.0
| |
| 14:18 | <joebob777as7_> vagrantc, do i need the else filename as well
| |
| 14:19 | <vagrantc> joebob777as7_: so, you've got an /opt/ltsp/nvidia ... and anything else in /opt/ltsp ? you've got a /var/lib/tftpboot/ltsp/nvidia .... anything else in /var/lib/tftpboot/ltsp ?
| |
| 14:20 | <joebob777as7_> in /opt/ltsp i've got amd64 i386 images and nvidia
| |
| 14:20 | and in /var/lib/tftpboot/ltsp i've got amd64 and i386
| |
| 14:20 | and nvidia
| |
| 14:20 | <vagrantc> joebob777as7_: so you need some way in dhcp to distinguish between using i386, amd64 and nvidia ... (though why not just install the nvdidia stuff into the i386 and amd64 dirs?)
| |
| 14:21 | Amaranth has quit IRC | |
| 14:21 | <joebob777as7_> because ogra told me that by install the graphics drivers i would overwrite xgl in that dir. so i wanted to leave it open for someone to be able to bring in a laptop and plug in...
| |
| 14:21 | <vagrantc> ok
| |
| 14:22 | <joebob777as7_> so in dhcpd.conf i put hotst ws001 {
| |
| 14:22 | <vagrantc> joebob777as7_: so, you just need to set up appropriate root-path and filename entries for each machine ...
| |
| 14:22 | <joebob777as7_> hardware ethernet 00:19:db:d2:ac;
| |
| 14:22 | fixed address 10.0.0.5;
| |
| 14:22 | option root-path "/opt/ltsp/nvidia";
| |
| 14:22 | filename "/ltsp/nvidia/pxelinux.0";
| |
| 14:22 | <vagrantc> !pastebot
| |
| 14:22 | <ltspbot> vagrantc: "pastebot" is The LTSP pastebot is at http://pastebot.ltsp.org. Please paste all text longer than a line or two to the pastebot, as it helps to reduce traffic in the channel. A link to the content will be pasted in the channel.
| |
| 14:22 | <vagrantc> joebob777as7_: please use the pastebot, and just paste your entire dhcpd.conf
| |
| 14:22 | <joebob777as7_> sry
| |
| 14:23 | <vagrantc> joebob777as7_: i'd also recommend exploring the group option in dhcp ...
| |
| 14:23 | <ltsppbot> "joebob777as7" pasted "dhcpd.conf" (29 lines) at http://pastebot.ltsp.org/325
| |
| 14:24 | <joebob777as7_> group option?
| |
| 14:24 | <vagrantc> joebob777as7_: well, if you only have one machine for ws001, then no big deal, but if you have say 5 machines that all need the same root-path and filename parameters, you set up a group for them ...
| |
| 14:25 | <joebob777as7_> yeah i do have 6 machines but i only one for the moment...
| |
| 14:25 | for the others would i just put hardware ethernet multiple times in ws001?
| |
| 14:25 | <vagrantc> joebob777as7_: i *think* you want to put your host declaration inside the subnet declaration
| |
| 14:26 | joebob777as7_: do you need a static ip ?
| |
| 14:27 | <joebob777as7_> no not really
| |
| 14:28 | <ltsppbot> "vagrantc" pasted "dhcpd.conf using group" (29 lines) at http://pastebot.ltsp.org/326
| |
| 14:28 | <vagrantc> joebob777as7_: i think that should work for you, and then you only need to add one line for each new machine
| |
| 14:29 | <joebob777as7_> oh ok awesome vagrantc thanks i'll try it now
| |
| 14:30 | <vagrantc> joebob777as7_: probably also make a group for amd64 machines, although you could probably use i386 on those.
| |
| 14:30 | <joebob777as7_> yeah i'll just use i386 on all other machines
| |
| 14:34 | <cliebow_> x
| |
| 14:35 | <joebob777as7_> vagrantc, still gave it i386 image...
| |
| 14:36 | <vagrantc> ah.
| |
| 14:36 | <joebob777as7_> do i need to restart anything?
| |
| 14:36 | <vagrantc> odd.
| |
| 14:36 | oh, yeah, you'll need to restart dhcp3-server
| |
| 14:36 | <cliebow_> vagrantc:you coming up to Maine??
| |
| 14:36 | <vagrantc> invoke-rc.d dhcp3-server restart
| |
| 14:36 | cliebow_: alas, no
| |
| 14:36 | <cliebow_> dang..
| |
| 14:38 | cliebow_ has quit IRC | |
| 14:40 | <vagrantc> joebob777as7_: that doesn't look like a valid mac addresss
| |
| 14:40 | joebob777as7_: looks like you're missing a bit
| |
| 14:40 | <joebob777as7_> vagrantc, didn't boot properly was left with /bin/sh: can't access tty; job control turned off
| |
| 14:40 | yeah i updated that already
| |
| 14:41 | so i updated sshkeys now i need to rebuild image right?
| |
| 14:41 | how do i do that for a different chroot?
| |
| 14:42 | * vagrantc has no love of the image stuff | |
| 14:42 | <vagrantc> you'll have to ask the ubuntu folks on that part
| |
| 14:43 | <joebob777as7_> think i found it... i just specified the port
| |
| 14:44 | shoot it raised the portnumber to 2002... so i screwed up?
| |
| 14:44 | ogra, you on?
| |
| 14:44 | Gadi has left #ltsp | |
| 14:46 | <joebob777as7_> vagrantc, still takes me to /bin/sh: can't access tty; job control turned off... any idea?
| |
| 14:46 | <vagrantc> no
| |
| 14:46 | <joebob777as7_> crap!
| |
| 14:47 | MagicStorm has joined #ltsp | |
| 14:47 | <joebob777as7_> sbalneav, ?
| |
| 14:47 | you on?
| |
| 14:53 | ok vagrantc i figured out the port deal after updating the sshkeys but it's not booting properly... it goes to the nvidia dir i can see that but as soon as the ubuntu upsplash comes up it goes to that /bin/sh black screen
| |
| 14:57 | <vagrantc> joebob777as7_: try editing /var/lib/tftpboot/ltsp/nvidia/pxelinux.cfg/default and removing options like quiet and splash
| |
| 14:58 | cliebow has quit IRC | |
| 15:00 | edgarin has joined #ltsp | |
| 15:05 | <joebob777as7_> there are a lot of directory not founds...
| |
| 15:05 | sepski has quit IRC | |
| 15:05 | <joebob777as7_> want me to type them vagrantc ?
| |
| 15:05 | i just rebuilt the image with no issue
| |
| 15:06 | * vagrantc would like to hand this off to people who actually understand the software being used | |
| 15:06 | <edgarin> Hi
| |
| 15:06 | <vagrantc> i don't use ubuntu, and i don't use the nbd stuff
| |
| 15:06 | <edgarin> hi jammcq
| |
| 15:06 | vagrantc, hi!! how are you?
| |
| 15:06 | <vagrantc> edgarin: pretty good
| |
| 15:06 | <edgarin> vagrantc, cool! :)
| |
| 15:10 | captain_magnus has quit IRC | |
| 15:11 | captain_magnus has joined #ltsp | |
| 15:15 | Clown-MG has quit IRC | |
| 15:21 | Amaranth has joined #ltsp | |
| 15:33 | cyberorg is now known as cyberorg_ | |
| 15:47 | Barbosa has quit IRC | |
| 15:49 | tiagovaz__ has joined #ltsp | |
| 15:50 | tiagovaz_ has quit IRC | |
| 15:56 | MagicStorm has quit IRC | |
| 16:03 | <sbalneav> Ah
| |
| 16:03 | ah ha
| |
| 16:03 | sshfs:
| |
| 16:03 | -o ssh_command=CMD
| |
| 16:03 | so, theoretically:
| |
| 16:04 | <ogra> pfft
| |
| 16:04 | -o ControlPath=/var/run/ldm_socket
| |
| 16:04 | <sbalneav> sshfs -o ssh_command='ssh -S /var/run/ldm_control_scoket"
| |
| 16:04 | <ogra> ;)
| |
| 16:04 | why the extra ssh call ?
| |
| 16:05 | <sbalneav> Well, they don't have that option now.
| |
| 16:05 | <ogra> scp -o ControlPath=/var/run/ldm_socket /tmp/$(echo $DISPLAY| sed s/\\./\-/g|cut -d':' -f1)_small.png server:/tmp/.thin-client-manager/
| |
| 16:05 | works perfect
| |
| 16:05 | i'd expect that to work with sshfs as well
| |
| 16:05 | <sbalneav> Well, the question was, if we were using sshfs to mount home dirs from the server, could we use the control socket from the existing ssh
| |
| 16:06 | -o SSHOPT=VAL
| |
| 16:06 | <ogra> yeah, like every other ssh based command
| |
| 16:06 | <sbalneav> looks like it'll work
| |
| 16:06 | <ogra> so -o ControlPath=/var/run/ldm_socket should be fine
| |
| 16:06 | <sbalneav> so, no patches required to sshfs
| |
| 16:07 | just a layer on top of sshfs to implement the statfs call
| |
| 16:07 | <ogra> do we really need it ?
| |
| 16:07 | <sbalneav> yeah
| |
| 16:07 | <ogra> i mean has anyone proven its broken ?
| |
| 16:07 | <sbalneav> yeah
| |
| 16:08 | That's why ltspfs exists :)
| |
| 16:08 | * ogra cant imagine firefox being evil everywhere else but caring about *that* | |
| 16:08 | <sbalneav> If that had worked, we would have used sshfs for localdevs
| |
| 16:08 | <ogra> yeah, i know
| |
| 16:08 | we should probably check if thats still the case though
| |
| 16:08 | two years passed :)
| |
| 16:08 | <sbalneav> I can verify it, to be sure.
| |
| 16:10 | oin$ mkdir foo
| |
| 16:10 | oin$ sshfs oin.ath.cx: foo
| |
| 16:10 | sbalneav@oin.ath.cx's password:
| |
| 16:10 | oin$ df -k
| |
| 16:10 | Filesystem 1K-blocks Used Available Use% Mounted on
| |
| 16:10 | /dev/sda1 147581948 29072076 111013144 21% /
| |
| 16:10 | varrun 517864 248 517616 1% /var/run
| |
| 16:10 | varlock 517864 0 517864 0% /var/lock
| |
| 16:10 | procbususb 517864 108 517756 1% /proc/bus/usb
| |
| 16:10 | udev 517864 108 517756 1% /dev
| |
| 16:10 | devshm 517864 0 517864 0% /dev/shm
| |
| 16:10 | sshfs#oin.ath.cx: 7999999992 0 7999999992 0% /usr/legal/home/sbalneav/foo
| |
| 16:10 | I'd say it's still borken :)
| |
| 16:11 | <joebob777as7_> not booting properly... it goes to the nvidia dir i can see that but as soon as the ubuntu upsplash comes up it goes to that /bin/sh black screen
| |
| 16:11 | bobby_C has joined #ltsp | |
| 16:11 | <sbalneav> but that's easy enough to fix
| |
| 16:11 | <ogra> !pastebot
| |
| 16:11 | <ltspbot> ogra: "pastebot" is The LTSP pastebot is at http://pastebot.ltsp.org. Please paste all text longer than a line or two to the pastebot, as it helps to reduce traffic in the channel. A link to the content will be pasted in the channel.
| |
| 16:11 | <ogra> :P
| |
| 16:11 | WOOOPEEEE !!!!!
| |
| 16:11 | * ogra dances | |
| 16:12 | <ogra> sbalneav, did you upgrade yelp today ?
| |
| 16:12 | tiagovaz__ has quit IRC | |
| 16:12 | <sbalneav> No, not yet.
| |
| 16:12 | <ogra> DOIT !
| |
| 16:12 | handbook is on the front now
| |
| 16:12 | jammcq has quit IRC | |
| 16:12 | <sbalneav> Have to wait 'till I get home :)
| |
| 16:12 | OK, that'll be the first thing I do :)
| |
| 16:13 | ok, I'm gonna head home.
| |
| 16:13 | see you all on later tonight.
| |
| 16:13 | sbalneav has quit IRC | |
| 16:13 | mikkel has quit IRC | |
| 16:15 | Guaraldo has left #ltsp | |
| 16:16 | <joebob777as7_> ogra, can you give me a hand?
| |
| 16:28 | fernando110 has quit IRC | |
| 16:33 | joebaker has left #ltsp | |
| 16:34 | nf1 has quit IRC | |
| 16:56 | muh2000 has quit IRC | |
| 17:01 | <joebob777as7_> says it can't find a bunch of dirs when it trys to boot....
| |
| 17:12 | nicoAMG has joined #ltsp | |
| 17:19 | <joebob777as7_> ogra, you on?
| |
| 17:22 | nicoAMG has quit IRC | |
| 17:30 | bobby_C has quit IRC | |
| 18:12 | fernando110 has joined #ltsp | |
| 18:36 | staffencasa has quit IRC | |
| 18:44 | joebob777as7 has joined #ltsp | |
| 18:45 | joebob777as7_ has quit IRC | |
| 19:02 | <vagrantc> otavio: might you have time for an ltsp upload tomorrow?
| |
| 19:04 | <otavio> vagrantc: sure
| |
| 19:04 | vagrantc: send it to me
| |
| 19:10 | <joebob777as7> ogra, you on? i'm having boot issues. taking me to a /bin/sh: can't access tty; job control turned off screen
| |
| 19:22 | vagrantc has quit IRC | |
| 19:44 | blackosprey has joined #ltsp | |
| 19:53 | K_O-Gnom has quit IRC | |
| 20:03 | jammcq has joined #ltsp | |
| 20:03 | <jammcq> hey kidz
| |
| 20:03 | sbalneav has joined #ltsp | |
| 20:03 | <sbalneav> Evening all
| |
| 20:04 | <jammcq> Scotty !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
| |
| 20:06 | <sbalneav> Evening jammcq
| |
| 20:28 | <blackosprey> hey i have an issue with my dhcpd server...
| |
| 20:28 | anyone care to help?
| |
| 20:34 | <sbalneav> What's the problem?
| |
| 20:37 | blackosprey has quit IRC | |
| 20:38 | <jammcq> heh
| |
| 20:38 | guess he figured it out
| |
| 20:40 | chris_punches has quit IRC | |
| 20:43 | edgarin has quit IRC | |
| 20:58 | cliebow has joined #ltsp | |
| 21:39 | monteslu has joined #ltsp | |
| 21:47 | <jcastro> moquist: around?
| |
| 21:48 | Lumiere has quit IRC | |
| 21:49 | <moquist> jcastro: yep
| |
| 21:50 | <jcastro> moquist: you coming to uds and the following ltsp party in maine?
| |
| 21:50 | <moquist> jcastro: just rolled out puppet yesterday & today, for one school.
| |
| 21:50 | jcastro: ltsp party? :(
| |
| 21:50 | jcastro: I'll be at UDS.
| |
| 21:50 | <jammcq> well, bts
| |
| 21:50 | bts-2007
| |
| 21:50 | * moquist didn't know about BTS... | |
| 21:50 | <jcastro> moquist: oh man, AWESOME. I just saw luke at the ohiolinuxfest. I was keen about pointing him towards you.
| |
| 21:50 | <jammcq> moquist: we're just getting it organized
| |
| 21:50 | <jcastro> moquist: I will have a car, so if you need a lift ...
| |
| 21:50 | <moquist> jammcq: :)
| |
| 21:51 | <jammcq> nov 3-5 (Sat-Mon) in SW.Harbor
| |
| 21:51 | <jcastro> moquist: I have the cool car, with ogra and stuff, don't ride with jammcq. :p
| |
| 21:51 | captain_magnus has quit IRC | |
| 21:51 | <moquist> jcastro: ha ha! :-)
| |
| 21:51 | captain_magnus has joined #ltsp | |
| 21:52 | <moquist> jammcq: I'd love to be there. I'll have to make sure that works for my work&home calendars.
| |
| 21:52 | <jcastro> moquist: if it's not too much trouble, can you document your puppet stuff and bring it with you? I'm keen on getting it in as server candy for LTS, I'd like some more data on it.
| |
| 21:52 | <moquist> jcastro: would it be enough to bring a tarball of /etc/puppet? :)
| |
| 21:52 | <jcastro> I find it a perfect match for people deploying LTSP, I'd like to get that better supported in ubuntu
| |
| 21:52 | moquist: yep yep, it would
| |
| 21:53 | <moquist> that's easily done, then.
| |
| 21:53 | muh2000 has joined #ltsp | |
| 21:53 | <moquist> right now I'm just pushing out dhcp configuration (with load balancing), sudoers, /etc/cups/ppd/* and printers.conf
| |
| 21:53 | printers.conf is awkward. I'm not happy with it.
| |
| 21:53 | <jcastro> yeah
| |
| 21:53 | you have my config right?
| |
| 21:54 | <moquist> I overwrote a teachers' printers.conf on his laptop, which had a one-off printer configured for it.
| |
| 21:54 | jcastro: er, maybe. haven't been looking at it.
| |
| 21:54 | <jcastro> k, feel free to holler at me
| |
| 21:54 | <moquist> jcastro: Oh, and my server got cracked and then 2/3 of my RAID array died in the same day, so I lost a bunch of stuff. Not sure yet whether I'm going to be able to recover my data.
| |
| 21:55 | <jcastro> I had a raid die my last day at the university
| |
| 21:55 | I feel you man
| |
| 21:55 | <moquist> Oh, painful. :(
| |
| 21:55 | <jcastro> luckily never been owned
| |
| 21:55 | <moquist> I managed to grab ~/docs before it went, which is VERY good.
| |
| 21:55 | I lost ~/projects, but most of that is already somewhere else.
| |
| 21:55 | <jcastro> the last few months I took a "linus" view on backups, as much shit copied to as many places as possible
| |
| 21:56 | Lumiere has joined #ltsp | |
| 21:56 | <jcastro> moquist: I gotta tell you my new sunray story
| |
| 21:56 | <moquist> jcastro: I think it was the cracked canonical servers, but I'm not sure. (I logged back into my machine from those.)
| |
| 21:56 | <jcastro> moquist: ouch.
| |
| 21:56 | <moquist> jcastro: new sunray story...?
| |
| 21:56 | <jcastro> so like we've been an ltsp shop for going on a year now, so sun sent down some "experts".
| |
| 21:57 | <moquist> "Expert Senior Sales", that is.
| |
| 21:57 | <jcastro> they want to donate $20k of server equipment for a cluster if we'll come back to sunray in our labs.
| |
| 21:58 | pure slime. unfortunately I left the job before I could squash that. Lucky for me the Dean is an ubuntu maniac, so when I tell him "Don't do that, seriously." he listened
| |
| 21:58 | <moquist> nice
| |
| 21:58 | jcastro: so you left your job...
| |
| 21:58 | <jcastro> but seriously, how scummy. "Donating" equipment with conditions to move back to their platform.
| |
| 21:59 | moquist: yeah, I work for canonical now
| |
| 21:59 | <moquist> Sun Scum.
| |
| 21:59 | jcastro: yay! so apparently that worked out. Congratulations!
| |
| 21:59 | <jcastro> moquist: thanks.
| |
| 21:59 | MagicStorm has joined #ltsp | |
| 22:00 | <moquist> Hmm. Just to clarify -- when I complained about printers.conf, I was actually complaining about handling it in puppet.
| |
| 22:00 | <jcastro> unfortunately, I was one of their lead dudes in .edu deployment stuff, so now they need a new rockstar who works in .edu, deploys LTSP, and is generally awesome.
| |
| 22:00 | moquist: it's a good thing you're around. :p
| |
| 22:00 | vagrantc has joined #ltsp | |
| 22:00 | <moquist> I would rather have printer { ppd => <source>, name => "name" } than file { printers.conf...}.
| |
| 22:00 | <jcastro> split it into classes
| |
| 22:00 | that's what I did for everything
| |
| 22:01 | <moquist> jcastro: No, I'm saying that I want a printer Puppet Type. (I hope there is one, I haven't been able to check yet.)
| |
| 22:01 | <jcastro> well, if you need one, let's just ask luke at uds.
| |
| 22:02 | <moquist> jcastro: well, do you agree?
| |
| 22:02 | jcastro: the point is that I want to be able to add printers without blowing away whatever may have been configured locally.
| |
| 22:02 | <jcastro> moquist: dude I hated cups so much I ran a windows print server
| |
| 22:02 | that's how much I hated cups.
| |
| 22:03 | <moquist> Oh, I don't care so much about cups. I'm not using it for print serving; we have JetDirect everywhere.
| |
| 22:03 | ...but that means I have to tell every machine about every printer.
| |
| 22:03 | <jcastro> moquist: document your issues and I'll bring them to luke, I have to take off.
| |
| 22:03 | * moquist nods | |
| 22:03 | <jcastro> moquist: but I'll see you in boston
| |
| 22:03 | <3
| |
| 22:03 | <moquist> I've been thinking of piping up in #puppet about it.
| |
| 22:03 | <jcastro> do so
| |
| 22:03 | <moquist> yep
| |
| 22:03 | <jcastro> it's a friendly place
| |
| 22:04 | <moquist> I've been in there a bit and gotten help.
| |
| 22:04 | <jcastro> you want to talk to lak, that's luke; use my name in vain to get help
| |
| 22:04 | heh
| |
| 22:04 | nite all!
| |
| 22:04 | <moquist> night owl
| |
| 22:09 | chad has joined #ltsp | |
| 22:12 | <chad> anybody here tonight?
| |
| 22:12 | <vagrantc> !question
| |
| 22:12 | <ltspbot> vagrantc: "question" is if you have a question about ltsp, just ask it, and people will respond if they can.
| |
| 22:13 | <vagrantc> ltspbot: forget question
| |
| 22:13 | <ltspbot> vagrantc: The operation succeeded.
| |
| 22:13 | <vagrantc> ltspbot: learn question as if you have a question about ltsp, please go ahead and ask it, and people will respond if they can.
| |
| 22:13 | <ltspbot> vagrantc: The operation succeeded.
| |
| 22:13 | <chad> Same one I posted last night -- any clues how to make printers available in edubuntu's ltsp kiosk mode?
| |
| 22:13 | Joris has joined #ltsp | |
| 22:18 | joebaker has joined #ltsp | |
| 22:18 | <joebaker> ogra are you there?
| |
| 22:25 | Joris_ has quit IRC | |
| 22:26 | <vagrantc> joebaker: ogra's probably sleeping
| |
| 22:31 | chad has left #ltsp | |
| 22:37 | F-GT has quit IRC | |
| 22:37 | <lns> who out there experiences random thin-client freezes with LTSP, say, while online w/Firefox?
| |
| 22:38 | I've got a class that was online today for 20 minutes, and 5 machines hard-locked while using Firefox.
| |
| 22:51 | <sbalneav> lns: There's been a huge discussion on this lately
| |
| 22:52 | Check out the ltsp-discuss, edubuntu-users, and edubuntu-devel mailing lists.
| |
| 22:53 | <lns> sbalneav, great - thank you
| |
| 22:53 | <sbalneav> lns: Bottom line: it's firefox grabbing all your ram.
| |
| 22:54 | <lns> sbalneav, wow...ok i'll check the lists, is there any light at the end of the tunnel?
| |
| 22:57 | <sbalneav> Strategies:
| |
| 22:57 | 1) Wait for firefox to decide to fix it (some work being done, no clue when finished)
| |
| 22:57 | 2) Add huge amounts of nbd swap
| |
| 22:58 | 3) Add more ram
| |
| 22:58 | 4) Use something else, like Konqeror, or Opera.
| |
| 23:03 | <lns> ...wow
| |
| 23:04 | I guess I don't understand the way firefox handles ram, but it sure seems like it should be restricted/recycled somehow instead of just eating it all up and causing the whole session to crash
| |
| 23:04 | <sbalneav> Read the threads.
| |
| 23:04 | <lns> in the process
| |
| 23:04 | <sbalneav> it's a complex issue.
| |
| 23:20 | Headin' to bed. Night all
| |
| 23:20 | cyberorg_ has quit IRC | |
| 23:21 | cyberorg_ has joined #ltsp | |
| 23:22 | <lns> night sbalneav thanks for the pointers!!
| |
| 23:28 | cyberorg_ has quit IRC | |
| 23:30 | cyberorg_ has joined #ltsp | |
| 23:38 | cyberorg_ has quit IRC | |
| 23:39 | cyberorg_ has joined #ltsp | |