00:00 | roshan08 has joined #ltsp | |
00:03 | <roshan08> i have just now installed KIWI
| |
00:03 | how to add clients
| |
00:04 | i am able to boot the client but i have not set any username or password how to do that
| |
00:06 | <Ryan52> roshan08: it's just your servers username and password.
| |
00:06 | roshan08: once people log in to the client, they are really on the server.
| |
00:07 | so just add users on your server as you normally would.
| |
00:07 | <roshan08> Ryan52: ok it means simply adding normal users to the server computer
| |
00:08 | <Ryan52> yes
| |
00:08 | alkisg has joined #ltsp | |
00:08 | <roshan08> Ryan52: ok i thought they were some other users
| |
00:08 | Ryan52: thanks
| |
00:09 | gonzaloaf_laptop has joined #ltsp | |
00:09 | <Ryan52> no problem :)
| |
00:15 | gonzaloaf_laptop has quit IRC | |
00:35 | alkisg1 has joined #ltsp | |
00:35 | alkisg has quit IRC | |
00:48 | kwak has joined #ltsp | |
00:58 | roshan08 has quit IRC | |
01:19 | edoardo has quit IRC | |
01:32 | gate_keeper_ has joined #ltsp | |
01:45 | gate_keeper_ has quit IRC | |
01:46 | kwak has quit IRC | |
01:47 | kwak has joined #ltsp | |
01:53 | Q-FUNK has joined #ltsp | |
01:55 | alkisg has joined #ltsp | |
01:58 | <alkisg> Intrepid changed the default panel icons (e.g. added fusa). To reset all user panels to the defalts, I delete .gconf/apps/panel, i.e. (IFS=$'\n'; for f in $(ls /home); do rm -rf "/home/$f/.gconf/apps/panel/"; done). Is this a proper way to do it?
| |
02:03 | <kwak> hi, im installing xubuntu/ltsp and I get installation step on build ltsp chroot.
| |
02:03 | any suggestions on what to do?
| |
02:06 | alkisg has quit IRC | |
02:06 | <Ryan52> kwak: can you rephrase that?
| |
02:08 | <kwak> i mean installation step problem
| |
02:08 | the step is build ltsp chroot
| |
02:08 | <Ryan52> okay, what is the problem?
| |
02:09 | <kwak> it says building ltsp chroot, it reaches 50% then i get the the screen with red background (error)
| |
02:11 | tjikkun_work has joined #ltsp | |
02:13 | <kwak> can i try doing that step later on and just proceed with grub installation? i don't why ltsp chroot installation fails
| |
02:13 | mikkel has joined #ltsp | |
02:14 | <Ryan52> you might be able to do "ctrl-alt-f4" and see more errors.
| |
02:14 | gate_keeper_ has joined #ltsp | |
02:16 | <kwak> ok,
| |
02:16 | errors
| |
02:16 | <Ryan52> and those errors would be what?
| |
02:17 | <kwak> /opt/lts/i386 already exists
| |
02:17 | WARNING:: Configuring 'ltsp-client-builder' failed with error code 1
| |
02:17 | main-menu WARNING: menu item 'lts-client-builder; failed
| |
02:18 | <Ryan52> that's probably because you've choosen the retry already. can you scroll up and see other errors?
| |
02:18 | <kwak> DEBUG: Resolver (libnewt0.52): package doesn't exists (ignore) and also efi-modules
| |
02:20 | alkisg1 has quit IRC | |
02:20 | <Ryan52> hrm. I'm not sure then.
| |
02:21 | if it lets you then you can just skip setting up the chroot and build it later.
| |
02:22 | <kwak> k
| |
02:22 | trying
| |
02:23 | grub installation successful. restarting now
| |
02:24 | <Ryan52> then once you boot up you'll want to delete /opt/ltsp/i386 and run ltsp-build-client
| |
02:28 | Q-FUNK has quit IRC | |
02:45 | xmagixx has quit IRC | |
02:48 | alkisg has joined #ltsp | |
03:01 | kwak_ has joined #ltsp | |
03:03 | kwak has quit IRC | |
03:04 | alkisg1 has joined #ltsp | |
03:05 | mr_claus has quit IRC | |
03:09 | <Appiah> Does anyone have a good solution for setting the same startpage(Firefox) for all users? and lock it? Dont think you can do it with gconf
| |
03:09 | kwak__ has joined #ltsp | |
03:11 | <Appiah> oh nevermind , firefox.txt ;)
| |
03:11 | alkisg1 has quit IRC | |
03:11 | <Appiah> .cfg*
| |
03:13 | Egyptian[Home]1 has quit IRC | |
03:15 | alkisg has left #ltsp | |
03:17 | gate_keeper_ has quit IRC | |
03:19 | <kwak__> what does this mean.
| |
03:20 | ALERT! /dev/disk/by-uuid.,,, does not exist. Dropping to shell!
| |
03:20 | my installation still not booting
| |
03:21 | <Ryan52> is that from the client or from the server?
| |
03:22 | either way, it's probably a question for #ubuntu...you've got something messed up.
| |
03:22 | <kwak__> from the server. i was in initramfs prompt then issued "return" command. then the login screen came up
| |
03:23 | ltsp chroot is still not functiong. tying to build
| |
03:25 | generalsnus has joined #ltsp | |
03:25 | <generalsnus> using edubuntu 8.10 ltsp server, when thin clients boots, they halt at this message: ltsp disconnecting: que, disconnect, sock, done
| |
03:27 | kwak_ has quit IRC | |
03:38 | Egyptian[Home] has joined #ltsp | |
03:39 | Egyptian[Home] has quit IRC | |
03:40 | <generalsnus> using edubuntu 8.10 ltsp server, when thin clients boots, the booting stops at this message: "ltsp disconnecting: que, disconnect, sock, done" whats wrong?
| |
03:42 | Davor has joined #ltsp | |
03:46 | Q-FUNK has joined #ltsp | |
03:49 | Nubae has joined #ltsp | |
03:53 | <generalsnus> using edubuntu 8.10 ltsp server, when thin clients boots, the booting stops at this message: "ltsp disconnecting: que, disconnect, sock, done" whats wrong?
| |
04:05 | ltspbot has joined #ltsp | |
04:06 | kwak_ has quit IRC | |
04:06 | ltspbot` has quit IRC | |
04:06 | kwak_ has joined #ltsp | |
04:06 | <Davor> Does anybody know how to run LTSP (x session) without ssh tunneling to increase performance/speed(Ubuntu 7.04)?
| |
04:06 | sbalneav has quit IRC | |
04:08 | <kwak_> someone in this list suggested to add LDM_DIRECTX=T in lts.conf to solve that.
| |
04:08 | <Appiah> True
| |
04:09 | <Davor> But that wot
| |
04:09 | <Appiah> it works
| |
04:09 | The login will still be tunneld (secured)
| |
04:09 | <Davor> But that setting works only with the latest LDM (Ububtu 8.04)
| |
04:09 | <Appiah> naa
| |
04:09 | Where does it say that?
| |
04:09 | <kwak_> ic
| |
04:10 | <Nubae> works since 7.10 I believe
| |
04:10 | and the latest LDM is Intrepid, not Hardy
| |
04:10 | <Appiah> I'm sure it works Davor
| |
04:12 | <Davor> OK if I change my lts.cong with LDM_DIRECTX=True do I need to run ltsp-update-kernels or something else?
| |
04:13 | <Appiah> depends on where you put it
| |
04:13 | Egyptian[Home] has joined #ltsp | |
04:14 | <Appiah> if its LTSP 5.0 you dont need to update
| |
04:14 | just reboot the client
| |
04:14 | /var/lib/tfptboot/ltsp/i386/lts.conf (for LTSP 5.0)
| |
04:18 | kwak__ has joined #ltsp | |
04:19 | <Davor> there is no lts.conf in /var/lib/tfptboot/ltsp/i386/ . I changed /opt/ltsp/i386/etc/lts.conf but avi files are still streaming through ssh (this however works well on 8.04)
| |
04:20 | <Appiah> there is no lts.conf per default
| |
04:20 | <Nubae> make the lts.conf
| |
04:20 | <Appiah> an easy way to see if its going thru SSH or not is to open up a terminal and do echo $DISPLAY
| |
04:21 | it should say <ipnr> instead of localhost:11
| |
04:21 | [Default]
| |
04:21 | LDM_DIRECT=True
| |
04:21 | thats enough
| |
04:26 | Q-FUNK has quit IRC | |
04:27 | <Davor> Sorry, it does not work. You can check www.ltsp.org/twiki/bin/view/Ltsp/LtsConf#LDM_DIRECTX
| |
04:27 | <Appiah> check it for what?
| |
04:29 | updates has joined #ltsp | |
04:30 | <updates> question: Will ltsp server work without GUI installed?
| |
04:31 | <Nubae> updates: no
| |
04:32 | <Appiah> What would be the point then updates ?
| |
04:32 | I mean what are you planing to use?
| |
04:35 | Vince-0 has joined #ltsp | |
04:35 | kwak_ has quit IRC | |
04:43 | kwak_ has joined #ltsp | |
04:46 | kwak__ has quit IRC | |
04:56 | Vince-0 has quit IRC | |
05:00 | Vince-0 has joined #ltsp | |
05:00 | <Davor> Output is "localhost:10.0". If I look at /opt/ltsp/i386/usr/share/doc/ltsp-client/examples/lts-parameters.txt.gz, there is no LDM_DIRECTX option at all. sshd is still taking over 30% of CPU. Once again, LDM_DIRECTX=True is working fine on 8.04, but we need to test 7.04 because end-user needs it.
| |
05:02 | <Nubae> why are u using 7.04?
| |
05:02 | thats totally unsupported
| |
05:02 | not even LTS
| |
05:02 | <updates> Appiah, thanks. sorry for the late reply. i have to go home. talk to you later when i get home. thanks!
| |
05:02 | updates has quit IRC | |
05:03 | <Nubae> currently supported are 6.06 (for a couple months more), 8.04, and 8.10
| |
05:04 | <Davor> I understand, but end-user is educational institution and we expect over 4000 thin-client deployments (20-30 per class)
| |
05:05 | End-user does not want to upgrade Edubumtu
| |
05:05 | <Nubae> then u should definetly NOT be using 7.04
| |
05:05 | they do realise, that means no security update
| |
05:06 | u should be using 8.04 which is LTS and supported until 2013
| |
05:08 | Guaraldo has joined #ltsp | |
05:08 | <Davor> Unfortunately I can not change their decision
| |
05:09 | <Nubae> well I'm quite sure 7.04 had directx=True as well, but like I said, it being unsuppored, you'll find difficulty getting help
| |
05:11 | Q-FUNK has joined #ltsp | |
05:11 | kwak_ has quit IRC | |
05:11 | kwak_ has joined #ltsp | |
05:15 | <Davor> I'll check once again, I'll reinstall everything from the scratch if I need to, your information is usefull anyway. I'll also talk to decision making guys and tell them what you suggested since I noticed several bugs as well.
| |
05:17 | <Nubae> yeah running 7.04 is asking for trouble
| |
05:17 | a hackers paradise
| |
05:17 | might as well be running windows :-)
| |
05:17 | kwak__ has joined #ltsp | |
05:22 | <ogra> Davor, 7.04 is EOL since some weeks
| |
05:22 | no security support or upgrades
| |
05:23 | <Nubae> ogra: did directx=true work for 7.04?
| |
05:23 | <ogra> no
| |
05:24 | <Nubae> when did it come in 7.10?
| |
05:24 | <ogra> and there is no easy way to make it work apart from rewriting half the code
| |
05:24 | iirc 7.10, yes
| |
05:24 | Guaraldo is now known as Guest33356 | |
05:24 | <Nubae> well, as both of us just told him, he shouldnt be running that :-)
| |
05:24 | <ogra> but 7.10 will only get security support for another 6 months
| |
05:25 | i'd go with 8.04 or 8.10 for production if building up something fresh
| |
05:25 | <Nubae> yeah, don't quite get the logic behind, they dont want to upgrade...
| |
05:26 | <ogra> well, if Davor wants to take responsibility for something insecure, up to him
| |
05:26 | Vince-0 has quit IRC | |
05:26 | <Nubae> yeah, people can run warty if they like too :p
| |
05:27 | Vince-0 has joined #ltsp | |
05:27 | <ogra> i wouldnt base a business on it :) but yes, people can
| |
05:28 | <Nubae> ogra: u gonna be at the edubuntu meeting Wednesday?
| |
05:28 | <ogra> not sure
| |
05:28 | <Nubae> will RichEd be there u think?
| |
05:32 | <ogra> no idea
| |
05:32 | i havent seen him for a while
| |
05:34 | Guest33356 has quit IRC | |
05:35 | kwak_ has quit IRC | |
05:37 | Q-FUNK has quit IRC | |
05:39 | kwak_ has joined #ltsp | |
05:40 | Q-FUNK has joined #ltsp | |
05:41 | kwak__ has quit IRC | |
05:41 | Q-FUNK has quit IRC | |
05:43 | Vince-0_ has joined #ltsp | |
05:46 | Q-FUNK has joined #ltsp | |
05:55 | Q-FUNK has quit IRC | |
05:55 | Q-FUNK has joined #ltsp | |
05:56 | Vince-0 has quit IRC | |
06:06 | Vince-0_ has quit IRC | |
06:08 | Egyptian[Home] has quit IRC | |
06:09 | Egyptian[Home] has joined #ltsp | |
06:13 | Q-FUNK has quit IRC | |
06:13 | Q-FUNK1 has joined #ltsp | |
06:13 | Q-FUNK1 is now known as Q-FUNK | |
06:15 | alkisg has joined #ltsp | |
06:17 | Davor has quit IRC | |
06:19 | pmatulis has joined #ltsp | |
06:27 | Q-FUNK has quit IRC | |
06:28 | Q-FUNK has joined #ltsp | |
06:30 | otavio has joined #ltsp | |
06:47 | Blinny has joined #ltsp | |
06:49 | Q-FUNK has joined #ltsp | |
06:50 | <Blinny> Hrm. I have one user that cannot login. His username changed while being logged into a thin terminal. All others (who weren't logged in when the script changed usernames) work fine. Where should I start looking? Permissions in his home look OK with the new username. /tmp files look OK too (though I have rebooted the server)
| |
06:50 | <alkisg> Blinny, any messages if you try ssh -l username from the TC console?
| |
06:52 | Has anyone used an (e.g. cisco) router instead of the dhcp3-server package? I got it working just fine, but it doesn't send dynamic dns updates to my dns server. Are router supposed to do that, or their dhcp server ...sucks?
| |
06:52 | <Blinny> alkisg: Thanks - Let me check
| |
06:55 | alkisg: No errors, logged in just fine
| |
06:56 | <alkisg> Blinny, did you chown -R username:group /home/userdir? Maybe some file in some subdirectory doens't have the right permissions...
| |
06:56 | <Blinny> alkisg: Yes.
| |
06:56 | alkisg: Though I should mention that the uid didn't change, just the name
| |
06:57 | So theoretically, none of the file permissions changed either. But yes, I did a chown -R user:group /home/user
| |
06:57 | <alkisg> Blinny, I can't think of anything else, wait for someone more experienced! :)
| |
06:57 | (log files etc)
| |
06:57 | <Blinny> Dig. Thanks anyway.
| |
07:00 | This whole .gtk-bookmarks thing is dumb way of storing the Places links
| |
07:00 | I need to suggest a better way.
| |
07:01 | jammcq has quit IRC | |
07:04 | <alkisg> Blinny, why don't you try usermod -l newname oldname two more times? (one to change it to a fake one, one to restore it to the real one)
| |
07:04 | (But I did several times change the username of students while they where logged in, and I never got this problem...) :(
| |
07:07 | <Blinny> Good idea, thank you.
| |
07:11 | Q-FUNK has quit IRC | |
07:16 | Q-FUNK has joined #ltsp | |
07:21 | rcy has quit IRC | |
07:24 | rcy has joined #ltsp | |
07:36 | generalsnus has quit IRC | |
07:37 | rcy has quit IRC | |
07:37 | six2one has joined #ltsp | |
07:38 | rcy has joined #ltsp | |
07:41 | alkisg has quit IRC | |
07:41 | alkisg has joined #ltsp | |
07:42 | Q-FUNK has quit IRC | |
08:02 | chrisinajar|work is now known as chrisinajar | |
08:04 | kwak__ has joined #ltsp | |
08:14 | alkisg has left #ltsp | |
08:15 | alekibango has quit IRC | |
08:17 | kwak_ has quit IRC | |
08:19 | Gadi has joined #ltsp | |
08:29 | ccherret1 is now known as ccherrett | |
08:35 | sbalneav has joined #ltsp | |
08:35 | <sbalneav> Morning all
| |
08:35 | <Gadi> !s
| |
08:35 | <ltspbot> Gadi: "s" is Scotty!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
| |
08:35 | <sbalneav> Hey Gadi!
| |
08:36 | <Gadi> has ragnar been in the channel recently?
| |
08:36 | <sbalneav> I just got here, dunno.
| |
08:36 | <Gadi> !seen laprag
| |
08:36 | <ltspbot> Gadi: laprag was last seen in #ltsp 2 weeks, 1 day, 17 hours, 58 minutes, and 59 seconds ago: <laprag> enough hacking tonight. good night all.
| |
08:36 | <sbalneav> Wasn't in on Saturday when I was here
| |
08:36 | <Gadi> !seen mistik1
| |
08:36 | <ltspbot> Gadi: mistik1 was last seen in #ltsp 2 weeks, 1 day, 18 hours, 44 minutes, and 42 seconds ago: <mistik1> take care Jim
| |
08:36 | <Gadi> *sigh*
| |
08:36 | everybody's on a 2 week vacation ;)
| |
08:37 | <sbalneav> Gadi: So, I came up with a fix for the X server race condition that Warren identified on friday.
| |
08:37 | <Gadi> catch me up a bit
| |
08:37 | I dunno how much I was paying attention
| |
08:37 | :)
| |
08:38 | <stgraber> moin sbalneav
| |
08:38 | <sbalneav> Well, we have a problem sometimes, on slow machines, where we spawn off the X server, then start issuing commands to it.
| |
08:38 | <stgraber> hi Gadi
| |
08:38 | <Gadi> hola, stgraber
| |
08:38 | <sbalneav> Meanwhile, it hasn't even got it's input channels plumbed yet
| |
08:38 | hey stgraber
| |
08:39 | <Gadi> ah - so what symptoms do you see?
| |
08:39 | unresponsive keyboard?
| |
08:39 | <sbalneav> So, th the proper solution is to set SIGUSR1 to SIG_IGN before we spawn off the X server.
| |
08:40 | This tells the X server that, once it's got it's act together, and is ready to serve clients, it'll send a SIGUSR1 to the parent process
| |
08:40 | which is ldm
| |
08:40 | <Gadi> ah
| |
08:40 | * Gadi likes simple fixes | |
08:41 | <sbalneav> So, we just wait for that signal, via the sigsuspend() call
| |
08:41 | When we get it, then, we've got a functioning X server, no race.
| |
08:41 | <Gadi> brilliant work, professor
| |
08:42 | :)
| |
08:42 | now, if you could only get the coconut radio going - we could get off this island
| |
08:43 | <ltsppbot> "sbalneav" pasted "Fix X race condition." (165 lines) at http://pastebot.ltsp.org/88
| |
08:43 | gate_keeper_ has joined #ltsp | |
08:44 | <gate_keeper_> how to disable pulse audio in X11 for ltsp?
| |
08:44 | <sbalneav> gate_keeper_: SOUND=false?
| |
08:45 | <gate_keeper_> hmm
| |
08:45 | i'm having problems with X11
| |
08:45 | <warren> sbalneav: how well tested?
| |
08:45 | <gate_keeper_> huge lo on X11
| |
08:45 | *traffic
| |
08:46 | <sbalneav> Tested on two terminals at home :)
| |
08:46 | Worked there.
| |
08:46 | <gate_keeper_> ubuntu server by default does not install pulse audio
| |
08:46 | as far i can see, ltsp is configured to call pulse audio :-/
| |
08:46 | /etc/X11
| |
08:46 | <sbalneav> gate_keeper_: You've been here before with this. Have you Identified what's actually GENERATING the traffic, via wireshark or something?
| |
08:47 | <Gadi> sbalneav: and the devnull redirects fix Xsession -right?
| |
08:47 | no more killing
| |
08:47 | :D
| |
08:47 | <gate_keeper_> yes, we've talked for this
| |
08:47 | <sbalneav> Gadi: Yup, that eliminates the need for the kill -1.
| |
08:48 | <Gadi> \o/
| |
08:48 | <sbalneav> Plus, TONS of processes that used to hang around before now exit.
| |
08:48 | That part I have tested.
| |
08:48 | <Gadi> and now the K* scripts work?
| |
08:48 | <Blinny> Yea!!
| |
08:48 | <sbalneav> I've backported the /dev/null bits to hardy, works a treat.
| |
08:48 | * Blinny applauds loudly | |
08:49 | <gate_keeper_> http://pastebin.com/d581de38d
| |
08:49 | <sbalneav> gate_keeper_: ...aaaaaaand? What's generating the traffic? Pulse audio for sure? or you're just guessing
| |
08:49 | <gate_keeper_> guessing
| |
08:49 | <warren> how does the /dev/null fix work?
| |
08:49 | ssh held open because of stdin or stdout?
| |
08:49 | <sbalneav> localhost:ldap localhost:41233 ESTABLISHED openldap 128789 19000/slapd
| |
08:50 | You running openldap?
| |
08:50 | <gate_keeper_> yes, with or without ldap
| |
08:50 | it's the same
| |
08:50 | <sbalneav> warren: correct
| |
08:50 | <gate_keeper_> http://pastebin.com/d78072ab4
| |
08:50 | <sbalneav> gate_keeper_: have you run wireshark on lo to see what's generating the traffic?
| |
08:50 | <gate_keeper_> just using iptraf
| |
08:51 | and standard in-out net tools
| |
08:51 | <sbalneav> A huge chunk of those connections are ldap.
| |
08:51 | <gate_keeper_> i've made auth for the users via ldap
| |
08:51 | <sbalneav> Using openldap for your passwd and group lookups?
| |
08:51 | <gate_keeper_> localhost:50017 localhost:6027 ESTABLISHED
| |
08:52 | most of the traffic on lo is happing on X11 ports
| |
08:52 | 6011-6047
| |
08:52 | tcp 0 0 localhost:6011 localhost:36407 ESTABLISHED
| |
08:52 | <warren> sbalneav: are you committing both /dev/null and the X race fix at the same time?
| |
08:52 | sbalneav: probably best to have people test them independently
| |
08:52 | <sbalneav> I haven't committed anything.
| |
08:52 | I don't touch -trunk
| |
08:52 | I leave that to others.
| |
08:53 | I'll update my branch today.
| |
08:53 | <Gadi> sbalneav: of all people, I would want you to touch ldm-trunk most :)
| |
08:53 | but not in a creepy kind of way
| |
08:53 | <sbalneav> gate_keeper_: can you paste your lts.conf file, and your /etc/nsswitch.conf file?
| |
08:54 | * sbalneav offers ldm-trunk some candy from the back of his van. | |
08:56 | <gate_keeper_> sbalneav, fresh install
| |
08:57 | without openldap
| |
08:57 | http://pastebin.com/d7d32e7ee
| |
08:57 | lts.conf is empty afaik in /opt/ltsp/i386/etc/lts.conf
| |
08:57 | <sbalneav> Oh, well if lts.conf is empty, no LDM_DIRECTX=true
| |
08:58 | <gate_keeper_> 8.10 ubuntu-server
| |
08:58 | ltsp 5.1
| |
08:58 | <sbalneav> then all the traffic on lo's the encrypted X going though ssh
| |
08:58 | X talks to the ssh session, which encrypts the X traffic.
| |
08:58 | that'll happen over lo
| |
09:00 | <gate_keeper_> hm..
| |
09:00 | :-/
| |
09:00 | <sbalneav> gate_keeper_: Should be /var/lib/tftboot/ltsp/i386 anyway for lts.conf
| |
09:00 | gate_keeper_: Do you want your X to be encrypted?
| |
09:01 | <gate_keeper_> whatever, what's best ..
| |
09:01 | <sbalneav> Dude.
| |
09:01 | That's up to you.
| |
09:01 | You want encrypted X, then leave it as it us.
| |
09:02 | as it is, sorry
| |
09:02 | <gate_keeper_> encrypted :)
| |
09:02 | <sbalneav> if you want your X faster, but unencrypted, then set LDM_DIRECTX=true
| |
09:02 | Then you're going to have a lot of traffic on lo.
| |
09:02 | there's your answer.
| |
09:03 | <gate_keeper_> btw in /var/lib.... there's no lts.conf
| |
09:03 | <sbalneav> right
| |
09:03 | there isn't, be default
| |
09:03 | <ogra> create one
| |
09:03 | <sbalneav> s/be/by/
| |
09:03 | warren has quit IRC | |
09:03 | <sbalneav> he wants his X encrypted. He can just leave it as-is.
| |
09:03 | <gate_keeper_> k,thx
| |
09:03 | <sbalneav> morning Ogra!
| |
09:04 | <ogra> as described in the documentation
| |
09:04 | !docs
| |
09:04 | <ltspbot> ogra: "docs" is For the most current documentation, see http://wiki.ltsp.org/twiki/bin/view/Ltsp/LtspDocumentationUpstream
| |
09:04 | <sbalneav> ogra: http://pastebot.ltsp.org/88
| |
09:04 | Fixed the race.
| |
09:04 | <ogra> where's the bot ?
| |
09:04 | !docs
| |
09:04 | <ltspbot> ogra: "docs" is For the most current documentation, see http://wiki.ltsp.org/twiki/bin/view/Ltsp/LtspDocumentationUpstream
| |
09:05 | <sbalneav> showing up for me
| |
09:05 | I see the docs factoid
| |
09:05 | <gate_keeper_> ok, added
| |
09:08 | aojd has joined #ltsp | |
09:09 | Faithful has quit IRC | |
09:10 | rcy has quit IRC | |
09:10 | mikkel has quit IRC | |
09:10 | tjikkun_work has quit IRC | |
09:10 | |Paradox| has quit IRC | |
09:10 | CaScAdE^1arAway has quit IRC | |
09:10 | ccherrett has quit IRC | |
09:10 | captain_magnus has quit IRC | |
09:10 | ogra has quit IRC | |
09:10 | daduke_ has quit IRC | |
09:10 | Hyperbyte has quit IRC | |
09:10 | ninkendo has quit IRC | |
09:10 | jbrett has quit IRC | |
09:10 | dberkholz has quit IRC | |
09:10 | artista_frustrad has quit IRC | |
09:10 | jcastro has quit IRC | |
09:10 | aojd is now known as |Paradox| | |
09:11 | rcy has joined #ltsp | |
09:11 | mikkel has joined #ltsp | |
09:11 | tjikkun_work has joined #ltsp | |
09:11 | CaScAdE^1arAway has joined #ltsp | |
09:11 | ccherrett has joined #ltsp | |
09:11 | captain_magnus has joined #ltsp | |
09:11 | ogra has joined #ltsp | |
09:11 | Hyperbyte has joined #ltsp | |
09:11 | artista_frustrad has joined #ltsp | |
09:11 | daduke_ has joined #ltsp | |
09:11 | dberkholz has joined #ltsp | |
09:11 | jbrett has joined #ltsp | |
09:11 | ninkendo has joined #ltsp | |
09:11 | jcastro has joined #ltsp | |
09:14 | ccherrett has quit IRC | |
09:15 | ccherrett has joined #ltsp | |
09:22 | <sbalneav> So, I've been trying to find a worthy recipient of a bunch of our old Via EPIA 5000 mobo's
| |
09:22 | <chrisinajar> me
| |
09:22 | <sbalneav> Turns out the local university's robotics lab uses them.
| |
09:22 | So, I'm donating 50 of them, in exchange for a tour :)
| |
09:23 | So, tomorrow afternoon, I'm gonna go see robots :)
| |
09:23 | chrisinajar: I can drive 'em to the robotcs lab, where are you?
| |
09:24 | <ogra> hey sbalneav (sorry, was in a conf call)
| |
09:24 | <sbalneav> NP
| |
09:24 | <ogra> sbalneav, really they use slow VIA ? not arm ?
| |
09:25 | <sbalneav> Dunno, guess they use what they can get their hands on :)
| |
09:25 | <ogra> heh
| |
09:30 | mikkel has quit IRC | |
09:34 | ccherrett has quit IRC | |
09:34 | daduke_ has quit IRC | |
09:34 | Hyperbyte has quit IRC | |
09:34 | ninkendo has quit IRC | |
09:34 | jbrett has quit IRC | |
09:34 | tjikkun_work has quit IRC | |
09:34 | ogra has quit IRC | |
09:34 | CaScAdE^1arAway has quit IRC | |
09:34 | dberkholz has quit IRC | |
09:34 | artista_frustrad has quit IRC | |
09:34 | captain_magnus has quit IRC | |
09:34 | jcastro has quit IRC | |
09:34 | rcy has quit IRC | |
09:35 | ccherrett has joined #ltsp | |
09:35 | rcy has joined #ltsp | |
09:35 | tjikkun_work has joined #ltsp | |
09:35 | CaScAdE^1arAway has joined #ltsp | |
09:35 | captain_magnus has joined #ltsp | |
09:35 | ogra has joined #ltsp | |
09:35 | Hyperbyte has joined #ltsp | |
09:35 | artista_frustrad has joined #ltsp | |
09:35 | daduke_ has joined #ltsp | |
09:35 | dberkholz has joined #ltsp | |
09:35 | jbrett has joined #ltsp | |
09:35 | ninkendo has joined #ltsp | |
09:35 | jcastro has joined #ltsp | |
09:38 | <gate_keeper_> guys, it's working
| |
09:38 | thx
| |
09:38 | :)
| |
09:39 | <sbalneav> You went with LDM_DIRECTX?
| |
09:39 | <gate_keeper_> yup
| |
09:40 | <sbalneav> Problem solved.
| |
09:42 | <gate_keeper_> yup
| |
09:42 | thanks a lot!
| |
09:42 | <Gadi> another LDM_DIRECTX success story
| |
09:42 | <sbalneav> NP
| |
09:42 | <gate_keeper_> :)
| |
09:42 | <sbalneav> Oh!
| |
09:42 | ohohoho!
| |
09:42 | ogra: See the updates today?
| |
09:42 | They fixed the gvfs-fuse hanging thingy
| |
09:43 | So, if we can push out the /dev/null thingy, that should eliminate most of the hung jobs issues.
| |
09:44 | stgraber: Lets you and I sit down this hackyfest, and coordinate with ogra, and get SRU's for a couple of these bugs I've sqwashed, and get 'em in the queue.
| |
09:45 | I think that'll make a lot of Hardy users happy
| |
09:48 | alkisg has joined #ltsp | |
09:55 | <gate_keeper_> laters..
| |
09:55 | gate_keeper_ has quit IRC | |
09:59 | <Gadi> sbalneav: what cipher do we default to for ssh?
| |
09:59 | in ldm
| |
09:59 | tell me it aint 3des
| |
10:02 | cyberorg has joined #ltsp | |
10:04 | <stgraber> sbalneav: sounds good
| |
10:06 | mccann has joined #ltsp | |
10:10 | warren has joined #ltsp | |
10:13 | <warren> what is the plan for hte hackfest? arrive 6th?
| |
10:13 | depart 9th?
| |
10:16 | warren has quit IRC | |
10:17 | warren has joined #ltsp | |
10:19 | <sbalneav> Gadi: Not sure.
| |
10:20 | I think it just defaults to whatever ssh defaults to
| |
10:23 | warren: I think so
| |
10:24 | <warren> does anyone need a ride to/from Boston?
| |
10:24 | * Gadi asks the same thing - s/Boston/NY/ | |
10:25 | CAN-o-SPAM has joined #ltsp | |
10:25 | <sbalneav> Gadi: Are you brining Mistik?
| |
10:26 | <Gadi> if I can find him to twist his arm
| |
10:26 | but-
| |
10:26 | !seen mistik1
| |
10:26 | <ltspbot> Gadi: mistik1 was last seen in #ltsp 2 weeks, 1 day, 20 hours, 34 minutes, and 18 seconds ago: <mistik1> take care Jim
| |
10:26 | <sbalneav> Gadi: looks like we don't have any ssh options other than OVERRIDE_PORT
| |
10:26 | * Gadi has no reggae CDs of his own :( | |
10:27 | <Gadi> sbalneav: yeah, saw that
| |
10:27 | <sbalneav> Suggestions for a generic SSH option varaible that we can use to pass any custom ssh commands to the session?
| |
10:27 | <Gadi> also saw that default = 3des for protocol1 and aes128-cbc for 2
| |
10:27 | LDM_SSH_OPTIONS
| |
10:27 | <sbalneav> LDM_SSH_OPTIONS or just SSH_OPTIONS
| |
10:27 | <Gadi> ;)
| |
10:27 | no - needs LDM
| |
10:28 | <sbalneav> okie
| |
10:28 | easy to add.
| |
10:28 | <Gadi> I like knowing that all LDM_ params affect only the ldm session
| |
10:28 | <sbalneav> Sure, makes sense.
| |
10:28 | <Gadi> at some point, I would like to explore passing LDM_ options via ldminfod
| |
10:28 | or some such
| |
10:29 | <warren> i'm not sure that's a good idea
| |
10:29 | easily interceptable unencrypted directive to run with arbitrary parameters?
| |
10:29 | <Gadi> ie, having the LDM config being controlled by the app server and not by the boot server
| |
10:29 | warren: ah, good point
| |
10:29 | maybe if we integrate stgraber's cluster stuff
| |
10:29 | we can get some config info via https
| |
10:30 | <chrisinajar> warren: what's hackfest?
| |
10:30 | <Gadi> prior to session launch
| |
10:30 | <warren> chrisinajar: sounds sinister huh?
| |
10:30 | <chrisinajar> warren: hackfest == ltsp bts?
| |
10:30 | staffencasa has joined #ltsp | |
10:30 | <Gadi> chrisinajar: yes
| |
10:30 | <chrisinajar> ok
| |
10:30 | I'll be there :)
| |
10:30 | <Gadi> well, in the fall ;)
| |
10:31 | I should say: this hackfest == ltsp bts
| |
10:31 | but not all hackfests are ltsp bts
| |
10:31 | <chrisinajar> ok, that makes sense..
| |
10:31 | <Gadi> anyone ready for the LSAT? :P
| |
10:32 | <chrisinajar> so what exactly is hackfest? I haven't heard of it before, but I'm going to LTSP bts
| |
10:32 | <ogra> LSAT ? new TV transmission system ?
| |
10:32 | <Gadi> ogra: standardized test for Law school in the US - lots of logic problems
| |
10:32 | :)
| |
10:32 | <ogra> heh
| |
10:33 | <pscheie> warren, do we have any sort of autologin on K12?
| |
10:33 | <chrisinajar> google: Did you mean hawkfest?
| |
10:33 | <warren> pscheie: LDM itself has autologin things, but I have no clue how it works
| |
10:33 | <chrisinajar> no google, no i did not
| |
10:34 | <ogra> chrisinajar, at the hackfests we usually focus on developing new features ... the BTS is additionally a traditional thing to get together loosely and socialize
| |
10:34 | <pscheie> even if it's possible at this time, I don't think it's documented anywhere
| |
10:34 | <warren> sbalneav: IIRC there's no whiteboard at that motel right?
| |
10:34 | <chrisinajar> ogra: oh, that's cool, well looking at the registration there aren't very many people going to bts, are the hackfest people not included in this registration?
| |
10:34 | <pscheie> bring a chalkboard
| |
10:35 | <sbalneav> warren: Last time McQuillan brought a blank paper flipchart
| |
10:35 | Want me to ask him to bring it again?
| |
10:35 | <Gadi> if I remember, I will throw a whiteboard in my trunk
| |
10:36 | but, I prolly won't remember
| |
10:36 | and it may not fit
| |
10:36 | tjikkun_work has quit IRC | |
10:36 | <Gadi> sbalneav: is Jim gonna still bring a shuttle box?
| |
10:38 | <warren> i'm bringing a few laptops to act as thin clients
| |
10:38 | <sbalneav> Gadi: not sure.
| |
10:38 | I'll ask him today.
| |
10:43 | <chrisinajar> so how many people do you think will be at bts/hackfest?
| |
10:50 | <stgraber> we'll bring a server with ltsp-cluster on it and at least two netbooks (Atom based) as thin clients (and probably some networking too, depending on what we can put in the car)
| |
10:52 | (the server is installed as an openvz host, so if someone bring a network card we'll be able to run more than one LTSP network on it)
| |
10:54 | alkisg has quit IRC | |
10:55 | <CAN-o-SPAM> maybe a post on the wiki ... hardware your bringing / hardware you'd like brought?
| |
10:55 | spectra has joined #ltsp | |
10:55 | <pscheie> is anyone bringing switches to tie this all together?
| |
10:57 | <CAN-o-SPAM> DisklessWorkstations can bring some thin clients, if anyone is interested ...
| |
10:58 | <chrisinajar> how is the conference aranged? like, would i want to bring a laptop or would it be acceptable to bring a workstation?
| |
10:58 | jammcq has joined #ltsp | |
10:59 | <jammcq> Goood morning LTSP
| |
10:59 | <sbalneav> jammcq says yes to the switch, no to the whiteboard, and no to the shuttle.
| |
11:00 | <brendan0powers> i could bring a whiteboard or two
| |
11:01 | there not that big, but fine for colaborative doodling
| |
11:02 | <jammcq> I think a whiteboard would be great to have
| |
11:02 | <sbalneav> chrisinajar: I think you're welcome to bring whatever the heck you can carry :)
| |
11:02 | <jammcq> somewhere, I've got a flipchart easle type thinggy, but dunno where it is
| |
11:03 | <chrisinajar> sbalneav: well, i've never been to one of these (me and brendan are going together) so i don't really know physically how it's aranged
| |
11:03 | i believe we have 2 pretty small whiteboards... like, 3 feet by 4 feet...
| |
11:05 | <Gadi> 42" plasma?
| |
11:05 | anybody?
| |
11:05 | makes a great whiteboard
| |
11:05 | <chrisinajar> got one of those too :-P
| |
11:05 | <Gadi> cool - I'll bring the dry-erase markers ;)
| |
11:05 | <sbalneav> chrisinajar: LTSP hackfests don't tend to be highly structured affairs. We just all bring some hardware, usually talk for a bit about whatever happens to be the burning issue/problem/feature/s we'd like to work on, then just hack.
| |
11:06 | Y'know, the way a hackfest SHOULD be.
| |
11:06 | <Gadi> don't worry - I'll have an agenda
| |
11:06 | :)
| |
11:06 | and alcohol
| |
11:06 | <chrisinajar> sbalneav: that's cool. I work at a place that has an ltsp-ish solution and as of late we've been trying to get more into the community...
| |
11:07 | <sbalneav> I've got 3 things I'd like to work on:
| |
11:07 | <chrisinajar> bbl
| |
11:07 | chrisinajar is now known as chrisinajar|lunc | |
11:07 | <pscheie> I think someone should hack together one of Johnny Chung Lee's wiimote-whiteboard markers
| |
11:07 | <sbalneav> 1) pushing recent bugfixes, and identifying any other burning bugs that need my attention
| |
11:08 | 2) more doco cleanup
| |
11:08 | 3) finishing up on Gadi's suggestion of moving X startup outside of LDM to the screen script.
| |
11:08 | alkisg has joined #ltsp | |
11:08 | <Gadi> sbalneav: *cough* wiki *cough* http://wiki.ltsp.org/twiki/bin/view/Ltsp/ByTheSea2008
| |
11:08 | * sbalneav hands gadi some cough syrup | |
11:09 | <Gadi> man, thanks
| |
11:09 | that's gettin nasty
| |
11:10 | Lns has joined #ltsp | |
11:13 | <Nubae> so its totally US and Canada centric :-)
| |
11:13 | <ogra> yeah
| |
11:13 | evil them
| |
11:13 | <Nubae> heh
| |
11:13 | <jammcq> hey, we've got a norwegian coming
| |
11:13 | that counts for something
| |
11:13 | <ogra> pseudo nrwegian
| |
11:14 | or did he move ?
| |
11:14 | <jammcq> he's living in Oslo. how norwegian does he need to be?
| |
11:14 | <sbalneav> Gadi: That make you happy? :)
| |
11:14 | <ogra> i thought he lived in the US now
| |
11:14 | <jammcq> he does alot of business in the US, but his home base is still the other side of the pond
| |
11:14 | <Nubae> well I was just reading the coming from part
| |
11:15 | <pscheie> ogra, are you coming?
| |
11:15 | <ogra> pscheie, sadly not
| |
11:15 | <pscheie> bummer
| |
11:15 | <ogra> the first one i'm not attending since 3 years :(
| |
11:16 | <Nubae> who's the Norwegian?
| |
11:16 | <Gadi> sbalneav: it is beautiful - but here in the US, we "impact" we don't "inpact"
| |
11:16 | :)
| |
11:16 | <jammcq> Nubae: Ragnar Wisloff
| |
11:16 | <ogra> Nubae, ragnar
| |
11:17 | <Nubae> ah
| |
11:17 | <sbalneav> It's impact in Canada as well. That's just my bad sppeling
| |
11:18 | Gadi: Fixeed
| |
11:18 | * jammcq wonders if there'll be a hurricane again this year | |
11:19 | * sbalneav hopes so. | |
11:19 | <jammcq> be careful what you wish for
| |
11:19 | <sbalneav> cliebow promised me one.
| |
11:19 | jammcq: We BOTH loved the hurricane last year :)
| |
11:20 | What's the temp down there looking like? Anyone looked already to save me the google? Should I brink parka + snowboots?
| |
11:21 | <jammcq> sbalneav: yeah, but this summer we had what seemed like a tornado in our town. not sure i'd like to see the full wrath of mother nature
| |
11:21 | forcast for SWHarbor calls for rain and 58F from thursday through sunday
| |
11:21 | <ogra> it was beautiful
| |
11:22 | <jammcq> dunno about snow boots, but a parka might be good for those long walks on the beach
| |
11:23 | <sbalneav> 58? I'll just bring sweaters & raingear then. Layering should handle that if it's gonna be above zero.
| |
11:24 | <jammcq> k
| |
11:24 | <sbalneav> 0C that is.
| |
11:29 | <warren> how many people are going?
| |
11:29 | do we have a website to point at?
| |
11:29 | with details
| |
11:29 | <sbalneav> Gadi: So, I've been doing some investigating, and it turns out that the xinit command actually does the waitfor SIGUSR1. So, I'm thinking, if we're going to start X from outside LDM, instead of calling X directly, we should go through the xinit command, and start ldm as the client that gets started up after X.
| |
11:29 | !bts
| |
11:29 | <ltspbot> sbalneav: "bts" is LTSP By The Sea 2008: http://wiki.ltsp.org/twiki/bin/view/Ltsp/ByTheSea2008
| |
11:30 | <sbalneav> warren: That should have it.
| |
11:30 | <warren> sbalneav: I think we should use the waitfor SIGUSR1 ourselves, everything else like gdm or rhgb does it themselves
| |
11:31 | <sbalneav> warren: Only problem with that is, I'm not sure if we can do that from within the shell
| |
11:32 | Gadi wanted to move the starting of X outside of LDM to the screen script
| |
11:32 | <warren> sbalneav: we could build a small C binary
| |
11:32 | <sbalneav> certainly, within the shell, we can trap SIGUSR1, but I'm not shu...
| |
11:32 | yeah, I was just going to suggest that.
| |
11:32 | heh
| |
11:32 | <warren> sbalneav: ideally we want a binary to run and monitor X so we can cleanly handle cleanup
| |
11:33 | it can detect more than just SIGUSR1 of "i'm ready"
| |
11:33 | also early death
| |
11:33 | etc.
| |
11:33 | <Gadi> sbalneav: yeah, calling thru xinit was my plan
| |
11:33 | <sbalneav> Well, really, what we want from LDM is to actually make it X aware.
| |
11:33 | So, have it do an XOpenDisplay, and monitor X events.
| |
11:34 | That way, if X disappears, we'll see that, and can do the cleanup.
| |
11:34 | Let's hash it out in Bah Hahbah
| |
11:35 | I'll add some notes to the wikikiki
| |
11:36 | <Gadi> sbalneav: plus, if we call through xinit, we can add a special xinitrc.d/ directory that can be common with other screen scripts to support non-LDM-specific init scripts
| |
11:37 | which can be useful for more proper support of rdesktop and other screen scripts
| |
11:37 | Q-FUNK has joined #ltsp | |
11:37 | <warren> as we discussed in thie last hackfest
| |
11:37 | <Gadi> warren: correcto
| |
11:38 | <sbalneav> Sounds like xinit's looking to be the way to go.
| |
11:38 | I think we need like a 1/2 hour of discussion, then maybe Gadi and I can sit down together and code it up.
| |
11:39 | <Gadi> ooh - it's only Monday and I already have a date!
| |
11:39 | * Gadi doesn't mean to brag | |
11:40 | <Q-FUNK> :D
| |
11:40 | <warren> sbalneav: is xinit really common for all distros?
| |
11:40 | sbalneav: including location?
| |
11:41 | sbalneav: I still think we need a binary that runs X so we can monitor all signals
| |
11:41 | sbalneav: there is no better way to detect failures and cleanup
| |
11:43 | * Gadi wonders how a binary running X helps with that | |
11:43 | <Gadi> Can we not cleanup if X dies?
| |
11:43 | Q-FUNK has quit IRC | |
11:43 | <warren> I'll try to get input from our desktop team
| |
11:43 | Gadi: you mean like what we do now? =)
| |
11:44 | <sbalneav> Well, if X dies, and LDM has the X display open, we'll get an X error, we can trigger cleanup from that.
| |
11:45 | * Lns waves happy Monday to chan.. And ogra: I'm here whenever you can work with me on the nbd_swap bug(s) | |
11:45 | <sbalneav> Certainly, we don't handle some things as gracefully as we should, and I think we've got a great opportunity here to clean some of this stuff up, and make it more robustified.
| |
11:45 | <Gadi> right - the X client can be informed of X dying without having a parent binary to the Xserver
| |
11:46 | <sbalneav> I think the key is making LDM an actual X application that monitors the X server via XOpenDisplay, and receiving X errors.
| |
11:47 | Rather than what it is now, which is just an app that starts a bunch of stuff, and crosses it's fingers and hopes everything works :)
| |
11:48 | CAN-o-SPAM has quit IRC | |
11:48 | <Gadi> yummy - app fingers....
| |
11:50 | <sbalneav> Lunch time. Be back in an hour
| |
11:58 | K_O-Gnom has joined #ltsp | |
11:59 | <ogra> /join #ubuntu-classroom
| |
11:59 | bah
| |
12:05 | vagrantc has joined #ltsp | |
12:08 | <warren> pscheie: you need a pick up from Bangor?
| |
12:11 | <pscheie> warren, no, I'm catching a ride with fgiraldeau & stgraber
| |
12:12 | I looked at the map and saw they have to come through Bangor, whereas I wasn't sure about anyone else actually driving that far north
| |
12:13 | warren, how many people are coming with you?
| |
12:14 | <warren> pscheie: currently nobody
| |
12:14 | pscheie: are you staying the entire time?
| |
12:15 | <pscheie> yep, my plane leaves Bangor at around 3:30pm on Sunday
| |
12:16 | on the wiki page, David Johnston is looking for a ride from Boston
| |
12:16 | I don't know him, but I think jammcq does
| |
12:16 | warren, you going to be there the whole time?
| |
12:17 | chrisinajar|lunc is now known as chrisinajar | |
12:17 | <warren> pscheie: apparently
| |
12:17 | if they're coming from north then probably better
| |
12:17 | my route doesn't go past there
| |
12:18 | <stgraber> yeah and we'd have been around Bangor at the same time pscheie land so it's really no problem :)
| |
12:18 | * pscheie thanks stgraber for his hospitality | |
12:19 | <pscheie> and his great sense of timing
| |
12:23 | laprag has joined #ltsp | |
12:24 | <Gadi> ragnar!
| |
12:25 | chrisinajar has quit IRC | |
12:26 | <laprag> hey gadi
| |
12:26 | <Gadi> still need a lift from NY?
| |
12:26 | (on thursday)
| |
12:26 | <laprag> ah. thanks a ton, but it looks like i won't come to ny.
| |
12:27 | so no thanks!
| |
12:27 | <Gadi> hope it wasn't something I said
| |
12:27 | ;)
| |
12:27 | <laprag> appreciate the offer though
| |
12:27 | heh, nothing you said :)
| |
12:27 | <Gadi> does that mean you will miss out on the Papy Van Winkles this weekend?
| |
12:27 | <laprag> nono
| |
12:28 | just that i managed to wriggle out of the meeting in ny
| |
12:28 | <Gadi> ah, so where do you fly into?
| |
12:28 | <laprag> now i'm working on flying out thursday morning rather than night
| |
12:28 | boston
| |
12:29 | it's a six hour drive from boston, isn't it?
| |
12:29 | but i'm just loading the laptop with ltsp now
| |
12:29 | so i can do something useful, possibly
| |
12:29 | chrisinajar has joined #ltsp | |
12:31 | <Gadi> laprag: Im driving from NY, so if you need a lift from Boston, lemme know
| |
12:31 | when do you land?
| |
12:32 | <chrisinajar> So, hypothetically, if I were to go to ltsp BTW/hackfest for only 1 day, which would be the most interesting?
| |
12:32 | *bts
| |
12:33 | <laprag> Gadi: hopefully around 1pm
| |
12:33 | chrisinajar has quit IRC | |
12:33 | <laprag> Gadi: so you will drive through boston?
| |
12:33 | <Gadi> I can
| |
12:33 | its on the way
| |
12:34 | chrisinajar has joined #ltsp | |
12:34 | <Gadi> so, 1pm, then customs, so you'll be out by ~2pm
| |
12:34 | unless they sniff your suitcase....
| |
12:35 | * Gadi should give you my cell | |
12:35 | <Gadi> :)
| |
12:41 | Ahmuck has quit IRC | |
12:45 | <laprag> hmmm, what's the easiest way to get the dhcp config right? mine looks quite default. symlink to somewhere?
| |
12:45 | ah
| |
12:45 | <warren> jammcq: OK, confirmed I'm coming to the hackfest. Possibly with another RH engineer. Could you please be sure I have a room with 2 beds reserved?
| |
12:45 | * laprag finds it | |
12:46 | TheBS has joined #ltsp | |
12:47 | <TheBS> Hey all. I have LTSP up'n running without issues. Pretty much straight-forward. I used Warren's Fedora 9 chroot (client) on both a Fedora 9 server, as well as a RHEL 5 server (with the appropriate additives).
| |
12:48 | What I'm having difficulty figuring out now is some of the LTSP init/upstart aspects -- such as where LDM is called
| |
12:48 | <pscheie> warren, http://www.seawallmotel.com/
| |
12:48 | <laprag> hmm
| |
12:49 | <TheBS> I didn't see it in prefdm (and LDM seems to be more like a "startx" init), and didn't see any related environment aspects, but I think I'm just overlooking some things
| |
12:49 | <laprag> ubuntu is a bit too helpful. it kills off the network interface when i turn off the client
| |
12:49 | <warren> pscheie: err... I figured it would be better for jammcq to handle the group? I dunno.
| |
12:50 | who is David Johnston?
| |
12:51 | <laprag> warren: he's from Charlotte NC, was there last year
| |
12:51 | <warren> I vaguely recall, was he the really annoying guy?
| |
12:52 | He wrote that he wants to share a car from Boston
| |
12:52 | <laprag> heh, probably me that
| |
12:52 | <warren> but if that is the guy I'm thinking of, I do NOT want to drive with him
| |
12:52 | Now if that person I'm referring to is in this channel, I'm sorry.
| |
12:52 | laprag: oh, not you
| |
12:52 | laprag: you're cool
| |
12:52 | * Blinny goes to sulk in the corner | |
12:53 | <laprag> warren: i doubt if he's the one, he's quite calm and reserved
| |
12:53 | <Blinny> j/k (l
| |
12:53 | <johnny> we need to send warren to charm school
| |
12:53 | <warren> screw that
| |
12:53 | I know how to curtsy.
| |
12:54 | <johnny> more flys with honey than vinegar and all that..
| |
12:54 | * johnny hangs out with warren in a dress | |
12:54 | <warren> fruit flies are really attracted to vinegar...
| |
12:54 | <laprag> ooooh, it's as sweet as ever to see those nice startup lines on the mointor :)
| |
12:54 | <warren> makes a great trap
| |
12:55 | <johnny> lol.. i must be the really annoying guy then.. oops :(
| |
12:55 | <pscheie> warren, maybe; I made my own reservation
| |
12:55 | <warren> johnny: no, wasn't you
| |
12:55 | It might have been me.
| |
12:55 | <johnny> hehe
| |
12:56 | <TheBS> Anyone wanna talk technical specifics with a noob? Most directly, the Fedora 9 LTSP 5 concepts around how LDM is launched in Init?
| |
12:56 | <johnny> it's launched by ltsp-client-launch in fedora
| |
12:56 | <warren> TheBS: were you the guy that e-mailed me directly?
| |
12:56 | <TheBS> Yes
| |
12:56 | ;)
| |
12:57 | Where is ltsp-client-launch called?
| |
12:57 | During init?
| |
12:57 | <johnny> yes
| |
12:57 | <warren> Bryan?
| |
12:57 | <TheBS> I didn't locate it in Upstart/init
| |
12:57 | Yes
| |
12:57 | <johnny> it should be in there..
| |
12:57 | <TheBS> BS = Bryan Smith
| |
12:57 | I'm blind
| |
12:57 | ;)
| |
12:57 | It comes up
| |
12:57 | Just can't find it
| |
12:57 | <warren> TheBS: You e-mailed me during some terribly crazy travel and 13 hour timezone shift
| |
12:57 | <TheBS> In the chroot, or is it dynamically created?
| |
12:57 | <johnny> in the chroot
| |
12:57 | <warren> TheBS: you might be interested to hear that we're working on an official LTSP5 backport to RHEL5 son
| |
12:58 | <TheBS> Oh, no need to explain Warren. Totally understand.
| |
12:58 | <warren> TheBS: with RHEL5 as the client chroot as well
| |
12:58 | <TheBS> Oh, I was kinda doing that already. ;)
| |
12:58 | I already started hacking your latest package. So it builds the client (although not vm)
| |
12:58 | <chrisinajar> So, hypothetically, if I were to go to ltsp BTS/hackfest for only 1 day, which would be the most interesting?... I have a friend who might only be able to make it for 1 day.
| |
12:58 | <warren> TheBS: please send any notes or patches to the list...
| |
12:58 | <TheBS> I have it working other than a few details
| |
12:58 | Yes, I'm working on an unified initrd creator for both diskless and LTSP
| |
12:59 | Really simple stuff
| |
12:59 | Q-FUNK has joined #ltsp | |
12:59 | <johnny> warren, so.. i finally came up with a reason why we should do tftp in the initramfs/initrd
| |
12:59 | <warren> [root@newcaprica ~]# cat /opt/ltsp/i386/etc/sysconfig/desktop
| |
12:59 | DISPLAYMANAGER=/usr/sbin/ltsp-client-launch
| |
12:59 | johnny: ?
| |
12:59 | <johnny> shows how much i know about fedora..
| |
12:59 | <TheBS> I also found I can build a chroot just using YUM with an alternative root
| |
12:59 | <johnny> :(
| |
12:59 | <warren> TheBS: that's what prefdm runs at the end of bootup
| |
13:00 | <TheBS> Right, understand prefdm, but didn't catch where LDM is called, it is a variable set?
| |
13:00 | <johnny> warren, the reason. is that you can run a different chroot/nfs/nbd for different clients
| |
13:00 | <warren> TheBS: oh, could you please provide the recipe that you use for yum?
| |
13:00 | <TheBS> ?
| |
13:00 | <Lns> Has anyone been having issues obtaining IP addresses for TCs after the latest Ubuntu updates of the past week or so?
| |
13:00 | <warren> TheBS: /var/lib/tftpboot/i386/lts.conf SCREEN_07=ldm
| |
13:00 | <TheBS> Just a package list
| |
13:00 | I'm actually going to scrape it from a Kickstart (and eventually use PyKickstart calls proper)
| |
13:01 | <warren> TheBS: read the scripts from /usr/sbin/ltsp-client-launch which runs screen_session which runs specified screen scripts from lts.conf
| |
13:01 | <johnny> so you could provide [client] CHROOT=amd64fat
| |
13:01 | <warren> TheBS: there's actually other people in RH wanting to backport the necessary stuff to make a supported version of livecd-tools for RHEL5.
| |
13:01 | <TheBS> Yes, I caught the parameters in lts.conf
| |
13:02 | But where does it insert that for prefdm? Or call it directly?
| |
13:02 | <warren> TheBS: I would suggest helping that group, because there are benefits to relying on kickstart files.
| |
13:02 | <TheBS> What program, process, etc...?
| |
13:02 | I'm really not into doing full up puppet right now
| |
13:02 | <warren> TheBS: read /usr/sbin/ltsp-client-launch
| |
13:02 | <TheBS> This has to be working by end-of-week
| |
13:02 | ;)
| |
13:02 | <johnny> warren, wouldn't it be nice to have different clients use different roots without modifying the dhcp server?
| |
13:02 | <TheBS> Including some way to automate, which I've already started by using Perl to scrape a kickstart
| |
13:02 | <warren> johnny: yes, but I think lts.conf is the wrong way to do that.
| |
13:03 | <johnny> hmm.. what would be another way?
| |
13:03 | it has to be something in the initrd
| |
13:03 | <warren> TheBS: is this weekend deadline for proof of concept or the actual thing the customer would use?
| |
13:03 | TheBS: you could make something that works for the weekend, and help the official way after
| |
13:03 | <johnny> unless we all used the same union mounting
| |
13:03 | which we don't..
| |
13:03 | <TheBS> A working prototype that is semi-deployable
| |
13:03 | Yes, that's what I'm doing
| |
13:03 | There are other people that are waiting on the solution
| |
13:03 | <warren> TheBS: you could also provide the client chroot as a tarball for now
| |
13:04 | <TheBS> Then we'll come back
| |
13:04 | They won't accept F9 as a client
| |
13:04 | <warren> a LOT less trouble than the customer finding a network source for the RHEL packages
| |
13:04 | TheBS: I mean RHEL5 client
| |
13:04 | <TheBS> Oh, yes, I know, but there still needs to be updates
| |
13:04 | <warren> that's fine
| |
13:04 | initial install I mean
| |
13:04 | from tarball
| |
13:04 | <TheBS> They are more keen on RHEL 6 for later
| |
13:04 | I know
| |
13:04 | I'm just trying to get the LDM aspects up
| |
13:04 | <warren> dude
| |
13:04 | I told you what you need
| |
13:04 | /usr/sbin/ltsp-client-launch read this script
| |
13:05 | <TheBS> I'm trying to figure out where in F9's init that the lts.conf settings are applied to what files
| |
13:05 | And /usr/sbin/ltsp-client-launch is called where
| |
13:05 | <warren> none of the lts.connf settings are applied
| |
13:05 | <TheBS> ??
| |
13:05 | In the initrd?
| |
13:05 | <warren> no
| |
13:05 | lts.conf in Fedora isn't downloaded from the server until /usr/sbin/ltsp-client-launch
| |
13:05 | which is at the end of the boot
| |
13:05 | <TheBS> Where is that called at the end of the boot?
| |
13:05 | <warren> johnny: you coming to the hackfest? better to discuss that in person witih everyone
| |
13:06 | <johnny> warren, i can't afford it
| |
13:06 | <ogra> phew ...
| |
13:06 | <TheBS> Trying to find the exact init
| |
13:06 | <johnny> maybe next time
| |
13:06 | <ogra> my fingers are glowing
| |
13:06 | * TheBS is thickheaded | |
13:06 | <johnny> i'll be online tho
| |
13:06 | <warren> TheBS: [root@newcaprica ~]# cat /opt/ltsp/i386/etc/sysconfig/desktop
| |
13:06 | DISPLAYMANAGER=/usr/sbin/ltsp-client-launch
| |
13:06 | * TheBS is blind too | |
13:06 | <johnny> maybe you can export your gobby sessions somehow ..
| |
13:06 | <TheBS> That's exactly what I needed!
| |
13:06 | Thanx!
| |
13:06 | <warren> TheBS: after all initscripts are done, and rc.local, I think it runs this from prefdm
| |
13:06 | <TheBS> Totally missed it
| |
13:06 | Yep
| |
13:06 | japerry has joined #ltsp | |
13:06 | <TheBS> It spawns it
| |
13:06 | Told you I was blind
| |
13:06 | ;)
| |
13:07 | <ogra> johnny, they could use gobby.ubuntu.com, its a public gobby server
| |
13:07 | <johnny> is it reliable?
| |
13:07 | <TheBS> I should have done a grep -i ltsp on /etc/sysconfig/*
| |
13:07 | * TheBS smacks himself | |
13:07 | <warren> TheBS: My suggestion for shipping something this weekend 1) provide a tarball for client chroot, the chroot creation tool will be fixed by other people real soon 2) focus on mkinitrd which is really the only other problem
| |
13:08 | TheBS: we might even fix mkinitrd for RHEL5 this weekend
| |
13:08 | <TheBS> Really?
| |
13:08 | <warren> yes
| |
13:08 | <TheBS> Hmmm, okay
| |
13:08 | I was just modifying the mkinitrd from stateless/diskless
| |
13:08 | <warren> stateless/diskless has a separate mkinitrd tool?
| |
13:08 | <TheBS> With a script modification ("updateDiskless")
| |
13:08 | <ogra> johnny, well, we use it with some thousand people during UDS usually ... currently it's rather idling
| |
13:08 | <warren> TheBS: there's like three separate departments in RH working on the same fucking problem now. separately.
| |
13:09 | <TheBS> Which includes using some newer mkinitrd support in RHEL 5 (seems it is based on RHEL 3/4 timeframe developments)
| |
13:09 | <warren> TheBS: My goal is to get the LTSP5 stuff backported to RHEL5.4 mkinitrd
| |
13:09 | <TheBS> Yeah, I totally understand
| |
13:09 | Hey man, the more you can feed me from future RHEL 5, the better
| |
13:09 | <ogra> johnny, sbalneav and jammcq know about it and have used it before, its free for public use so if there are specs we should work together one gobby.ubuntu.com might be a good thing to use
| |
13:09 | <warren> TheBS: now you can do MOST of the nfs root boot stuff without modifying RHEL5 mkinitrd
| |
13:09 | <TheBS> Right
| |
13:09 | <warren> it is a little unreliable though
| |
13:09 | <TheBS> Right, it won't be perfect
| |
13:09 | <warren> TheBS: how many different types of client hardware for this customer?
| |
13:10 | <TheBS> But it should work for most PCs
| |
13:10 | <warren> TheBS: priv msg
| |
13:10 | <TheBS> Not many, and they would test before any purchase
| |
13:10 | Oh, sorry
| |
13:10 | <warren> do you see my priv msgs?
| |
13:10 | <TheBS> Yep
| |
13:10 | <warren> k
| |
13:12 | <laprag> if the authentication using ldm takes a long time, what might be wrong?
| |
13:15 | <Gadi> laprag: dns
| |
13:16 | <laprag> k
| |
13:16 | * laprag tries to fix that | |
13:17 | <ogra> laprag, !
| |
13:17 | Nubae has quit IRC | |
13:17 | <laprag> hey there herr oliver
| |
13:19 | <ogra> :)
| |
13:22 | <Lns> Anyone wanna help me fix some nbd swap issues? :)
| |
13:23 | <Gadi> so, if I have a USB-serial adapter connected to a serial printer -whats the cups uri?
| |
13:23 | serial?
| |
13:23 | usb?
| |
13:23 | * Gadi is so confused | |
13:25 | <ogra> ipp ?
| |
13:25 | * ogra isnt serious :) | |
13:25 | <Gadi> and I wont call you shirley
| |
13:25 | <ogra> i'd guess some ttyUSB device
| |
13:25 | <Gadi> right - but i dunno what backend to use
| |
13:26 | <johnny> lol
| |
13:26 | <ogra> serial i guess
| |
13:26 | <johnny> probably the serial..
| |
13:26 | <Gadi> ill try
| |
13:26 | <ogra> why the heck do you even have serial printers
| |
13:26 | <Lns> ogra: POS receipt printers i'd assume
| |
13:27 | <Gadi> because I have a customer requesting to use one with a thin client
| |
13:27 | (oh and we have label printers here that are USB or serial)
| |
13:28 | but, I need to get the serial working
| |
13:28 | <ogra> get new customers then
| |
13:28 | <Gadi> lol
| |
13:28 | <Lns> lol
| |
13:28 | * Gadi forgot to put on his developer-screw-the-user hat | |
13:28 | <Gadi> :)
| |
13:28 | <ogra> :)
| |
13:29 | * ogra just tries to recover from his talk ... sorry i'm in sarcastic mood :) | |
13:29 | <chrisinajar> ogra: if only solutions were that simple.... //get new customers then//
| |
13:30 | <ogra> heh
| |
13:34 | laprag has quit IRC | |
13:38 | alkisg has quit IRC | |
13:44 | <sbalneav> Lns: So, remind me, what's the problem again?
| |
13:46 | <Lns> sbalneav: thanks for responding :) see LP #281498
| |
13:46 | hrm..thought the bot posted urls - https://bugs.launchpad.net/ubuntu/+source/ltsp/+bug/281498
| |
13:49 | nubae has joined #ltsp | |
13:51 | <sbalneav> Lns: OK, so you've made the changes, does it give any error messages now, or is it just a case that the dd is hung up?
| |
13:51 | <Lns> sbalneav: no more messages other than hanging up on the TC at "Negotiation: ", and yes, dd is hung up on the server
| |
13:51 | <jammcq> hey, it's laprag
| |
13:53 | warren has quit IRC | |
13:54 | <jammcq> err, was laprag
| |
13:54 | <sbalneav> Lns: OK, so at a command prompt on the server, what happens if you "telnet localhost 9572"
| |
13:55 | <Lns> sbalneav: during the hang-up or does it matter?
| |
13:55 | <sbalneav> DOesn't matter
| |
13:55 | <Lns> sbalneav: lns@austincreekubuntu:~$ telnet localhost 9572 Trying 127.0.0.1... Connected to localhost. Escape character is '^]'.
| |
13:56 | bobby_C has joined #ltsp | |
13:56 | <sbalneav> Do you eventually see a string that says NBDMAGIC then some gobbdleygook?
| |
13:56 | <Lns> lesse...
| |
13:56 | <sbalneav> Should only take 1 second
| |
13:56 | <Lns> then no
| |
13:56 | <sbalneav> ok
| |
13:57 | This is on hard, right?
| |
13:57 | hardy
| |
13:57 | <Lns> yes
| |
13:57 | <sbalneav> k
| |
13:57 | what's the contents of your . /etc/ltsp/nbdswapd.conf
| |
13:57 | have one?
| |
13:57 | <Lns> lns@austincreekubuntu:~$ cat /etc/ltsp/nbdswapd.conf SIZE=256
| |
13:57 | yep
| |
13:58 | (i'm removing CRs btw)
| |
13:58 | <sbalneav> k, for the moment, lets remove it. or move it out of the way.
| |
13:58 | <Lns> k
| |
13:58 | sbalneav: oh, you know what there is a msg, let me get it real quick
| |
13:59 | laprag has joined #ltsp | |
13:59 | <Lns> sbalneav: It was that "could not parse config file nothing to do, bye!" msg - though I remember hearing that was moot
| |
14:00 | <sbalneav> what, telnet localhost 9572 gives you that?
| |
14:00 | <Lns> no
| |
14:00 | bootup on the server actually
| |
14:01 | <sbalneav> So what gives you that message?
| |
14:01 | <Lns> nbd-server
| |
14:01 | <sbalneav> ok, you got the nbdswapd.conf file out of the way?
| |
14:01 | Try re-telnetting again
| |
14:01 | ogra has quit IRC | |
14:02 | <Lns> yep.. ok, got the "NBDMAGICB ?? S
| |
14:02 | <sbalneav> ok, try booting a terminal
| |
14:02 | <Lns> ok lemme call the onsite tech
| |
14:02 | <sbalneav> that'll give you a 32 meg swap.
| |
14:02 | lets see if that works, then we need to find out why it hangs with SIZE=256
| |
14:03 | <Lns> ok
| |
14:05 | sbalneav: ok, the tech is rebooting a few thin clients right now
| |
14:06 | mikkel has joined #ltsp | |
14:07 | * Lns can't believe HP T5530 thin-clients still ship with only 128MB RAM | |
14:10 | <Lns> sbalneav: sorry, still waiting, she has to leave the phone to check on status
| |
14:10 | <sbalneav> NP
| |
14:13 | CAN-o-SPAM has joined #ltsp | |
14:16 | ogra has joined #ltsp | |
14:21 | laprag has quit IRC | |
14:23 | <Lns> sbalneav: ok, will be ~20min before we can continue - there's now a rogue trendnet device repeatedly asking for DHCP addys, causing thin clients to fail DHCP config. This is unrelated to the nbd issue though
| |
14:23 | Can I ping you when we get that off the network to continue troubleshooting?
| |
14:24 | nubae has quit IRC | |
14:25 | * Lns is starting to despise the lack of public school network guidelines for teachers | |
14:28 | pmatulis has quit IRC | |
14:30 | alkisg has joined #ltsp | |
14:32 | <sbalneav> Lns: No problem
| |
14:32 | I'll be on the channel till 4:15 or so local time.
| |
14:32 | <Lns> ok
| |
14:32 | thank you =)
| |
14:33 | Blinny has quit IRC | |
14:34 | Nubae has joined #ltsp | |
14:36 | spectra has quit IRC | |
14:40 | cliebow has joined #ltsp | |
14:42 | Nubae has left #ltsp | |
14:43 | nubae has joined #ltsp | |
14:44 | subhodip has joined #ltsp | |
14:51 | cliebow has quit IRC | |
15:00 | warren has joined #ltsp | |
15:14 | FuriousGeorge has joined #ltsp | |
15:19 | six2one has quit IRC | |
15:27 | K_O-Gnom has quit IRC | |
15:50 | subhodip has quit IRC | |
15:52 | jammcq has quit IRC | |
16:10 | <Lns> sbalneav: still there?
| |
16:10 | <sbalneav> For about 3 minutes
| |
16:10 | Got a result?
| |
16:11 | <Lns> yikes :p ok.. well w/no nbdswapd.conf, the bootups seem to be a bit faster, but still around 2-3 minutes, mostly hanging at the "hardware detection" (I'm assuming it's still at nbd swap file creation)
| |
16:11 | <sbalneav> If you set NBD_SWAP=false then it's fast?
| |
16:12 | <Lns> hmm i'm pretty sure, lemme try
| |
16:13 | Testing..one min
| |
16:14 | quick Q while waiting..if you have a LIKE stanza in lts.conf for thinclients, but no actual options for the section it goes to, it shouldn't present issues right?
| |
16:14 | the section exists but the options for them are commented out
| |
16:14 | <sbalneav> Dunno
| |
16:15 | If it does, it's a bug :)
| |
16:15 | <Lns> ok
| |
16:16 | <sbalneav> It's still booting?
| |
16:16 | <Lns> yeah...
| |
16:16 | <sbalneav> You've got something else wrong
| |
16:16 | <Lns> so it looks like osme other issue right now
| |
16:17 | ugh..it sucks cuz i know it's something new that came up over the weekend now
| |
16:17 | <sbalneav> try something like acpi=force on the pxe kernel command line.
| |
16:17 | anyway, I gotta go. I'll be on later tonight.
| |
16:17 | <Lns> ok, thanks for your help :)
| |
16:17 | <sbalneav> talk to you later.
| |
16:18 | alekibango has joined #ltsp | |
16:20 | <Gadi> Lns: you cannot have a stanza with no options
| |
16:21 | that will produce a parse error
| |
16:21 | and none of the lts.conf file params will be read
| |
16:21 | bobby_C has quit IRC | |
16:21 | <Gadi> at least put a dummy variable in that stanza
| |
16:21 | like: LNS=cool
| |
16:21 | <Lns> Gadi: so .. i have a complete "LIKE = hpt5530" stanza, with an existing [hpt5530] line..just no options listed under that
| |
16:21 | <Gadi> right
| |
16:21 | <Lns> oh ok
| |
16:22 | <Gadi> under [hpt5530] put:
| |
16:22 | LNS=cool
| |
16:22 | <Lns> lol
| |
16:22 | * Lns wants that to be a real option | |
16:22 | <Lns> and only = true will be accepted =p
| |
16:22 | ok done..i'm off to the site to troulbeshoot more
| |
16:23 | thx Gadi =)
| |
16:29 | CAN-o-SPAM has quit IRC | |
16:30 | klausade has joined #ltsp | |
16:32 | alkisg has left #ltsp | |
16:34 | chrisinajar is now known as chrisinajar|work | |
16:39 | mikkel has quit IRC | |
16:40 | Q-FUNK has quit IRC | |
17:04 | rjune__ has joined #ltsp | |
17:05 | rjune has quit IRC | |
17:20 | Gadi has left #ltsp | |
17:46 | Lns_onsite has joined #ltsp | |
17:48 | <Lns_onsite> Ok, here's an interesting q.. after, say, 10-15min of not booting any dhcp clients, i can boot a handful at a time and they'll boot to LTSP fine. If I go Options -> Reboot and they attempt PXE/DHCP again, it fails. Any ideas?
| |
17:48 | daemon.log shows dhcp is attempting DHCPOFFER but the client never accepts. Again, this is after rebooting the client after obtaining a valid lease
| |
17:56 | reboot time
| |
17:56 | Lns_onsite has quit IRC | |
17:59 | artista_frustrad has quit IRC | |
18:00 | chrisinajar has joined #ltsp | |
18:03 | Lns_onsite has joined #ltsp | |
18:05 | <Lns_onsite> ok this is really weird
| |
18:05 | Is it normal for 'tail -f' and/or auth.log to delay in logging?
| |
18:05 | on a TC, i open gnome-terminal, su to root, the first time it takes like 15 seconds for me to get a root prompt.
| |
18:06 | subsequent attempts are instantaneous, but all the while I'm 'tail -f auth.log' and it takes 10-20 sec for the auth lines to pop up after su'ing.
| |
18:07 | Also, after rebooting after a kernel upgrade to .21, I tried logging on ON THE SERVER on tty1, tty2, tty3 and GDM....all of them hung after logging in. MOTD displayes on ttys, and then ...nothing. No prompt.
| |
18:07 | I rebooted the server after seeing that and everything on the server console itself seemed fine..but wtf would cause that, and now this what i'm seeing on TCs ?
| |
18:08 | warren has quit IRC | |
18:14 | Lns_onsite has quit IRC | |
18:32 | Lns_onsite has joined #ltsp | |
18:33 | <Lns_onsite> Ok who wants to kill me
| |
18:33 | * Lns_onsite raises hand | |
18:33 | <Ryan52> what did you break? ;)
| |
18:34 | <Lns_onsite> It seems as though after the last batch of kernel upgrades (amongst other upgrades), there is some sort of I/O, or other system starvation going on to where even logging onto a tty is causing a ~30sec (minimum) delay
| |
18:34 | This isn't just LTSP anymore
| |
18:36 | on the server, I reboot - logging onto tty1, tty2, tty3...and logging out, causes a major delay with no rampant processes, no disk i/o... loggign onto GDM does the same thing (hangs after password is entered and character dots gray-out). Also, same thing after booting ~8 thin-clients. 8 boot up fine (after about 10 minutes of network/system idle time), and then if i reboot, the same clients won't obtain an IP address. Something crazy is going on on the serv
| |
18:36 | er.
| |
18:36 | It's hanging on *something*, and it doesn't seem to be auth specific, since dhcp clients are acting the same way....
| |
18:38 | now the ttys are logging in fine...
| |
18:39 | <vagrantc> DNS?
| |
18:40 | cyberorg has quit IRC | |
18:41 | <Lns_onsite> vagrantc: would that cause ttys on the server console to hang on login and logout?
| |
18:42 | now that the ttys were logging in, i went back out into the lab and booted 8 thinclients...they booted fine.
| |
18:42 | i immediately rebooted them via ldm options -> reboot, and now they won't get an IP.
| |
18:43 | and now the ttys on the server console are hanging on login.
| |
18:43 | something is exhausting itself on the server.
| |
18:44 | this is the weirdest f*cking thing iv'e ever seen. pardon my french..
| |
18:46 | from my experience it seems similar to entropy running out, but i still have 3k
| |
18:47 | nbd_swap is even turned off, so that's not the issue.
| |
18:48 | at least as far as exausting i/o or some crazy thing
| |
18:51 | <vagrantc> well, DNS with network auth ...
| |
18:51 | could definitely cause issues on the console, too.
| |
18:52 | <loather-work> yeah, if dns isn't working it will cause console logins to take a while
| |
18:52 | dns pretty much sucks
| |
18:54 | <Lns_onsite> well dns seems to be fine.. it's querying, on the booted-up terminals and server i never got a resolve issue when visiting pages/here
| |
18:54 | staffencasa has quit IRC | |
18:55 | <Lns_onsite> and it seems completely reproducible when booting about 8 pxe/dhcp clients...i'm wondering what apic/something like that has to do with it
| |
19:02 | FuriousGeorge has left #ltsp | |
19:02 | petre has joined #ltsp | |
19:07 | Lns_onsite has quit IRC | |
19:28 | Gadi has joined #ltsp | |
19:30 | <sbalneav> Evening all!
| |
19:31 | <Lns> sbalneav!
| |
19:32 | <stgraber> evening sbalneav
| |
19:32 | <sbalneav> Hey hey
| |
19:33 | <stgraber> sbalneav: I just saw yout update to the wiki, it's interesting because it's something we'd have wanted to discuss with Francis :) (having X started from the screen script)
| |
19:34 | sbalneav: I'm not really a fan of having everything hardcoded in C unless there is a real improvement in doing so (shell is so much easier to read/understand :))
| |
19:34 | <Lns> Ok I'm logging out...but anyone who wants to think about this in the bg.. With whatever Hardy kernel/other updates in the past 1 1/2 weeks or so, my servers have been able to boot a set (8 normally) thinclients fine. As soon as I turn them off and back on, DHCP will not give them an address, as well as at the server console, logging on (tty/gdm/doesn't matter) produces incredible lag. MOTD on TTYs show, prompt will not show for at least 2 minutes. After sitting idl
| |
19:34 | e for 10-15min, everything is back to normal. Fight!
| |
19:34 | * Lns waves goodnight | |
19:34 | Lns has quit IRC | |
19:37 | * petre agrees with stgraber | |
19:39 | <sbalneav> Well, C ldm was a huge speed improvement over the python one.
| |
19:40 | And things like cdpinger, and ltspfs would be impossible to do in sh.
| |
19:41 | Sh's fine, but all it does is make it easier on US. C makes it easer for the people running low end clients.
| |
19:41 | Who, after all, is who we've traditionally tried to support.
| |
19:47 | Gadi: You about big guy?
| |
19:58 | <loather-work> i like the fact that it was moved to C
| |
19:59 | the python LDM was such a dog
| |
19:59 | <sbalneav> It got us started, so it was a good first go around.
| |
20:00 | <loather-work> definitely. python for proof-of-concept is a great idea
| |
20:00 | i just wouldn't write anything production in it
| |
20:00 | <sbalneav> I think it makes sense to do things in C that will give us the performance boost/memory savings we need, and anything that's non-critical, leave as much as we can in a HLL
| |
20:02 | <loather-work> python is *so* slow and has a resource footprint the size of the moon though. i'd say anything that can't be done in shell should be done in C.
| |
20:03 | ...and you can do a *lot* of stuff in shell.
| |
20:08 | dberkholz_ has joined #ltsp | |
20:09 | chrisinajar_ has joined #ltsp | |
20:09 | ccherret1 has joined #ltsp | |
20:10 | <sbalneav> I'm working on adding serial support to jetpipe. Sockets, serial geometry, devices. There's something that a PITA to do in C, but will be fairly short in python.
| |
20:11 | <Ryan52> what's jetpipe?
| |
20:11 | <sbalneav> The print server
| |
20:11 | <Ryan52> ah
| |
20:11 | <sbalneav> for hanging a printer off a terminal
| |
20:11 | Gadi has quit IRC | |
20:11 | chrisinajar has quit IRC | |
20:11 | rjune__ has quit IRC | |
20:11 | mccann has quit IRC | |
20:11 | daduke_ has quit IRC | |
20:11 | Hyperbyte has quit IRC | |
20:11 | ninkendo has quit IRC | |
20:11 | jbrett has quit IRC | |
20:11 | CaScAdE^1arAway has quit IRC | |
20:11 | dberkholz has quit IRC | |
20:11 | captain_magnus has quit IRC | |
20:11 | jcastro has quit IRC | |
20:11 | rcy has quit IRC | |
20:11 | ccherrett has quit IRC | |
20:11 | <sbalneav> We used to use lp_server, which did absolutely everything we wanted
| |
20:12 | Gadi has joined #ltsp | |
20:12 | chrisinajar has joined #ltsp | |
20:12 | rjune__ has joined #ltsp | |
20:12 | mccann has joined #ltsp | |
20:12 | rcy has joined #ltsp | |
20:12 | CaScAdE^1arAway has joined #ltsp | |
20:12 | captain_magnus has joined #ltsp | |
20:12 | Hyperbyte has joined #ltsp | |
20:12 | daduke_ has joined #ltsp | |
20:12 | jbrett has joined #ltsp | |
20:12 | ninkendo has joined #ltsp | |
20:12 | jcastro has joined #ltsp | |
20:12 | <sbalneav> except come with a license that we could redistribute it legally
| |
20:13 | jammcq has joined #ltsp | |
20:13 | <jammcq> helloooooo
| |
20:14 | <sbalneav> Hey jammcq
| |
20:14 | <Gadi> hello
| |
20:15 | <sbalneav> Hey Gadi !
| |
20:15 | <Gadi> sbalneav: got gobby installed and branching ltsp-trunk
| |
20:16 | will host jetpipe in a sec
| |
20:16 | japerry has quit IRC | |
20:16 | loather has quit IRC | |
20:17 | <sbalneav> I'm already editing, let me know and I'll load in what I've done so far
| |
20:19 | <Gadi> hehe
| |
20:19 | sure... start playing without me
| |
20:20 | * Gadi taps microphone | |
20:20 | <Gadi> testing. 1, 2, 3
| |
20:21 | chrisinajar_ has quit IRC | |
20:22 | <sbalneav> I'm here
| |
20:24 | <rjune__> !g
| |
20:24 | <ltspbot> rjune__: "g" is Gadi!!!!!!!!!!!!!!!!!!!!!!!!
| |
20:25 | warren has joined #ltsp | |
20:26 | artista_frustrad has joined #ltsp | |
20:29 | chrisinajar has quit IRC | |
20:36 | <Ryan52> warren: if I just want two virtual machines to talk to each other (and not my main internet), can I use a bridge and use tap devices like your example script does?
| |
20:38 | <warren> yes
| |
20:38 | Ryan52: and you can add multiple tap devices to a VM
| |
20:39 | <Ryan52> I mean can I just have a bridge with 2 tap devices?
| |
20:39 | japerry has joined #ltsp | |
20:39 | <Ryan52> cause it's not working...
| |
20:40 | japerry_cat has joined #ltsp | |
20:41 | <Ryan52> using tcp sockets I can get the thin client to get dhcp from the server vm, but then there's just moving dots for a while until I loose my patience.
| |
20:50 | <vagrantc> Ryan52: i've had to bring the interfaces up with "ifconfig $IFACE up" and such ...
| |
20:51 | Ryan52: my scripts not working for you?
| |
20:53 | * vagrantc has been making extensive use of bridges and virtualbox the last week or so | |
20:55 | japerry has quit IRC | |
20:58 | <Ryan52> vagrantc: yes, I tried that, and I also tried "ifconfig $IFACE up 0.0.0.0"..
| |
20:58 | <vagrantc> Ryan52: this with kvm ?
| |
20:59 | <Ryan52> yup
| |
21:13 | screw it; I'll just use a real diskless terminal at freegeek on saturday :P
| |
21:13 | <vagrantc> heh
| |
21:14 | vagrantc has quit IRC | |
21:35 | cyberorg has joined #ltsp | |
21:40 | mccann has quit IRC | |
21:43 | petre has quit IRC | |
21:55 | kwak__ has quit IRC | |
22:02 | siki has joined #ltsp | |
22:06 | CaScAdE^FarAway has joined #ltsp | |
22:18 | kwak__ has joined #ltsp | |
22:19 | <kwak__> help. i'm still geting the initramfs prompt while ubuntu boots
| |
22:19 | CaScAdE^1arAway has quit IRC | |
22:20 | <kwak__> there's an error usplash: no usable theme found for 640x800
| |
22:20 | screen init failed
| |
22:20 | then i get initramfs prompt. in order to proceed i have to type 'return'
| |
22:26 | <johnny> 640x800??
| |
22:26 | is that your screen?
| |
22:28 | <kwak__> nope. i didn't set any screen res yet
| |
22:28 | ubuntu is in bootup process when this came out
| |
22:33 | kwak_ has joined #ltsp | |
22:34 | <kwak_> any suggestions?
| |
22:36 | kwak has joined #ltsp | |
22:36 | <kwak> my clients are getting this error.
| |
22:37 | after setting up ltsp client
| |
22:37 | Receive control failed (-32)
| |
22:46 | gentgeen__ has joined #ltsp | |
22:47 | <sbalneav> it's When are they getting this error exactly?
| |
22:47 | What kind of clients are they, BTW?
| |
22:51 | <kwak> dell optiplex 100
| |
22:51 | <sbalneav> How much ram?
| |
22:51 | <kwak> they get the error after the line - building ltsp client
| |
22:51 | 3GB
| |
22:51 | sorry
| |
22:51 | the clients 128mb
| |
22:51 | the server is 3Gb
| |
22:52 | <sbalneav> You're building the chroot for the clients on the server, right?
| |
22:53 | kwak__ has quit IRC | |
22:53 | Gadi has left #ltsp | |
22:54 | <sbalneav> is the server getting this error message during the ltsp-build-client, or are the clients getting it during boot. I'm not clear on that.
| |
22:55 | kwak_ has quit IRC | |
22:57 | kwak2 has joined #ltsp | |
22:58 | <kwak2> sbalneav. I'm getting the error from the clients. Now I'm not getting any response from the client. Can't get any dhcp response.
| |
22:58 | I have 3 DELL optiplex GX150. i tried on 1 earlier, and it work but not for the GX100
| |
22:59 | <sbalneav> What's the server, 64 bit or 32 bit?
| |
23:02 | kwak has quit IRC | |
23:03 | kwak2 has quit IRC | |
23:18 | alkisg has joined #ltsp | |
23:39 | dberkholz_ is now known as dberkholz | |
23:48 | alkisg has left #ltsp | |