00:02 | FuriousGeorge_ has joined #ltsp | |
00:05 | FuriousGeorge has quit IRC | |
00:16 | waldo323 has quit IRC | |
00:24 | alincoln has quit IRC | |
00:25 | alincoln has joined #ltsp | |
00:32 | Selveste1 has quit IRC | |
00:35 | FuriousGeorge_ has left #ltsp | |
00:36 | FuriousGeorge has joined #ltsp | |
00:36 | <FuriousGeorge> grrr, i keep losing my connection
| |
00:36 | anyway, aside from the logistical danger of someone accidentally shutting off the server or spilling coffee on it (but not in my case)
| |
00:36 | i see no reason why i cant use my server as a workstation for a local user
| |
00:37 | of course, google searching "ltsp user login server workstation" or something like that isn't really helping ;)
| |
00:40 | alkisg has joined #ltsp | |
00:42 | rhodan has quit IRC | |
00:46 | F-GT has quit IRC | |
00:52 | map7 has joined #ltsp | |
00:52 | alkisg has quit IRC | |
00:53 | alincoln has quit IRC | |
00:55 | alincoln has joined #ltsp | |
01:02 | F-GT has joined #ltsp | |
01:05 | alkisg has joined #ltsp | |
01:11 | Kicer86 has joined #ltsp | |
01:12 | alkisg has quit IRC | |
01:14 | bobby_C has joined #ltsp | |
01:17 | sene has quit IRC | |
01:21 | shogunx has quit IRC | |
01:21 | Selveste1 has joined #ltsp | |
01:22 | bobby_C has quit IRC | |
01:27 | shogunx has joined #ltsp | |
01:35 | Selveste1 has quit IRC | |
01:40 | map7 has quit IRC | |
02:03 | gnunux has joined #ltsp | |
02:08 | alkisg has joined #ltsp | |
02:25 | map7 has joined #ltsp | |
02:37 | F-GT has quit IRC | |
02:37 | alkisg has quit IRC | |
02:38 | alkisg has joined #ltsp | |
02:48 | FuriousGeorge_ has joined #ltsp | |
02:49 | pts has joined #ltsp | |
02:49 | korcan has quit IRC | |
02:50 | FuriousGeorge has quit IRC | |
02:54 | FuriousGeorge_ has left #ltsp | |
02:54 | F-GT has joined #ltsp | |
03:14 | pts has quit IRC | |
03:15 | pts_ has joined #ltsp | |
03:17 | FuriousGeorge has joined #ltsp | |
03:17 | <FuriousGeorge> hey all
| |
03:17 | i installed ltsp over a year ago (kubuntu). Are there any good guides to updating it, and is it worth it?
| |
03:18 | well, i guess the first question is: what, if any, are good reasons to upgrade kubuntu ltsp which was installed circa 11/08?
| |
03:24 | FuriousGeorge_ has joined #ltsp | |
03:26 | FuriousGeorge has quit IRC | |
03:44 | alkisg has quit IRC | |
04:19 | FuriousGeorge has joined #ltsp | |
04:22 | FuriousGeorge_ has quit IRC | |
04:27 | Faithful has quit IRC | |
04:44 | Faithful has joined #ltsp | |
04:45 | alkisg has joined #ltsp | |
04:59 | artista_frustrad has joined #ltsp | |
05:03 | Selveste1 has joined #ltsp | |
05:05 | hersonls has joined #ltsp | |
05:06 | avolkovi has quit IRC | |
05:06 | Selveste1 has quit IRC | |
05:07 | Selveste1 has joined #ltsp | |
05:17 | FuriousGeorge_ has joined #ltsp | |
05:17 | frederickjh has joined #ltsp | |
05:20 | FuriousGeorge has quit IRC | |
05:24 | * alkisg notes that `apt-get install ltsp-server` on lucid uses "/srv/tftp" as the tftpboot directory! We should use preseeding instead... | |
05:24 | <alkisg> stgraber: ^^
| |
05:27 | Urm, and also it's started as a daemon, not from inetd... hmm...
| |
05:33 | alkisg has quit IRC | |
05:35 | alkisg has joined #ltsp | |
05:45 | mikkel has joined #ltsp | |
05:54 | Selveste1 has quit IRC | |
06:04 | pmatulis has joined #ltsp | |
06:06 | Kicer86 has quit IRC | |
06:07 | lucascoala has joined #ltsp | |
06:20 | artista_frustrad has quit IRC | |
06:26 | Kicer86 has joined #ltsp | |
06:29 | Kicer86 has quit IRC | |
06:30 | Kicer86 has joined #ltsp | |
06:36 | Kicer86 has quit IRC | |
06:36 | Kicer86 has joined #ltsp | |
06:39 | klausade has joined #ltsp | |
06:55 | GodFather has joined #ltsp | |
07:21 | vvinet has joined #ltsp | |
07:29 | alkisg has quit IRC | |
07:35 | johnny has quit IRC | |
07:39 | Selveste1 has joined #ltsp | |
07:44 | japerry has quit IRC | |
07:56 | etyack has joined #ltsp | |
08:16 | Gadi has joined #ltsp | |
08:28 | japerry has joined #ltsp | |
08:35 | jammcq has joined #ltsp | |
08:35 | <jammcq> good morning friends
| |
08:40 | highvoltage has quit IRC | |
08:41 | japerry has quit IRC | |
08:42 | highvoltage has joined #ltsp | |
08:44 | vbundi has joined #ltsp | |
08:46 | lucascoala has quit IRC | |
08:50 | alexqwesa has quit IRC | |
08:56 | alexqwesa has joined #ltsp | |
08:56 | pts_ has quit IRC | |
09:00 | etyack has left #ltsp | |
09:09 | shawnp0wers has joined #ltsp | |
09:17 | astbis has joined #ltsp | |
09:17 | etyack has joined #ltsp | |
09:19 | astbis has quit IRC | |
09:19 | etyack has quit IRC | |
09:19 | etyack has joined #ltsp | |
09:24 | hery_ has joined #ltsp | |
09:29 | hery_ has left #ltsp | |
09:29 | hery_ has joined #ltsp | |
09:31 | hery_ has quit IRC | |
09:32 | heryatmadja has joined #ltsp | |
09:33 | heryatmadja has quit IRC | |
09:34 | hery_ has joined #ltsp | |
09:35 | hery_ has quit IRC | |
09:35 | mikkel has quit IRC | |
09:36 | kevin__ has joined #ltsp | |
09:38 | etyack has left #ltsp | |
09:38 | Selveste1 has quit IRC | |
09:40 | hery_ has joined #ltsp | |
09:41 | etyack has joined #ltsp | |
09:42 | japerry has joined #ltsp | |
09:42 | shawnp0wers has quit IRC | |
09:47 | johnny has joined #ltsp | |
09:49 | <sbalneav> Morning all
| |
09:50 | shawnp0wers has joined #ltsp | |
09:53 | <vvinet> morning
| |
09:56 | CAN-o-SPAM has joined #ltsp | |
09:57 | <shawnp0wers> Test
| |
09:58 | <sbalneav> Morning shawnp0wers
| |
09:58 | <shawnp0wers> kevin__: you can't comment here?
| |
09:58 | 'mornin sbalneav :)
| |
09:59 | kevin__ has quit IRC | |
10:00 | kevin__ has joined #ltsp | |
10:02 | kevin__ has left #ltsp | |
10:05 | <CAN-o-SPAM> shawnp0wers: registered nicks now required
| |
10:05 | <shawnp0wers> OH, I see
| |
10:05 | <sbalneav> CAN-o-SPAM: That something permanent, do you know, or is that just while freenode's under attack?
| |
10:06 | <CAN-o-SPAM> sbalneav: not sure, know it was to help combat against those attacks
| |
10:06 | <johnny> i wish freenode would open up an xmpp interfae :(
| |
10:06 | * shawnp0wers didn't realize not everyone was registered. lol | |
10:06 | <CAN-o-SPAM> don't see why they wouldn't keep it
| |
10:06 | hery_ has left #ltsp | |
10:06 | <sbalneav> Yeah, only problem is, for irc first-timers, it's a bit of a hurdle
| |
10:07 | hery_ has joined #ltsp | |
10:07 | gnunux has quit IRC | |
10:08 | hery_ has quit IRC | |
10:08 | <CAN-o-SPAM> sbalneav: i must admit i struggled for a minute, but the last time i registered an IRC nick was many many years ago, and i thought this one was unique enough :)
| |
10:13 | herytjb has joined #ltsp | |
10:13 | <herytjb> testing
| |
10:13 | <johnny> CAN-o-SPAM, /msg nickserv help
| |
10:13 | <herytjb> finally
| |
10:13 | morning all (actually night here)
| |
10:16 | herytjb has quit IRC | |
10:17 | hery-infinity has joined #ltsp | |
10:18 | hery-infinity has quit IRC | |
10:19 | <shawnp0wers> I wonder if the /topic could reflect the need to have a registered nick...
| |
10:19 | for those unfamiliar with IRC
| |
10:20 | <johnny> how can you get here if you need a registered nick?
| |
10:20 | considered switching to xmpp instead then..
| |
10:20 | and let folks use speeqe to get support if they have no irc client..
| |
10:20 | damianos_ has joined #ltsp | |
10:21 | <shawnp0wers> johnny: You can get here, just not "say" anything. :)
| |
10:21 | <johnny> ah
| |
10:21 | herytjb has joined #ltsp | |
10:23 | <herytjb> nickserv help
| |
10:23 | herytjb has left #ltsp | |
10:23 | herytjb has joined #ltsp | |
10:26 | damianos_ has left #ltsp | |
10:27 | frederickjh has quit IRC | |
10:28 | damianos has joined #ltsp | |
10:32 | frederickjh has joined #ltsp | |
10:32 | Kicer86 has quit IRC | |
10:47 | bulletxt has joined #ltsp | |
10:48 | <bulletxt> hi, there is something I'm trying to understand. is it faster to load a program through server or local app?
| |
10:48 | <johnny> depends..
| |
10:48 | on network connection .. cpu speed of the client, ram
| |
10:49 | <bulletxt> if the network is good but the client has a slow cpu, is it better to load via server right?
| |
10:52 | Damianos has joined #ltsp | |
10:54 | <johnny> the ram is a bigger deal often
| |
10:54 | yes.. would be better to load via server
| |
10:55 | that'ts why ltsp does that by default :)
| |
10:59 | ogra has joined #ltsp | |
11:00 | <herytjb> what's the pro and con between ltsp solution or 'virtual desktop' solution (e.g. NComputing)?
| |
11:02 | <sbalneav> herytjb: GIYF
| |
11:02 | <johnny> because that requires you to have a nicer computer
| |
11:02 | to be able to launch such an app
| |
11:02 | for a client
| |
11:03 | altho you can combine them
| |
11:03 | <sbalneav> herytjb: also, the docs may help you
| |
11:03 | !docs
| |
11:03 | <ltspbot> sbalneav: "docs" :: For the most current documentation, see https://sourceforge.net/apps/mediawiki/ltsp/index.php?title=Ltsp_LtspDocumentationUpstream
| |
11:03 | <johnny> the apps run on the server so you can take advantage of faster speeds
| |
11:04 | rjune has joined #ltsp | |
11:04 | <herytjb> I googled GIYF to find out that Google is INDEED Your Friend. :)
| |
11:04 | thanks
| |
11:05 | Managed to download LTSP manual. will check
| |
11:06 | rjune has quit IRC | |
11:10 | Damianos has quit IRC | |
11:10 | Damianos has joined #ltsp | |
11:13 | staffencasa has quit IRC | |
11:14 | ogra has quit IRC | |
11:17 | staffencasa has joined #ltsp | |
11:20 | rhodan has joined #ltsp | |
11:27 | vagrantc has joined #ltsp | |
11:31 | herytjb has quit IRC | |
11:32 | ogra has joined #ltsp | |
11:43 | cliebow has joined #ltsp | |
11:49 | <vbundi> If I wanted to run 2 dhcp servers with one doing *only* the thin clients (ltsp server) how do I stop the other dhcp server from offering ip's when the thin client is booting
| |
11:49 | <vagrantc> why do you need to run 2 dhcp servers?
| |
11:49 | <vbundi> the ltsp dhcp server gives the first ip okay, but as it's booting it seems to grab it again and it gets it from the other server... It screws everything up and I get an initramfs prompt/busybox
| |
11:50 | mostly for testing right now
| |
11:50 | <johnny> why not let your other server serve ltsp?
| |
11:51 | easier that way
| |
11:51 | serve the kernel that is..
| |
11:51 | <vbundi> because at this time, that server is currently serving LTSP4
| |
11:52 | <vagrantc> you could configure pxelinux to use IPAPPEND... then it would get the same ip information both times...
| |
11:53 | <vbundi> vagrantc: sounds like what I want to do, never heard of ipappend
| |
11:53 | <vagrantc> alkisg is the master of IPAPPEND, but apparently not around at the moment...
| |
11:53 | <vbundi> hmm
| |
11:54 | looks like I can just add ipappend 1 to my pxelinux.cfg
| |
11:55 | or maybe not quite that easy..
| |
11:57 | <vagrantc> nothing like trying it out to find out :)
| |
11:59 | <vbundi> yea
| |
11:59 | I'll let you know if I get it workin, thanks for the help
| |
12:04 | aha that did it first try
| |
12:04 | exit
| |
12:04 | oops
| |
12:04 | cliebow has quit IRC | |
12:04 | cliebow has joined #ltsp | |
12:05 | <vbundi> vagrantc: ip=dhcp:192.168.119:192.168.1.1:255.255.255.0
| |
12:05 | vagrantc: ipappend 1
| |
12:05 | added those to /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default
| |
12:05 | <vagrantc> vbundi: i thought ipappend added the ip address for you?
| |
12:06 | <vbundi> syntax is ip=<client-ip>:<boot-server-ip>:<gateway>:<netmask>
| |
12:06 | <vagrantc> shouldn't need to specify ip= in pxelinux.cfg/default
| |
12:06 | <vbundi> no? I'll try removing it then.
| |
12:07 | <vagrantc> yes, i'm well versed in the syntax of the ip= kernel parameter :)
| |
12:07 | <vbundi> oh ok
| |
12:07 | hmm yea you're right I didn't need it
| |
12:08 | <vagrantc> another option would be to use dnsmasq DHCP proxy configuration on the LTSP server... but if you have two ltsp servesr, this won't likely help so much
| |
12:08 | <vbundi> I thought I'd have to specify my boot-server because I have 2 dhcp servers serving 2 images
| |
12:10 | vagrantc: hmm actually I'm getting a blank screen when it boots up... I see my cursor and everything but nothin else.
| |
12:10 | Damianos has quit IRC | |
12:11 | <vbundi> adding the ip= line back in didn't help either... wonder why it worked the first time and not now
| |
12:14 | restupi has joined #ltsp | |
12:16 | <frederickjh> Hi all! I recently upgraded my Ubuntu LTSP Intrepid to Jaunty to Karmin, and now I am having a strange problem with the sound. As soon as the client machine says " Starting LTSP Client . . . " the speakers start putting out white noise (like an FM radio not tuned to any station).
| |
12:17 | <vagrantc> set your *VOLUME= parameters...
| |
12:17 | <frederickjh> This white noise can not be muted on any audio channel and only stops when the client shuts down.
| |
12:17 | <Gadi> frederickjh: try adding MIC_VOLUME=0 and FRONT_MIC_VOLUME=0 to your lts.conf
| |
12:17 | <vagrantc> frederickjh: it's probably picking up from the mic and generating a feedback loop
| |
12:17 | <frederickjh> Ok will try.
| |
12:18 | <vagrantc> and it will considerably reduce the pain in your ears to lower VOLUME=20
| |
12:18 | while troubleshooting
| |
12:18 | <highvoltage> .wub 12
| |
12:21 | <frederickjh> interesting vagrantc that you say that as I find the volume on linux systems a bit low.
| |
12:26 | <cliebow> a
| |
12:27 | <frederickjh> vagrantc what is the default setting for the VOLUME= parameter?
| |
12:28 | Thanks everyone that did the trick.
| |
12:29 | Do you know if the autologin feature in the lts.conf is still working? I tried setting it up for testing stuff to fix this but was never able to get it to work.
| |
12:30 | <johnny> should be working
| |
12:30 | you probably didn't set it right
| |
12:30 | <frederickjh> I added a section with the mac address in [ ]s and then the LDM_USERNAME=
| |
12:30 | LDM_PASSWORD= parameters with the relavant details
| |
12:30 | <johnny> ah.. that shouldwork
| |
12:31 | cliebow has quit IRC | |
12:31 | <frederickjh> Do the letters in the mac address need to be capitals or lowercase
| |
12:31 | ?
| |
12:31 | cliebow has joined #ltsp | |
12:31 | <frederickjh> or does it matter
| |
12:32 | restupi has left #ltsp | |
12:32 | Q-FUNK has joined #ltsp | |
12:32 | <Q-FUNK> re
| |
12:33 | <frederickjh> Does this look right:
| |
12:33 | [00:13:90:02:D9:83]
| |
12:33 | LDM_USERNAME=test
| |
12:33 | LDM_PASSWORD=test1234
| |
12:35 | cliebow has quit IRC | |
12:35 | cliebow has joined #ltsp | |
12:36 | <stgraber> frederickjh: LDM_AUTOLOGIN=true
| |
12:38 | vbundi has quit IRC | |
12:40 | <Q-FUNK> stgraber: would you have time to test the current -geode in Lucid on GX2 hardware?
| |
12:40 | vbundi has joined #ltsp | |
12:42 | vbundi_ has joined #ltsp | |
12:44 | <vbundi_> .
| |
12:44 | vbundi_ has quit IRC | |
12:47 | jhutchins has quit IRC | |
12:48 | jhutchins has joined #ltsp | |
12:50 | <frederickjh> I will try that stgraber
| |
12:51 | Does that go under the [Default] section in lts.conf?
| |
12:53 | I guess it does it worked.
| |
12:53 | Where is the best guide to the lts.conf parameters on the internet?
| |
12:59 | etyack has quit IRC | |
13:00 | <cliebow> !docs
| |
13:00 | <ltspbot> cliebow: "docs" :: For the most current documentation, see https://sourceforge.net/apps/mediawiki/ltsp/index.php?title=Ltsp_LtspDocumentationUpstream
| |
13:04 | jammcq has quit IRC | |
13:06 | frederickjh has quit IRC | |
13:13 | <cliebow> kj
| |
13:15 | Q-FUNK has left #ltsp | |
13:21 | <Gadi> _UsUrPeR_: ping
| |
13:21 | johnny has left #ltsp | |
13:24 | dro has joined #ltsp | |
13:25 | <dro> CAN-o-SPAM: ltsp-discuss is busy today :P
| |
13:26 | <CAN-o-SPAM> Gadi: _UsUrPeR_ is out, whats up?
| |
13:26 | dro: hahaha
| |
13:28 | jhutchins has quit IRC | |
13:28 | johnny has joined #ltsp | |
13:28 | jhutchins has joined #ltsp | |
13:29 | dro is now known as treyh | |
13:30 | treyh is now known as dro | |
13:30 | Ahmuck has quit IRC | |
13:31 | frederickjh has joined #ltsp | |
13:32 | frederickjh has left #ltsp | |
13:32 | frederickjh has joined #ltsp | |
13:33 | dro has quit IRC | |
13:33 | dro has joined #ltsp | |
13:34 | alexqwesa has quit IRC | |
13:36 | alexqwesa has joined #ltsp | |
13:37 | <Gadi> CAN-o-SPAM: just wondering (since you guys play a lot with the via driver) if you know whether the driver supports cloning the display on two outputs with different resolutions
| |
13:38 | say, CRT @ 1280x1024 and DVI @ 1920x1080
| |
13:38 | but mirrored
| |
13:38 | not xinerama
| |
13:38 | * Gadi suspects the driver is not capable | |
13:48 | alkisg has joined #ltsp | |
13:51 | <knipwim> how come i'm seeing blocks instead of characters in my ldm?
| |
13:52 | ldm-2.0.50
| |
13:52 | <vagrantc> knipwim: don't have the appropriate fonts installed?
| |
13:52 | <Gadi> compiz?
| |
13:52 | <vagrantc> a number of asian locales require fonts i don't typically install by default
| |
13:53 | ogra_ has joined #ltsp | |
13:53 | ogra has quit IRC | |
13:54 | johnny has left #ltsp | |
13:54 | johnny has joined #ltsp | |
13:55 | <CAN-o-SPAM> Gadi: not sure ATM, my initial guess would be no
| |
13:56 | touch base with _UsUrPeR_ on monday, but as far as i know hes never looked at that
| |
13:56 | <knipwim> vagrantc: which are appropriate for an english locale?
| |
13:56 | corefonts? freefonts?
| |
13:56 | <vagrantc> knipwim: no idea... that always "just works" for me :)
| |
13:58 | knipwim: ttf-devavu-core ?
| |
13:59 | i can't spell that for the life of me.
| |
13:59 | ttf-dejavu-core
| |
14:01 | <knipwim> i'll give it a try
| |
14:02 | akuepker has joined #ltsp | |
14:02 | FuriousGeorge_ has quit IRC | |
14:04 | ogra_ has quit IRC | |
14:05 | dro has quit IRC | |
14:06 | bobby_C has joined #ltsp | |
14:07 | ogra has joined #ltsp | |
14:10 | otavio has quit IRC | |
14:13 | alkisg has quit IRC | |
14:19 | <knipwim> vagrantc: it worked!
| |
14:20 | thanks
| |
14:20 | alkisg has joined #ltsp | |
14:20 | <vagrantc> great
| |
14:22 | alkisg has quit IRC | |
14:24 | alkisg has joined #ltsp | |
14:25 | cliebow has quit IRC | |
14:26 | alkisg has quit IRC | |
14:26 | dro has joined #ltsp | |
14:28 | <jhutchins_lt> Gadi: Does the card have enough ram for both buffers?
| |
14:29 | <Gadi> jhutchins_lt: yup
| |
14:29 | I can Xinerama it just fine
| |
14:29 | just cannot clone
| |
14:29 | at different res
| |
14:29 | alkisg has joined #ltsp | |
14:30 | <jhutchins_lt> Kinda makes sense given the idea of "clone".
| |
14:33 | <Gadi> gotta run
| |
14:33 | ciao
| |
14:33 | Gadi has left #ltsp | |
14:38 | <knipwim> so the [Default] header is the minimum content of an lts.conf?
| |
14:38 | ogra has quit IRC | |
14:39 | alkisg has joined #ltsp | |
14:39 | ogra has joined #ltsp | |
14:39 | shawnp0wers has quit IRC | |
14:42 | <dro> knipwim: the minimum content?
| |
14:42 | <knipwim> for a valid default file
| |
14:42 | kusznir has joined #ltsp | |
14:43 | <dro> knipwim: yea that would work
| |
14:43 | knipwim: but if you'e not going to use it, you can delete it
| |
14:43 | <kusznir> Hi all: I've got a user who is trying to use my ltsp setup (ubuntu 9.04 server), but whenever they start applications (such as the gnome terminal), it takes minutes before it actually appears on the desktop. The visual updates seem to be timely enough (moving windows around is near real-time); its just starting grapical apps takes a while before the window shows up.
| |
14:44 | <dro> kusznir: hardware specs of thin client? and is it only affecting one user?
| |
14:44 | <kusznir> All users have issues with some apps, such as inkscape. Many users have reported issues with apps such as "calculator", while other users don't have these issues. This particular user can't seem to load much of anything timely.
| |
14:45 | <dro> kusznir: does it happen to that user no matter what client they connect from?
| |
14:45 | <kusznir> dro: the one user right now is actually using a full PC (Pentium D 3.2Ghz, gig-E, nvidia 6100 video). Most of our users use Devon IT TC5's (Intel 945-based systems)
| |
14:45 | <dro> kusznir: a generic question but have you updated your chroot?
| |
14:47 | <knipwim> dro: so no lts.conf is actually necessary?
| |
14:47 | <dro> knipwim: correct
| |
14:47 | <vagrantc> dro: lts.conf shouldn't be necessary, no.
| |
14:48 | <knipwim> johnny: i'm wondering why a default lts.conf is included in the ebuild
| |
14:48 | <kusznir> dro: I have made changes to the chroot some time ago, but I have not done a os-style update to the chroot, and I haven't made any changes to the chroot in some time.
| |
14:49 | dro: the user just tried from another thin client and is able to open gnome terminal without issue.
| |
14:51 | <dro> kusznir: sounds like some of your issues are from needing to update the chroot, and the other issue that user had might be from his client
| |
14:51 | <kusznir> So what is the procedure for updating my chroot?
| |
14:51 | <dro> kusznir: u running ubuntu?
| |
14:51 | <jhutchins_lt> kusznir: How's your swap activity on the server?
| |
14:51 | <kusznir> Yep.
| |
14:51 | <dro> kusznir: https://help.ubuntu.com/community/UbuntuLTSP/UpdatingChroot
| |
14:52 | <sbalneav> 9.04, you'll probably want to make sure stgrabers updated X library's installed.
| |
14:52 | libxcb, IIRC
| |
14:53 | Egyptian[Home] has quit IRC | |
14:57 | <johnny> no.. lts.conf is not necessary, i just think it's a god idea knipwim
| |
15:02 | hersonls has quit IRC | |
15:11 | shamino has quit IRC | |
15:15 | <dro> johnny: i agree
| |
15:16 | <johnny> actually.. dro i don't even care about lts.conf existing by default except for the required [default] header
| |
15:16 | if somebody knows how to modify our lexer file to remove the need for default.. that would be great..
| |
15:16 | ignore it if it exists
| |
15:17 | and pretend that data under no header is the global config
| |
15:17 | <dro> johnny: would probably increase the bootup time by a tiny tiny bit
| |
15:17 | <johnny> huh?
| |
15:17 | what would?
| |
15:18 | providing a default lts.conf ?
| |
15:18 | <dro> johnny: if there was an option to not try and load a lts.conf whether it was there or not, save 0.1ms lol
| |
15:19 | <johnny> no.. i was never suggesting that
| |
15:19 | if it's there.. it should be loaded
| |
15:19 | <sbalneav> Actually, providing an lts.conf with things like SERVER defined would greatly speed up an ltsp boot.
| |
15:19 | <johnny> sbalneav, then why doesn't ubuntu do it?
| |
15:19 | dro, my only suggestion was to make the lts.conf not require [default] header at all
| |
15:20 | <sbalneav> Because ubuntu wanted things to autoconfig, so that users wouldn't need to worry about lts.conf
| |
15:20 | i.e. they traded speed for ease of use.
| |
15:20 | <johnny> sbalneav, well i don't see why ltsp-build-client couldn't do it
| |
15:20 | and ltsp-update-image
| |
15:20 | <sbalneav> They certainly could.
| |
15:20 | <johnny> also.. using http would be nice :)
| |
15:20 | <sbalneav> All someone's gotta do is write the code.
| |
15:20 | <johnny> sbalneav, no.. that's the case at all
| |
15:20 | <sbalneav> Using http for what?
| |
15:21 | <johnny> first somebody has to decide whether they will accept it
| |
15:21 | then code can be written :)
| |
15:21 | i have 3 terminals.. i really don't care that much
| |
15:21 | using http instead of tftp for loading lts.conf
| |
15:22 | <sbalneav> Why would you do that? You already have to have a tftp server on the network somewhere.
| |
15:22 | frederickjh has quit IRC | |
15:23 | <johnny> to generate the lts.conf dynamically
| |
15:23 | could even tell the server some info about our clients that way
| |
15:23 | and set variables based on that
| |
15:23 | <sbalneav> Could do that easily with tftp
| |
15:24 | <johnny> how?
| |
15:24 | pretty sure tftp has nothing like cgi..
| |
15:24 | only upload/download
| |
15:24 | <sbalneav> write a fuse filesystem. Path elements to the file would autogenerate the lts.conf file
| |
15:24 | <johnny> uhmm..easier to use http
| |
15:24 | i know how to code http apps
| |
15:24 | i don't know how to code fuse fs
| |
15:25 | what you said doesn't sound easy
| |
15:25 | writing http apps is easy
| |
15:25 | shamino has joined #ltsp | |
15:25 | <sbalneav> ok, so, go for it :)
| |
15:26 | all you'd have to do us just get wget in the initramfs
| |
15:26 | <johnny> once again.. it's not worth writing if nobody uses it
| |
15:26 | i don't need it personally
| |
15:26 | so it would need to have guaranteed upstream acceptance
| |
15:26 | herytjb has joined #ltsp | |
15:26 | <johnny> probably need to discuss this with alkisg
| |
15:26 | <sbalneav> dude, if you WRITE it, it'll go in upstream. :)
| |
15:26 | <alkisg> johnny: hmmm what if we made a custom tftp server?
| |
15:26 | <johnny> to see if it will benefit anybody actually at all
| |
15:27 | alkisg, why?
| |
15:27 | <alkisg> I think we could code one in python
| |
15:27 | <johnny> you can code an http app in python
| |
15:27 | <sbalneav> Since when have we ever turned down a patch with added functionality? :)
| |
15:27 | <alkisg> We could make it read the mac address and stuff :)
| |
15:27 | <johnny> without running something like apache
| |
15:27 | alkisg, seriously.. just write http app
| |
15:27 | <alkisg> johnny: while keeping it COMPLIANT with all the ethernet cards out there...
| |
15:27 | <johnny> i am only speaking of the lts.conf stuff
| |
15:27 | <alkisg> johnny: how are you gonna boot the clients with http?
| |
15:27 | <johnny> not the initial kernel downloading
| |
15:28 | <sbalneav> alkisg: he's just talking about the grabbing lts.conf part
| |
15:28 | <alkisg> It would be better if we could send custom kernel parameters
| |
15:28 | Why limit it on lts.conf?
| |
15:28 | <johnny> perhaps an tftp + http..
| |
15:28 | alkisg, i think it would be best to run such a thing as a always running daemon tho.. not via inetd
| |
15:28 | since it would be best to right itin a scripting language
| |
15:28 | write it*
| |
15:29 | <alkisg> We'd need wget in the initramfs... would that be a problem?
| |
15:29 | (for the http lts.conf part)
| |
15:30 | Hmmm or we could also provide it with own out tftp, the protocol doesn't matter much...
| |
15:31 | E.g . instead of passing GET parameters, we could pass /path/parameters...
| |
15:31 | <vagrantc> just use ltsp_config.d to do whatever dynamic configuration you want.
| |
15:32 | * alkisg thinks that vagrantc, johnny and alkisg had this talk before :D | |
15:32 | <vagrantc> and that's why i committed the ltsp_config.d patch :)
| |
15:33 | <alkisg> How does that work, exactly?
| |
15:33 | We just put client scripts there?
| |
15:33 | <vagrantc> alkisg: it's just a run-parts hook to ltsp_config
| |
15:34 | alkisg: so it's sourced fairly early in ltsp_config ...
| |
15:34 | <alkisg> Hmmm yup that sounds cool
| |
15:34 | <vagrantc> alkisg: drop you script snippets in /opt/ltsp/i396/usr/share/ltsp/ltsp_config.d/* and do what ever you want.
| |
15:35 | though requires rebuilding the image to make changes, if you're *that* kind of distro, i guess. :)
| |
15:35 | <alkisg> :)
| |
15:36 | <vagrantc> could populate it from initramfs somehow, if you really wanted to get fancy.
| |
15:36 | there's so many options, really.
| |
15:37 | <sbalneav> what kind of things would you like to set? Any lts.conf variable?
| |
15:37 | <alkisg> I'd like to be able to generate an lts.conf on the fly, based on parameters that the client sends
| |
15:38 | <sbalneav> ok, and how do you send those parameters in the initramfs?
| |
15:38 | <johnny> wget sounded like a good idea
| |
15:38 | <alkisg> I think tftp is even cooler
| |
15:38 | <johnny> a stripped down version is available in busybox iir
| |
15:38 | iirc*
| |
15:38 | <sbalneav> right, but HOW DO YOU SEND THE PARAMETERS?
| |
15:38 | <johnny> well.. i geuess you could add a set command to go with options
| |
15:39 | what do you mean? wget can send parameters..
| |
15:39 | <vagrantc> sbalneav: via first class parcel post!
| |
15:39 | <alkisg> E.g. the client would run tftp server; bin; get my-mac-address/my-ip/my-whatever-info, and our tftp server would reply with an appropriate, dynamically generated file
| |
15:39 | <sbalneav> ok, and how is "my-whatever-info" specified in the initramfs?
| |
15:40 | <alkisg> for wget, we'd use: wget http://server?mac=XX&ip=YY&otherinfo=ZZZ
| |
15:40 | <sbalneav> how's it getting that info to the tftp/http server
| |
15:40 | so you JUST want to pass mac and IP?
| |
15:40 | that's easy, the initramfs can find out the mac and ip
| |
15:40 | <vagrantc> alkisg: that sounds like the random package i saw uploaded to debian the other day...
| |
15:41 | <sbalneav> I mean, are you going to set a dhcp variable that the initramfs is going to grab and pass that along, or what?
| |
15:41 | <vagrantc> http://packages.qa.debian.org/f/fts.html
| |
15:42 | <alkisg> vagrantc: gosa? that's something like ebox, isn't it?
| |
15:42 | <vagrantc> alkisg: i cannot make heads or tails of what the heck this thing is, to tell you the honest truth.
| |
15:43 | <alkisg> sbalneav: well, the client info would be: mac, ip, cpu, ram and other devices...
| |
15:43 | So the server would reply with an lts.conf specifically tailored for this client
| |
15:43 | E.g. "if RAM > 500" then fatclient = true...
| |
15:44 | getltscfg wouldn't be needed
| |
15:44 | (no parsing involved from the client side)
| |
15:44 | <vagrantc> trivial to do with ltsp_config.d snippets.
| |
15:44 | * vagrantc ends the marketing hype | |
15:45 | <sbalneav> ok, so, that's not hard. fuse filesystem.
| |
15:46 | <alkisg> Why would fuse be needed?
| |
15:46 | For http, a simple php script could do it,
| |
15:46 | <vagrantc> to get it to work with tftpd
| |
15:46 | <alkisg> I don't think it would be needed
| |
15:46 | We could handle it at the protocol level
| |
15:46 | <sbalneav> Load from /ltsconf/MA:CA:DD:RE:SS:00/ip.ad.dr.ss/RAM/CPU/VIDEOCARD
| |
15:46 | <alkisg> Not at the file system level
| |
15:47 | Network is involved, so we could handle it there
| |
15:47 | <sbalneav> calls a shell script, passes it mac, ip, ram, cpu, video
| |
15:47 | returns with a file
| |
15:47 | 200 lines of code.
| |
15:47 | no, you'd just use tftp
| |
15:48 | tftp would look for: /ltsconf/MA:CA:DD:RE:SS:00/ip.ad.dr.ss/RAM/CPU/VIDEOCARD/lts.conf
| |
15:48 | <alkisg> What would be easier to write? A slightly modified tftp server or a fuse filesystem?
| |
15:48 | <sbalneav> I know how to write fuse filesystems :)
| |
15:48 | <alkisg> E.g. tftpd-hpa has a "read file" function somewhere
| |
15:48 | If we just modified that to call a script instead, we'd be done...
| |
15:49 | <sbalneav> right, but you somehow have to also get it to handle the fake path
| |
15:49 | <alkisg> Isn't that all that's needed? A "read file" override?
| |
15:49 | (along with the path, in the same function...)
| |
15:49 | <sbalneav> the existing tftp server's going to try to cd into those dirs
| |
15:49 | <alkisg> Why?
| |
15:50 | <vagrantc> i must say, seems like running an http server is simpler
| |
15:50 | <sbalneav> because that's how tftpd-hpa works
| |
15:50 | <alkisg> It CDs to the dirs where the files to be "get" are? weird...
| |
15:50 | akuepker has quit IRC | |
15:51 | <sbalneav> it does a chdir
| |
15:52 | You'd still have to be modifying the initramfs
| |
15:52 | <alkisg> Isn't that for "-s"?
| |
15:52 | <sbalneav> because you can't make this scheme the default
| |
15:52 | <alkisg> Sure, we'd need to get something with tftp, but we wouldn't need any binaries, just the ltsp* scripts
| |
15:54 | <sbalneav> right, but the question becomes, how do you differentiate, in the chroot, that you're going to do a normal tftpget of lts.conf, versus you're going to look up the mac, ip, ram, etc, and to a modified tftpget to get different info
| |
15:54 | or, do a wget.
| |
15:54 | highvoltage has quit IRC | |
15:55 | <alkisg> One way is to use a special dir: <tftpboot>/specialdir/<everything here is a parameter>
| |
15:55 | <sbalneav> icky, but workable.
| |
15:57 | Well, I'll make the offer. if you guys want me to hack together a fuse filesystem this weekend that'll do that, I'll see what I can do. Can probably do it in python real easy-like
| |
15:57 | * alkisg guesses that we could even persuade Simon to provide us a hook in dnsmasq for that | |
15:57 | dro has quit IRC | |
15:57 | <sbalneav> so from the mountpoint: mac/ip/ram/cpu/video/lts.conf, seem ok for the ordering of the path elements?
| |
15:58 | <vagrantc> sbalneav: it seems too static...
| |
15:58 | <alkisg> Maybe with equals
| |
15:58 | <sbalneav> what else is the chroot going to know about?
| |
15:58 | <alkisg> mac=xxx/ip=xxx/ram=xxx
| |
15:58 | <vagrantc> sbalneav: i.e. what if you don't care about one of the elements in the middle?
| |
15:58 | <alkisg> This way it could easily be reversed
| |
15:58 | ip=xx/mac=yyy
| |
15:58 | <vagrantc> on some machines ... but you do on others...
| |
15:58 | <sbalneav> Tell you what.
| |
15:59 | I'm going to do what I first said.
| |
15:59 | You guys wanna widen the scope after that, you'll have the code :)
| |
16:00 | <vagrantc> i'm almost wondering if we shouldn't tweak getltscfg to support more configuration parameters...
| |
16:00 | right now it just supports configuration by mac, host name and ip address ...
| |
16:00 | * alkisg really think getlstcfg isn't needed | |
16:00 | <alkisg> At least on the client
| |
16:00 | <vagrantc> alkisg: well, there's that too ...
| |
16:01 | <alkisg> All the parsing could be done on the server, and only the client-specific values could be sent
| |
16:01 | <vagrantc> alkisg: but it's the infrastructure we already have ...
| |
16:01 | <sbalneav> Anywho, carry on, I'm off for the day. Be on later tonight.
| |
16:01 | <alkisg> vagrantc: it could stay compatible
| |
16:01 | Bye sbalneav
| |
16:01 | <vagrantc> it seems like generating an lts.conf though any number of means to work around limitations in getltscfg ... might not necessarily be the best way to go.
| |
16:01 | * vagrantc waves to sbalneav | |
16:02 | <alkisg> vagrantc: if it returns a [Default] with the client specific parameters in it, it would be compatible with the current code
| |
16:02 | <vagrantc> alkisg: sure ... but why not just make getltscfg do what we want?
| |
16:02 | <johnny> how about somebody drop getltscfg from needing [default] :_
| |
16:02 | <alkisg> One reason is because I don't want to send sensitive info to other clients
| |
16:03 | E.g. I don't want client02 to know the LDM_PASSWORD of client01
| |
16:03 | <vagrantc> fair enough.
| |
16:03 | <alkisg> johnny: I can send you the patch if you're willing to submit it :D
| |
16:04 | * alkisg doesn't want to deal with holy relics :D | |
16:04 | <vagrantc> alkisg: you just want getltscfg to work with an empty lts.conf ?
| |
16:05 | <johnny> i do..
| |
16:05 | no
| |
16:05 | <vagrantc> don't see anything wrong with including that
| |
16:05 | <johnny> i want it easier for people to just drop
| |
16:05 | LDM_DIRECTX=Y and be done
| |
16:05 | <ogra> stop dropping people !
| |
16:05 | <johnny> not having to remember to put [default] above it
| |
16:05 | <vagrantc> johnny: you have to get under their center of gravity... that's pretty much key ...
| |
16:05 | <johnny> hmm?
| |
16:06 | <alkisg> heh
| |
16:06 | <johnny> /me missed the joke :)
| |
16:06 | hehe
| |
16:06 | <vbundi> Hey so I'm having some trouble getting my VM Terminal to boot while there is another DHCP server (also serving images)
| |
16:06 | <alkisg> vagrantc: for the empty lts.conf I've sent a patch in the devs list... maybe 9 months ago :)
| |
16:06 | <vagrantc> alkisg: oh.
| |
16:06 | <ogra> *grin*
| |
16:06 | <vbundi> I've set my test DHCP server to be authoritative, and have added my VM MAC to the dhcp config file
| |
16:07 | <vagrantc> alkisg: apparently it wasn't nudged enough :)
| |
16:07 | <alkisg> It's just a line that needs to be changed - but noone's willing to commit it, it seems :D
| |
16:07 | <johnny> somebody should commit it..
| |
16:07 | <vbundi> when I boot the terminal, it talks to the right DHCP server, but once it starts loading it switches over and starts loading the image from the other server
| |
16:07 | <johnny> so it can get into lucid potentially
| |
16:07 | <vagrantc> yeah. someone. commit it!
| |
16:07 | <johnny> if it's not too late..
| |
16:07 | <alkisg> I don't think it is, 5.2 isn't out yet
| |
16:07 | <johnny> i'm too new to want to commit something that touches lts.conff :)
| |
16:08 | <vagrantc> pfft
| |
16:08 | CAN-o-SPAM has quit IRC | |
16:08 | * alkisg is too old to want to commit something that touches lts.conff :) | |
16:08 | <vagrantc> it could be put in the release notes for ltsp 5.2.x
| |
16:08 | RADICAL NEW CHANGES!
| |
16:09 | <alkisg> [Default] isn't needed! It also works with empty lts.conf! :)
| |
16:09 | <ogra> * we added an f to lts.conf
| |
16:09 | <vagrantc> ogra: yes!!!
| |
16:09 | <alkisg> ogra is on fire today :P
| |
16:09 | <ogra> :)
| |
16:09 | <vagrantc> deprecated lts.conf, replaced with ltsp.conf ... please update your code within the next decade before we drop support for lts.conf entirely.
| |
16:10 | and we mean it!
| |
16:10 | <ogra> hehe
| |
16:10 | <alkisg> Ooooh so you'll force people to give up on sarge?
| |
16:11 | (or woody?)
| |
16:11 | <vagrantc> alkisg: hey, look at this... it was almost a year ago that you submitted the patch :)
| |
16:12 | <alkisg> Hmmm maybe _I_ should be the one to commit it... "see people? that's what happens when you ignore patches, one becomes a developer 1 year later and he commits them himself" :P
| |
16:12 | highvoltage has joined #ltsp | |
16:13 | * vagrantc is testing it *right now* | |
16:13 | <vagrantc> alkisg: looks good to me!
| |
16:13 | alkisg: but really... you should do the honors.
| |
16:14 | <alkisg> uh, ok, but should I also remove the requirement for [default]?
| |
16:14 | Or is that too radical?
| |
16:15 | <vagrantc> alkisg: i'd say commit that two-line patch immediately. if you want to try for more than that, go for it. :)
| |
16:15 | <alkisg> OK. /me searches his emails...
| |
16:15 | <vagrantc> http://ltsp.pastebin.com/m1b8e7559
| |
16:16 | that one works for me...
| |
16:16 | <alkisg> got it
| |
16:17 | * vagrantc misses ltsppbot | |
16:22 | bulletxt has left #ltsp | |
16:23 | herytjb has quit IRC | |
16:33 | japerry has quit IRC | |
16:39 | bobby_C has quit IRC | |
16:45 | <alkisg> vbundi: to bypass your problem, you can put IPAPPEND=3 in pxelinux.cfg/default
| |
16:46 | vbundi: to properly resolve it, you need to ditch one of your dhcp servers...
| |
16:51 | mushroomtwo has quit IRC | |
17:00 | <vbundi> alkisg: alright, I've tried ipappend=1 and that worked... sort of I think...
| |
17:01 | alkisg: I'll have to read up on ipappend a bit, I don't know what the difference between them is
| |
17:10 | <alkisg> vbundi: this is a better way to solve it: https://help.ubuntu.com/community/UbuntuLTSP/ProxyDHCP
| |
17:10 | It's a way to keep using one of the dhcp servers only
| |
17:11 | Documentation for IPAPPEND is here: http://syslinux.zytor.com/wiki/index.php/SYSLINUX#IPAPPEND_flag_val_.5BPXELINUX_only.5D
| |
17:12 | <vbundi> yeah thats the documentation I found, I haven't fully read it yet
| |
17:12 | <alkisg> vbundi: why do you need 2 dhcp servers? What's your situation?
| |
17:12 | <vagrantc> vbundi's situation is more complicated... two LTSP servers.
| |
17:12 | <vbundi> alkisg: mostly because I'm testing at the moment I have an old LTSP4 server in production right now that I'm working around
| |
17:13 | I won't be running 2 ltsp servers forever
| |
17:13 | <alkisg> vbundi: and you want some clients, based on their mac addresses, to boot from the new ltsp server?
| |
17:14 | <vbundi> alkisg: yes
| |
17:14 | <alkisg> Then next-server in dhcpd.conf would be a good way to do it...
| |
17:15 | <vbundi> how would that work... all devices boot to server X, some devices get handed off for extra instructions (ie filename) to server Y?
| |
17:16 | <alkisg> No. All devices pxe boot from one of the ltsp servers (select whichever you want), and then they continue (from the kernel loading with tftp and on) with the server you select with next-server
| |
17:17 | So you'll be using 1 dhcp server, but that's all. Each client will use the ltsp server you specified.
| |
17:17 | <vbundi> ahh
| |
17:18 | <alkisg> dhcp server==only the initial IP. next-server == anything from that point on
| |
17:19 | So if you send a different next-server based on the clients mac addresses, you're done
| |
17:21 | <vbundi> I'm running HP T5700 terminals (eden 800mhz 512MB ram) and on LTSP they boot in 50 seconds, in LTSP 5 it's 4:45
| |
17:22 | LTSP4 = 50 seconds
| |
17:22 | 5 minute boot time can't be normal can it?
| |
17:23 | <alkisg> Nah
| |
17:23 | Try a boot graph
| |
17:23 | GGD has joined #ltsp | |
17:24 | Faithful has quit IRC | |
17:24 | <vbundi> apt-get install bootchart in my chroot?
| |
17:33 | alkisg has quit IRC | |
17:35 | alkisg has joined #ltsp | |
17:36 | alkisg has joined #ltsp | |
17:39 | GGD_ has joined #ltsp | |
17:39 | GGD has quit IRC | |
17:40 | alkisg has quit IRC | |
17:41 | GGD_ has left #ltsp | |
17:42 | GGD has joined #ltsp | |
17:43 | alkisg has joined #ltsp | |
18:05 | alkisg has quit IRC | |
18:06 | Faithful has joined #ltsp | |
18:12 | vvinet has quit IRC | |
18:38 | GGD has quit IRC | |
18:43 | GGD has joined #ltsp | |
18:44 | staffencasa has quit IRC | |
18:45 | vvinet has joined #ltsp | |
18:49 | lucascoala has joined #ltsp | |
18:51 | johnny has left #ltsp | |
19:17 | johnny has joined #ltsp | |
19:26 | pmatulis has quit IRC | |
19:27 | gentgeen__ has quit IRC | |
19:29 | lucascoala has quit IRC | |
19:51 | ogra has quit IRC | |
19:54 | GodFather has quit IRC | |
19:54 | try2free has joined #ltsp | |
19:56 | try2free has quit IRC | |
19:56 | pmatulis has joined #ltsp | |
20:04 | Ahmuck-Sr has joined #ltsp | |
20:08 | pmatulis has quit IRC | |
20:11 | sbalneavbb has joined #ltsp | |
20:12 | sbalneavbb has quit IRC | |
20:19 | Faithful has quit IRC | |
20:24 | johnny has left #ltsp | |
21:04 | johnny has joined #ltsp | |
21:13 | Faithful has joined #ltsp | |
21:31 | try2free has joined #ltsp | |
21:34 | try2free has left #ltsp | |
21:35 | mushroomblue has joined #ltsp | |
22:23 | Egyptian[Home] has joined #ltsp | |
22:47 | rhodan_ has joined #ltsp | |
22:48 | rhodan has quit IRC | |
22:56 | cyberorg has quit IRC | |
23:21 | try2free has joined #ltsp | |
23:56 | sutula has quit IRC | |
23:56 | jbrett has quit IRC | |