00:11 | us3r has joined #ltsp | |
00:12 | johnny has joined #ltsp | |
00:33 | us3r has quit IRC | |
00:44 | us3r has joined #ltsp | |
00:50 | stiev3 has quit IRC | |
01:35 | vagrantc has quit IRC | |
01:35 | vagrantc has joined #ltsp | |
01:39 | pts_ has joined #ltsp | |
01:54 | <pts_> The link for the upstreeam documentations gives me error 404
| |
01:55 | <alkisg> !docs
| |
01:55 | <ltspbot> alkisg: "docs" :: For the most current documentation, see https://sourceforge.net/apps/mediawiki/ltsp/index.php?title=Ltsp_LtspDocumentationUpstream
| |
01:55 | <Ryan52> /topic lies.
| |
01:55 | where's jammcq when you need him?
| |
01:55 | mikkel has joined #ltsp | |
01:56 | <Ryan52> sbalneav: ^----
| |
01:56 | <pts_> The link on the Ubuntu LTSP page should also be updated, but I guess thats out of your juristriction :)
| |
01:57 | <alkisg> pts_: you can change that yourself :)
| |
01:57 | (please do)
| |
01:58 | <pts_> alkisg: ofc, not used to this new fashion Internet yet
| |
02:57 | vagrantc has quit IRC | |
03:08 | bobby_C has joined #ltsp | |
03:56 | ogra_ has joined #ltsp | |
03:58 | ogra has quit IRC | |
04:03 | us3r has quit IRC | |
04:19 | alkisg has quit IRC | |
04:25 | ogra_ has quit IRC | |
04:25 | ogra_ has joined #ltsp | |
04:30 | ogra_ is now known as ogra | |
04:31 | tarbo has quit IRC | |
04:33 | alexqwesa__ is now known as alexqwesa | |
04:35 | alkisg has joined #ltsp | |
05:05 | lucascoala has joined #ltsp | |
05:55 | wwx has quit IRC | |
06:02 | vicd has quit IRC | |
06:03 | hersonls has joined #ltsp | |
06:08 | wwx has joined #ltsp | |
06:09 | pmatulis has joined #ltsp | |
06:13 | pmatulis has quit IRC | |
06:15 | pmatulis has joined #ltsp | |
06:41 | alkisg has quit IRC | |
06:44 | spectra has quit IRC | |
07:04 | alkisg has joined #ltsp | |
07:20 | coordinador has joined #ltsp | |
07:20 | <coordinador> Hi, i have a problem with sound, sometimes sound mutes itself
| |
07:21 | i have to open alsamixer and back to normal the volume
| |
07:22 | Sarten-X has quit IRC | |
07:32 | Sarten-X has joined #ltsp | |
07:33 | bobby_C has quit IRC | |
07:34 | GodFather has joined #ltsp | |
08:00 | cmm has joined #ltsp | |
08:01 | cmm is now known as Guest38881 | |
08:08 | <dmarkey_> where is Gideon Romm again?
| |
08:15 | Gadi has joined #ltsp | |
08:22 | coordinador has quit IRC | |
08:32 | Ahmuck_ has joined #ltsp | |
08:36 | pts_ has quit IRC | |
08:38 | mikkel has quit IRC | |
08:40 | <johnny> [09:12] dmarkey_: where is Gideon Romm again?
| |
08:40 | beep beep Gadi
| |
08:42 | <Gadi> no honking! can't you read the sign
| |
08:43 | <dmarkey_> Gadi: hey again
| |
08:43 | <Gadi> g'morining
| |
08:44 | or something
| |
08:44 | * Gadi shouldnt type with his mouth full | |
08:44 | <dmarkey_> have we any projects splicing LTSP with a desktop virtualisation system
| |
08:45 | <Gadi> lots of adjectives and verbs in that sentence
| |
08:45 | how do you mean?
| |
08:46 | * Gadi has run LTSP with VMWare View, Citrix XenApp, Quest vWorkspace, Ericom, etc. | |
08:46 | <Gadi> if thats what you mean
| |
08:46 | tarbo has joined #ltsp | |
08:46 | <dmarkey_> as in, an LTSP client corresponds with a virtual machine on a hypervisor somewhere, 1-1 relationship rather than many clients to one instance
| |
08:46 | <Gadi> sure
| |
08:46 | but its mostly the backend that takes care of that
| |
08:46 | <dmarkey_> so a broker?
| |
08:47 | <Gadi> you still connect to the virtual machine with RDP/ICA
| |
08:47 | right
| |
08:47 | you need a broker
| |
08:47 | thats what most of the above products I mention are
| |
08:48 | you just need to install the clients in the chroot and make an appropriate screen script
| |
08:48 | <sbalneav> Ryan52: Yes?
| |
08:48 | <dmarkey_> yea, but it wouldnt be hard to implement an open source solution from scratch
| |
08:49 | using either VirtualBox or Xen
| |
08:49 | <sbalneav> Morning all
| |
08:49 | <Gadi> why do you need an open source solution for Xen?
| |
08:49 | <dmarkey_> is there already one?
| |
08:49 | <Gadi> ah, you mean you want to make a free broker
| |
08:49 | :)
| |
08:50 | <dmarkey_> yea
| |
08:50 | <Gadi> thats a different story
| |
08:50 | yeah, a free broker would be a nice project
| |
08:51 | <dmarkey_> but im thinking pure RDP with windows
| |
08:51 | dunno how i'd do it with other OS's
| |
08:51 | <Gadi> well, windows has a session broker built in now
| |
08:51 | we just dont have an OSS client that works with it yet
| |
08:51 | <dmarkey_> but rdesktop doesnt support it
| |
08:51 | <Gadi> thats a failing of rdesktop
| |
08:51 | a new project just started this year: FreeRDP
| |
08:51 | its a fork of rdesktop
| |
08:52 | <dmarkey_> but will the broker manage lets say 100 XP VM's sitting on Xen?
| |
08:52 | <Gadi> bec many devs are unhappy with rdesktop upstream
| |
08:53 | not the Microsoft broker
| |
08:53 | you need XenApp
| |
08:53 | <dmarkey_> Right
| |
08:53 | <Gadi> or Ericom
| |
08:53 | or some such
| |
08:53 | <dmarkey_> but, its not that hard..
| |
08:53 | <Gadi> right - ltsp-cluster is kinda getting there
| |
08:54 | for what you want to do
| |
08:54 | but, I dont think they are thinking along those lines
| |
08:54 | it would prolly be best to nudge them in that direction
| |
08:54 | hersonls has quit IRC | |
08:55 | <Gadi> we could have LDM contact a broker rather than a server
| |
08:55 | and then be redirected to the server
| |
08:55 | that would be nice
| |
08:55 | <ogra> LDM stock exchange ?
| |
08:55 | <Gadi> hmm...
| |
08:55 | hersonls has joined #ltsp | |
08:55 | <dmarkey_> but i am talking more windows
| |
08:55 | <Gadi> maybe our little LDM can grow up a bit
| |
08:55 | who cares what you connect to
| |
08:56 | we just need to do it with the right protocol
| |
08:56 | <dmarkey_> does ldm interface with rdestop?
| |
08:56 | * ogra would love to see someone implement ldm-standalone | |
08:56 | <Gadi> actually, these days it can with the "nomad" patch
| |
08:57 | ogra: it would be real easy, just cp screen.d/ldm ldm-launcher and modify it
| |
08:57 | :)
| |
08:57 | ldm-launcher can use xephyr instead of X
| |
08:57 | <ogra> Gadi, no, the hard part is the policy compliance for display managers in debian7ubuntu
| |
08:57 | you have to jump to 100s of hoops for that
| |
08:58 | *through
| |
08:58 | * Gadi does not understand | |
08:58 | * Gadi will code it up and throw it in multiverse | |
08:58 | <Gadi> :)
| |
08:59 | <ogra> you need to comply with certain standards to be allowed to provide it as display manager
| |
08:59 | <Gadi> isnt multiverse where Homer Simpson is floating around eating potato chips?
| |
08:59 | <ogra> vagrant knows more about it than me, he spent some weeks on it and gave up
| |
08:59 | <Gadi> I wont call it a display manager
| |
08:59 | Ill call it an LDM client
| |
08:59 | done
| |
08:59 | <ogra> heh
| |
09:01 | <dmarkey_> ok so LDM contacts the broker, checks username/password and directs to whichever VM is available, or and old session
| |
09:01 | an*
| |
09:01 | <Gadi> right - it basically sends back: use RDP to connect to server XYZ or use ssh -X to connect to server ABC
| |
09:01 | ...
| |
09:02 | GodFather has quit IRC | |
09:02 | <dmarkey_> hmm.. no session mobility on the linux side
| |
09:02 | <Gadi> we would be a multilingual broker
| |
09:02 | yeah, but thats a different topic
| |
09:02 | you want session mobility, send back: use NX to connect to server FGH
| |
09:03 | <dmarkey_> so, if someone is logged in in location A, moves to location B, but still logged into location A, what happens
| |
09:03 | mikkel has joined #ltsp | |
09:03 | <Gadi> if you set LDM_LIMIT_ONE_SESSION = True, he gets kicked off A
| |
09:03 | <dmarkey_> so that kills the forked sshd?
| |
09:03 | <Gadi> as we do now
| |
09:04 | yeah
| |
09:04 | it kills all his old processes
| |
09:04 | <dmarkey_> even ones not related to the old session? say if he had an old screen
| |
09:07 | <Gadi> dont make me look back at the code
| |
09:07 | :)
| |
09:08 | <dmarkey_> then you have to have a provisioning system etc.. it could get complicated
| |
09:08 | <Gadi> well, ltsp-cluster has a load-balancer piece
| |
09:08 | that could be expanded to be a rules-based provisioning system
| |
09:09 | but, yeah, if it were a one-line patch, it wouldnt be another project
| |
09:09 | :)
| |
09:09 | <dmarkey_> i must look into ltsp-cluster
| |
09:10 | <Gadi> but, I do like the idea of an ldm-broker
| |
09:10 | <dmarkey_> yea, should be really easy with some xmlrpc
| |
09:10 | althougfh it is pure C isnt it
| |
09:13 | * dmarkey_ was thinking python | |
09:16 | <Ryan52> sbalneav: the link for docs in /topic points to a 404
| |
09:19 | <sbalneav> Ryan52: Hm, I don't have chanops, I'll have to see if I can get jammcq to pop in.
| |
09:20 | thanks for the heads up
| |
09:20 | <Ryan52> sbalneav: chanserv says you do...
| |
09:20 | <sbalneav> I do?!
| |
09:20 | buh
| |
09:21 | um, ok, how do I become op? :)
| |
09:21 | <Ryan52> sbalneav: /msg chanserv op #ltsp
| |
09:21 | johnny has left #ltsp | |
09:21 | ChanServ sets mode: +o sbalneav | |
09:21 | johnny has joined #ltsp | |
09:21 | <sbalneav> Wow
| |
09:21 | CAN-o-SPAM has joined #ltsp | |
09:21 | <Ryan52> ta da!
| |
09:21 | :)
| |
09:21 | <sbalneav> When the heck did I get that!?
| |
09:22 | Ok, well, if I'm gonna reset the topic, what would people like to see up there?
| |
09:22 | <ogra> proper urls :)
| |
09:22 | i think the rest is still valid
| |
09:24 | <sbalneav> Hmmm, wonder what IRC command I use to list the topic... /list topic?
| |
09:24 | <Ryan52> /topic
| |
09:25 | <sbalneav> ah, ok, and to set? (Geez, I've never done this before!)
| |
09:25 | <Ryan52> /topic Blah Blah blah blah blah...
| |
09:25 | <sbalneav> ok
| |
09:26 | uno momento
| |
09:26 | sbalneav changes topic to "LTSP-5 is the current version that you should be using. Check out Ubuntu, Debian and Fedora for awesome integration. Gentoo is getting very close and Opensuse has kiwi-ltsp. IRC logs at: http://www.nubae.com/logs/, LTSP Docs: https://sourceforge.net/apps/mediawiki/ltsp/index.php?title=Ltsp_LtspDocumentationUpstream" | |
09:26 | <ogra> in xchat you just click into the topic field and change it :)
| |
09:27 | <sbalneav> Can't run xchat in screen :)
| |
09:27 | <ogra> pfft
| |
09:27 | commandline junkie
| |
09:27 | <sbalneav> gui goon
| |
09:27 | :)
| |
09:27 | <ogra> did you ever try byobou ?
| |
09:27 | :)
| |
09:28 | screen with bottom panel :)
| |
09:28 | oh, its byobu
| |
09:28 | <sbalneav> No, I haven't! I'll have to check it out.
| |
09:28 | <ogra> not ou
| |
09:29 | <sbalneav> I thought it was a contraction of "Bring Your Own Bouze" :)
| |
09:29 | <ogra> heh
| |
09:29 | no, its something japanese
| |
09:29 | <sbalneav> yeah, folding screens or something geisha's hide behind.
| |
09:30 | <ogra> yeah
| |
09:30 | <sbalneav> Lets see if I can deop. I don't like having this much power
| |
09:30 | ChanServ sets mode: -o sbalneav | |
09:30 | <sbalneav> foo
| |
09:31 | <ogra> clap clap clap
| |
09:31 | <sbalneav> Whew, that was scary
| |
09:31 | I think we need to change the interface.
| |
09:31 | should be /msg chanserv sudo /topic blah Blah
| |
09:42 | bobby_C has joined #ltsp | |
09:54 | <dmarkey_> so, where does LTSP developmnent talk take place, here? mailing lists?
| |
09:56 | <johnny> both yes..
| |
09:56 | <sbalneav> dmarkey_: Prolly more here than mailing lists
| |
09:56 | we tend to discuss here, then post findings to the mailing list.
| |
09:56 | <dmarkey_> Cool so how do i get the ball rolling on an LTSP broker :)
| |
10:00 | <cyberorg> sbalneav, http://tr.im/ltspdoc this would be better :)
| |
10:00 | <johnny> why?
| |
10:00 | url shorteners are not the greatest thing to happen to the internet..
| |
10:01 | who knows when the service might disappear and leave millions of broken links
| |
10:01 | <Gadi> dmarkey_: *LDM* broker
| |
10:01 | <johnny> cyberorg, please use full urls
| |
10:01 | <cyberorg> johnny, of course not, but at least we can see other stuff in the /topic bar in xchat
| |
10:01 | <johnny> unless you're on twitter
| |
10:01 | or some other micro blog
| |
10:01 | <dmarkey_> Gadi: Surwe
| |
10:01 | <johnny> cyberorg, better to add a redirect at the source instead
| |
10:01 | <dmarkey_> Gadi: but if possible id like a standalone client aswell
| |
10:02 | <Gadi> dmarkey_: I have a sworn duty to protect and defend the independent integrity of LDM
| |
10:02 | <johnny> ttps://sourceforge.net/apps/mediawiki/ltsp/LTSPDocs is possible at the very least
| |
10:02 | <Gadi> dmarkey_: yes - we can make ldm-client ;)
| |
10:02 | <johnny> or..
| |
10:02 | <Gadi> dmarkey_: easy enough
| |
10:02 | <dmarkey_> So, thats basically ssh -X + Xnest?
| |
10:02 | <johnny> whoever owns ltsp.org can make docs.ltsp.org go to that..
| |
10:02 | that would be a better solution
| |
10:03 | <Gadi> dmarkey_: if we have a broker, it would be more than that
| |
10:03 | <dmarkey_> but as it stands at the minutwe
| |
10:03 | <Gadi> mostly
| |
10:03 | <cyberorg> johnny, yes there are many better options, but tr.im is available right now to use till better alternative is done, saves some space in topic
| |
10:03 | <Gadi> it would be in the same ldm package
| |
10:03 | we would just have a script calld ldm-client that gets placed in /usr/bin
| |
10:03 | <johnny> cyberorg, changing the dns entry can be done now too
| |
10:04 | <cyberorg> johnny, yup, but you and i cannot :)
| |
10:04 | <dmarkey_> and would this install all the dependencies like Xnest, rdesktop, nxclient?
| |
10:04 | <Gadi> actually, maybe a different package that relies on ldm
| |
10:04 | alexqwesa has quit IRC | |
10:04 | <Gadi> so it can pull in all the other deps
| |
10:05 | <dmarkey_> hmm.. nxclient is a pain to try to script as far as i remember
| |
10:05 | <Gadi> nah
| |
10:06 | pain is a four letter word
| |
10:06 | <dmarkey_> has anyone deployed FreeNX with LTSP, with nxclient as the client?
| |
10:06 | <Gadi> yup
| |
10:07 | <dmarkey_> nxagent is a heavy load on the server as far as i remember
| |
10:07 | <Gadi> do you always worry this much?
| |
10:07 | :)
| |
10:09 | I promise, there's a bridge out there, you just need to step off the cliff - the bridge blends into the cliff wall to make it look invisible
| |
10:09 | alexqwesa has joined #ltsp | |
10:09 | <sbalneav> Only the Penitent man shall pass...
| |
10:09 | <dmarkey_> :)
| |
10:10 | <sbalneav> He chose..... poorly.
| |
10:10 | <dmarkey_> I'm not sure a 1-1 realtionship works with LTSP on the linux side
| |
10:10 | or if its needed
| |
10:10 | lucascoala_ has joined #ltsp | |
10:11 | _UsUrPeR_ has quit IRC | |
10:11 | monteslu_ has quit IRC | |
10:11 | Appiah has quit IRC | |
10:11 | vmlintu has quit IRC | |
10:11 | vlt has quit IRC | |
10:11 | moldy has quit IRC | |
10:11 | Sarten-X has quit IRC | |
10:11 | GiantPickle has quit IRC | |
10:11 | NeonLicht has quit IRC | |
10:11 | Wastrel_ has quit IRC | |
10:11 | zamba has quit IRC | |
10:11 | elias_a has quit IRC | |
10:11 | lupine_85 has quit IRC | |
10:14 | vlt has joined #ltsp | |
10:14 | vmlintu has joined #ltsp | |
10:14 | moldy has joined #ltsp | |
10:14 | lucascoala_ has joined #ltsp | |
10:14 | _UsUrPeR_ has joined #ltsp | |
10:14 | Appiah has joined #ltsp | |
10:14 | monteslu_ has joined #ltsp | |
10:14 | Sarten-X has joined #ltsp | |
10:14 | Wastrel_ has joined #ltsp | |
10:14 | elias_a has joined #ltsp | |
10:14 | zamba has joined #ltsp | |
10:14 | GiantPickle has joined #ltsp | |
10:14 | NeonLicht has joined #ltsp | |
10:14 | lupine_85 has joined #ltsp | |
10:16 | lucascoala has quit IRC | |
10:19 | din_os has joined #ltsp | |
10:20 | _UsUrPeR_ has quit IRC | |
10:23 | johnny has left #ltsp | |
10:24 | johnny has joined #ltsp | |
10:29 | <dmarkey_> has anyone experienced a ceiling of sessions a server can handle, that possibly could be more with individual machines on a hypervisor?
| |
10:31 | <sbalneav> Ceiling as in a hardcoded limit, or ceiling as in machine's cpu pukes on the floor from overload?
| |
10:32 | <dmarkey_> the latter
| |
10:32 | as in everone's session starts to get affected
| |
10:32 | <sbalneav> Last time I checked, hypervisors don't magically make cpu cycles out of thin air.
| |
10:33 | if you have a machine with 20 ltsp sessions, it uses 20 * x cpu cycles
| |
10:33 | if you have a machine running a hypervisor with 20 virtual machines, it's going to use 20 * (x + y) where y = overhead associated with running a hypervisor
| |
10:34 | so people are going to be affected WORSE that way.
| |
10:34 | <dmarkey_> im not sure that's entirely true
| |
10:34 | <sbalneav> Please explain.
| |
10:35 | I'm just dying to hear how adding more things for the cpu to do is going to make it faster :)
| |
10:36 | <johnny> well, it would isolate insane processes
| |
10:36 | so other people could continue to work
| |
10:36 | <sbalneav> How?
| |
10:36 | <dmarkey_> Because each user has their own machine, they dont get affected quote as severely from another VM hogging the CPU, if they wer all on the same kernel instance the affect would be dramatic
| |
10:36 | <johnny> because you can restrict the upper bound on cpu usage i think?
| |
10:37 | per instance
| |
10:37 | so.. only the people or person on that one instance would get affected by their own insane flash :)
| |
10:37 | alexqwesa has quit IRC | |
10:37 | <johnny> other people who weren't using flash could continue doing what they were doing
| |
10:37 | <dmarkey_> johnny: exactly
| |
10:37 | * sbalneav rolls eyes | |
10:37 | <johnny> dmarkey_, i think sbalneav thinks it's the wrong place to solve the problem
| |
10:37 | and he's right
| |
10:38 | but still.. not much we can really do
| |
10:38 | <sbalneav> You know, instead of everybody theorizing about how wonderful things would be in this regard
| |
10:38 | <johnny> i'm not theorizin nothin
| |
10:38 | my 3 clients are fine :)
| |
10:38 | lol
| |
10:38 | <sbalneav> I wish someone would ACTUALLY DO IT AND POST SOME HARD DATA SOMEWHERE
| |
10:38 | <johnny> -> dmarkey_ :)
| |
10:39 | i just think it's too annoying to manage
| |
10:39 | <dmarkey_> its very hard to find a metric for user experience
| |
10:39 | <johnny> you could video teh difference tho
| |
10:39 | screencast*
| |
10:39 | as how it affected responsiveness on other terminals
| |
10:39 | <sbalneav> And besides, you can also set cpu limits in a plain single instance as well
| |
10:39 | via the nice command.
| |
10:40 | <johnny> nice often doesn't work as nicely as i wish
| |
10:40 | <sbalneav> and with 20 hypervisors, you've got 20 inits
| |
10:40 | <johnny> it's still not the same aes forcing it at the hypervisor
| |
10:40 | <sbalneav> 20 buffer-flushers
| |
10:40 | <johnny> sbalneav, yes.. i think it's the wrong way to solve the problem
| |
10:40 | <dmarkey_> johnny: what's your suggestion?
| |
10:40 | <johnny> the right way is to fix the apps
| |
10:40 | :)
| |
10:40 | <dmarkey_> well, thats never going to happen
| |
10:40 | <johnny> for flash? or other?
| |
10:41 | for others.. it seems possible
| |
10:41 | <dmarkey_> you're always going to get badly designed apps that hog CPU
| |
10:41 | <sbalneav> Apps get fixed. It just takes someone willing to invest the time and effort to fix the app.
| |
10:41 | <johnny> sbalneav, except flash.. :)
| |
10:41 | <dmarkey_> thwe only way to contain them is via virtualisation
| |
10:41 | <sbalneav> johnny: gnash
| |
10:41 | <johnny> that requires coding experience far beyond most people
| |
10:42 | including myself
| |
10:42 | altho.. hopefully all this new fangled gstreamer cairo stuff will help
| |
10:42 | <dmarkey_> what about a user that compiles an infinate loop and runs it.
| |
10:42 | <sbalneav> How's virtualization going to fix that?
| |
10:43 | <johnny> there was a nice daemon to stop that kidna thing
| |
10:43 | sbalneav, only make it affect a limited number of people
| |
10:43 | <sbalneav> you either have to manage a runaway process, or a runaway virtual machine.
| |
10:43 | <johnny> that's what..
| |
10:43 | the virutal machine won't run away
| |
10:43 | <dmarkey_> scales back their VM instance so its doesnt get CPU time, easy
| |
10:43 | <johnny> or shouldn't at least..
| |
10:43 | <sbalneav> Then howcome I get vmware machines that end up sucking down my vmware farm?
| |
10:43 | when something goes awry?
| |
10:44 | <dmarkey_> i've been using xen for years, never got a VM that affected another VM, except for maybe a bit of disk IO
| |
10:44 | <sbalneav> Of course they do
| |
10:45 | if you've got 2 virtual machines
| |
10:45 | 1 using 100% cpu
| |
10:45 | the other using 50% cpu
| |
10:45 | the host didn't magically get 50% more cpu
| |
10:45 | They're being throttled
| |
10:46 | alkisg has quit IRC | |
10:46 | <sbalneav> virtualization doesn't magically make cpu cycles apear out of nowhere.
| |
10:47 | all you're doing is moving the problem/issue up (or down, depending on your viewpoint) one level.
| |
10:48 | johnny has left #ltsp | |
10:48 | johnny has joined #ltsp | |
10:49 | CAN-o-SPAM has quit IRC | |
10:50 | <sbalneav> Now, if you want to say, "I'd rather MANAGE the issue at a vm level, rather than on a per-process level" that's ok.
| |
10:50 | but vm'ing things doesn't make the problem go away, it just chages where you deal with the problem.
| |
10:52 | norbi has joined #ltsp | |
10:53 | <norbi> Hello, I currently have an LTSP set-up. Wanted to know how can I check client info from the client machine. If I do cat /proc/meminfo it gives me server info. Thanks for help.
| |
10:56 | <sbalneav> norbi: Probably want to use ltsp_localapp to run a terminal so you can look at /proc/meminfo
| |
10:57 | johnny: I notice I killed the conversation. I hope people aren't taking me wrong
| |
10:57 | I'm just trying to point out that vm's don't SOLVE the problem, they just make the problem DIFFERENT, I'm not trying to be obstreperous.
| |
10:59 | <johnny> i think everybody knows that sbalneav
| |
10:59 | it's just to confine the problem
| |
10:59 | not solve it
| |
11:00 | confining it makes it easier to not affect other people
| |
11:00 | when nice doesn't quite do the trick
| |
11:03 | <norbi> sbalneav: Sorry I'm not sure what you mean, is that an app that i download? I am running LTSP built into Edubuntu.
| |
11:04 | <Gadi> norbi: type: ltsp-localapps xterm in a terminal window
| |
11:04 | that pops up another terminal window that is local to the client
| |
11:07 | <norbi> Excellent, that worked, thank you. Another question, not so much related to ltsp. Is there a utility which i can test HDD on client and Memory on client?
| |
11:07 | alexqwesa has joined #ltsp | |
11:11 | alexqwesa has quit IRC | |
11:18 | johnny has left #ltsp | |
11:18 | johnny has joined #ltsp | |
11:19 | norbi has quit IRC | |
11:26 | alexqwesa has joined #ltsp | |
11:34 | scottmaccal has joined #ltsp | |
11:35 | <scottmaccal> Snow day today.
| |
11:37 | Lns has joined #ltsp | |
11:47 | timoteoramos has joined #ltsp | |
11:47 | Brian_H has joined #ltsp | |
11:53 | anivair has joined #ltsp | |
11:53 | <anivair> Afternoon all. I hope someone is in hte mood to help me out a bit, because I'm sort of stumped.
| |
11:53 | I have a fresh install of ltsp running on a fresh install of ubuntu (karmic).
| |
11:54 | vrand new server that is ten times the hardware of our old one, but it's running at something like half the speed
| |
11:54 | everything is moving at a crawl. I'm wondering if it's to do with the processor, since it's across the board. I've never used ltsp on an x64 machine before
| |
11:55 | <johnny> well.. check the system monitor or top
| |
11:55 | see what is using up your resources (or not)
| |
11:59 | hersonls has quit IRC | |
11:59 | hersonls has joined #ltsp | |
12:00 | ehrenpr has joined #ltsp | |
12:00 | <ehrenpr> is there anyone around who can help me setup italc?
| |
12:06 | squarepeg_ has joined #ltsp | |
12:10 | _UsUrPeR_ has joined #ltsp | |
12:11 | ehrenpr has quit IRC | |
12:12 | <anivair> nada. resources don't look bad. some firefox at the top, but that's common
| |
12:13 | nothing that should be bnogging down this server
| |
12:13 | it's sort of a beast
| |
12:14 | <johnny> install iotop
| |
12:15 | check to see what's banging your disk then?
| |
12:15 | <anivair> will do
| |
12:16 | I was hoping that it was tracker. like maybe it was just a problem when they first logged on because tracker was being a jerk, but it didn't get better
| |
12:18 | squarepeg_ has quit IRC | |
12:18 | Eghie has joined #ltsp | |
12:19 | <johnny> anivair, tracker doesn't stop being a jerk
| |
12:19 | i disabled tracker on all my installs until 0.7 comes out
| |
12:19 | 0.7 will be the first one that is any good
| |
12:19 | anivair, disable it in /etc/xdg/autostart
| |
12:19 | edit both tracker files
| |
12:19 | there's an autostart key in there.. forget what it is called
| |
12:19 | <Eghie> hello
| |
12:19 | <anivair> well, i checked for waus to disable tracker, but the ubuntu pages suggest adding an auto-hide option, which doesn't sound like it actually disables anything
| |
12:19 | <johnny> i just told you how
| |
12:20 | <Eghie> are there any script to automaticly copy the LTSP client to the harddrive of the thin client?
| |
12:20 | <johnny> why?
| |
12:20 | <anivair> yeah, saw that after I responded
| |
12:20 | <johnny> why would you want to do that?
| |
12:20 | <Eghie> so you can offload the nbd
| |
12:20 | offload network
| |
12:20 | <johnny> it won't just work..
| |
12:20 | you would have to edit it
| |
12:20 | your problem isn't there Eghie ..
| |
12:20 | i doubt nbd is causing you too much traffic
| |
12:20 | <Eghie> don't have any problem currently though
| |
12:20 | <johnny> it's gonna be more hassle than it is wroth to make it work as you think you want it
| |
12:21 | <Eghie> nah, clients do have a harddrive on board
| |
12:21 | so I was thinking of using that some sort
| |
12:22 | <johnny> well you could use the local swap
| |
12:22 | that would be more useful
| |
12:22 | <anivair> odd ... I seem not to have any tracker files. Was it maybe removed from karmic?
| |
12:22 | huh
| |
12:22 | <Eghie> yeah, doing that already
| |
12:23 | <johnny> Eghie, that's the best way to use it.. to use local swap
| |
12:23 | <Eghie> nah, was more thinking of a way as when network is not working, a small kind of offline ubuntu for small things or something similar
| |
12:23 | <johnny> anivair,
| |
12:23 | [root@falling ~]# ls /etc/xdg/autostart/tracker*
| |
12:23 | /etc/xdg/autostart/tracker-applet.desktop /etc/xdg/autostart/trackerd.desktop
| |
12:23 | Eghie, you wouldn't be able to use ltsp anyways if you did that
| |
12:23 | as you need network anyways to get a desktop
| |
12:24 | just install a regular ubuntu if you want offline access
| |
12:24 | <Eghie> nah, more as backup solution
| |
12:24 | yeah, could do that as well
| |
12:24 | <johnny> the chroot wouldn't be able to do anything without network
| |
12:24 | eve nif you did get it on the hard drive
| |
12:24 | better to just install ubuntu
| |
12:24 | or some other distro
| |
12:25 | <Eghie> well, debootstrapping ubuntu should work I guess
| |
12:25 | <anivair> that's what I'm saying. I have no tracker files is autostart
| |
12:25 | which, i guess, means that it's not my problem, at least
| |
12:26 | <johnny> anivair, well is it surely starting up ?
| |
12:27 | <Eghie> I guess I could debootstrap edubuntu-desktop
| |
12:28 | <johnny> if you feel like it
| |
12:28 | but are you sure it's worth the hassle..
| |
12:28 | <Eghie> nah, network/server can sometimes get slow on many clients with heavy programs
| |
12:29 | <anivair> must not be. ps aux|grep tracker and trackerd give me nothing
| |
12:29 | <Eghie> and sometimes they want to use offline capabilities on such occasions
| |
12:29 | <anivair> for the time being, I won't look a gift horse in hte mouth
| |
12:29 | less tracker = better
| |
12:33 | <Eghie> johnny is swap to only thing I can use those harddisk for, to speedup the clients?
| |
12:34 | or can I lay a sort of caching layer on top of parts of the chroot?
| |
12:35 | <johnny> too much hassle
| |
12:35 | if you wanna speed up certain apps you can install them locally in the chroot
| |
12:35 | <Eghie> yeah, but I am a programmer liking actually those things ;)
| |
12:35 | <johnny> if your clients are powerful enough
| |
12:35 | <Eghie> yeah they are
| |
12:35 | <johnny> then install firefox and flash for example
| |
12:35 | in the chroot
| |
12:36 | <Eghie> using ltsp-localapps on FF openoffice and such
| |
12:36 | <johnny> i think we have docs for that..
| |
12:36 | yes
| |
12:48 | Barbosa_ has quit IRC | |
12:48 | barbosa has quit IRC | |
13:04 | Guest38881 has quit IRC | |
13:04 | Cris1 has joined #ltsp | |
13:08 | alkisg has joined #ltsp | |
13:08 | jammcq has joined #ltsp | |
13:08 | <jammcq> hey friends
| |
13:10 | spectra has joined #ltsp | |
13:13 | shawnp0wers has joined #ltsp | |
13:16 | <alkisg> Hey jammcq! An idea was dropped earier here, that maybe docs.ltsp.org could point to the wiki site, so that it would be shorter for the irc /topic...
| |
13:16 | johnny has left #ltsp | |
13:16 | johnny has joined #ltsp | |
13:17 | <shawnp0wers> Or http://tinyurl.com/ltspdocs
| |
13:18 | <_UsUrPeR_> hey all. I'm trying to figure out what happened to etherboot support in 9.10.
| |
13:18 | any pointers?
| |
13:18 | <shawnp0wers> Yeah -- it's broken
| |
13:18 | :D
| |
13:19 | <_UsUrPeR_> nuuuuu -_-
| |
13:19 | <shawnp0wers> hehehehe
| |
13:19 | and low priority to fix, due to gPXE
| |
13:19 | timoteoramos has quit IRC | |
13:19 | <shawnp0wers> I have to re-do most of my older clients that have a rom-o-matic image on them
| |
13:19 | and put gPXE on them
| |
13:20 | timoteoramos has joined #ltsp | |
13:20 | <_UsUrPeR_> yeah, well there's a lot of units out there, and we can't rom-o-matic all of them :)
| |
13:20 | <shawnp0wers> Also: SNOW DAY! W00T!
| |
13:20 | <_UsUrPeR_> srsly?
| |
13:20 | <shawnp0wers> heheheheh
| |
13:20 | oh yeah
| |
13:20 | <_UsUrPeR_> what does it take to get a snow day up there?
| |
13:20 | <shawnp0wers> Something broke in making the nbi etherboot image
| |
13:20 | Gadi knew more about it than I did
| |
13:20 | timoteoramos2 has joined #ltsp | |
13:20 | <_UsUrPeR_> down here, it seemed to require the Pearl Harbor of snowshowers to get my school to give us a day off.
| |
13:20 | <shawnp0wers> Um -- like a foot of snow and 35MPH winds...
| |
13:20 | timoteoramos2 has quit IRC | |
13:21 | <_UsUrPeR_> gadi: ping?
| |
13:21 | <shawnp0wers> Poor Gadi has given me so much help with my 9.10 rollout.
| |
13:21 | * shawnp0wers sends virtual pizza and beer to Gadi | |
13:22 | alexqwesa has quit IRC | |
13:23 | <_UsUrPeR_> how's clustering going?
| |
13:23 | <shawnp0wers> Seems good. I still get that intermittent problem where a user hangs during login.
| |
13:24 | I can't seem to get a pattern though, nor any error messages that make sense
| |
13:24 | Do appservers need to be identical to balance correctly?
| |
13:24 | my 4 are, but if I add more to the cluster, they won't be...
| |
13:25 | timoteoramos has quit IRC | |
13:26 | timoteoramos has joined #ltsp | |
13:26 | <shawnp0wers> _UsUrPeR_: I'm actually thinking we might get a two-fer, and get tmorrow off too
| |
13:26 | the snow is really coming down
| |
13:26 | and is supposed to continue
| |
13:26 | :D
| |
13:26 | <_UsUrPeR_> shawnp0wers: lucky duck
| |
13:26 | <shawnp0wers> it helps that we have LOTS of back roads
| |
13:27 | <_UsUrPeR_> I thought that was the entire upper-michigan road system up there
| |
13:27 | <shawnp0wers> well yeah
| |
13:27 | <anivair> shawnp0wners: did you have a lot of trouble with 9.10? i'm having some issues now myself, and if there's something you can suggest, i'm all ears.
| |
13:27 | <shawnp0wers> but our district is very wide spread
| |
13:28 | anivair: well, not lots of problems really -- what sort do you mean?
| |
13:28 | I am using a single NIC setup, which eliminates a lot of issues
| |
13:28 | which _UsUrPeR_ reminded me of yesterday. :)
| |
13:28 | <anivair> I'm having issues with the machine running at a crawl (despite a LOT more ram and processing power than the old server)
| |
13:29 | <shawnp0wers> the server is running slow, or the thin clients?
| |
13:29 | <anivair> I'm also single nic (I have 2 installed, but only one is live)
| |
13:29 | <shawnp0wers> might it be trying to use the 2nd NIC?
| |
13:29 | DNS issue?
| |
13:29 | timeout issue?
| |
13:29 | <anivair> hmm ... the clients, I suppose, though the server does seem to hang occasionally when a client is stuck on something
| |
13:29 | it hangs temporarily. 30 seconds tops
| |
13:29 | both are possible
| |
13:29 | <shawnp0wers> that sounds like DNS or network issues
| |
13:30 | <anivair> drat. networking is not my strong suite.
| |
13:30 | <shawnp0wers> I bond my two ethernet ports together, and use them both with one IP address
| |
13:30 | <anivair> I didn't even know you could do that
| |
13:30 | <shawnp0wers> yeah -- those 30 second hangs really sound like DNS
| |
13:30 | <anivair> *hands in his geek card*
| |
13:30 | <shawnp0wers> oh yeah, bonding ethernet ports is so cool
| |
13:30 | it's not hard either, but can require a switch that supports it
| |
13:30 | <anivair> any good way to check to see if it's DNS?
| |
13:31 | <shawnp0wers> for the best performance
| |
13:31 | um...
| |
13:31 | well, not a surefire way
| |
13:31 | <anivair> heh
| |
13:31 | <shawnp0wers> check /etc/resolv.conf on the server
| |
13:31 | see what's in there
| |
13:31 | make sure it's pointing to somehting that works
| |
13:31 | also make sure your DHCP server is handing out DNS info to the clients
| |
13:32 | <anivair> it's pointing to my primary and secondary DNS, definately
| |
13:32 | <shawnp0wers> hmm
| |
13:32 | is DNS firewalled off for the thin clients?
| |
13:32 | <anivair> with the exception of some whitespace, it's the same file that I have on two other working servers
| |
13:32 | it is
| |
13:32 | well, the whole network is behind a firewall
| |
13:33 | <shawnp0wers> but can the thin clients reach the dns servers they are assigned via DHCP?
| |
13:33 | if they can't reach the servers, they might hang while trying
| |
13:33 | I'm really just guessing here though
| |
13:33 | alexqwesa has joined #ltsp | |
13:33 | <shawnp0wers> is your second ethernet port configured at all?
| |
13:34 | maybe something is trying to use it...
| |
13:34 | <anivair> DHCP
| |
13:34 | <shawnp0wers> wait, your second ethernet port is running DHCP?
| |
13:34 | <anivair> I suppose it could be that the thin clients are trying to use the disabled post
| |
13:34 | <shawnp0wers> it depends on how it is disabled I guess
| |
13:35 | <anivair> er, no. the thin clients are assigned via DHCP (standard LTSP style)
| |
13:35 | there is no other dhcp running in the office
| |
13:35 | the second port ... let me check
| |
13:35 | <shawnp0wers> also, what gateway do the thin clients get assigned?
| |
13:35 | perhaps they can't access directly, but the server can?
| |
13:35 | <anivair> the primary is eth0 and it's configured right. the second is eth1 and i is blank, i believe
| |
13:35 | hmm
| |
13:35 | I'm not sure what gateway they get assigned
| |
13:36 | I assumed they inherited the gateway info from the server
| |
13:36 | <shawnp0wers> it might be doing something as simple as trying to reach an NTP server, and hanging while trying
| |
13:36 | single nic setups get complicated, for completely other reasons that 2 NIC setups are complicated. :D
| |
13:36 | <anivair> heh
| |
13:36 | at least for this switch I kept hte old server running. that was a bad mistake last time!
| |
13:37 | <alkisg> "(09:32:06 μμ) shawnp0wers: also make sure your DHCP server is handing out DNS info to the clients" ==> Is that working in 9.10? I thought we fixed it (it wasn't supported before, we had to use DNS_SERVER in lts.conf), but I didn't test it...
| |
13:38 | <anivair> alkisg: if it were not, would it hang before reaching the outside internet, or would it not reach it at all, so you think?
| |
13:38 | the thin clients can access the net, though slowly
| |
13:38 | (they seem ot be doing everyhting slowly)
| |
13:38 | <alkisg> anivair: well, run ltsp-localapps xterm and see :)
| |
13:38 | <anivair> if I didn't know better, I'd say it was low ran, but I know that cant be the case
| |
13:39 | hrm ... I'll have to jump on a thin client to run that
| |
13:39 | I'm ssh'd into the server ATM
| |
13:39 | <alkisg> Well the clients access the internet in the same way as the server
| |
13:39 | alexqwesa has quit IRC | |
13:40 | <alkisg> DNS would be only used for localapps, do you use localapps?
| |
13:40 | <shawnp0wers> alkisg: I wasn't sure if anything else needed DNS
| |
13:40 | <anivair> I have not configured it, so if it's not configured by default, I am not
| |
13:40 | <alkisg> Then dns, gateways etc are not an issue,
| |
13:40 | <anivair> (it's been a few versions since I installed ltsp last, so I'm not sure what has changed exactly)
| |
13:40 | 8.04 was the last version of ubuntu I installed on
| |
13:40 | <alkisg> except for reverse dns lookups that the server does when it sees the IPs of the clients
| |
13:40 | <anivair> hmm
| |
13:41 | <alkisg> So if you don't have dns entries for them, maybe you can put them in /etc/hosts
| |
13:42 | <anivair> wait, I should put my dns servers in /etc/hosts? I don't think I've ever heard of that.
| |
13:42 | <alkisg> No no
| |
13:42 | <anivair> oh, good
| |
13:42 | (I was confused)
| |
13:42 | <alkisg> When the server is contacted by "10.123.123.123", it tries to find the dns name for that ip
| |
13:43 | So if your dns server knows that that IP corresponds to ltsp123, then it replies promptly
| |
13:43 | <anivair> I see.
| |
13:43 | <alkisg> If your dns server doesn't know about it, then you can put "10.123.123.123 ltsp123" in /etc/hosts
| |
13:43 | <anivair> the ltsp server is listed in /etc/hosts
| |
13:44 | I've actually got en entry for localhost and for the hostname of hte server
| |
13:44 | I've got localhost at 127.0.0.1 and the hostname at 127.0.1.1
| |
13:46 | alexqwesa has joined #ltsp | |
13:52 | pmatulis has quit IRC | |
13:52 | <dmarkey_> sbalneav: sorry got pulled into a meeting
| |
13:55 | Kicer86 has joined #ltsp | |
13:55 | davidj has quit IRC | |
13:57 | <dmarkey_> i just think it might be a good experement to swee the results
| |
13:57 | see*
| |
13:59 | <johnny> nobody is stoppin ya dmarkey_ :)
| |
14:00 | <dmarkey_> i might a student to do it as a final year project
| |
14:00 | get*
| |
14:03 | shawnp0wers has quit IRC | |
14:13 | anivair has quit IRC | |
14:15 | shawnp0wers has joined #ltsp | |
14:22 | CAN-o-SPAM has joined #ltsp | |
14:23 | rjune has joined #ltsp | |
14:26 | lucascoala_ has quit IRC | |
14:34 | shawnp0wers has quit IRC | |
14:39 | shawnp0wers has joined #ltsp | |
14:46 | Kicer86 has quit IRC | |
14:48 | Ahmuck has joined #ltsp | |
14:56 | alexqwesa has quit IRC | |
15:01 | timoteoramos has quit IRC | |
15:01 | _UsUrPeR_ has quit IRC | |
15:02 | hersonls has quit IRC | |
15:02 | timoteoramos has joined #ltsp | |
15:12 | ehrenpr has joined #ltsp | |
15:12 | shawnp0wers has quit IRC | |
15:13 | <ehrenpr> can anyone help me get italc set up? I can connect if I am logged into the server but, I cant connect from the same usere on a thin client. It complains about service not installed or broken keys.
| |
15:13 | vagrantc has joined #ltsp | |
15:14 | <sbalneav> dmarkey_: I'd be delighted to see some hard data. I have no vested interest in being proved right or wrong, my only issue is one of dealing with the symptom vs. dealing with the problem. I think vm'ing everything creates a LOT of overhead, and in the end, doesn't actually solve anything.
| |
15:14 | <johnny> it solves some peoplee getting screwed by others? :)
| |
15:14 | one method of virtuallization that would have less overhead would be something like linux-vserver or openvz
| |
15:15 | as you're not virtualzing the kernel when you do that
| |
15:15 | <sbalneav> No, they just get screwed differently
| |
15:15 | f'rinstance. Lets say you cap cpu usage.
| |
15:15 | <johnny> the people who didn't launch the insane job are the ones who aren't getting screwed anymore
| |
15:15 | <sbalneav> ok, so one runaway process can't bog down anyone else.
| |
15:15 | <johnny> and that is the major benefit
| |
15:16 | <sbalneav> but, lets say you've got 19 people typing in openoffice, using up 10% cpu
| |
15:16 | and one guy who needs to do a fourier transform, which will take 1 minute of cpu usage.
| |
15:16 | Egyptian[Home] has quit IRC | |
15:16 | <sbalneav> he's now going to take 10 minutes, 'cuz he's throttled down so hard.
| |
15:17 | _UsUrPeR_ has joined #ltsp | |
15:17 | <sbalneav> whereas before, he wouldn't have affected the other 19, since the cpu's loafing on them.
| |
15:17 | <johnny> well, you can of course set some thresholds for bursting
| |
15:17 | so as to minimize that problem
| |
15:17 | <sbalneav> And then someone exceeds them
| |
15:18 | and your management of the system becomes much harder.
| |
15:18 | <johnny> at least they don't bog down anybody else anymoer :)
| |
15:18 | of course..
| |
15:18 | <sbalneav> your kindergarten teacher who wants to show videos of fluffy bunnies now has to become a kvm expert
| |
15:18 | <jammcq> mmmm, fluffy bunnies
| |
15:19 | makes an awesome stew
| |
15:19 | <sbalneav> Like I say, you haven't *solved the problem*, you've *just made the problem different*
| |
15:19 | jammcq!!!!!!!!!!!!!!!!!!!!!!!!!
| |
15:19 | <johnny> no.. not solved..
| |
15:19 | it's not about solving.. only minimizing the impact
| |
15:20 | <sbalneav> Taking aspirin for a brain tumor minimizes the impact too.
| |
15:20 | CAN-o-SPAM has quit IRC | |
15:20 | <sbalneav> Brain don't hurt so much :)
| |
15:22 | Plus, I'd wager given machine X that could handle 20 *real* ltsp clients, is probably only going to be able to handle 15 kvm sessions.
| |
15:22 | But I'm pulling numbers out of my *ss on that one, we'd need hard data to know for sure.
| |
15:23 | <dmarkey_> sbalneav: clients doing what
| |
15:23 | <sbalneav> "normal stuff", for whatever value of "normal" we have.
| |
15:23 | Here's a real hard number:
| |
15:23 | <dmarkey_> so, on youtube :)
| |
15:24 | <sbalneav> 15:24:11 up 79 days, 1:46, 35 users, load average: 0.96, 1.49, 1.44
| |
15:24 | k, this is a 2 dual core xeon box
| |
15:24 | 8 gigs of ram
| |
15:24 | <dmarkey_> how many read users
| |
15:24 | real
| |
15:25 | <sbalneav> 35
| |
15:25 | like it says at the top
| |
15:25 | <dmarkey_> 35 seperate sessions?
| |
15:25 | <sbalneav> yes
| |
15:25 | so, how many virtual machines am I going to be able to run on that box?
| |
15:26 | <dmarkey_> 36
| |
15:26 | :)
| |
15:26 | <sbalneav> You are so full of grade A certified b********
| |
15:26 | and you know it.
| |
15:26 | try MAYBE 10
| |
15:26 | <vagrantc> depends on what your bottlenecks are
| |
15:26 | <dmarkey_> its going to be memory
| |
15:27 | <sbalneav> I don't HAVE any, machine's running fine.
| |
15:27 | <vagrantc> linux-vserver and openvz are basically glorified chroots
| |
15:27 | <dmarkey_> yea, you'll still have the same probs i think
| |
15:27 | <vagrantc> if it ain't broke, and the economy sucks, go ahead and fix it
| |
15:28 | Egyptian[Home] has joined #ltsp | |
15:29 | <vagrantc> there's no problem like no problem
| |
15:30 | <dmarkey_> well maybe that model doesnt work for LTSP
| |
15:30 | <sbalneav> No, no, VM's solve everything!
| |
15:31 | <dmarkey_> thank god.. seeing sense at last!
| |
15:31 | <sbalneav> They're like magical moonbeams! they make cpu cycles and memory appear out of thin air!
| |
15:31 | <vagrantc> virtual machines rock as far as backups, recovery, remote maintenance, etc. but there's some overhead.
| |
15:31 | <sbalneav> so, dmarky_, your challenge, should you choose to accept it...
| |
15:32 | * vagrantc could use a few magical moonbeams | |
15:32 | <sbalneav> is to tell me how much hardware I'm going to need in order to run 35 separate VM's with an aggregated loadavg of 1.5, for 35 users.
| |
15:32 | to do openoffice, thunderbird, and firefox (no youtube) :)
| |
15:33 | <Eghie> sbalneav: do you count ltsp-localapps?
| |
15:33 | <sbalneav> no
| |
15:33 | I do NO localapps
| |
15:33 | <vagrantc> well, openvz and linux-vserver take virtually no additional overhead
| |
15:33 | <Eghie> hehe, ok
| |
15:33 | <sbalneav> all. on. the. server.
| |
15:33 | <Eghie> ah ok
| |
15:33 | <vagrantc> uses the same ram and disk space (if you choose to)
| |
15:33 | <dmarkey_> vagrantc: but very little advantage
| |
15:34 | <vagrantc> dmarkey_: i wouldn't say that.
| |
15:35 | <dmarkey_> would an openvz jail deal with a runaway process better than a vanilla kernel?
| |
15:35 | <vagrantc> yes
| |
15:35 | well, it could
| |
15:35 | <johnny> there are quite a few things to tweak in regards to that dmarkey_ ..
| |
15:35 | <vagrantc> takes some administrative overhead to set some limits
| |
15:36 | <johnny> bit alot simpler than xen or kvm..
| |
15:36 | <vagrantc> and to make sure those limits aren't evil. but it gives some additional tools to use for very little cost.
| |
15:37 | i'm not saying it's definitely going to solve all your problems, but i wouldn't completely write it off either.
| |
15:37 | <dmarkey_> i suppose each chroot could be an lvm snapshot, so deployment would be fast
| |
15:38 | <johnny> gotta be a way to take less space..
| |
15:38 | share the / somewhow
| |
15:38 | as you really only care about the process segregation
| |
15:38 | <dmarkey_> hmm..
| |
15:40 | suppose you could do a mount --bind
| |
15:40 | otavio_ has joined #ltsp | |
15:41 | <alkisg> ehrenpr: distro / version?
| |
15:44 | <ehrenpr> edubuntu karmic
| |
15:44 | <alkisg> ehrenpr: and you installed italc on the server only, or did you put it at the chroot as well?
| |
15:45 | alexqwesa has joined #ltsp | |
15:45 | otavio__ has joined #ltsp | |
15:46 | <ehrenpr> at first on the server then couldnt connect to italc master from client. and in chroot. Right now I am tring again with these instructions for chroot install https://help.ubuntu.com/community/UbuntuLTSP/iTalc
| |
15:47 | johnny has left #ltsp | |
15:47 | johnny has joined #ltsp | |
15:48 | otavio has quit IRC | |
15:49 | <alkisg> ehrenpr: ok, choose which method you want so that we can go on with debugging it. Just on the server is much, much simpler, but if for some reason you prefer the chroot method, np...
| |
15:50 | <ehrenpr> ok i am almost done with chroot option.
| |
15:50 | <alkisg> First, lets check the keys: sudo diff -r /etc/italc/keys /opt/ltsp/i386/etc/italc/keys
| |
15:50 | What does this say?
| |
15:58 | otavio_ has quit IRC | |
16:00 | <ehrenpr> i have to reboot
| |
16:00 | ehrenpr has quit IRC | |
16:00 | <Eghie> how much does nbd actually use on average?
| |
16:00 | as network traffic
| |
16:03 | ehrenpr has joined #ltsp | |
16:06 | <ehrenpr> it lists a ton of keys, admin/key, ate/supporter/key etc
| |
16:07 | stevesmith1983 has joined #ltsp | |
16:07 | <alkisg> ehrenpr: then the keys are not copied correctly. You need to copy the server keys to the chroot.
| |
16:08 | (the public ones)
| |
16:08 | <ehrenpr> How?
| |
16:08 | _UsUrPeR__ has joined #ltsp | |
16:09 | <ehrenpr> just cp /etc/italc/keys/public to chroot?
| |
16:09 | <alkisg> Yes, cp -a
| |
16:09 | <ehrenpr> ok let me see
| |
16:09 | <alkisg> But if the wiki doesn't have the correct instructions, you'd better use the simple method (=NOT on the chroot)
| |
16:11 | _UsUrPeR_ has quit IRC | |
16:13 | <ehrenpr> sudo cp -a /etc/italc/keys/public /opt/ltsp/i386/etc/italc/keys/public still says broken keys, do i need to rebuild?
| |
16:17 | <alkisg> Who says broken keys? iTalc? Yes, you need to rebuild.
| |
16:18 | <ehrenpr> I figured Im currently rebuilding, Yes Italc
| |
16:21 | map7_ has joined #ltsp | |
16:22 | <alkisg> ehrenpr: what does this give you? ls -d /opt/ltsp/i386/etc/italc/keys/public/public
| |
16:22 | If that directory exists, you didn't copy the keys correctly...
| |
16:22 | <map7_> If I'm using the ltsp-build-image, which on my ubuntu system creates an image in /opt/ltsp/i386, what is the best way to backup that directory?
| |
16:23 | I've tried copying that directory to another name, but when I restore it, then update the image it has errors relating to ICEauthority
| |
16:23 | <Eghie> cp -Ra /opt/ltsp/i386 /opt/ltsp/i386_backup
| |
16:24 | <ehrenpr> rebooting..
| |
16:24 | <alkisg> Why?
| |
16:24 | <johnny> just cp -a
| |
16:24 | no need for -R with -a
| |
16:24 | <Eghie> ah ok
| |
16:24 | ehrenpr has quit IRC | |
16:25 | pmatulis has joined #ltsp | |
16:25 | <map7_> cool thanks for that, i've been running a few tests and thought I was missing something simple
| |
16:25 | <Eghie> could also use rsync -aHAX /opt/ltsp/i386 /opt/ltsp/i386_backup (if even the cp -a doesn't work)
| |
16:26 | but I do think cp -a does work ok
| |
16:26 | <map7_> what about tar with compression is that ok?
| |
16:26 | I plan on having a few backups and don't want them to take up much space
| |
16:27 | <Eghie> if you really want an ok backup solution, use rdiff-backup or bacula
| |
16:27 | or rsnapshot
| |
16:27 | or something
| |
16:27 | <johnny> tar with p is important..
| |
16:27 | tar cvzpf ltsp_chroot.tar.gz /opt/ltsp/i386
| |
16:27 | the p is important for the same reason -a is
| |
16:27 | keep the permissions
| |
16:28 | <map7_> thanks johnny
| |
16:29 | Eghie has quit IRC | |
16:30 | <alkisg> I think you can also backup the image in /opt/ltsp/images, and un-squashfs it if you want the chroot again...
| |
16:30 | This is already compressed...
| |
16:31 | <map7_> that's cool
| |
16:31 | <johnny> alkisg, that is'nt necessarily true
| |
16:31 | some distros trim stuff before compression
| |
16:31 | before squashing i mean
| |
16:31 | as in --exclude or whatever arg
| |
16:32 | so.. you always want to backup the full system
| |
16:32 | the full chroot that is..
| |
16:32 | <alkisg> But isn't what they exclude unnecessary? E.g. /root etc?
| |
16:34 | But sure, I've never tried to un-squashfs it, so I'm just guessing here... :)
| |
16:34 | <johnny> alkisg, sure.. but if you have stuff in the chroot like man pages
| |
16:34 | wouldn't you like to read them?
| |
16:35 | for a program that doesn't exist on the server
| |
16:35 | or whatever
| |
16:35 | <alkisg> Which distro deletes man pages? :O
| |
16:38 | * alkisg just saw that there's a switch to *not* compress the image, and wonders if that's faster on gigabit network + very old clients.. | |
16:43 | mikkel has quit IRC | |
16:43 | vagrantc has quit IRC | |
16:47 | alkisg has quit IRC | |
16:59 | jammcq has quit IRC | |
17:04 | din_os has quit IRC | |
17:05 | Cris1 has quit IRC | |
17:07 | bobby_C has quit IRC | |
17:11 | <johnny> uggh.. how can i find out what is using so much io on an rhel 5 box :(
| |
17:11 | i'm used to iotop :(
| |
17:14 | evilx__ has joined #LTSP | |
17:22 | evilx_ has quit IRC | |
17:24 | tarbo has quit IRC | |
17:29 | shawnp0wers has joined #ltsp | |
17:31 | Gadi has left #ltsp | |
17:49 | vagrantc has joined #ltsp | |
17:51 | kusznir has quit IRC | |
18:00 | kusznir has joined #ltsp | |
18:11 | Egyptian[Home] has quit IRC | |
18:13 | Egyptian[Home] has joined #ltsp | |
18:19 | shawnp0wers has quit IRC | |
18:24 | stevesmith1983 has quit IRC | |
18:25 | staffencasa has quit IRC | |
18:37 | map7_ has quit IRC | |
18:38 | map7_ has joined #ltsp | |
18:41 | tstafford has quit IRC | |
18:45 | tstafford has joined #ltsp | |
18:55 | shawnp0wers has joined #ltsp | |
18:56 | madcat__ has joined #ltsp | |
18:56 | Lns has quit IRC | |
18:59 | madcat__ is now known as madcat | |
19:01 | <madcat> Hello, how do you save this transcript? When I save it it's in Chinese characters, what the?
| |
19:06 | map7_ has quit IRC | |
19:07 | map7_ has joined #ltsp | |
19:09 | scottmaccal has quit IRC | |
19:14 | vagrantc has quit IRC | |
19:39 | Brian_H has quit IRC | |
19:43 | <moldy> jbernard: /aa
| |
19:44 | oops sorry :)
| |
19:59 | pmatulis has quit IRC | |
20:01 | <madcat> bye
| |
20:01 | madcat has quit IRC | |
20:44 | shawnp0wers has quit IRC | |
21:11 | madcat has joined #ltsp | |
21:13 | <madcat> Hi, why can't I run localapps on my thin client, I choose: Applications>Office>OpenOffice.org Word Processor, then I ltsp-localapps xterm, ps ax | grep office -> I don't see OpenOffice running on the thin client! Can anybody help me?
| |
21:14 | I already followed this guide: https://wiki.ubuntu.com/LTSPLocalAppSetup
| |
21:15 | My goal is to launch the application without having to type ltsp-localapps (program)
| |
21:15 | everytime - because I have 20 thin clients!
| |
21:15 | Thanks...
| |
21:16 | Also this guide: https://help.ubuntu.com/community/UbuntuLTSP/LTSPLocalAppsJaunty
| |
21:18 | Selveste1____ has joined #ltsp | |
21:19 | map7_ has quit IRC | |
21:19 | Egyptian[Home] has quit IRC | |
21:19 | otavio__ has quit IRC | |
21:19 | shogunx has quit IRC | |
21:19 | Selveste1___ has quit IRC | |
21:19 | loather-work has quit IRC | |
21:20 | map7_ has joined #ltsp | |
21:20 | Egyptian[Home] has joined #ltsp | |
21:20 | otavio__ has joined #ltsp | |
21:20 | shogunx has joined #ltsp | |
21:20 | Selveste1___ has joined #ltsp | |
21:20 | nobios has joined #ltsp | |
21:20 | loather-work has joined #ltsp | |
21:20 | loather-work has joined #ltsp | |
21:27 | madcat has quit IRC | |
21:27 | madcat has joined #ltsp | |
21:28 | <madcat> Hi is someone here?
| |
21:32 | foxbuntu has quit IRC | |
21:33 | madcat has quit IRC | |
21:34 | Selveste1___ has quit IRC | |
21:42 | johnny has left #ltsp | |
21:43 | johnny has joined #ltsp | |
21:44 | alincoln has quit IRC | |
21:47 | alincoln has joined #ltsp | |
21:56 | tarbo has joined #ltsp | |
21:59 | nobios has quit IRC | |
22:26 | yanu has quit IRC | |
22:28 | yanu has joined #ltsp | |
22:31 | Brian_H has joined #ltsp | |
22:44 | Brian_H has quit IRC | |
22:58 | foxbuntu has joined #ltsp | |
23:01 | vicd has joined #ltsp | |
23:05 | alkisg has joined #ltsp | |
23:12 | johnny has left #ltsp | |
23:12 | johnny has joined #ltsp | |
23:46 | johnny has left #ltsp | |
23:46 | alkisg has quit IRC | |
23:46 | johnny has joined #ltsp | |
23:50 | vagrantc has joined #ltsp | |