IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 7 June 2007 (all times are UTC)
| 00:31 | meetpai has quit IRC | |
| 00:34 | mistik1 has quit IRC | |
| 00:35 | cyberorg has joined #ltsp | |
| 00:37 | mistik1 has joined #ltsp | |
| 01:00 | Nikopoika has quit IRC | |
| 01:00 | Nikopoika has joined #ltsp | |
| 01:14 | <klausade> n #g
| |
| 01:35 | Burgundavia has quit IRC | |
| 01:57 | nf1 has joined #ltsp | |
| 02:33 | daya has joined #ltsp | |
| 02:42 | Q-FUNK has joined #ltsp | |
| 02:47 | vanya has joined #ltsp | |
| 03:01 | daya_ has joined #ltsp | |
| 03:22 | lambda_ has joined #ltsp | |
| 03:32 | subir has quit IRC | |
| 03:35 | vagrantc has joined #ltsp | |
| 03:38 | pere_away is now known as pere | |
| 03:47 | <daya_> does ltsp 4.2 or 4.1 fully supports sound and usb in debian etch,
| |
| 03:47 | ?
| |
| 03:48 | <vagrantc> hah
| |
| 03:48 | daya_: the server side will be no more easy with 4.x than with 5
| |
| 03:48 | <daya_> vagrantc, now I am thinking to migrate to 4.1 or 4.2
| |
| 03:49 | <vagrantc> you're going to have the same problems.
| |
| 03:49 | at least on the server side
| |
| 03:49 | <daya_> vagrantc, hey ha haha ha
| |
| 03:49 | vagrantc, :)
| |
| 03:49 | <vagrantc> the client side *might* work better.
| |
| 03:50 | muh2000 has quit IRC | |
| 03:50 | <vagrantc> daya_: if you want something that works out of the box, edubuntu is your best bet at this point.
| |
| 03:50 | muh2000 has joined #ltsp | |
| 03:51 | <vagrantc> daya_: or maybe debian-edu
| |
| 03:53 | <daya_> vagrantc, actually we are going to deploy our distros as I told u earlier, along with ltsp, so at this time I can't go to the alternatives,
| |
| 03:53 | muh2000 has quit IRC | |
| 03:54 | <vagrantc> daya_: sorry. i troubleshoot lots of problems and forget who does what
| |
| 03:55 | <daya_> vagrantc, http://nepalinux.org,
| |
| 03:55 | <vagrantc> daya_: in 3-4 days, i can prepare a backport of the newer ltsp for etch, and then you should be able to set it up sound the same way ubuntu does (pulseaudio + alsa)
| |
| 03:56 | <daya_> vagrantc, hey great,
| |
| 03:56 | <vagrantc> daya_: or, if the packages make it into sid, you can grab them from there.
| |
| 03:56 | <daya_> vagrantc, In next week I am going to deploy ltsp in one of the remote villages,
| |
| 03:57 | vagrantc, one questions, do I need only separate packages, or whole ltsp , :D
| |
| 03:57 | <vagrantc> ok. gotta go.
| |
| 03:57 | i'll be back later
| |
| 03:57 | vagrantc has quit IRC | |
| 03:58 | _ringlej has quit IRC | |
| 04:13 | kamalparyani has joined #ltsp | |
| 04:13 | <kamalparyani> hi
| |
| 04:14 | when i do sudo ltsp-build-client , can it pick the packages from the cdrom instead of downloading them
| |
| 04:17 | :)
| |
| 04:26 | Avatara has joined #ltsp | |
| 04:31 | Avatara has quit IRC | |
| 04:31 | Avatara has joined #ltsp | |
| 04:38 | daya_ has quit IRC | |
| 04:42 | <ogra> kamalparyani, mount the cdrom in /cdrom ... sudo ltsp-build-client --mirror file:///cdrom (yes, three slashes :) )
| |
| 04:43 | <kamalparyani> thanks
| |
| 04:45 | <ogra> if files are missing and it stops, add an --extra-mirror option with an online mirror to pull the missing stuff from ... see ltsp-build-client --extra-help
| |
| 04:47 | <kamalparyani> got this error W: Failure trying to run: chroot /opt/ltsp/i386 mount -t proc proc /proc
| |
| 04:47 | error: LTSP client installation ended abnormally
| |
| 04:47 | <ogra> wipe the old cruft before running that comannd
| |
| 04:48 | umount /opt/ltsp/i386/proc && rm -rf /opt/ltsp/i386
| |
| 04:50 | <kamalparyani> i did that
| |
| 04:50 | still the same error
| |
| 04:50 | Failure trying to run: chroot /opt/ltsp/i386 mount -t proc proc /proc
| |
| 04:53 | hi i tried the alternate install cd instead of the normal cd and it is working now
| |
| 04:53 | <lambda_> ltsp-build-client return E: Sub-process /usr/bin/dpkg returned an error code (1)
| |
| 04:53 | for xserver-xorg-input-wacom
| |
| 04:53 | xserver-xorg-input-all
| |
| 04:54 | <ogra> heh, right, sorry i could have told you that :)
| |
| 04:54 | the desktop CD has no packages but a livefs that gets installed
| |
| 04:55 | lambda_, if you scroll up to the download section in the output i bet you find a failed package download
| |
| 04:55 | <kamalparyani> thanks
| |
| 04:56 | shogunx has quit IRC | |
| 04:56 | kamalparyani has quit IRC | |
| 04:57 | <lambda_> I try that== ltsp-build-client --extra-mirror "http://pkg-ltsp.alioth.debian.org/debian etch-ltsp-backports main" --apt-key pkg-ltsp-keyring --late-packages ltspfsd
| |
| 04:58 | <ogra> you could probably add a --mirror option and point to a local mirror near you
| |
| 04:59 | <lambda_> the ltsp client build fail every time
| |
| 05:00 | <ogra> might be a broken debian package then, i'm on ubuntu so i cant tell ...
| |
| 05:00 | subir has joined #ltsp | |
| 05:01 | gu has quit IRC | |
| 05:08 | <lambda_> dpkg: dependency problems prevent configuration of xserver-xorg-input-all:
| |
| 05:08 | xserver-xorg-input-all depends on xserver-xorg-input-wacom; however:
| |
| 05:08 | Package xserver-xorg-input-wacom is not configured yet.
| |
| 05:09 | :(
| |
| 05:10 | plamengr has joined #ltsp | |
| 05:10 | plamengr has left #ltsp | |
| 05:10 | gonzaloaf has quit IRC | |
| 05:16 | gu has joined #ltsp | |
| 05:26 | subir has quit IRC | |
| 05:40 | tsurc has joined #ltsp | |
| 05:43 | otavio has quit IRC | |
| 05:51 | shogunx has joined #ltsp | |
| 06:22 | Q-FUNK has quit IRC | |
| 06:29 | cliebow has quit IRC | |
| 06:36 | meduxa has joined #ltsp | |
| 06:45 | Q-FUNK has joined #ltsp | |
| 06:49 | jammcq has quit IRC | |
| 07:13 | <lambda_> TFTP arp timeout
| |
| 07:13 | pfouu
| |
| 07:13 | vagrantc has joined #ltsp | |
| 07:18 | Gad1 has left #ltsp | |
| 07:34 | MoL0ToV has joined #ltsp | |
| 07:38 | jammcq has joined #ltsp | |
| 07:38 | <jammcq> g'morning kidz
| |
| 07:40 | <vagrantc> heya
| |
| 07:42 | cliebow has joined #ltsp | |
| 07:44 | Q-FUNK has quit IRC | |
| 07:54 | estong_ has joined #ltsp | |
| 07:56 | <estong_> after the ltsp build, i have these: Errors were encountered while processing:
| |
| 07:56 | xserver-xorg-input-wacom
| |
| 07:56 | xserver-xorg-input-all
| |
| 07:56 | <ogra> vagrantc, seems there is something broken in debian with the wacom driver, thats not te first complaint today
| |
| 07:56 | <estong_> i don't need wacom but i am not so sure on xserver-xorg-input-all
| |
| 07:57 | a5benwillis has joined #ltsp | |
| 07:58 | <estong_> hi ogra.
| |
| 07:58 | would it be ok to proceed to the next step -- configuration?
| |
| 07:59 | <ogra> jammcq, i'm pretty sure i can get HW like the T1220 to ~40 secod boots (yes, after yur measuring method) with the suqashfs/nbd stuff i just built my first initramfs and its fast like hell
| |
| 07:59 | <jammcq> ogra: ummm, you really like that nbd thing, eh?
| |
| 08:00 | <ogra> well, we use it anyway, and it operates on a lower level in the kernel than a filesystem
| |
| 08:00 | yes, i love it :)
| |
| 08:01 | <jammcq> lets just say that I'm skeptical, but watching with great interest
| |
| 08:01 | <ogra> way more important is that we can get rid of all the bind mounting of tmpfses
| |
| 08:01 | <jammcq> have you tried the e2300 yet?
| |
| 08:01 | <ogra> through simply doing a union mount of the nbd device and a tmpfs
| |
| 08:02 | that will gain a lot since we dont have to ar up and move around directories like we do now
| |
| 08:02 | *tar
| |
| 08:02 | jammcq, yes but not measured
| |
| 08:02 | the 2300 is nothin to use with the current ldm
| |
| 08:02 | no matter what you have below
| |
| 08:03 | <jammcq> my guess is that the e2300 won't handle the squashfs very well
| |
| 08:03 | <ogra> oh, thats no prob
| |
| 08:03 | <jammcq> but i'd love to be proven wrong on that
| |
| 08:03 | <ogra> its also way faster
| |
| 08:03 | than nfs
| |
| 08:03 | but the ldm still takes a minute or so to show the login screen
| |
| 08:03 | i'm hoping for sbalenavs C wrapper here
| |
| 08:03 | <jammcq> yeah, me too
| |
| 08:04 | I'm getting lots of questions about the speed
| |
| 08:04 | <ogra> since the slowest part seems the wrapper
| |
| 08:04 | <jammcq> people wanting to know if it's being fixed
| |
| 08:04 | <ogra> with the C greeter its already 1/3 faster here
| |
| 08:04 | it will be fixed in gutsy
| |
| 08:04 | i cant promise for the e2300, but for real HW it will be
| |
| 08:04 | <jammcq> but what do you think people should deploy today?
| |
| 08:04 | <ogra> ltsp5 indeed
| |
| 08:05 | lambda_ has quit IRC | |
| 08:05 | <jammcq> e2300 MUST be faster for gutsy, I'll have 1,000 of those units on June 18th
| |
| 08:05 | <ogra> did you *ever* use it ?
| |
| 08:05 | it is not usable for me !
| |
| 08:05 | <jammcq> people WILL NOT deploy ltsp-5 today, in it's current form
| |
| 08:05 | <ogra> even on 4.2
| |
| 08:05 | <jammcq> it works fine here
| |
| 08:05 | <ogra> i can watch the cursor being redrawn on both ltsp's
| |
| 08:06 | nly if i force it to something like 1204x786 it gets roughly usable
| |
| 08:06 | <jammcq> I used it quite a bit here on ltsp-4.2, with no problem
| |
| 08:06 | <ogra> *1024
| |
| 08:06 | <jammcq> well yeah, at 1024x768
| |
| 08:06 | that's the monitor I have
| |
| 08:06 | <ogra> well
| |
| 08:06 | <jammcq> well
| |
| 08:07 | <ogra> you should have at least ordered the 300Mhz variant ... :/
| |
| 08:07 | jsgotangco has joined #ltsp | |
| 08:08 | <jammcq> they don't have a 300Mhz in that footprint
| |
| 08:08 | ph call, be back in a bit
| |
| 08:08 | <ogra> did i say that the shielding isnt usable in european environments as well ?
| |
| 08:08 | i got heavy diagonal stripes ... 50Hz ...
| |
| 08:08 | on the display
| |
| 08:08 | kaos01 has quit IRC | |
| 08:13 | otavio has joined #ltsp | |
| 08:16 | <estong_> does "ltsp-build-client" create /tftpboot ?
| |
| 08:16 | <ogra> o
| |
| 08:16 | no
| |
| 08:16 | thats from ltsp 4.2
| |
| 08:16 | <vagrantc> estong_: is this on etch or lenny or sid ?
| |
| 08:17 | <estong_> etch
| |
| 08:17 | <vagrantc> hm.
| |
| 08:17 | ogra has quit IRC | |
| 08:17 | <vagrantc> as far as i know, there's been no updates to etch that would affect it, and it's always "worked for me".
| |
| 08:18 | <estong_> the wacom thing? or the ltsp5 install.
| |
| 08:18 | <vagrantc> estong_: was this a wacom problem during an ltsp5 install?
| |
| 08:18 | <estong_> i don't need the wacom, but yes, there was an installation error because of it.
| |
| 08:19 | <vagrantc> right.
| |
| 08:24 | <estong_> and so, the /tftpboot is now gone for ltsp5. The sample dhcpd.conf has /ltsp/i386 and this folder is not there yet in my filesystem.
| |
| 08:25 | when does this dir created?
| |
| 08:25 | ogra has joined #ltsp | |
| 08:25 | <vagrantc> estong_: /var/lib/tftpboot/
| |
| 08:26 | estong_: but if any step in the ltsp-build-client fails, the install will not be functional.
| |
| 08:27 | well, most likely won't be functional.
| |
| 08:27 | estong_: so you need to fix that first.
| |
| 08:27 | <estong_> i git errors on the wacom and the input-all.
| |
| 08:27 | i got
| |
| 08:27 | <vagrantc> estong_: did you get any more specific information about why it fails?
| |
| 08:27 | estong_: you didn't run out of disk space or something?
| |
| 08:28 | <estong_> Setting up xserver-xorg-input-wacom (0.7.4.1-5) ...
| |
| 08:28 | ln: creating symbolic link `/etc/udev/rules.d/z60_xserver-xorg-input-wacom.rules' to `../xserver-xorg-input-wacom.rules': File exists
| |
| 08:28 | dpkg: error processing xserver-xorg-input-wacom (--configure):
| |
| 08:28 | subprocess post-installation script returned error exit status 1
| |
| 08:28 | dpkg: dependency problems prevent configuration of xserver-xorg-input-all:
| |
| 08:28 | xserver-xorg-input-all depends on xserver-xorg-input-wacom; however:
| |
| 08:28 | Package xserver-xorg-input-wacom is not configured yet.
| |
| 08:28 | dpkg: error processing xserver-xorg-input-all (--configure):
| |
| 08:28 | dependency problems - leaving unconfigured
| |
| 08:28 | <vagrantc> !pastebot
| |
| 08:28 | <ltspbot`> vagrantc: "pastebot" is The LTSP pastebot is at http://pastebot.ltsp.org. Please paste all text longer than a line or two to the pastebot, as it helps to reduce traffic in the channel. A link to the content will be pasted in the channel.
| |
| 08:28 | <vagrantc> estong_: please use the pastebot in the future
| |
| 08:28 | <estong_> ops!
| |
| 08:28 | sori
| |
| 08:29 | i hope u can get something from that.
| |
| 08:29 | <vagrantc> estong_: did you re-install over an existing install?
| |
| 08:30 | ogra: we should have never removed the "error out if directory exists" code.
| |
| 08:30 | <estong_> i retried the ltsp build about 3 times because of disconnection to the ltsp server.
| |
| 08:30 | <vagrantc> ogra: since it's in a plugin, it's easy to work around for those who know what they're doing.
| |
| 08:31 | estong_: ok, so ... mv /opt/ltsp/i386 /opt/ltsp/i386.old
| |
| 08:31 | estong_: and start over.
| |
| 08:31 | <estong_> will i have to go through that long download process again?
| |
| 08:31 | <ogra> vagrantc, its in feisty ...
| |
| 08:32 | and i kept the warning in edgy iirc
| |
| 08:32 | jsgotangco has quit IRC | |
| 08:32 | <vagrantc> ogra: yeah, i know.
| |
| 08:32 | ogra: otavio wanted to remove it... and i didn't scream loud enough
| |
| 08:32 | estong_: yes.
| |
| 08:33 | estong_: if you keep getting timeouts, you might want to try a different mirror
| |
| 08:33 | <ogra> vagrantc, well, it will be in sid again
| |
| 08:34 | <vagrantc> ogra: yes.
| |
| 08:34 | someday ...
| |
| 08:34 | * vagrantc waits impatiently | |
| 08:34 | <estong_> looks like the 2 errors are fatal enough to warrant another 1 hour of downloading. :(
| |
| 08:34 | lambda_ has joined #ltsp | |
| 08:34 | <lambda_> hi
| |
| 08:35 | <estong_> but what can i do, i am just a mere mortal.
| |
| 08:35 | <vagrantc> estong_: ok ... give me a second. i know how to work around the problem.
| |
| 08:36 | estong_: ok ... so cd /opt/ltsp/i386
| |
| 08:36 | estong_: as root
| |
| 08:36 | <estong_> done
| |
| 08:36 | <vagrantc> estong_: and be absolutely sure you in /opt/ltsp/i386
| |
| 08:37 | <estong_> yup.
| |
| 08:37 | <vagrantc> estong_: and quadrupal-triple check that you're in /opt/ltsp/i386
| |
| 08:37 | <lambda_> I am a little worry
| |
| 08:37 | <estong_> he he he
| |
| 08:37 | <vagrantc> estong_: rm -rf [b-u]* var/[bl-t]* var/cache/[dlf]* var/cache/apt/*.bin
| |
| 08:37 | <jammcq> be afraid, be very afraid
| |
| 08:37 | <lambda_> I don't understand shared-network THINCLIENTS { ...... declaration
| |
| 08:37 | <estong_> so that's why.
| |
| 08:37 | i hit enter and i closed my eyes
| |
| 08:37 | done
| |
| 08:38 | <vagrantc> estong_: then you can try again and most of the packages will already be downloaded, but nothing is installed
| |
| 08:38 | superkaybee has joined #ltsp | |
| 08:38 | <estong_> oh!
| |
| 08:38 | <lambda_> it assume there's a 192.168.0.0/255.255.255.0 network
| |
| 08:38 | <estong_> ltsp-build-client
| |
| 08:38 | right?
| |
| 08:38 | <lambda_> in my dhcpd.conf
| |
| 08:38 | <vagrantc> estong_: unless your downloads are corrupt
| |
| 08:38 | estong_: yes
| |
| 08:38 | <estong_> ok. tnx
| |
| 08:38 | <jammcq> vagrantc: won't ltsp-build-client refuse to run if there's already a /opt/ltsp/i386 ?
| |
| 08:38 | <vagrantc> jammcq: not in etch.
| |
| 08:39 | <jammcq> oooh, good to know
| |
| 08:39 | <lambda_> there's NOT such physical network on my server
| |
| 08:39 | <rjune> howdy jim
| |
| 08:39 | <jammcq> rjune: hey
| |
| 08:39 | lambda_: what distro are you running?
| |
| 08:39 | <vagrantc> jammcq: when the next upload hits sid, it will, though.
| |
| 08:39 | <lambda_> debian etch
| |
| 08:39 | <rjune> got an lbe question if you have a few
| |
| 08:39 | <lambda_> just eth1 Lien encap:Ethernet HWaddr 00:17:31:63:26:6A
| |
| 08:39 | inet adr:192.168.2.27 Bcast:192.168.2.255 Masque:255.255.255.0
| |
| 08:39 | <jammcq> lambda_: where did you get that dhcpd.conf file from?
| |
| 08:39 | rjune: shoot
| |
| 08:39 | <vagrantc> lambda_: is it debian-edu, or just plain etch ?
| |
| 08:40 | <lambda_> debian-edu
| |
| 08:40 | <rjune> in theory, I should be able to update glibc and gcc in lbe-src without too much trouble, yes?
| |
| 08:40 | <lambda_> other hardware ethernet device: eth0 Lien encap:Ethernet HWaddr 00:0E:2E:AB:84:E1
| |
| 08:40 | inet adr:10.0.2.2 Bcast:10.0.3.255 Masque:255.255.254.0
| |
| 08:40 | <vagrantc> lambda_: have you asked in the debian-edu channel ? they have very specific networking setup.
| |
| 08:40 | <jammcq> rjune: ummm, that's a tall order. glibc and gcc are VERY fussy
| |
| 08:40 | daya has quit IRC | |
| 08:41 | <rjune> That's what I wanted to check. would it be better to start at crosscomp-src and update there then cascade down? or ltsp-src and build back?
| |
| 08:41 | <jammcq> umm, yes
| |
| 08:41 | you'll prolly have to try both ways
| |
| 08:42 | <superkaybee> does anyone know why a diskless terminal might come up as root's desktop when logging in as a regular user? (ie: :0 on ltsp server)
| |
| 08:46 | <ogra> woah
| |
| 08:47 | with the /cow tmpfs its even 1/3 faster
| |
| 08:47 | <vagrantc> ogra: how much memory does the /cow eat up once it's booted?
| |
| 08:48 | <ogra> as much as is needed for writeable files
| |
| 08:48 | <jammcq> ogra: keep it up, and someday, it might even be almost as fast as ltsp-4.2 :)
| |
| 08:48 | <vagrantc> ogra: well, i know that :P
| |
| 08:48 | <ogra> i.e. exactly the amount we used with the tar and gzip methids for bind mounts
| |
| 08:49 | but its a lot faster due to the in kernel decompression vs tar and gzip for the rw files/dirs
| |
| 08:49 | <vagrantc> ogra: are you saying that because it's theoretically what it will take, or have you actually looked at how much it uses? :P
| |
| 08:49 | <ogra> jammcq, i think i can pass 4.2 with scotties ldm
| |
| 08:49 | <jammcq> that'll be cool
| |
| 08:50 | <ogra> vagrantc, i'm still playing around give me some toying time before i come up with something :)
| |
| 08:50 | <vagrantc> ogra: i'm just curious if more stuff gets cached into ram and such
| |
| 08:50 | unexpected things
| |
| 08:50 | <ogra> strangely it doesnt accep my password
| |
| 08:50 | so something i still wrong
| |
| 08:50 | <vagrantc> that's what happened when i used translucencyfs with lessdisks years ago ...
| |
| 08:51 | <ogra> my rootpw for the chroot i mean
| |
| 08:51 | <vagrantc> more stuff got pulled into ram that didn't happen when the entire filesystem was read-write
| |
| 08:51 | er, read-only
| |
| 08:51 | <ogra> i dont think ram is an issue
| |
| 08:51 | the only issues i saw were compression related
| |
| 08:51 | but apparently the squashfs stuff is way faster here
| |
| 08:52 | <vagrantc> did you try nbd without squashfs ?
| |
| 08:52 | * ogra switches to XDMCP and goes to try the e2300 | |
| 08:53 | meduxa has quit IRC | |
| 08:54 | <ogra> the annoing part is to run mksqashfs after making changes :/
| |
| 08:54 | vagrantc, not yet
| |
| 08:55 | i'll try all variants my fanatsy can come up with in all combos i can imagine :)
| |
| 08:55 | <vagrantc> estong_: you might also want to run ltsp-build-client with --keep-packages
| |
| 08:55 | <ogra> the t1220 is surely below a minute even with the old ldm now
| |
| 08:55 | <estong_> but the process has started already.
| |
| 08:56 | i am on retrieving libc6 now. :)
| |
| 08:56 | <vagrantc> estong_: oh well. hopefully it'll just work this time.
| |
| 08:56 | C1ph3r has joined #ltsp | |
| 08:56 | <estong_> we can learn from this.
| |
| 08:56 | i'm sure of that.
| |
| 08:57 | * ogra laughs about nbd-client saying "que ?" if the server dies | |
| 08:58 | <ogra> vagrantc, the whole thig uses 54M with running X
| |
| 08:58 | thats similar to the old variant
| |
| 08:58 | * vagrantc still wonders how big the /cow is | |
| 08:59 | <rjune> ogra: does ubuntu include any sort of groupware software?
| |
| 08:59 | something ala phpgroupware, or citadel, etc?
| |
| 08:59 | <ogra> rjune, well, there is a lot stuff in universe
| |
| 08:59 | ie.e kolab, phpgroupware, hula etc etc
| |
| 08:59 | <rjune> kolab? interesting
| |
| 08:59 | <ogra> but i dont think we have something in main
| |
| 09:00 | <rjune> I'm wanting to setup a primary LDAP datastore for user/group information, then have everything auth to it.
| |
| 09:00 | <ogra> vagrantc, well, i could tell you if i handt forgotten to move mount /cow in initramfs, i can only see the unionfs :)
| |
| 09:01 | the general used tmpfs is 46M
| |
| 09:01 | <vagrantc> ogra: you can't even see the raw mount? crazy.
| |
| 09:01 | <ogra> no if it was mounted in initramfs
| |
| 09:02 | df complains that it cant find /cow
| |
| 09:02 | * ogra fixes the initramfs | |
| 09:03 | <estong_> btw, what's the minimum client config to run openoffice writer comfortably?
| |
| 09:03 | <rjune> 1. a metric boatload of ram
| |
| 09:04 | 2. a P3-1Ghz
| |
| 09:04 | <estong_> how about 128MB?
| |
| 09:04 | <ogra> the client only needs to run X and the login
| |
| 09:04 | be concerned about server RAM ;)
| |
| 09:04 | <rjune> I don't think that qualifies. OO.o really likes the ram, I would say 256MB or more for a single user.
| |
| 09:04 | wols_ has joined #ltsp | |
| 09:04 | <rjune> doh! ogra is right of course, I forgot this was #ltsp
| |
| 09:05 | <estong_> yes
| |
| 09:05 | so what's the minimum for X then.
| |
| 09:05 | <vagrantc> estong_: 64-128 should be plenty
| |
| 09:05 | <estong_> wow!
| |
| 09:06 | <wols_> hi. got a question where I don't even know ltsp is the right answer. but I ask it here anyways:
| |
| 09:06 | <estong_> and the video card?
| |
| 09:06 | <vagrantc> estong_: on the clients
| |
| 09:06 | <estong_> yes
| |
| 09:06 | <vagrantc> estong_: 4-8MB of video ram should be plenty
| |
| 09:06 | for most purposes
| |
| 09:06 | <wols_> I have a PC with 2 monitors (single videocard), 2 keyboards, 2 mice. is LTSP the leanest way to make two people wrok at that PC?
| |
| 09:06 | <estong_> finally the processor?
| |
| 09:06 | <vagrantc> !serversize
| |
| 09:06 | <ltspbot`> vagrantc: "serversize" is http://wiki.ltsp.org/twiki/bin/view/Ltsp/ServerSizing
| |
| 09:06 | <vagrantc> estong_: check that out
| |
| 09:07 | <rjune> Once upon a time there was a php script
| |
| 09:08 | wols_: nope
| |
| 09:08 | ltsp isn't appropriate at all
| |
| 09:08 | <estong_> am just wondering if a 233MHz client could do openoffice.
| |
| 09:08 | <vagrantc> estong_: depends on the server more than the client
| |
| 09:08 | <ogra> estong_, as i said, worry about the server
| |
| 09:08 | the client only displays :)
| |
| 09:09 | <wols_> rjune: can you tell me what is then please?
| |
| 09:09 | <estong_> wud this be the case still if i have a 486 machine for a client?
| |
| 09:09 | <vagrantc> estong_: though 233MHz is a little on the slow side.
| |
| 09:09 | MoL0ToV has quit IRC | |
| 09:09 | <rjune> wols_: I wish I could remember the name of it. LTSP makes terminals out of systems, you want to share a single system with two local consoles
| |
| 09:10 | <ogra> 400 upwards should be fine
| |
| 09:10 | <vagrantc> estong_: you'd probably need to run the sessions unencrypted for machines that slow.
| |
| 09:10 | * vagrantc kicks an e2300 | |
| 09:10 | <rjune> ogra: but RAM on the client is important for the number of windows being displayed
| |
| 09:10 | <estong_> ok. i'll remember all that.
| |
| 09:10 | <ogra> rjune, indeed
| |
| 09:10 | <wols_> rjune: yes that
| |
| 09:10 | * rjune is installing ubuntu-server now. | |
| 09:11 | <wols_> rjune: yes that's it
| |
| 09:12 | petre has joined #ltsp | |
| 09:12 | <ogra> vagrantc, /cow has 2.6M
| |
| 09:12 | <petre> morning all
| |
| 09:13 | <ogra> s/has/uses
| |
| 09:13 | <vagrantc> ogra: that seems considerably more than before.
| |
| 09:13 | <ogra> and 43M available from the tmpfs
| |
| 09:13 | <vagrantc> ogra: like, usually, it's only 80k or so
| |
| 09:13 | <ogra> yeah
| |
| 09:13 | <vagrantc> with the tmpfs+bind mounts
| |
| 09:13 | so i was *right* !
| |
| 09:13 | <ogra> syslog created a bunch of files
| |
| 09:13 | * vagrantc sticks out tounge | |
| 09:14 | <ogra> how did you measure the size before ?
| |
| 09:14 | <vagrantc> du -smc /var/lib/ltsp-client-setup
| |
| 09:14 | <ogra> we never had a central point like /cow
| |
| 09:14 | <vagrantc> or something like that
| |
| 09:14 | <ogra> ah, indeed we had one
| |
| 09:14 | <vagrantc> the bind mouts just make it look like lots
| |
| 09:14 | <ogra> even though thats only the tarred up stuff
| |
| 09:14 | <vagrantc> but it's really one filesystem
| |
| 09:14 | <ogra> not the directly created tmpfs thigs
| |
| 09:15 | <vagrantc> it's everything
| |
| 09:15 | <ogra> but that might be 10k or so
| |
| 09:15 | nope there are sone files that get created iirc
| |
| 09:15 | <vagrantc> well, i think the /var/run /var/lock stuff is separate, since that's handled by something else
| |
| 09:15 | <ogra> right
| |
| 09:16 | <estong_> vagrantc, the rm trick did not do it. It's re-downloading the whole thing again. May http.us.debian.org forgive us. :)
| |
| 09:16 | <ogra> hmm
| |
| 09:16 | it copies stuff from /lib
| |
| 09:16 | * ogra looks what that is | |
| 09:16 | <vagrantc> estong_: well, maybe the files weren't there, or were corrupted. in which case it would download them again.
| |
| 09:16 | superkaybee has left #ltsp | |
| 09:16 | <vagrantc> ogra: /lib/modules ?
| |
| 09:17 | <ogra> vagrantc, 2.4M are modules
| |
| 09:17 | yeah
| |
| 09:17 | <vagrantc> hah!
| |
| 09:17 | * vagrantc has been through this before | |
| 09:17 | <vagrantc> it's like a hazy dream come to life
| |
| 09:17 | <ogra> gah, its the set of initramfs mods
| |
| 09:17 | all NIC drivers
| |
| 09:18 | well, actually its all modules that are loaded
| |
| 09:18 | makes sense
| |
| 09:18 | <vagrantc> how's that ?
| |
| 09:19 | <ogra> well, likely copied over from initramfs
| |
| 09:19 | i dont really mind
| |
| 09:19 | at least atm
| |
| 09:19 | wols_ has left #ltsp | |
| 09:20 | <ogra> thats another area where we can speed up later if we want
| |
| 09:20 | the output of the mount command alone is worth to change the implementation :P
| |
| 09:21 | it fits on *one* screen
| |
| 09:22 | <vagrantc> heh
| |
| 09:23 | <ogra> man we should really drop syslog
| |
| 09:23 | why is it in the chroot at all ?
| |
| 09:23 | <vagrantc> to enable remote syslogging ?
| |
| 09:24 | <ogra> hmm, the e2300 doesnt start X now :/
| |
| 09:24 | vagrantc, doesnt logger suffice for that ?
| |
| 09:24 | <vagrantc> ogra: never tried
| |
| 09:24 | <ogra> i dont see a reason why tons of files get created o boot in /var/log
| |
| 09:24 | <vagrantc> what's in /etc/syslog.conf ?
| |
| 09:24 | <ogra> on the e2300 its a significant slowdow
| |
| 09:25 | lambda_ has quit IRC | |
| 09:25 | <ogra> oh, crap the defaults
| |
| 09:25 | indeed, thast a mess
| |
| 09:25 | <vagrantc> it's probably just some stupid bug from the changes i did to syslog handling
| |
| 09:26 | * ogra makes note to wipe that file from ltsp-build-client and replace it with something empty | |
| 09:26 | <vagrantc> maybe a bunk merge
| |
| 09:26 | <ogra> well, it just needs to be empty and get filled fom ltsp-client-setup
| |
| 09:26 | <vagrantc> if [ -z "$SYSLOG" ] || [ "$SYSLOG" = "remote" ]; then
| |
| 09:27 | you set SYSLOG=server
| |
| 09:27 | <ogra> yes
| |
| 09:27 | oh
| |
| 09:27 | when did you add that ?
| |
| 09:27 | * ogra fixes | |
| 09:27 | <vagrantc> probably the last merge
| |
| 09:27 | ogra: how are you fixing it?
| |
| 09:28 | <ogra> changing server to remote
| |
| 09:28 | <vagrantc> i discussed it on the list
| |
| 09:28 | <ogra> i should probably just drp it
| |
| 09:28 | from the default filr
| |
| 09:28 | file
| |
| 09:28 | <vagrantc> that's what i'd do.
| |
| 09:29 | <ogra> yeah
| |
| 09:29 | sigh
| |
| 09:29 | * ogra watches mksquashfs | |
| 09:30 | * vagrantc gets ready to head to a country where vagrantc doesn't understand the language at all | |
| 09:31 | * vagrantc prepares a few "bon jour"s and "merci boucoup"s | |
| 09:31 | <ogra> lol
| |
| 09:32 | nowadays they speak a bt eglish at least, thanks to the internet
| |
| 09:32 | C1ph3r has quit IRC | |
| 09:32 | <vagrantc> internet?
| |
| 09:32 | * ogra remembers when he was there 85 ... | |
| 09:32 | <ogra> nobody spoke english
| |
| 09:36 | jammcq, 63sec for the e2300 to console prompt, 110sec to X with XDMCP
| |
| 09:36 | does that sound good for you ?
| |
| 09:37 | * ogra wonders if jammcq fell off his chair | |
| 09:38 | vagrantc has quit IRC | |
| 09:41 | nf1 has quit IRC | |
| 09:44 | <estong_> some install warnings/errors:
| |
| 09:44 | Setting up xfonts-100dpi (1.0.0-3) ...
| |
| 09:44 | warning: /usr/lib/X11/fonts/100dpi does not exist or is not a directory
| |
| 09:44 | <ogra> thats fine
| |
| 09:44 | i just says that i didnt exist before
| |
| 09:44 | <estong_> i c
| |
| 09:45 | <ogra> you can ignore all the warnings ... there are surely some more
| |
| 09:45 | as long as they are not errors all is fine ;)
| |
| 09:45 | <estong_> there are a number of those -- all font related.
| |
| 09:45 | ok. how about this one:
| |
| 09:45 | Setting up ttf-dejavu (2.15-1) ...
| |
| 09:45 | Fontconfig error: Cannot load default config file
| |
| 09:46 | <ogra> fine as well
| |
| 09:47 | <estong_> ok. that seems to be the only error. the rest are warnings.
| |
| 09:47 | puetz has joined #ltsp | |
| 09:47 | <estong_> looks like vagrantc fixed that wacom issue.
| |
| 09:48 | finally!
| |
| 09:48 | info: LTSP client installation completed successfully
| |
| 09:49 | <jammcq> ogra: see, now you are getting somewhere
| |
| 09:49 | <ogra> :)
| |
| 09:49 | jammcq, no, you
| |
| 09:49 | <jammcq> ogra: you da man
| |
| 09:49 | <ogra> *i* dont care about the e2300 .... i dont stave if i cant sell it :P
| |
| 09:49 | <jammcq> you'll care, when people start saying nasty things about your favorite distro
| |
| 09:49 | <ogra> with scotties ldm it should be around 2min boots
| |
| 09:50 | right, indeed, i wasnt serious ;)
| |
| 09:50 | <jammcq> :)
| |
| 09:50 | I know, you care
| |
| 09:50 | <petre> "stave"?
| |
| 09:50 | <ogra> i'm not sure yet we want that squashfs thing though
| |
| 09:50 | <jammcq> oh?
| |
| 09:50 | <ogra> *starve indeed
| |
| 09:50 | <jammcq> finding some badness ?
| |
| 09:50 | <ogra> no
| |
| 09:50 | <petre> ah
| |
| 09:51 | <ogra> but you need to run mksquashfs afetr every change
| |
| 09:51 | it squashes up the chroot which takes about 3-5min
| |
| 09:51 | that even applies to lts.conf changes
| |
| 09:51 | <jammcq> ogra: I don't really consider that a problem
| |
| 09:51 | once setup, people don't make that many changes
| |
| 09:51 | <ogra> well, more documentation
| |
| 09:52 | and ltsp-manager wouldnt work with it
| |
| 09:52 | well, it would ... hmm
| |
| 09:52 | i could leave the chroot around and store te image inside it ...
| |
| 09:52 | and ave a script that automatically generates the squashfs
| |
| 09:53 | so after chrooting for changes that happens automatically
| |
| 09:53 | hmm, that could work
| |
| 09:54 | jammcq, i don think thats for gutsy as a default, but for gutsy+1 ... for gutsy i'd say lets keep it as --experimental option for ltsp-build-client so we get more testing
| |
| 09:54 | <jammcq> see... you've talked yourself into it working :)
| |
| 09:55 | <ogra> its a very big change and i'd like to find all drawbacks before such an intrusive change is made
| |
| 09:55 | cool is that you dont need nfs ...
| |
| 09:55 | neither root-path :)
| |
| 09:56 | and we could even have the squashfs encrypted easily for the paranoid people :)
| |
| 09:56 | jammcq, another thing is was thinking about is a ramfs mode for nbdswapd ...
| |
| 09:57 | <jammcq> what does that mean?
| |
| 09:57 | <ogra> that way we could actually use the swap instead f having it only as a parachute
| |
| 09:57 | well, create the swapfiles in a tmpfs on the server
| |
| 09:57 | you will still have network latency but swapping might actually become usable
| |
| 09:57 | * jammcq wonders why you'd want to take precious ram away from the server | |
| 09:58 | <ogra> well, indeed it would be optional
| |
| 09:58 | but if we use nbd from initramfs anyway, we could already swap from there
| |
| 09:58 | you could run 16M clients ;)
| |
| 09:58 | <jammcq> i'd love to see a test where you prove whether it's any better. I'd say let Linux caching deal with the swapfiles coming from disk or ram
| |
| 09:59 | the bottleneck isn't gonna be reading the swapfiles from disk. it'll be transfering over the network
| |
| 09:59 | <ogra> thats soethig for another day ... today is my / fixing day ;)
| |
| 09:59 | yeah, that might be
| |
| 09:59 | i didnt measure yet
| |
| 09:59 | <jammcq> ph call, bbiab
| |
| 09:59 | <ogra> and to be hinest i dodnt think i ever managed to get one of my clients to swap :P
| |
| 10:00 | *honest
| |
| 10:00 | <estong_> is it ok to use 172.29.0.0 instead of 192.168.0.0 for my subnet?
| |
| 10:02 | <ogra> sure
| |
| 10:02 | <estong_> is this recommended --> deb http://pkg-ltsp.alioth.debian.org/debian etch-ltsp-backports main
| |
| 10:02 | <ogra> just make sure to change t everywhere
| |
| 10:02 | no idea, vagrant would know ...
| |
| 10:02 | vanya has quit IRC | |
| 10:02 | <ogra> i can only recommend ubuntu stuff :)
| |
| 10:02 | <estong_> i won't upgrade for now then.
| |
| 10:03 | <puetz> hi. I want to start firefox as a local app on the client in 4.2. Does anybody know of a recent built of e.g. firefox 2.x? I tried to set up lbe but that didn't work
| |
| 10:05 | I cvs checked out head and LTSP-4_2_update_2, both bail out "16:41:00 Building: glibc *** Unable to configure glibc: Inappropriate ioctl for device"
| |
| 10:05 | <estong_> i'm done with the configs. is there an ssh way of checking that everything is well? am not in the computer lab right now and nobody is in there at the moment.
| |
| 10:05 | <puetz> this is an athlon64 with 32bit libs, I researched _a_lot_ but I can't find the error. host gcc is 3.3
| |
| 10:06 | system is ubuntu feisty. I *think* I installed all the devel stuff
| |
| 10:08 | tsurc has quit IRC | |
| 10:09 | <estong_> thanks ogra. i hope i won't be back tomorrow. :)
| |
| 10:10 | <puetz> the error that glibc threw was "error: cannot compute sizeof (unsigned long), 77". I googled that, there it seemed that 64bit/32bit issues were to blame but all I looked at was using 32 bit only
| |
| 10:12 | estong_ has quit IRC | |
| 10:15 | lambda has joined #ltsp | |
| 10:23 | yopla has joined #ltsp | |
| 10:39 | kchiefs has joined #ltsp | |
| 10:40 | <kchiefs> i have a question for anyone
| |
| 10:41 | we have purchased the 1220PXE term from diskless and it will not load the image at all...it just slowing goes through and errors..the older terms work fine..anyone know of this problem?
| |
| 10:41 | it is SLAX 11 with LTSP 4.2
| |
| 10:41 | lambda has quit IRC | |
| 10:42 | <kchiefs> anybody?
| |
| 10:43 | <yopla> what are errors?
| |
| 10:45 | lambda_ has joined #ltsp | |
| 10:45 | puetz has quit IRC | |
| 10:48 | <kchiefs> yopla it goes through to "trying to load pxelinux.conf/default" and sits there for about 10 mins then just says can't load image and goes to boot: in the prompt
| |
| 10:53 | <yopla> if you have the correct vmlinuz and initrd in the default file then it should be a pxelinux compatibility problem
| |
| 10:54 | in the tftpboot tree
| |
| 10:56 | What is your tftp server?
| |
| 10:56 | Egyptian[Home] has joined #ltsp | |
| 11:02 | <yopla> here is my question : Can we use rdesktop in ltsp5?
| |
| 11:04 | <cyberorg> yopla: yes
| |
| 11:06 | <yopla> I hven't found lts parameter
| |
| 11:06 | is it still in lts.conf?
| |
| 11:11 | Gadi has joined #ltsp | |
| 11:15 | <kchiefs> yopla: you want to know what tftpd i use?
| |
| 11:15 | Egyptian[Home] has quit IRC | |
| 11:15 | muh2000 has joined #ltsp | |
| 11:16 | <cyberorg> yopla: i am not sure how it is used, try manually launching it once client logs in
| |
| 11:17 | <kchiefs> yopla: in.tftpd
| |
| 11:17 | Egyptian[Home] has joined #ltsp | |
| 11:19 | <yopla> cyberorg: OK. But I need some more stuff on ltsp server.
| |
| 11:21 | <ogra> jammcq, i'm running a T1220 without swap and with mem=32M ... free shows 28M total ram :) isnt squashfs great ?
| |
| 11:21 | (XDMCP indeed)
| |
| 11:22 | <Gadi> ogra: is that over nbd?
| |
| 11:22 | <ogra> yep
| |
| 11:23 | <Gadi> do you know how well that scales?
| |
| 11:23 | <ogra> unionfs with tmpfs/squashfs over nbd
| |
| 11:23 | better than nfs i think
| |
| 11:23 | the squashfs image you mount is only 120M
| |
| 11:23 | vs the nfsroot which is 400M
| |
| 11:23 | <Gadi> really? I had considered it at one point for a project, but was concerned about the 'here be dragons' warning on their webpage
| |
| 11:24 | <ogra> the in-kernel compression is a lot faster than the bind mount function using tar and gzip we had for the RW stuff
| |
| 11:24 | <Gadi> how fast does your image boot?
| |
| 11:24 | <ogra> Gadi, te e2300 that boots 4 min or so in feisty is up in less than two with it
| |
| 11:25 | <Gadi> nice
| |
| 11:25 | thats pretty incredible
| |
| 11:25 | <ogra> (both XDMCP values, ldm isnt bearable in its current state on that thing)
| |
| 11:25 | <Gadi> right
| |
| 11:25 | <ogra> i'm counting on sbalneav for the ldm stuff
| |
| 11:25 | <Gadi> but even with all of the dpkg stuff?
| |
| 11:25 | wow
| |
| 11:25 | <ogra> gah, indeed after login the mouse freezes without swap
| |
| 11:26 | (on the 32M thing here)
| |
| 11:26 | Avatara has quit IRC | |
| 11:26 | <Gadi> it onlyu has 32M ram?
| |
| 11:26 | <yopla> cyberorg: my client runs ldm. my server has no dm. What does the client display with no lts.conf?
| |
| 11:26 | <ogra> but its quite impressive that it comes up with 28M
| |
| 11:26 | * Gadi didnt know you could buy anything like that anymore | |
| 11:26 | <ogra> well, i set mem=32M in the pxe config
| |
| 11:26 | <rjune> Gadi: !
| |
| 11:26 | <Gadi> cyberorg: it run /etc/X11/Xsession
| |
| 11:26 | <ogra> that way the kernel only sees 32
| |
| 11:27 | <Gadi> rjune!!
| |
| 11:27 | <cyberorg> yopla: what Gadi ^^ said :)
| |
| 11:27 | <yopla> lol
| |
| 11:27 | <Gadi> oops, sorry
| |
| 11:27 | :)
| |
| 11:27 | <cyberorg> yopla: ldm doesnt need DM on server
| |
| 11:27 | <Gadi> ogra: is there a limit to the number of nbd connections?
| |
| 11:28 | <ogra> Gadi, i dont think so, but there is a default set
| |
| 11:28 | <petre> Gadi: !!!!!!!!!!!!!!!!!!!
| |
| 11:28 | <ogra> even though
| |
| 11:28 | <Gadi> what is the default?
| |
| 11:28 | <kchiefs> anyone use the 1220 term from disklessworkstations on their ltsp network?
| |
| 11:28 | <ogra> irt seems to depend how you start the nbd server
| |
| 11:28 | i would start it fron an ltsp-server initscript so we're free to use what we want ;)
| |
| 11:29 | <Gadi> ogra: I guess one down side is updating the chroot while clients are on
| |
| 11:29 | <ogra> hmm, actually there is no default config
| |
| 11:29 | yeah
| |
| 11:29 | <Gadi> petre: sent you an email
| |
| 11:29 | <ogra> well, that will also depend on how nbd-server handles HUP
| |
| 11:29 | <yopla> cyberorg: what do I need to connect then? my session is rejected
| |
| 11:30 | * petre goes to read email | |
| 11:30 | <ogra> if its proper it will just serve the refreshed file after a HUP signal
| |
| 11:30 | <Gadi> that would be cool
| |
| 11:30 | <ogra> if not i'd consider it a bug
| |
| 11:30 | petre has quit IRC | |
| 11:30 | <cyberorg> yopla: you want to use rdesktop instead of ldm?
| |
| 11:30 | <rjune> how goes it Gadi?
| |
| 11:31 | <ogra> the annoyig part is the time mksquashfs takes after you changed something
| |
| 11:31 | <Gadi> rjune: busy busy
| |
| 11:31 | petre has joined #ltsp | |
| 11:31 | <Gadi> ogra: heh, yeah
| |
| 11:31 | <ogra> even lts.conf changes need a rebuild
| |
| 11:31 | petre has quit IRC | |
| 11:31 | <Gadi> well, there's a good reason to make lts.conf downloadable ;)
| |
| 11:31 | <yopla> cyberorg: I just want to open a session with my first working lts5 client
| |
| 11:32 | <kchiefs> anyone here running ltsp on slax?
| |
| 11:32 | <cyberorg> you want to login to the client's desktop from another machine?
| |
| 11:33 | <Gadi> yopla: what distro?
| |
| 11:33 | <yopla> gadi: my server distro is debian sid. My client tree is feisty
| |
| 11:33 | <ogra> Gadi, well, if we would have ldap by default that would be trivial
| |
| 11:33 | <Gadi> yopla: apt-get install ltspfs
| |
| 11:34 | pscheie has joined #ltsp | |
| 11:34 | <Gadi> on the server
| |
| 11:34 | ogra: trivial in some ways, yes
| |
| 11:35 | ogra: I still vote for tftp
| |
| 11:35 | its easy and we get it for free
| |
| 11:35 | <ogra> right
| |
| 11:35 | <Gadi> and can easily be used even in the initramfs
| |
| 11:36 | <ogra> well in the initramfs we have / :)
| |
| 11:36 | we can read it from the filesystem
| |
| 11:36 | <Gadi> only if we keep it in the filesystem
| |
| 11:36 | <yopla> gadi: can't acces /dev/fuse
| |
| 11:37 | dpkg-reconfigure problem
| |
| 11:37 | <ogra> are yu in the fuse group ?
| |
| 11:37 | <yopla> I'm root
| |
| 11:37 | <Gadi> yopla: you mean after installing ltspfs?
| |
| 11:37 | <ogra> udev is responsible for /dev/fuse ...
| |
| 11:38 | <Gadi> ls /dev/fuse
| |
| 11:38 | if it does not exist, modprobe fuse
| |
| 11:38 | <ogra> debian has a very weird post installation script in the fuse package tough ...
| |
| 11:38 | <Gadi> then, reinstall ltspfs
| |
| 11:38 | <ogra> i dropped that for ubuntu and made something saner
| |
| 11:38 | <yopla> lsmod : fuse 49072 0
| |
| 11:38 | <ogra> in debian you also need to add it to /etc/modules iirc
| |
| 11:39 | else it wont persist on reboot
| |
| 11:39 | <yopla> but still can't have /dev/fuse
| |
| 11:39 | <ogra> weird
| |
| 11:39 | <yopla> amd64 problem?
| |
| 11:39 | muh2000 has quit IRC | |
| 11:40 | <yopla> my test server is a xen3.1 pv machine....
| |
| 11:40 | <ogra> look in /etc/udev/rules.d/
| |
| 11:41 | there should be a rules file for fuse
| |
| 11:41 | that usually creates the device automatically if the module gets loaded
| |
| 11:46 | <yopla> orga: lol udev wasn't installed
| |
| 11:48 | now i have ltspfs
| |
| 11:49 | what next?
| |
| 11:49 | <ogra> jammcq, WOW !
| |
| 11:49 | jammcq, nbd swap with swapfiles in tmpfs is really great
| |
| 11:50 | i have a tmpfs based swapfile on a 32M boot ... firefow with 20 tabs and differnt huge sites, gimp, f-spot and rhythmbox running ...
| |
| 11:50 | 18M of the swapspace are used
| |
| 11:50 | <jammcq> ogra: and you are saying it's better than using nbd swap to a file?
| |
| 11:50 | <ogra> i dont see any slowdown in usage ... no mouse blocking or any weird stuff
| |
| 11:50 | yes, a lot
| |
| 11:51 | * jammcq believes more testing is in order | |
| 11:51 | <ogra> i can actually work even half of the swap s used
| |
| 11:51 | its really snappy
| |
| 11:51 | <jammcq> cuz it contradicts what I would consider "common sense". anyway, cool, and I'm anxious to see it in action
| |
| 11:51 | but first..... lunch time
| |
| 11:51 | <ogra> :)
| |
| 11:52 | btw i added the swap stuff to initramfs, else it wouldnt boot with 32M :)
| |
| 11:56 | <Gadi> ogra: I'd be interested in seeing that patch, if you have it
| |
| 11:56 | lambda_ has quit IRC | |
| 11:57 | <Gadi> is it *cough* in bzr?
| |
| 11:59 | yopla has quit IRC | |
| 12:00 | lambda_ has joined #ltsp | |
| 12:04 | <ogra> Gadi, nah its not something to unleash to the world yet ... i'm just playing around
| |
| 12:05 | http://paste.ubuntu-nl.org/24608/
| |
| 12:06 | http://paste.ubuntu-nl.org/24610/
| |
| 12:06 | and then you need to add squashfs and unionfs to auto_add_modules' base list in /opt/ltsp/$ROOT/usr/share/initramfs-tools/hook-functions
| |
| 12:07 | rebuild the initramfs, run ltsp-updae-kernels and boot a client
| |
| 12:07 | <Gadi> or add it to the modules file
| |
| 12:07 | <ogra> note that the speedups will rather come through the stuff we can drop through this :)
| |
| 12:08 | i.e. disable anything related to bind_mounts in ltsp-client-setup
| |
| 12:08 | <Gadi> indeed
| |
| 12:08 | <ogra> also note that i'm already on gutsy
| |
| 12:08 | <Gadi> well, aren't *you* gutsy ;)
| |
| 12:08 | <ogra> its significantly faster since i dont have X detection anymore
| |
| 12:09 | and the kernel "feels" snappier
| |
| 12:09 | <Gadi> ah, that is faster
| |
| 12:13 | cyberorg has left #ltsp | |
| 12:16 | <rjune> Gadi: you using ubuntu too?
| |
| 12:19 | <kchiefs> i am having problems getting the image with 2 terminals. a third works fine..the only difference is one is using realtek ethernet and it works the other 2 are using intel and they don't work. anyone heard of this?
| |
| 12:20 | my specs are Slackware 11.0 and ltsp 4.2
| |
| 12:21 | <Gadi> rjune: I've been known to ubuntu
| |
| 12:22 | <kchiefs> correction...the other 2 are using VIA not intel
| |
| 12:24 | <ogra> hmm
| |
| 12:25 | <kchiefs> maybe i should just try to make ltsp5 work with slack
| |
| 12:25 | it is a weird problem that I have not run into on any other distro
| |
| 12:26 | <ogra> Gadi, if /cow wouldnt be on a tmpfs but a local disk we could have fat clients with that ... only the delta to the readonly fs will be stored locally
| |
| 12:26 | <Gadi> yuck
| |
| 12:26 | sepski has joined #ltsp | |
| 12:27 | irule has joined #ltsp | |
| 12:27 | <ogra> oooh that opens so many opportunities :)
| |
| 12:28 | so sad that the work for migrqating from one implementation to another would make more work than actuallz implementing it
| |
| 12:29 | gah, i hate broken kezmaps
| |
| 12:32 | lambda_ has quit IRC | |
| 12:34 | Egyptian[Home1 has joined #ltsp | |
| 12:35 | Egyptian[Home] has quit IRC | |
| 12:41 | lambda_ has joined #ltsp | |
| 12:47 | pere is now known as pere_away | |
| 13:08 | kchiefs has quit IRC | |
| 13:09 | kaminski-ltsp-br has quit IRC | |
| 13:15 | Zenton_ has quit IRC | |
| 13:21 | <lambda_> pfou
| |
| 13:21 | Jun 7 20:29:24 tjener atftpd[16939]: Serving ltsp/i386/pxelinux.0 to 10.0.2.2:32984
| |
| 13:21 | tjener atftpd[16939]: recvmsg: Connection refused
| |
| 13:21 | tjener atftpd[16939]: tftpd_file.c: 926: recvfrom: Connection refused
| |
| 13:22 | my atftpd server seems quite bugged :( :(
| |
| 13:25 | <sepski> lambda_, have you tampered with hosts.allow or hosts.deny ?
| |
| 13:25 | <lambda_> I have checked these files
| |
| 13:25 | /etc/hosts.deny don't own any restriction
| |
| 13:28 | hosts.allow has in.tftpd: 10.0.2.200 (the diskless workstation)line
| |
| 13:33 | privet has quit IRC | |
| 13:36 | bill_c has joined #ltsp | |
| 13:36 | irule has quit IRC | |
| 13:39 | gigabytes has joined #ltsp | |
| 13:56 | privet has joined #ltsp | |
| 13:57 | <klausade> lambda_: how much of the config files on tjener have you changed?
| |
| 13:59 | <lambda_> I don't renember
| |
| 13:59 | remember
| |
| 14:00 | gloin has joined #ltsp | |
| 14:01 | <lambda_> klausade, can you set me on the solving way PLEASE :)
| |
| 14:01 | <gloin> hi all
| |
| 14:01 | <klausade> lambda_: it's suposed to work out-of-the-box, and it does so on all my tens of servers. if you don't know which files you have changed, we can't easily help you.
| |
| 14:02 | <lambda_> I have edited /Etc/hosts.allow dhcp.conf
| |
| 14:02 | <gloin> been a while, anyhow, new network, new setup. Ubuntu server, using ltsp-server (not -standalone), our network has a win2k server for dhcp. I've set 066 boot server host name to the ip of the ltsp server, but pxe booting doesn't work. Says 'tftp timeout'
| |
| 14:02 | lsof -i shows tftp listening over UDP on the server
| |
| 14:02 | <klausade> lambda_: can you start all over? And join us on debian-edu?
| |
| 14:03 | <gloin> (also set 067 bootfile name)
| |
| 14:03 | and root path
| |
| 14:03 | <klausade> gloin: there is an excelent guid on how to use windows dhcp-server for this at http://diet-pc.sourceforge.net/windows/etherboot-w2k.html
| |
| 14:04 | gloin: I actually used it with success last week.
| |
| 14:04 | * gloin checks it out | |
| 14:05 | <klausade> gloin: one thing that gave me a few minutes of grief was that I forgott about /etc/exports
| |
| 14:06 | <gloin> heh
| |
| 14:06 | got that
| |
| 14:06 | hmm
| |
| 14:06 | a lot more options required than I thought
| |
| 14:07 | <klausade> gloin: I got away with option 17,66 and 67
| |
| 14:07 | <gloin> doesn't much help that the reservation options window has a dinky little pane that only shows 1.5 list options
| |
| 14:07 | klausade: see, that's the thing - I have all those options correctly set
| |
| 14:08 | "PXE-E32: TFTP Open timeout" on the client
| |
| 14:09 | <klausade> gloin: are you realy sure that the pxe-stuff in the nic does work? Older cards , afaik 0.99 pxe version does not always work.
| |
| 14:10 | <gloin> it says intel pxe-2.1
| |
| 14:10 | build 082 though...
| |
| 14:11 | <klausade> gloin: so, the nic should work.
| |
| 14:11 | gloin: what about syslog on the server?
| |
| 14:11 | <gloin> nothing showing up there
| |
| 14:12 | J45p3r_ has joined #ltsp | |
| 14:14 | gigabytes has quit IRC | |
| 14:14 | nf1 has joined #ltsp | |
| 14:15 | <gloin> well, I can't help but think the client isn't getting to the server
| |
| 14:15 | but I have no friggin idea why
| |
| 14:15 | <bill_c> hud and wireshark can help shed some light on that
| |
| 14:16 | hub even
| |
| 14:16 | <gloin> ergh
| |
| 14:16 | a hub, huh?
| |
| 14:16 | I'd have to buy one
| |
| 14:16 | two actually, and have someone a couple hundred miles from me where the server is do the same thing
| |
| 14:17 | shouldn't tftp be listening on TCP as well as UDP?
| |
| 14:19 | PMantis has joined #LTSP | |
| 14:20 | <bill_c> can't sniff out traffic on a switch, unless the sniffer is running on one of the boxes the traffic is destined for
| |
| 14:20 | <gloin> aye
| |
| 14:37 | lambda_ has quit IRC | |
| 14:44 | <PMantis> jammcq: Ping
| |
| 14:48 | Gadi has left #ltsp | |
| 14:49 | pscheie has quit IRC | |
| 14:53 | <gloin> hmm, interesting
| |
| 14:53 | it's the tftp server
| |
| 14:53 | can't get anything from it
| |
| 14:54 | trying to do stuff manually and it times out
| |
| 15:04 | damn
| |
| 15:04 | something's truly hosed with the tftp server afaik
| |
| 15:06 | * gloin wonders if ubuntu has atftp in universe or somewhere | |
| 15:10 | <gloin> argh
| |
| 15:11 | rjune has quit IRC | |
| 15:27 | ltspbot has joined #ltsp | |
| 15:34 | Q-FUNK has joined #ltsp | |
| 15:35 | ltspbot` has quit IRC | |
| 15:53 | Q-FUNK has quit IRC | |
| 16:18 | Egyptian[Home1 is now known as Egyptian[Home] | |
| 16:21 | J45p3r_ has quit IRC | |
| 16:22 | bobby_C has quit IRC | |
| 16:22 | Burgundavia has joined #ltsp | |
| 16:30 | sepski has quit IRC | |
| 16:31 | Burgundavia has quit IRC | |
| 16:31 | oh207 has quit IRC | |
| 16:31 | oh207 has joined #LTSP | |
| 16:37 | Burgundavia has joined #ltsp | |
| 16:47 | Q-FUNK has joined #ltsp | |
| 17:02 | jammcq has quit IRC | |
| 17:07 | bill_c has quit IRC | |
| 17:09 | vmlintu has joined #ltsp | |
| 17:18 | Egyptian[Home] has quit IRC | |
| 17:26 | privet has quit IRC | |
| 17:30 | PMantis has quit IRC | |
| 17:54 | <cliebow> ogra:now you have Gadi saying "Indeed!"
| |
| 18:13 | <gloin> hmm
| |
| 18:13 | how, in muecow, can I disable X11 on the clients? There's no lts.conf anymore...
| |
| 18:14 | * gloin has an ubuntu box working now, but doesn't need/want X11 on the terminals | |
| 18:20 | privet has joined #ltsp | |
| 18:36 | <moquist> gloin: you can create an lts.conf
| |
| 18:37 | <gloin> moquist: oh? Might be worth doing that ;) After I reinstall ltsp :|
| |
| 18:37 | <moquist> oy!
| |
| 18:37 | good luck. :)
| |
| 18:37 | <gloin> ubuntu makes it easy enough
| |
| 18:42 | rm -rf /opt/ltsp and run ltsp-build-client
| |
| 18:42 | times have certainly changed
| |
| 18:52 | privet has quit IRC | |
| 18:53 | privet has joined #ltsp | |
| 18:58 | bill_c has joined #ltsp | |
| 19:04 | <whiprush> moquist: holy crap my email reply was way longer than I expected.
| |
| 19:04 | gonzaloaf has joined #ltsp | |
| 19:05 | <cliebow> heh
| |
| 19:11 | * moquist prepares to be shocked and amazed at the length of whiprush's...email reply. | |
| 19:16 | bill_c has quit IRC | |
| 19:20 | J45p3r has joined #ltsp | |
| 19:25 | gepatino has joined #ltsp | |
| 19:34 | artista_frustrad has joined #ltsp | |
| 19:37 | gepatino has quit IRC | |
| 19:50 | Burgundavia has quit IRC | |
| 19:51 | jammcq has joined #ltsp | |
| 19:52 | Q-FUNK has quit IRC | |
| 20:07 | J45p3r has quit IRC | |
| 20:34 | sahil has joined #ltsp | |
| 20:36 | sethbrown has joined #ltsp | |
| 20:36 | sethbrown has left #ltsp | |
| 20:40 | Burgundavia has joined #ltsp | |
| 21:25 | ogra has quit IRC | |
| 21:26 | ogra has joined #ltsp | |
| 21:48 | <sahil> there are both sis drm and sis video kernel modules both have the name sis how would i specificy which to use?
| |
| 21:49 | can i give the path in the MODULES_0# parameter?
| |
| 21:52 | estong_ has joined #ltsp | |
| 22:03 | yanu_ has joined #ltsp | |
| 22:11 | estong_ has quit IRC | |
| 22:17 | yanu has quit IRC | |
| 22:33 | otavio has quit IRC | |
| 22:40 | muh2000 has joined #ltsp | |
| 22:46 | sahil has quit IRC | |
| 23:49 | estong_ has joined #ltsp | |
| 23:58 | <estong_> am getting an nfs-premount, mount failed error on the client. The server syslog shows DHCP DISCOVER/OFFER/REQUEST/PACK in quick succession when the client boots.
| |
| 23:59 | client fails to complete the boot process.
| |