IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 1 January 2011 (all times are UTC)
| 01:00 | alkisg has joined #ltsp | |
| 01:02 | <alkisg> Happy new year to all
| |
| 01:09 | <muppis> Thank you and same to you.
| |
| 01:10 | alkisg, got boring evening so I played little with ltsp. Had rebuild whole image to get autologin to work.
| |
| 01:11 | <alkisg> Rebuilding should be necessary though...
| |
| 01:13 | <muppis> But now need to disable Gnome screensaver as xbmc has build-in one and it works better when not using keyboard nor mouse.
| |
| 01:15 | otavio has quit IRC | |
| 01:15 | <alkisg> Gnome screensaver has "blank only" hardcoded for ltsp clients afaik
| |
| 01:16 | <muppis> Cannot be disabled?
| |
| 01:18 | Well, hope it doesn't cause massive problems.
| |
| 01:23 | alkisg has quit IRC | |
| 01:23 | alkisg has joined #ltsp | |
| 01:24 | <alkisg> Urm vbox just crashed my kernel when I tried to change host-only networking settings...
| |
| 01:25 | <muppis> Maybe it got little hangover from last night?
| |
| 01:26 | <alkisg> :)
| |
| 01:26 | <muppis> :)
| |
| 01:27 | But I think I gotta go out with the kids. There plenty of snow there..
| |
| 01:30 | <alkisg> NIce!!!
| |
| 03:14 | komunista has joined #ltsp | |
| 03:56 | Trixboxer has joined #ltsp | |
| 05:34 | F-GT has quit IRC | |
| 05:46 | alkisg has quit IRC | |
| 06:30 | F-GT has joined #ltsp | |
| 06:35 | uday has joined #ltsp | |
| 06:37 | Kicer86 has joined #ltsp | |
| 07:13 | uday has quit IRC | |
| 07:29 | Mobe has quit IRC | |
| 07:30 | Mobe has joined #ltsp | |
| 08:35 | komunista has quit IRC | |
| 08:37 | litlebuda has joined #ltsp | |
| 09:24 | RiXtEr has quit IRC | |
| 09:25 | RiXtEr has joined #ltsp | |
| 09:35 | alkisg has joined #ltsp | |
| 09:41 | Mobe has quit IRC | |
| 09:45 | Mobe has joined #ltsp | |
| 09:52 | bobby_C has joined #ltsp | |
| 10:50 | vagrantc has joined #ltsp | |
| 11:37 | Mobe has quit IRC | |
| 11:52 | dr_data has quit IRC | |
| 12:19 | gentgeen__ has quit IRC | |
| 12:42 | <alkisg> nbd-server -c seems very stable, the diff is 1.2 Gb so far and it's still fast (I even used it to make a /swap file on the client...)
| |
| 12:42 | <vagrantc> yay
| |
| 12:43 | that's the cow option?
| |
| 12:43 | <alkisg> Yup
| |
| 12:44 | I netbooted a natty vm off of it by just installing nbd-client and setting eth0 inet manual in interfaces, and now I'm stressing it...
| |
| 12:44 | Also nfs+tmpfs with aufs was stable
| |
| 12:45 | Next I want to test nfs + cow nbd instead of tmpfs + aufs
| |
| 12:46 | <vagrantc> wow. last time i tried nfs with aufs it exploded badly
| |
| 12:46 | <alkisg> It needs mount --move, it doesn't like mount --bind
| |
| 12:47 | <vagrantc> do we do mount --bind anywhere?
| |
| 12:47 | <alkisg> (and I tested with the same paths, /rofs and /cow, I don't know if it matters)
| |
| 12:47 | I think so, yes
| |
| 12:47 | I.e. if ot
| |
| 12:47 | <vagrantc> mount --move is so much better if it works
| |
| 12:47 | <alkisg> I.e. if it's mounted on /cow and then made available on another path, maybe it has problems then
| |
| 12:47 | It makes mtab work ok (it's writeable then)
| |
| 12:48 | No need to symlink it to /proc/mounts
| |
| 12:48 | While mount --bind makes duplicate entries in mtab
| |
| 12:48 | <vagrantc> one of the problems with mtab is not writing to mtab itself, but that files assumed they could write to /etc/mtab?*
| |
| 12:49 | i think there are a few other files with similar issues ...
| |
| 12:49 | <alkisg> With nfs + [tmpfs | empty cow nbd] + aufs, it's not a problem, they can write wherever they want...
| |
| 12:49 | <vagrantc> passwd and group and such
| |
| 12:49 | that would be slick
| |
| 12:50 | <alkisg> After implementing it, I saw a web page that described exactly how I did it... http://www.logicsupply.com/blog/2009/01/27/how-to-build-a-read-only-linux-system/
| |
| 12:50 | <vagrantc> heh
| |
| 13:00 | <alkisg> Nope, after 40 minutes of stressing (the diff went to 1.3G) no noticable delay in nbd-server -c.
| |
| 13:00 | No failures in syslog, all seem fine
| |
| 13:02 | <vagrantc> very, very cool.
| |
| 13:03 | alkisg: did you just tweak the ltsp_nbd script to support it?
| |
| 13:03 | <alkisg> No, I'm testing without LTSP
| |
| 13:03 | Just a VM installation of natty, exporting the .vdi disk with nbd
| |
| 13:03 | <vagrantc> the .vdi disk is readable as a normal disk?
| |
| 13:04 | <alkisg> And I only installed nbd-client in the VM + added eth0 inet manual in interfaces so that network-manager doesn't touch it
| |
| 13:04 | vdfuse from virtualbox-ose-fuse can export the partitions as block devices
| |
| 13:04 | So they can be loop mounted or exported with nbd-server
| |
| 13:05 | Ah, so there was also the vdfuse overhead, heh, I didn't notice it
| |
| 13:06 | I bet with squashfs and without vdfuse it'll go even faster
| |
| 13:06 | (it's usually 2.5 times faster, measured on my fat client labs... cool)
| |
| 13:10 | Trixboxer has quit IRC | |
| 13:12 | * alkisg goes on to write an ltsp-update-image equivalent for vbox VMs... | |
| 13:27 | <alkisg> Heh, tmpfs/aufs is also handy there, it allows for chroot-installing nbd-client + tweaking interfaces/fstab without having to modify the VM beforehand
| |
| 13:30 | (either that or mounting the VM partition with nbd-server -c, which makes it cow)
| |
| 13:38 | _UsUrPeR_ has quit IRC | |
| 13:39 | _UsUrPeR__ has joined #ltsp | |
| 13:53 | mymrhelpdesk has quit IRC | |
| 14:25 | litlebuda has quit IRC | |
| 14:26 | litlebuda has joined #ltsp | |
| 15:09 | vmlintu has joined #ltsp | |
| 15:37 | Kicer86 has quit IRC | |
| 15:43 | gentgeen__ has joined #ltsp | |
| 15:52 | vagrantc has quit IRC | |
| 16:27 | <AndyGraybeal> what do you guys think of x2go and do you think it is a worthy replacement for LTSP as far as terminal server/clients go?
| |
| 16:28 | i hope no one takes offsense to the question, i just started to use it as a replacement for NX (NX started giving me some fit)
| |
| 16:29 | and now i am wondering if it would be better to replace ltsp with it instead.
| |
| 16:29 | there's nothing wrong with ltsp in my mind, and in fact i love it. i'm just questioning my setup.
| |
| 16:30 | i think x2go would give me the ability for 'session' management; and i like that. i want to be able to boot up a machine at work, suspend that session, go home and then connect to the same sesion from home.
| |
| 16:31 | if you have any thoughts i would be interested in hearing them.
| |
| 16:32 | <alkisg> I've never used it, but if it uses nx, then it's not very fast for video etc
| |
| 16:32 | X is the fastest there is for local network
| |
| 16:33 | (when direct, without ssh)
| |
| 16:34 | Unfortunately I don't know of any good x-proxy implementations that could combine both speed + session management...
| |
| 16:35 | With a good x-proxy, plain X could still be used locally for speed, and then something nx-based from home (adsl etc), with session pause/resume in both cases...
| |
| 16:37 | Also maybe nx-based solutions won't scale well for too many clients, I imagine they raise the server CPU too much
| |
| 16:50 | alkisg has quit IRC | |
| 17:06 | vmlintu has quit IRC | |
| 17:27 | ogra has quit IRC | |
| 17:46 | bobby_C has quit IRC | |
| 18:01 | cliebow has joined #ltsp | |
| 18:02 | cliebow has quit IRC | |
| 18:10 | otavio has joined #ltsp | |
| 19:07 | litlebuda has quit IRC | |
| 19:08 | litlebuda has joined #ltsp | |
| 19:17 | alkisg has joined #ltsp | |
| 19:24 | Mathis has joined #ltsp | |
| 19:24 | <Mathis> hello
| |
| 19:39 | <alkisg> Hello
| |
| 19:43 | <AndyGraybeal> thank you for your explanation alkisg.
| |
| 19:50 | <Mathis> what is this ltsp-clustering used for?
| |
| 20:13 | ogra has joined #ltsp | |
| 20:13 | ogra has quit IRC | |
| 20:13 | ogra has joined #ltsp | |
| 20:37 | cliebow has joined #ltsp | |
| 20:53 | cliebow has quit IRC | |
| 22:03 | alkisg has quit IRC | |
| 22:51 | litlebuda has quit IRC | |
| 23:41 | Mathis has quit IRC | |