00:14 | plamengr has joined #ltsp | |
00:45 | plamengr has quit IRC | |
00:51 | DonSilver has joined #ltsp | |
01:24 | DonSilver has quit IRC | |
01:27 | mikkel has joined #ltsp | |
01:31 | gate_keeper_ has joined #ltsp | |
01:38 | gate_keeper__ has joined #ltsp | |
01:52 | gate_keeper_ has quit IRC | |
01:55 | Pascal_1 has joined #ltsp | |
02:08 | Pascal_1 has left #ltsp | |
02:29 | Eghie has joined #ltsp | |
02:41 | <nubae1> any bash scripters that can help me out? I'm trying to get the internal server ip only and have a convoluted way of doing it, but wonder if there is a better way, as my script assumes the ip is always on eth1
| |
02:42 | <ltsppbot> "Nubae" pasted "catch ip" (7 lines) at http://pastebot.ltsp.org/9
| |
03:20 | Q-FUNK has joined #ltsp | |
03:28 | gate_keeper__ has quit IRC | |
03:31 | DawnLight has joined #ltsp | |
03:31 | <DawnLight> hello. i have an ibm netvista thin client here. it is an old thingy. the first problem with it is that it doesn't want to use pxe boot despite that i configure it's bios to do so. so it uses NFS. despite that i tell it not to do so. so i use nfs to give it the pxelinux.0 image and it fetches it and then it says that it is an unsupported file type... why?
| |
03:34 | <nubae1> DawnLight: doesn't the pxe network card have its own configuration panel (usually shift+f10 or some such)
| |
03:35 | <DawnLight> nubae1: i don't think so. this has a strange ibm configuration utility
| |
03:36 | <Q-FUNK> howdy!
| |
03:37 | <DawnLight> and it makes strange dhcp requests like 66 and 67 and two hundred or something
| |
03:37 | can you think of a reason?
| |
03:38 | i'll be back later
| |
03:38 | gate_keeper_ has joined #ltsp | |
03:48 | nubae1 is now known as nubae | |
03:49 | <nubae> hiu Q-FUNK
| |
03:53 | cyberorg has joined #ltsp | |
04:01 | siki has joined #ltsp | |
04:04 | <highvoltage> howdy Q-FUNK
| |
04:34 | Ryan52 has quit IRC | |
04:58 | nubae has quit IRC | |
05:34 | jonnor has joined #ltsp | |
06:10 | nubae has joined #ltsp | |
06:39 | nubae has quit IRC | |
06:39 | nubae has joined #ltsp | |
06:54 | jammcq has quit IRC | |
07:08 | psycodad has quit IRC | |
07:08 | psycodad has joined #ltsp | |
07:20 | ace_suares has joined #ltsp | |
07:23 | <ace_suares> how to make upstart recognize a new file/program that need to start at boot? Is rebooting the whole machine the only way ?
| |
07:37 | psycodad has quit IRC | |
07:39 | <Q-FUNK> ace_suares: it's suposed to be event-based, so there must be a way to launch it on the fly
| |
07:41 | randra has joined #ltsp | |
07:43 | psycodad has joined #ltsp | |
07:44 | <ogra> that question would probably better be asked in #upstart :)
| |
07:45 | though if the package/file/whatever uses a dbus service that will require a reboot anyway
| |
07:55 | Q-FUNK has quit IRC | |
08:11 | gate_keeper_ has quit IRC | |
08:16 | _UsUrPeR_ has joined #ltsp | |
08:22 | Blinny has joined #ltsp | |
08:31 | vmlintu has joined #ltsp | |
08:33 | mikkel has quit IRC | |
08:39 | moldy_ has joined #ltsp | |
08:40 | Warp4 has joined #LTSP | |
08:41 | <Warp4> ok a quick question...
| |
08:41 | we are running LTSP under Debian Etch here at work. I have a user that has an issue that everytime he clicks on the K menu it freezes and if he does it again it crashes.
| |
08:47 | bobby_C has joined #ltsp | |
08:51 | <sbalneav> Morning all
| |
08:51 | <Warp4> good morning
| |
08:55 | psycodad has quit IRC | |
08:55 | Blinny has quit IRC | |
08:56 | <DawnLight> so how does this localdev stuff works in the ubuntu hardy version?
| |
08:57 | moldy__ has quit IRC | |
08:57 | <sbalneav> DawnLight: Are you asking IF it works, or HOW it works?
| |
08:57 | <DawnLight> i have LOCALDEV=y and kde doesn't give me any notice
| |
08:58 | i have LOCALDEV=y and kde doesn't give me any notice
| |
08:58 | <sbalneav> KDE won't pop anything up on the desktop
| |
08:58 | <DawnLight> oh no
| |
08:58 | <sbalneav> You'll need to look on /media/<user>
| |
08:58 | <DawnLight> hmm
| |
08:58 | <sbalneav> the mount will be there.
| |
08:58 | <DawnLight> ahuh
| |
08:58 | <sbalneav> Also, have you added the user to the "fuse" group?
| |
08:58 | <DawnLight> yes
| |
08:59 | <sbalneav> ok
| |
08:59 | <DawnLight> so does the mounting happen in the server?
| |
08:59 | <nubae> when installing packages for a plugin (under /usr/share/ltsp/plugins/ltsp-build-client/Ubuntu/) I have some unsigned packages... I know I can use --accept-unsigned-packages but I'd like not to have to do that...
| |
08:59 | <DawnLight> ah, i see. there is a fuse.ltspfs mount in the server
| |
09:00 | <nubae> doh, guess I can look at the accept-unsigned-packages and use that code :-/
| |
09:01 | <DawnLight> i can't see my files on the usb
| |
09:01 | is it soppused to be mounted on the client?
| |
09:01 | because it isn't
| |
09:02 | <sbalneav> DawnLight: It's mounted on the server.
| |
09:02 | It will only mount on the client when it's being used
| |
09:02 | after 2 seconds of inactivity, it umounts
| |
09:02 | on the client
| |
09:04 | <DawnLight> but...
| |
09:04 | what does "used" mean?
| |
09:05 | <Warp4> we are running LTSP under Debian Etch here at work. I have a user that has an issue that everytime he clicks on the K menu it freezes and if he does it again it crashes.
| |
09:05 | <DawnLight> i go in /media/dawn/usbdisk-sda2 and do ls and nothing
| |
09:05 | <sbalneav> DawnLight: Looking at it, copying to it, reading from it
| |
09:05 | <DawnLight> 2 seconds? not 2 minutes?
| |
09:10 | so anyway... it doesn't work...
| |
09:10 | <_UsUrPeR_> warp4: this is a single client?
| |
09:10 | <Warp4> yeah just one user
| |
09:11 | <_UsUrPeR_> when you say "user" do you mean a single username, or a single client?
| |
09:11 | <Warp4> a single username
| |
09:11 | no matter what workstation he logs on to, the same thing happens
| |
09:12 | <_UsUrPeR_> ahh
| |
09:12 | hmm
| |
09:12 | that's a quandry for sure. Have you attempted to remove and re-add his account?
| |
09:13 | psycodad has joined #ltsp | |
09:13 | <Warp4> well id do that, but we use an LDAP backend for authentication
| |
09:13 | which also handles our dialup and email accounts
| |
09:15 | <_UsUrPeR_> ugh. That's my best advise, man. Sorry.
| |
09:19 | mikkel has joined #ltsp | |
09:21 | <sbalneav> DawnLight: At a command prompt, can you type the command "id" as the user, and paste the results to our pastbot?
| |
09:21 | !pastebot
| |
09:21 | <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.
| |
09:27 | Blinny has joined #ltsp | |
09:28 | gate_keeper_ has joined #ltsp | |
09:35 | <DawnLight> sbalneav: thanks. it is working suddenly. i wonder if there's an inconsistancy
| |
09:38 | the lts-parameters.txt in the 5.0.40 version does not seem to be complete. is there an online version?
| |
09:44 | psycodad has quit IRC | |
09:47 | psycodad has joined #ltsp | |
09:51 | Ryan52 has joined #ltsp | |
09:56 | gate_keeper_ has quit IRC | |
10:04 | sutula has joined #ltsp | |
10:06 | <nubae> hmmm, when creating a plugin, I use the following: APT_GET_OPTS="$APT_GET_OPTS --force-yes" to force unsigned packages... I put this in configure) but seems to be ignored
| |
10:08 | Eghie has quit IRC | |
10:11 | exodos has joined #ltsp | |
10:21 | warren has quit IRC | |
10:22 | psycodad has quit IRC | |
10:26 | CAN-o-SPAM has joined #ltsp | |
10:26 | <nubae> say I want to add an extra repo to grab packages from during ltsp-build-client, where would I do that?
| |
10:26 | Egyptian[Home] has joined #ltsp | |
10:27 | Egyptian[Home] has quit IRC | |
10:30 | Gadi has joined #ltsp | |
10:30 | <nubae> I suppose --accept-unsigned-packages overrides whatever u define for $APT_GET_OPTS
| |
10:30 | Egyptian[Home1 has quit IRC | |
10:31 | <nubae> even if I hardcode --force-yes, ltsp-build-client still switches it off at install time
| |
10:31 | hi Gadi
| |
10:31 | <Gadi> g'morning
| |
10:32 | rafaelmf has joined #ltsp | |
10:46 | vagrantc has joined #ltsp | |
10:48 | staffencasa has joined #ltsp | |
10:51 | mccann has joined #ltsp | |
11:00 | <nubae> so can someone help me... I'm trying to add an extra repo during ltsp-build-client build time, but not sure where I can do that, there must be a better way than doing echo "repo" >> /etc/apt/sources.list
| |
11:01 | <vagrantc> "ltsp-build-client --extra-help" should mention several commandline arguments...
| |
11:02 | CAN-o-SPAM has quit IRC | |
11:03 | <nubae> yeah but nothing about extra repos
| |
11:05 | <Warp4> ok another question...
| |
11:07 | <vagrantc> nubae: no --extra-mirror option?
| |
11:07 | rafaelmf has quit IRC | |
11:08 | elisboa has joined #ltsp | |
11:11 | pimpministerp has joined #ltsp | |
11:12 | <nubae> vagrantc: that might work, what I'm trying to do is get medibuntu (media repo with googleearth etc) to be included for a particular plugin
| |
11:12 | but requires signing too
| |
11:12 | <vagrantc> nubae: well, use the other *mirror plugins as a base for a new plugin
| |
11:13 | <nubae> might be easier just to include the code in the main plugin
| |
11:13 | <vagrantc> nubae: and i added a plugin on debian which will use whatever apt keyring your server uses to the client by default
| |
11:13 | <nubae> --copy-sources-list?
| |
11:13 | <vagrantc> no
| |
11:13 | it's not a commandline option, just a plugin. don't think ogra has pulled it into ubuntu yet, if that's what you're using.
| |
11:14 | <ogra> --copy-sources-list?
| |
11:14 | it comes from ubuntu
| |
11:14 | <nubae> well the ltsp source has all the plugins, debians as well
| |
11:14 | rafaelmf has joined #ltsp | |
11:14 | <nubae> maybe u mean --apt-keys
| |
11:15 | <vagrantc> just because it has the plugins, doesn't mean it uses them.
| |
11:15 | <ogra> it doesnt
| |
11:15 | <nubae> yeah but I can use the code :-)
| |
11:15 | look at and use
| |
11:15 | <ogra> it uses the Ubuntu plugins indeed, not the debian ones
| |
11:15 | <vagrantc> simple symlinking should do.
| |
11:15 | <ogra> right
| |
11:15 | <vagrantc> unless there's some incompatibility ...
| |
11:15 | <nubae> ogra, is it too late to include a --fatclient plugin in intrepid?
| |
11:15 | I've got it finished now
| |
11:16 | <ogra> nubae, its not to late but i wont have time
| |
11:16 | * ogra doesnt even have the time to chat here | |
11:16 | <nubae> ok :-)
| |
11:16 | <ogra> beta is on thursday
| |
11:16 | <nubae> ok, I'll bother u after
| |
11:16 | <ogra> i'll be lucky if i find some hours to even fix the worst ltsp things before that
| |
11:17 | and after only massively critical bugs get fixed
| |
11:17 | ltsp in intrepid will likely suck and i have to find someone to take care for it in jaunty
| |
11:17 | my time is way to limited
| |
11:18 | but stgraber already expressed some interest ... if debian wouldt always be so behind on stuff we could just use their packages :/
| |
11:18 | <nubae> so far the alpha 6 is ok... been using it for a while... just needed the usb fix...
| |
11:18 | which has already been posted... so I have no problems with it
| |
11:18 | <ogra> but it would break heavily in intrepid if i pulled just the lenny packages
| |
11:18 | i.e. no input at all etc
| |
11:19 | <nubae> hmmm
| |
11:19 | <ogra> yeah, filed bugs should still get some attention from my side
| |
11:19 | even after beta
| |
11:19 | but currently ubuntu-mobile keeps me massively busy
| |
11:19 | and jaunty will be worse
| |
11:19 | <nubae> well, I'm learning more and more, hopefully I can help out in the future
| |
11:20 | * ogra didnt expect to drown in 100s of mails a day after he released ubuntu-mobile | |
11:20 | <nubae> that popular eh?
| |
11:20 | <ogra> that would be best, getting a developer team together on ubuntu side
| |
11:20 | yeah, sadly
| |
11:21 | if i would have been able to predict that i would have released it a month earlier ... it was only supposed to show off some stuff we plan for jaunty ... but 100s of people use it
| |
11:21 | <nubae> I've been asked to put sugar onto ltsp... so that will be more learning experience, except the sugar packages are like 6 months behind on ubuntu
| |
11:21 | <ogra> http://www.umpcportal.com/2008/09/ubuntu-mobile-edition-news-and-first-boot-video/
| |
11:21 | yeah, jani didnt do many uploads of sugar after getting it in basically
| |
11:22 | <nubae> does he want someone else to take over packaging do u know?
| |
11:22 | gbolte has joined #ltsp | |
11:22 | <ogra> no idea
| |
11:22 | mail him :)
| |
11:23 | <nubae> well on launchpad there is a new sugar group set up 3 weeks ago
| |
11:23 | not sure the relation to jani
| |
11:23 | jbrett has joined #ltsp | |
11:24 | <nubae> but I'll mail him
| |
11:24 | cliebow has joined #ltsp | |
11:24 | <nubae> just one quick question... if I set up a cron job to copy across just non system users to the chroot daily, is there a security issue there?
| |
11:25 | <ogra> ah, i remember your paste ...
| |
11:25 | make sure to start at UID 1000
| |
11:26 | <nubae> ok, Fedora starts at 500...
| |
11:26 | its the simplest solution I could come up with and it works great, but wonder about security
| |
11:26 | <ogra> ah, i didnt know you implement that on fedora
| |
11:26 | its surey not massively secure
| |
11:26 | <nubae> well, I'm trying to keep it open, but the ubuntu one will start at 1000
| |
11:26 | <ogra> but i its only in your own network you can indeed close it down, put ssh keys in place etc
| |
11:27 | gonzaloaf_laptop has joined #ltsp | |
11:43 | <nubae> funny I just noticed the edubuntu logo has chn
| |
11:43 | changed
| |
11:49 | <ogra> did it ?
| |
11:49 | psycodad has joined #ltsp | |
11:53 | rafaelmf has quit IRC | |
11:56 | warren has joined #ltsp | |
11:57 | <nubae> maybe its just me... the one guy with his hand pointing up seems to be the official logo on the site, so maybe its always been like that, but I just noticed it on the top menu too, and my older edubuntu install doesnt have that
| |
11:58 | <ogra> well, it was always like that
| |
11:59 | i replace distributor-logo.png in edubuntu-artwork since day one
| |
11:59 | Warp4 has quit IRC | |
12:03 | <nubae> yeah I probably have a mix of ubuntu and edubuntu on the old system
| |
12:06 | cyberorg has quit IRC | |
12:13 | Lns has joined #ltsp | |
12:19 | psycodad has quit IRC | |
12:20 | <Lns> Is there any word on nbdrootd procs being left behind? I've been told before it doesn't affect stability/performance but was wondering if there is a way to clean them up (besides obviously rebooting).
| |
12:20 | gnome-watchdog doesn't help as it only deletes procs owned by the user logging in, and nbdrootd runs as 'nobody'.
| |
12:22 | psycodad has joined #ltsp | |
12:25 | cyberorg has joined #ltsp | |
12:27 | nubae has quit IRC | |
12:27 | <stgraber> I have pushed some of the fixes I have done here for ltsp-cluster in: https://code.launchpad.net/~stgraber/ltsp/ltsp-trunk
| |
12:28 | I haven't tried regenerating a new chroot with these but run individually they work fine
| |
12:32 | nubae has joined #ltsp | |
12:38 | juanpaul has quit IRC | |
12:40 | _UsUrPeR_ has quit IRC | |
12:44 | psycodad has quit IRC | |
12:46 | <Lns> Bueller?
| |
12:46 | =p
| |
12:50 | psycodad has joined #ltsp | |
12:50 | cyberorg has quit IRC | |
12:52 | spectra has joined #ltsp | |
12:52 | <nubae> hi Lns
| |
12:53 | <Lns> nubae: hey!
| |
12:53 | how's it going
| |
12:54 | <nubae> yah not bad... just finished writing fat client plugin... working nicely...
| |
12:56 | <Lns> nubae: nice. what does it do?
| |
12:59 | <nubae> builds a fat client in the chroot instead of thin client
| |
12:59 | basically a netbooted workstation
| |
13:00 | <Lns> nubae: ooh, *the* fat client plugin. =p
| |
13:00 | nubae: how is the performance?
| |
13:00 | compared to a TC
| |
13:00 | <nubae> pretty damn nice... if u have the ram
| |
13:01 | <Lns> nubae: how much would you say is required?
| |
13:01 | probably at least 256 i'd imagine, with gnome/firefox/etc
| |
13:01 | <nubae> the main thing is, u can run video, audio, flash, blender, etc without any issues
| |
13:01 | yeah but it will even run on 128
| |
13:01 | <Lns> nubae: how does authentication work?
| |
13:01 | <nubae> I have a thincan here and it runs on that
| |
13:01 | I chose the simples path... copy non system users to the chroot with cron
| |
13:02 | <Lns> ah
| |
13:02 | so does your fat client plugin use nbd or nfs?
| |
13:02 | <nubae> nbd
| |
13:02 | <Lns> nice
| |
13:02 | <nubae> could work with nfs too, but this is faster
| |
13:03 | <Lns> and all fat clients are still diskless?
| |
13:03 | <nubae> only thing is had to figure out how to do the user synch between server and client... since nbd needs to be rebuilt every time user is put in
| |
13:03 | yep, but they can access local disks too
| |
13:03 | home is mounted over nfs
| |
13:04 | <Lns> So even loading a hefty s/w like OOo is fairly quick?
| |
13:04 | wait...its all loaded in RAM isn't it if it's nbd?
| |
13:04 | the whole chroot
| |
13:04 | <nubae> I'm running it right now on my test machine which is a 500mhz celeron with 500 mb ram
| |
13:04 | and its quite usable
| |
13:05 | <Lns> oh nm
| |
13:05 | <nubae> I'll send u the plugin if u want to take a look
| |
13:05 | <Lns> its ok... i wish i had the time to test it out :)
| |
13:05 | * Lns needs to take care of his 7 LTSP TC sites first | |
13:06 | <nubae> yeah... I hear ya... if u ever need any help with maintenance... I'm here :-)
| |
13:06 | got a lot of time on my hands right now
| |
13:06 | <Lns> nubae: really
| |
13:06 | what do you mean by maintenance?
| |
13:07 | <nubae> help with the servers... any kind
| |
13:07 | I know ltsp like the back of my hand now
| |
13:07 | <vagrantc> Lns: you could either bind mount or unionfs mount /etc/group and /etc/passwd for the syncing ...
| |
13:08 | nubae: oops, that was to you ^^
| |
13:08 | <Lns> vagrantc: oh wow you can mount single files w/unionfs?
| |
13:08 | <nubae> thats the other option I was looking at... but looks complicated
| |
13:08 | * Lns doesn't know much about nbd | |
13:09 | <vagrantc> nubae: well, you could have a cron job that copies those files to an NFS share... and have a cron job or other script that looks for updates to those files and merges them into the thin-client as it's running.
| |
13:09 | or use something other than NFS
| |
13:10 | <nubae> merging how?
| |
13:10 | <vagrantc> well, how do you do it now?
| |
13:10 | clearly you already have a way to handle it, no?
| |
13:11 | <nubae> copy across non system users and rebuild image
| |
13:11 | its in a cron.daily script
| |
13:11 | _UsUrPeR_ has joined #ltsp | |
13:11 | <vagrantc> so, instead of re-building the whole image, copy the files somewhere the thin-clients can get at, and if they're updated, copy them into place on the thin-client
| |
13:11 | <Lns> Obviously something like LDAP would be a more elegant solution, although a whole other ball of wax
| |
13:11 | <vagrantc> indeed
| |
13:12 | simpler in some ways, more complicated in others.
| |
13:12 | CAN-o-SPAM has joined #ltsp | |
13:12 | <nubae> yeah thats possible, but then I have to make sure the server and client are installed and all the config files, its much more complex
| |
13:12 | <ogra> rsync through ssh would be an elegant solution
| |
13:12 | * ogra always fails to see whats elegant about ldap | |
13:12 | <Lns> vagrantc: definitely... but you know, LDAP seems to fit well with LTSP in many viewpoints
| |
13:12 | <vagrantc> nubae: if you're already mounting /home over nfs, you could copy /etc/group and /etc/passwd there...
| |
13:13 | * ogra always thinks about ldap as a sucks least solution, but the wrod elegant would never cdome to his mind | |
13:13 | <nubae> how do I get the thin client then to read it there?
| |
13:13 | <vagrantc> nubae: cron job on the thin client/
| |
13:13 | ?
| |
13:13 | <nubae> ok, so cronjob that copies from nfs mount to local?
| |
13:14 | <Lns> ogra: well IMHO, LDAP is a much more stable solution than having "copies" of user/group data in different places that can get out of sync
| |
13:14 | <vagrantc> nubae: right
| |
13:14 | nubae: or play with bind-mounting them...
| |
13:14 | <nubae> hmmm... yeah I'll look at that, but I think the other solution I can implement right away...
| |
13:14 | I'll keep bind mounting for version 2
| |
13:15 | :-)
| |
13:15 | ok... going to eat bbl
| |
13:15 | * ogra thought that was v2 :) | |
13:15 | <ogra> didnt you start that in gutsy ?
| |
13:15 | <nubae> yeah but this is one that is publicly usable
| |
13:15 | the other one needed lots of manual tweaks
| |
13:15 | this needs none, it just wokrs
| |
13:15 | works even
| |
13:17 | <Lns> ogra / vagrantc: can I pick your brain for a sec re: nbdrootd procs being left behind? Is there any way to clean them up nicely?
| |
13:17 | <ogra> well, lets look if we can pull it into the extra-plugins stuff in the ubuntu package so you can get some feedback
| |
13:17 | Lns, tcpd should care
| |
13:17 | gonzaloaf_laptop has quit IRC | |
13:17 | <ogra> tehy will stick around for the keepalive time it has compiled in
| |
13:18 | <Lns> ogra: what is the keepalive time? I'm sure it's not long, all of my hardy servers have stale nbdroot procs owned by 'nobody'
| |
13:18 | <ogra> might be 10h, i forgot
| |
13:18 | alkisg has joined #ltsp | |
13:18 | <Lns> oh wow
| |
13:18 | <ogra> probably vagrantc still remembers
| |
13:18 | we once looked that up
| |
13:19 | <Lns> i doubt there's an lts.conf/other option to change that, eh?
| |
13:19 | <Gadi> 2h
| |
13:19 | <Lns> Gadi!
| |
13:19 | <Gadi> lns!
| |
13:19 | <Lns> hmm.
| |
13:19 | =p
| |
13:19 | <Gadi> you can adjust keepalive time in procfs
| |
13:19 | <ogra> no, its a server side setting
| |
13:20 | nothing to adjust n lts.conf
| |
13:20 | <Lns> ogra: oh yeah, duh.
| |
13:20 | Gadi: any reason why it can't be shorter of a time? I'm assuming keepalive is so that if network connectivity is lost it doesn't have to respawn.. ?
| |
13:21 | psycodad has quit IRC | |
13:21 | psycodad has joined #ltsp | |
13:21 | <ogra> ask tcpd upstream ?
| |
13:21 | <Gadi> no - keepalive sends messages to the client and kills the daemon if the client does not respond after that time
| |
13:22 | depending upon the daemon you are having tcpd monitor, you may want long or short times
| |
13:22 | I think 2h is a compromise
| |
13:22 | <ogra> i really think its much longer
| |
13:22 | <Lns> Gadi: ok, so what if the client reboots? does it use a specific socket for that proc? in other words, if the client reboots and spawns a new nbdrootd process, will the old one timeout?
| |
13:22 | <Gadi> also depending upon the network
| |
13:23 | the old one should time out, I think
| |
13:23 | Pascal_1 has joined #ltsp | |
13:23 | <ogra> yeah
| |
13:23 | Pascal_1 has quit IRC | |
13:23 | <Lns> it's weird..I have 3 logged in users on one server w/72 nbdrootd procs running
| |
13:24 | in a 35 TC lab
| |
13:24 | <Gadi> well, are all thin clients on?
| |
13:24 | users don't matter
| |
13:24 | (I love saying that)
| |
13:24 | <Lns> lol
| |
13:24 | thats true. i'm sure the TCs are on, yes
| |
13:24 | <Gadi> nbd is about workstations that are on
| |
13:24 | irregardless of users
| |
13:24 | <ogra> so each uses one nbdswap and one nbdrootd
| |
13:24 | <Lns> right...but why have twice as may procs as TCs?
| |
13:25 | <ogra> in case you have swap enabled
| |
13:25 | <Lns> ogra: i grepped 'nbdrootd' specificaly
| |
13:26 | ok I think i get it more now... definitely, if the stations are on nbdrootd will be active (users don't matter!).. maybe the old procs were just because of a whole lab reboot recently.
| |
13:26 | All other sites have 30-40 nbdrootd's running, which is normal i'm sure.
| |
13:28 | <Gadi> Lns: cat /proc/sys/net/ipv4/tcp_keepalive_time
| |
13:28 | btw
| |
13:28 | cyberorg has joined #ltsp | |
13:28 | <Lns> Gadi: ty!
| |
13:29 | Gadi: 7200
| |
13:29 | <Gadi> in other words, you can adjust with /etc/sysctl.conf, if you like
| |
13:29 | 7200 = 120 * 60 seconds = 2 hours
| |
13:30 | <Lns> Gadi: cool. wasn't sure the units
| |
13:31 | * Lns loves his "district-wide script for running commands automatically" script | |
13:31 | <Lns> 7200 at all sites.
| |
13:31 | So, ok, it's a non-issue I guess. :) Thanks for the clarification all
| |
13:35 | gonzaloaf_laptop has joined #ltsp | |
13:45 | <ogra> yippiiiie
| |
13:45 | * ogra dances | |
13:45 | <stgraber> ogra: ?
| |
13:46 | <ogra> http://cdimage.ubuntu.com/ubuntu-mobile/intrepid/current/ ...
| |
13:46 | finally got the datacenter convinced to do what i want
| |
13:46 | <stgraber> s/cdimage/chromium/
| |
13:46 | * ogra was hacking on the build scripts for day snow | |
13:46 | <ogra> *days
| |
13:47 | <stgraber> doh, chromium is as slow as beryllium ... will be hard to do ISO testing for the beta
| |
13:47 | * stgraber dances too :) | |
13:48 | <stgraber> ltsp-build-client --arch i386 --ltsp-cluster --ltsp-cluster-control 192.168.0.4 --ltsp-cluster-lbserver 192.168.0.5
| |
13:48 | and you get ltsp-cluster installed and configured
| |
13:48 | <ogra> 706M ... not even that big
| |
13:48 | congrats !
| |
13:48 | cascade has joined #ltsp | |
13:48 | <stgraber> only bad point, it's using a PPA and then has to use -y --force-yes
| |
13:49 | when will we have signed PPAs ? :)
| |
13:49 | cascade is now known as CaScAdE^Bulgaria | |
13:50 | Q-FUNK has joined #ltsp | |
13:50 | <ogra> stgraber, there is a bug ...
| |
13:53 | hmm, cant find it
| |
13:53 | <stgraber> yeah, for a long time I guess
| |
13:54 | <ogra> yeap
| |
13:54 | <stgraber> https://bugs.launchpad.net/soyuz/+bug/125103 ?
| |
13:54 | <ogra> cjwatson subscribed me once but i lost it it seems
| |
13:55 | yep, thats the one
| |
13:56 | psycodad has quit IRC | |
14:00 | johnny has joined #ltsp | |
14:00 | irule has joined #ltsp | |
14:12 | alkisg has quit IRC | |
14:21 | nubae has quit IRC | |
14:22 | cliebow has quit IRC | |
14:26 | <vagrantc> stgraber: looks like i had a little time anyways ... so, regarding your patches ...
| |
14:26 | Blinny has quit IRC | |
14:26 | <vagrantc> stgraber: curious if for the password loop we couldn't do "while ! passwd root ; do ..." instead of set +e passwd root ... while $? = 0 ... passwd root ...
| |
14:28 | stgraber: the PULSE_SERVER thing should probably only happen if SOUND=True ... which i think should be available in the script. and then you'll need boolean_is_true from ltsp-common-functions (or wherever it is)...
| |
14:31 | stgraber: and it might be good practice to validate the sudoers file with visudo -c -f /etc/sudoers and revert if problems arise...
| |
14:33 | stgraber: i'd actually do things like adding ltsp-cluster-client to EARLY_PACKAGES or LATE_PACKAGES rather than installing directly... and using the mirror related functions to add that stuff
| |
14:33 | stgraber: but if ogra finds it acceptible as is, i don't see any reason not to merge it.
| |
14:34 | <ogra> i agree with your comments, and beyond that trust your judgement :)
| |
14:34 | if you want it usable on debian you should also make sure the package depends on sudo
| |
14:35 | <Gadi> if aplugin installs other ltsp-related packages, shouldn't we review the code there, too?
| |
14:35 | before merging it into upstream
| |
14:36 | <warren> Gadi: vagrantc: ogra: i have to tag things later today or early tomorrow
| |
14:36 | get stuff in
| |
14:37 | <ogra> feel free
| |
14:37 | <Gadi> warren: !!!
| |
14:37 | <ogra> i'm in beta freeze
| |
14:37 | <warren> ogra: did you folks disable the e1000e driver for the beta?
| |
14:37 | <ogra> and dont do much work on ltsp anyway (no time for that, mobile is to groundbraking )
| |
14:37 | <warren> ogra: pretty nasty!
| |
14:37 | destroying hardware and all that
| |
14:37 | <ogra> warren, yes
| |
14:37 | but i think there was a recent fix
| |
14:38 | so it might be enabled by the actual beta release
| |
14:38 | yeah
| |
14:38 | <warren> ogra: not according to my read upstream...
| |
14:38 | nubae has joined #ltsp | |
14:38 | <warren> ogra: there isn't a fix yet, only patches to help identify what is causing it
| |
14:39 | <ogra> ah, right, i mixed that up with a fix to e1000
| |
14:39 | e1000e is still broken
| |
14:39 | <warren> very
| |
14:40 | <ogra> the mean thing is that tesing can get really expensive if you work on that bug
| |
14:40 | <warren> yes, which is why everyone is deferring to Intel
| |
14:40 | everyone is disabling the driver for now
| |
14:40 | <Q-FUNK> hm
| |
14:40 | <warren> pretty bad given all intel hardware for almost 2 years needs e1000e
| |
14:40 | <ogra> who are not the fastes to do fixes by own experience
| |
14:40 | <warren> my older ICH7 hardware uses e1000e but isn't effected by the problem
| |
14:41 | <Q-FUNK> given the break on e1000, I wouldn't be surprised if a similar fuck up in the kernel is what prevents the deb60 from working
| |
14:41 | <warren> Q-FUNK: don't make assumptions
| |
14:41 | <ogra> and e1000 is fine :)
| |
14:41 | <warren> Q-FUNK: e1000e seems to be a stupid design decision in the ICH8 and ICH9 hardware
| |
14:41 | <ogra> e1000e isnt
| |
14:41 | <warren> ogra: e1000e on pre-ICH8 is fine
| |
14:41 | * ogra kicks hal | |
14:41 | <Q-FUNK> ditto for recent changes in the wifi driver
| |
14:42 | <ogra> evil thing
| |
14:42 | <warren> Q-FUNK: there were other big changes in 2.6.26+ that might be of concern
| |
14:42 | <Q-FUNK> yup
| |
14:42 | and many of those came from intel
| |
14:42 | <warren> Linus forced people to rip out an entire part of the VM for example, because it was ugly. It made performance really bad.
| |
14:42 | CaScAdE^Bulgaria has quit IRC | |
14:42 | <warren> He's forcing people to fix it.
| |
14:43 | * ogra is happy so far with 2.6.27 though | |
14:47 | pimpministerp has quit IRC | |
14:49 | captain_1agnus has joined #ltsp | |
14:52 | captain_magnus has quit IRC | |
14:52 | ace_suares has quit IRC | |
14:53 | ace_suares has joined #ltsp | |
15:01 | DawnLigh1 has joined #ltsp | |
15:03 | randra has quit IRC | |
15:09 | elisboa has quit IRC | |
15:14 | DawnLight has quit IRC | |
15:18 | captain_1agnus has quit IRC | |
15:19 | captain_magnus has joined #ltsp | |
15:22 | pimpministerp has joined #ltsp | |
15:27 | pimpministerp has quit IRC | |
15:27 | psycodad has joined #ltsp | |
15:28 | CAN-o-SPAM has quit IRC | |
15:28 | Q-FUNK has quit IRC | |
15:31 | vmlintu has quit IRC | |
15:33 | vmlintu has joined #ltsp | |
15:44 | mikkel has quit IRC | |
15:47 | johnny has left #ltsp | |
15:55 | psycodad has quit IRC | |
15:55 | bobby_C has quit IRC | |
16:00 | psycodad has joined #ltsp | |
16:13 | Gadi has left #ltsp | |
16:20 | Egyptian[Home] has joined #ltsp | |
16:23 | vagrantc has quit IRC | |
16:29 | twinprism has quit IRC | |
16:31 | twinprism has joined #ltsp | |
16:50 | petre has joined #ltsp | |
16:51 | spectra has quit IRC | |
16:54 | monteslu has quit IRC | |
17:07 | monteslu has joined #ltsp | |
17:28 | nubae1 has joined #ltsp | |
17:28 | nubae has quit IRC | |
17:29 | nubae1 is now known as Nubae | |
17:29 | mccann has quit IRC | |
17:38 | captain_magnus has quit IRC | |
17:38 | captain_magnus has joined #ltsp | |
17:45 | psycodad has quit IRC | |
17:49 | <Nubae> | |
17:51 | alekibango has quit IRC | |
17:52 | psycodad has joined #ltsp | |
18:18 | alekibango has joined #ltsp | |
18:35 | gonzaloaf_laptop has quit IRC | |
18:48 | siki has quit IRC | |
18:52 | Lns has quit IRC | |
18:58 | cliebow has joined #ltsp | |
18:59 | johnny has joined #ltsp | |
18:59 | <cliebow> petre:ping
| |
19:00 | staffencasa has quit IRC | |
19:03 | petre has quit IRC | |
19:08 | cliebow has left #ltsp | |
19:10 | jonnor has quit IRC | |
19:11 | irule has quit IRC | |
19:13 | warren has quit IRC | |
19:18 | psycodad has quit IRC | |
19:18 | psycodad has joined #ltsp | |
19:27 | DawnLigh1 has quit IRC | |
19:27 | gbolte has quit IRC | |
19:35 | cliebow has joined #ltsp | |
20:31 | cliebow has quit IRC | |
21:20 | warren has joined #ltsp | |
21:33 | ogra-maemo has quit IRC | |
21:34 | ogra has quit IRC | |
21:41 | cpunches has quit IRC | |
21:46 | topslakr has quit IRC | |
21:46 | ogra-maemo has joined #ltsp | |
21:47 | ogra has joined #ltsp | |
21:49 | topslakr has joined #ltsp | |
22:23 | rcy has quit IRC | |
22:42 | ace_suares has quit IRC | |
22:48 | nubae1 has joined #ltsp | |
22:49 | Nubae has quit IRC | |
23:32 | jbrett has quit IRC | |
23:32 | sutula has quit IRC | |
23:40 | ace_suares has joined #ltsp | |
23:57 | nubae has joined #ltsp | |
23:59 | nubae1 has quit IRC | |