IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 25 January 2011 (all times are UTC)
| 00:03 | MorningSon has quit IRC | |
| 00:58 | alkisg has joined #ltsp | |
| 02:08 | gnunux has joined #ltsp | |
| 02:08 | <gnunux> hi
| |
| 02:39 | mistik1 has quit IRC | |
| 02:46 | mistik1 has joined #ltsp | |
| 02:48 | dobber has joined #ltsp | |
| 02:53 | sep has joined #ltsp | |
| 03:09 | bobby_C has joined #ltsp | |
| 03:17 | toscalix has joined #ltsp | |
| 03:41 | dede_py has joined #ltsp | |
| 04:01 | dede_py has quit IRC | |
| 04:33 | Kicer86 has joined #ltsp | |
| 04:33 | Da-Geek has joined #ltsp | |
| 04:38 | dlezcano has quit IRC | |
| 04:58 | mistik1 has quit IRC | |
| 05:08 | alkisg has quit IRC | |
| 05:08 | toscalix has quit IRC | |
| 05:09 | toscalix has joined #ltsp | |
| 05:28 | bobby_C has quit IRC | |
| 06:16 | [GuS] has joined #ltsp | |
| 06:16 | [GuS] has joined #ltsp | |
| 06:35 | mistik1 has joined #ltsp | |
| 06:48 | MorningSon has joined #ltsp | |
| 06:50 | tech_dvo has joined #ltsp | |
| 06:51 | <tech_dvo> anybody knows how to trouble shoot -- "not enough memory to load image" error?
| |
| 06:56 | Da-Geek has quit IRC | |
| 06:57 | Da-Geek has joined #ltsp | |
| 06:58 | alkisg has joined #ltsp | |
| 07:10 | <Appiah> tech_dvo: when does this error show up?
| |
| 07:11 | <tech_dvo> during initram...wait
| |
| 07:13 | Loading initrd.img
| |
| 07:13 | Not enough memory to load specified image
| |
| 07:14 | <Appiah> at PXE boot...
| |
| 07:17 | tech_dvo: what client(s) gets this error?
| |
| 07:18 | <tech_dvo> celeron mainboard computers with 256mb ram
| |
| 07:19 | <Appiah> I relate this to a kernel issue... (think I've seen it before and has nothing to do with LTSP)
| |
| 07:19 | what dist are you using?
| |
| 07:19 | <tech_dvo> ubuntu 10.10
| |
| 07:20 | does that mean -- my thin client can't boot with ubunt 10.10?
| |
| 07:21 | or do I need to build a specific "kernel" for this kind of thin clients --w/c is more work as I have yet to experience building kernel :)
| |
| 07:21 | <muppis> No, I got same kind of client which works fine with 10.10.
| |
| 07:21 | <Appiah> did you change anything in the chroot?
| |
| 07:21 | <muppis> Nope.
| |
| 07:21 | <Appiah> asking tech_dvo :)
| |
| 07:22 | <muppis> :P
| |
| 07:22 | <mgariepy> tech_dvo, did you ran a memtest to make sure your memory is fine ?
| |
| 07:22 | <tech_dvo> am not yet very familiar with what I did with chroot but I may have touch chroot
| |
| 07:22 | <muppis> And I didn't read the nick.
| |
| 07:22 | <Appiah> can you try to rebuild it ? ltsp-build-image
| |
| 07:22 | <alkisg> Are you using etherboot or gpxe, and what version?
| |
| 07:22 | <tech_dvo> did not run memtest -- can i do that over ltsp?
| |
| 07:23 | alkisg: I am using the program you gave me last week --
| |
| 07:23 | <alkisg> Which program is that?
| |
| 07:24 | <tech_dvo> it runs fine on types of thin clients already: ltsp-loader_0.6.21.exe
| |
| 07:24 | <alkisg> Ah, so gpxe
| |
| 07:24 | <tech_dvo> two types
| |
| 07:25 | <alkisg> Maybe it's a problem with pxelinux, try loading the kernel directly
| |
| 07:25 | Press ctrl+b to gpxe to get a prompt
| |
| 07:25 | Run: dhcp net0
| |
| 07:25 | <tech_dvo> okay -- will do that
| |
| 07:25 | <Appiah> http://syslinux.zytor.com/archives/2008-July/010462.html
| |
| 07:26 | <alkisg> dhcp net0
| |
| 07:26 | # Put the LTSP server IP in the following line
| |
| 07:26 | set next-server 10.160.31.10
| |
| 07:26 | kernel tftp://${next-server}/ltsp/i386/vmlinuz ro initrd=initrd.img nbdroot=${next-server}:2000 nbd_proxy=false
| |
| 07:26 | initrd tftp://${next-server}/ltsp/i386/initrd.img
| |
| 07:26 | boot vmlinuz
| |
| 07:26 | With those commands you boot the kernel directly, bypassing pxelinux.0
| |
| 07:27 | So if it's a pxelinux bug, you can either search for a newer/patched version, or bypass it completely.
| |
| 07:27 | evil_root is now known as zz_evil_root | |
| 07:30 | <tech_dvo> ok
| |
| 07:34 | [GuS] has quit IRC | |
| 07:41 | <tech_dvo> alkisg: I get connection time out in kernel tftp:..... part
| |
| 07:43 | <alkisg> tech_dvo: did you put the correct IP there?
| |
| 07:43 | <tech_dvo> yup 10.160.31.10
| |
| 07:43 | did it twice
| |
| 07:43 | same error
| |
| 07:43 | <alkisg> That's my ip
| |
| 07:43 | Not yours
| |
| 07:43 | What's your server ip?
| |
| 07:43 | <tech_dvo> i see -- 192.168.1.200
| |
| 07:43 | <alkisg> Right, put that one
| |
| 07:47 | <tech_dvo> initrd=initrd.img
| |
| 07:47 | command not found
| |
| 07:47 | i erase the = sign
| |
| 07:48 | i get the error initrd.img File not found (0x2d12603b)
| |
| 07:48 | <alkisg> I didn't write such a command
| |
| 07:48 | I wrote this:
| |
| 07:48 | kernel tftp://${next-server}/ltsp/i386/vmlinuz ro initrd=initrd.img nbdroot=${next-server}:2000 nbd_proxy=false
| |
| 07:48 | That's ALL IN ONE LINE
| |
| 07:48 | <tech_dvo> i see -- its just one command ok
| |
| 07:51 | jd42 has joined #ltsp | |
| 07:51 | dlezcano has joined #ltsp | |
| 07:58 | <tech_dvo> alkisg: after boot vmlinuz ...... I get error
| |
| 07:59 | Begin:Running /scripts/nfs-bottom... chroot: can't execute '/usr/bin/test': Exec format error
| |
| 07:59 | alkisg has quit IRC | |
| 07:59 | <tech_dvo> chroot: can't execute '/usr/bin/getltscfg':Exec format error...
| |
| 08:12 | zz_evil_root is now known as evil_root | |
| 08:12 | wwx has quit IRC | |
| 08:15 | wwx has joined #ltsp | |
| 08:21 | Gadi has joined #ltsp | |
| 08:23 | NeonLicht has quit IRC | |
| 08:24 | NeonLicht has joined #ltsp | |
| 08:29 | tech_dvo has quit IRC | |
| 08:43 | bobby_C has joined #ltsp | |
| 08:53 | GodFather has joined #ltsp | |
| 08:56 | garymc has joined #ltsp | |
| 08:57 | <garymc> hi guys anyone know of an app for my macbook pro that I can load and I can be on my ltsp server aswell as my mac?
| |
| 08:57 | so i dont have to turn off the mac and pxe boot?
| |
| 09:00 | <evil_root> you could ssh to your server and then your on your ltsp
| |
| 09:00 | but you want a virtual thin client on your macbook?
| |
| 09:04 | garymc?
| |
| 09:04 | <garymc> sorry im here
| |
| 09:04 | yes virtual client
| |
| 09:05 | <evil_root> i am 99% sure vmware fusion allows network boot
| |
| 09:05 | PXE boot
| |
| 09:09 | <garymc> is that while I run all my other programs
| |
| 09:11 | <evil_root> you should be able to yes
| |
| 09:19 | GodFather has quit IRC | |
| 09:19 | GodFather has joined #ltsp | |
| 09:29 | fly23 has joined #ltsp | |
| 09:29 | <fly23> hi
| |
| 09:30 | <evil_root> hi
| |
| 09:31 | <fly23> i've got a problem with ltsp clients
| |
| 09:31 | dhcp works, pxe works, tftpd works
| |
| 09:31 | i386.img distributed via nbd, works fine
| |
| 09:32 | login seems to work too, but system hangs after user credentials
| |
| 09:33 | .xsession-errors in the users home dir says something about 'cannot open display: localhost:11.0'
| |
| 09:34 | <Gadi> fly23: have you tried disabling compiz?
| |
| 09:34 | !disable_compiz
| |
| 09:34 | <ltspbot> Gadi: "disable_compiz" :: To disable compiz, try: sudo gconftool-2 --direct --config-source xml:readwrite:/etc/gconf/gconf.xml.mandatory --type string --set /desktop/gnome/session/required_components/windowmanager metacity
| |
| 09:36 | <fly23> does it disable compiz for every user?
| |
| 09:37 | <Gadi> yes
| |
| 09:37 | <fly23> hm, nice
| |
| 09:37 | seems to work ;)
| |
| 09:38 | big thanks! you saved my evening
| |
| 09:38 | <Gadi> pleasure
| |
| 09:38 | sorry your thin clients won't be rotating cubes
| |
| 09:38 | :)
| |
| 09:39 | <fly23> na, no problem
| |
| 09:42 | bobby_C has quit IRC | |
| 09:48 | <mistik1> hello
| |
| 09:54 | <Gadi> mistik1: !!
| |
| 09:56 | shawnp0wers has joined #ltsp | |
| 10:02 | <garymc> so anyone know of a free software or something else that will let me login to my LTSP via my mac as you have to pay for vmware fusion
| |
| 10:03 | <fly23> is there any virtualbox/qemu for mac ?
| |
| 10:09 | hm, is nbd-server supposed to be sitting in every rc?.d ? i thought it has to be summoned via inetd like tftp
| |
| 10:10 | Da-Geek has quit IRC | |
| 10:11 | gnunux has quit IRC | |
| 10:20 | <mistik1> hey Gadi
| |
| 10:27 | garymc has quit IRC | |
| 10:30 | xet7 has joined #ltsp | |
| 10:41 | dobber has quit IRC | |
| 11:29 | ogra has quit IRC | |
| 11:30 | ogra has joined #ltsp | |
| 11:30 | ogra is now known as Guest34014 | |
| 11:35 | RiXtEr has quit IRC | |
| 11:35 | RiXtEr has joined #ltsp | |
| 11:49 | staffencasa has joined #ltsp | |
| 12:12 | Guest34014 is now known as ogra_ | |
| 12:43 | shawnp0wers has quit IRC | |
| 12:53 | Nubae is now known as Nubae_away | |
| 13:10 | toscalix has quit IRC | |
| 13:14 | robehend1 has joined #ltsp | |
| 13:15 | <robehend1> woo, ltsp saved the day again today
| |
| 13:16 | windows machines crapped out on me, had to have a class of 30 kids come into the lab in 20 minutes. Just booted to the network, logged them into LTSp, and it worked slick
| |
| 13:26 | pmatulis_ has joined #ltsp | |
| 13:32 | <robehend1> say, random question. Running some 10.04 fat clients, and they keep randomly freezing. any idea?
| |
| 13:40 | alkisg has joined #ltsp | |
| 13:41 | <robehend1> alkisg: say, alkisg, you ever have difficulties with your fat-clients freezing up/
| |
| 13:41 | <alkisg> Only with nfs /home with locks
| |
| 13:42 | statd problems, I solved them with "-o nolock"
| |
| 13:42 | <robehend1> fat-clients, by default, use sshfs homes, correct?
| |
| 13:42 | <alkisg> Correct
| |
| 13:42 | <robehend1> hmm, weird.
| |
| 13:43 | <alkisg> Ah, in a lab a person had problems with atl cards. Hmm but this too was nfs-specific.
| |
| 13:43 | <robehend1> I'll have to figure out what the kids mean by "freezing up". Perhaps its just a program acting stupid.
| |
| 13:43 | <alkisg> I've seen several freezes with KMS too
| |
| 13:43 | <robehend1> these are all onboard intel, seems to run slick
| |
| 13:43 | <alkisg> Many of them solved with "nomodeset"
| |
| 13:43 | <robehend1> kms?
| |
| 13:43 | <alkisg> Which intels?
| |
| 13:43 | 8xx have a *lot* of problems with Kernel Mode Setting (kms) etc
| |
| 13:43 | <robehend1> i'd have to dig into one to tell you, but there 5+ years old
| |
| 13:44 | which would be around the 8xx/9xx series..
| |
| 13:44 | <alkisg> If it's 855, the main problem was solved a week ago upstream
| |
| 13:44 | If it's 845, try maverick, I don't know why but it's more stable there, it crashes on Lucid
| |
| 13:44 | <robehend1> hmm
| |
| 13:44 | <alkisg> 9xx should be fine though
| |
| 13:45 | <robehend1> i've been meaning to rebuild my chroot anyways, i guess. only reason I havent is becuase of having difficulties with Sabayon, and i dont want to deal with profiles again
| |
| 13:45 | <alkisg> Start with: lspci -nn -k | grep -A 2 VGA
| |
| 13:46 | If you have 855, you'll need a kernel from a PPA in the chroot, so rebuilding won't help
| |
| 13:46 | <robehend1> dont have a client in front of me, sadly. i'll have to wait until the kids are done for the day before I can commandeer one
| |
| 13:46 | <alkisg> Didn't you say you installed sch-scripts?
| |
| 13:47 | Right click on a client and select "root terminal, locally"
| |
| 13:47 | <robehend1> *smacks self in head*
| |
| 13:47 | <alkisg> This way you'll get a client root shell in your own client
| |
| 13:47 | <robehend1> woah. ran sch-scripts, big erros
| |
| 13:49 | alkisg: http://ltsp.pastebin.com/dfm3dsyC
| |
| 13:50 | <alkisg> How did you run it? ssh to the server and exec sch-scripts?
| |
| 13:50 | <robehend1> yep
| |
| 13:50 | <alkisg> No display there :)
| |
| 13:50 | <robehend1> bah
| |
| 13:50 | <alkisg> Try ssh -X
| |
| 13:51 | <robehend1> did that, same thing
| |
| 13:51 | <alkisg> Does gedit work?
| |
| 13:51 | <robehend1> nope. i've screwed something up with my displays it seems
| |
| 13:52 | probably when i was attempting to learn how to do vnc over ssh
| |
| 13:52 | <alkisg> Erm, meaning?
| |
| 13:53 | Won't neatx or freenx fit the bill better there?
| |
| 13:53 | Or you want to share an *existing* X screen?
| |
| 13:53 | <robehend1> it was more of a "wonder if i could do this" mentallity
| |
| 13:53 | * alkisg doesn't go into such places :D | |
| 13:53 | <robehend1> i'm learning not to ;)
| |
| 13:54 | now to figure out how to stop and redo what i've done. after i figure out what i've done ;)
| |
| 13:58 | on the plus side, i'd been meaning to roll from ubuntu to edubuntu, so that works nicely ;0
| |
| 14:02 | laprag has joined #ltsp | |
| 14:13 | Nubae_away is now known as Nubae | |
| 14:13 | Nubae has joined #ltsp | |
| 14:22 | <robehend1> alkisg: do you know of any way to limit the amount of ram a fat-client can consume? I remember when you used Localapps that firefox and such could eat up all the ram of the thin-client, and that caused crashes..wondering if it could be the same here
| |
| 14:23 | <alkisg> robehend1: try putting a 512MB NBD swap
| |
| 14:23 | That should solve all your ram-related problems
| |
| 14:23 | <robehend1> alkisg: and i would do that by..
| |
| 14:23 | <alkisg> NBD_SWAP=True in lts.conf
| |
| 14:23 | <robehend1> hmm, alright. easy enough
| |
| 14:23 | <alkisg> And: $ cat /etc/ltsp/nbdswapd.conf
| |
| 14:23 | SIZE=512
| |
| 14:24 | <robehend1> ah, ok. do i have to create a new nbd sharepoint on the server as well?
| |
| 14:24 | <alkisg> No it should already be there
| |
| 14:24 | <highvoltage> alkisg: talking of ram on thin clients...
| |
| 14:24 | vagrantc has joined #ltsp | |
| 14:24 | <robehend1> the cat command gets run on the server, or chrooted into the client
| |
| 14:24 | <alkisg> And those files are sparse so don't get scared if you see 50 x 500 MB in /tmp :D
| |
| 14:24 | Server
| |
| 14:24 | No chroot involved
| |
| 14:24 | <mgariepy> alkisg, does the hosts.allow is necessary ?
| |
| 14:25 | <highvoltage> alkisg: didn't you have a fix or workaround for limiting X memory usage on thin clients as well? (I might have the wrong end of the stick here so forgive me if I do)
| |
| 14:25 | <mgariepy> nbdswapd: ALL: keepalive
| |
| 14:25 | <robehend1> ah, ok. I'm assuming that if it starts swapping, it'll slow down dramatically?
| |
| 14:25 | <alkisg> robehend1: not for the first 50-100 Mb (depending on use)
| |
| 14:25 | <robehend1> hmm. alright. worth a shot, anyways
| |
| 14:25 | <highvoltage> robehend1: it's only slightly slow while it's actually swapping, but having something slightly slow for half a second is a lot better than your whole session being destroyed :)
| |
| 14:25 | <alkisg> After some critical amount of swapping, it gets reeeeally slow
| |
| 14:26 | As in, 2 minutes to open a menu
| |
| 14:26 | <robehend1> hmm
| |
| 14:26 | hopefully it wont get that bad then ;)
| |
| 14:26 | so uh..that nbdswap.conf file doesnt exist ;)
| |
| 14:26 | <alkisg> mgariepy: there are 2 uses for hosts.allow, one is for security and the other (keepalive) is to allow the daemon to shutdown if the client hangs
| |
| 14:26 | robehend1: create it
| |
| 14:27 | <robehend1> wow. I am slow today, aren't I.
| |
| 14:27 | does it need [Default] like lts.conf, or just SIZE=512
| |
| 14:27 | <alkisg> highvoltage: limiting X memory usage will get firefox to crash, right? I prefer to increase NBD swap and make it a little slower but not crash
| |
| 14:27 | robehend1: just SIZE=512
| |
| 14:27 | (it's a shell file, don't put spaces)
| |
| 14:27 | <robehend1> alrighty, thanks :D
| |
| 14:28 | yep, no spaces. hopefully this fixes the "oh gods it breaking call IT call IT" messages
| |
| 14:28 | <highvoltage> alkisg: yeah, it would eventually cause firefox to crash, but I'd give it a very high limit (as in, close to system memory), if I'd do something like that
| |
| 14:28 | <alkisg> highvoltage: e.g. I tested firefox with some web pages that needed 500 MB RAM, and the client had only 256
| |
| 14:29 | After 30 secs or 1 minute firefox was done rendering, and everything was fast again
| |
| 14:29 | Removing compcache also helps there
| |
| 14:29 | <robehend1> ooo, how you do that ;)
| |
| 14:29 | i'm all for speed
| |
| 14:29 | <alkisg> As it gives an extra 25% of RAM to the clients
| |
| 14:29 | <highvoltage> ah, I think that was what I was looking for
| |
| 14:29 | <alkisg> nocompcache in pxelinux.cfg/default
| |
| 14:29 | <highvoltage> how do I do that?
| |
| 14:29 | <robehend1> nice
| |
| 14:30 | <highvoltage> ok, just as a kernel line parameter?
| |
| 14:30 | <alkisg> Right. We put that specifically to help LTSP clients, but I think it does more harm than good, maybe stgraber should remove it (it's in the packaging section)
| |
| 14:30 | compcache is good for live CDs but not for LTSP where we have other swapping methods...
| |
| 14:31 | <robehend1> what does it do, exactly
| |
| 14:31 | <highvoltage> it dedicates some of your ram for swap
| |
| 14:31 | <alkisg> it compresses a part of the memory
| |
| 14:31 | <highvoltage> so your ram gets swapped to compressed ram
| |
| 14:32 | <robehend1> ok. and, by default, it doesnt do that?
| |
| 14:32 | <alkisg> So yeah it provides swap but at the same time it consumes that memory. And 25% is very significant, NBD_SWAP could be used instead (for tmpfs etc) so that all real memory is available
| |
| 14:32 | On real systems compcache isn't used
| |
| 14:32 | For LTSP, we put it in the initramfs
| |
| 14:32 | <robehend1> ah, ok. that makes sense then
| |
| 14:32 | <alkisg> It's good for live CDs, but imho it's bad for LTSP
| |
| 14:33 | (and even for live CDs it's only used for systems with < 512 MB RAM)
| |
| 14:33 | <highvoltage> hmm, on my thin client at the office I don't have that in my cmdline, and it's not using it, I guess it was taken out another way in our setup
| |
| 14:33 | <alkisg> highvoltage: lsmod | grep ramz
| |
| 14:33 | <robehend1> hmm. well, that works then. I'm just all for speed and stability with my users, so ya.
| |
| 14:33 | <alkisg> (locally)
| |
| 14:34 | lsmoz ramzswap
| |
| 14:34 | *lsmod
| |
| 14:35 | <robehend1> oh! before I forget, another thing i was going to ask you, all knowing alkisg. If I'm doing fat clients, is there a difference in server-sizing? i figured since its all on the client, i could go drastically smaller
| |
| 14:35 | <highvoltage> alkisg: it was loaded, but I could unload it without complaint, and my amount of swap is still the same
| |
| 14:35 | <alkisg> robehend1: sure e.g. you don't need a fast cpu on the server, any i386 would do
| |
| 14:36 | <robehend1> hmm, ok. i'm just gonna virtualize it, so thats not the concern. more of the Ram alottment
| |
| 14:36 | <alkisg> highvoltage: did it really unload?
| |
| 14:36 | robehend1: RAM is mainly used to cache the image
| |
| 14:36 | So e.g. for a 2 Gb virtual disk, maybe caching 512 Mb of it is enough
| |
| 14:36 | So in that case a 1 Gb server would suffice
| |
| 14:37 | <highvoltage> alkisg: yep
| |
| 14:37 | <alkisg> Not sure why/how/etc. We do instruct it to use 25% of the available RAM, though
| |
| 14:39 | <robehend1> hmm
| |
| 14:40 | alkisg: so, if i'm understanding it right, I could setup a virtual server with 4 gigs of Ram, and it'd boot upwards of 200 clients easily, since its not really using its ram at all
| |
| 14:40 | rpm_ has joined #ltsp | |
| 14:40 | <alkisg> robehend1: sure, the network is the main limit there
| |
| 14:40 | RAM and CPU are no problem
| |
| 14:41 | <robehend1> its on gig network, best i can do for the moment. it wont be booting 200 clients anyways, as i'd need fiber for that i'm sure
| |
| 14:41 | <alkisg> highvoltage: I don't see any code in $CHROOT/usr/share/initramfs-tools/hooks/compcache that would make compcache not create a swap file, except (1) the nocompcache parameter, and (2) RAM > 512 on the live cd (casper)
| |
| 14:41 | laprag has quit IRC | |
| 14:41 | <alkisg> robehend1: more nics on the server
| |
| 14:42 | <robehend1> alkisg: well, this server does have 2 nics. i've never bonded nics on ubuntu before though. off to google!
| |
| 14:42 | <alkisg> No need for bonding, if you feel uncomfortable with it
| |
| 14:42 | <robehend1> how else would one do it?
| |
| 14:42 | <alkisg> 2 different subnets or some iptable rules would also do
| |
| 14:43 | <robehend1> ah. its not so much a comfort level, as a 'havent done before' kind of thing
| |
| 14:43 | <alkisg> Ah, lets start with the basics, pci nics or pci-e?
| |
| 14:43 | <robehend1> ..onboard..
| |
| 14:43 | <rpm_> Gadi: I found a workaround so gnome-panel comes up with likewiseOpen. What I did was log in as a new user and then when it hangs use "killall gnome-panel" to get the system to load. Then I logged out and got back in as an administrator and copied the contents of the newly created users home profile to /etc/skel. After doing that new users are able to get in the first time without killing gnome-panel.
| |
| 14:43 | <robehend1> its a server class machine, 2 onboard nics, believe their intels
| |
| 14:43 | <alkisg> robehend1: ...that doesn't say much :) pci bus speed is limited to 1 Gbps, so there's no point in putting 2 pci nics
| |
| 14:43 | Ah pci-e then, nice
| |
| 14:43 | lauri has quit IRC | |
| 14:44 | <Gadi> rpm_: nice hack
| |
| 14:44 | <rpm_> hope it helps
| |
| 14:44 | <robehend1> alkisg: yep. the box its on was bought in 2008. 2 dualcore 3 ghz xeons, 32 gigs of ram, and 2 TB of memory
| |
| 14:44 | <alkisg> robehend1: OK go with bonding, I written a page for it last year... https://help.ubuntu.com/community/UbuntuLTSP/Trunking
| |
| 14:44 | <Gadi> thx
| |
| 14:44 | <robehend1> oo thanks
| |
| 14:45 | alkisg: i'm assuming that it still applies with 10.04?
| |
| 14:45 | <alkisg> robehend1: I believe so, haven't tried it though
| |
| 14:46 | <robehend1> well, then I'll let you know, and edit the wiki accordingly ;0
| |
| 14:46 | <alkisg> Nice
| |
| 14:49 | <robehend1> now hopefully this wont overburden my wee 3com switches ;)
| |
| 14:52 | alkisg: had a thought on your virtualbox-ltsp deployment dealio. I thought on it, and wondered how would one handle DHCP, since the teacher machine more than likely only has a single Nic in it. Sure, you can setup virtual nics galore, but it'll still pump out the same physical pipe..
| |
| 14:52 | best way i figured it would work was to setup some kind of program that would allow the teacher to input the MACs of the client machines, and then add those to the dhcp config, so it only provides IP's to those.
| |
| 14:54 | <stgraber> highvoltage: IIRC compcache is only enabled if you have less than X MB of RAM (at least we didn't do anything to disable it at the office)
| |
| 14:56 | Gadi has left #ltsp | |
| 14:59 | <robehend1> if one wanted to build a fat-client with the ubuntu-edu-* packages, i'm assuming you'd add them with the "late packages" feature?
| |
| 15:00 | <alkisg> robehend1: https://help.ubuntu.com/community/UbuntuLTSP/ProxyDHCP
| |
| 15:00 | The easiest way to provide dhcp services
| |
| 15:00 | <robehend1> you know, i always forget that exists
| |
| 15:00 | <alkisg> Completely seamless with existing dhcp setups
| |
| 15:00 | Also, it's possible to use regular dhcp with 1 nic
| |
| 15:00 | (within vbox)
| |
| 15:01 | There's no rule saying that you need 2 nics. It's just the "out of the box setup", nothing more
| |
| 15:01 | stgraber: that's only for the live cd, it checks for "casper"
| |
| 15:01 | <robehend1> makes sense. i was over thinking it, i think
| |
| 15:01 | <alkisg> (and more than 512 RAM)
| |
| 15:02 | if [ "\${BOOT}" = "casper" ]; then
| |
| 15:02 | if [ "\${TOTAL_RAM}" -gt 524288 ]; then
| |
| 15:02 | exit 0
| |
| 15:02 | ...but we have BOOT=ltsp_nbd
| |
| 15:04 | <robehend1> alkisg: can you bridge a bonded nic, so that i could use it with KVM?
| |
| 15:06 | <alkisg> No idea. If you do that check for the effective bandwidth, I think that when I tried something similar with bonding + vbox, I lost the speed benefit
| |
| 15:06 | But it was 3 years ago, I forgot the details
| |
| 15:06 | <robehend1> would make sense, since its being handed off to a virtual nic afterwards
| |
| 15:07 | <alkisg> If it makes sense for you network, you can just use 2 subnets instead of bonding
| |
| 15:08 | <robehend1> how so/
| |
| 15:08 | <alkisg> 192.168.0.x and 192.168.1.x
| |
| 15:08 | static IPs, update dhcpd.conf and ready
| |
| 15:08 | <robehend1> hmm
| |
| 15:08 | thats doable as well
| |
| 15:08 | <alkisg> "makes sense for your network" ==> if it can be separater into 2 physical subnets, check your switches etc
| |
| 15:09 | <robehend1> thats not a problem, i've got two subnets running right now.
| |
| 15:09 | one for office staff, one for students
| |
| 15:09 | <alkisg> You also need NATting for at least one of your subnets, to give them internet access
| |
| 15:09 | You'd also...
| |
| 15:10 | bobby_C has joined #ltsp | |
| 15:10 | rpm_ has quit IRC | |
| 15:15 | ogra_ has quit IRC | |
| 15:15 | ogra_ has joined #ltsp | |
| 15:16 | <robehend1> makes sense, makes sense
| |
| 15:30 | xet7 has quit IRC | |
| 15:32 | Kicer86 has quit IRC | |
| 16:09 | alkisg has quit IRC | |
| 16:09 | robehend1 has left #ltsp | |
| 16:11 | Mip5 has joined #ltsp | |
| 16:11 | Mip5 has left #ltsp | |
| 16:32 | evil_root is now known as zz_evil_root | |
| 17:00 | vagrantc has quit IRC | |
| 17:19 | Gremble has joined #ltsp | |
| 17:24 | Gremble has quit IRC | |
| 18:08 | bobby_C has quit IRC | |
| 18:39 | Indy has joined #ltsp | |
| 18:45 | Indy has quit IRC | |
| 18:47 | staffencasa has quit IRC | |
| 18:47 | zz_evil_root is now known as evil_root | |
| 18:50 | alexqwesa has quit IRC | |
| 18:51 | alexqwesa has joined #ltsp | |
| 20:23 | pmatulis_ has quit IRC | |
| 20:51 | jd42 has quit IRC | |
| 21:29 | evil_root is now known as zz_evil_root | |
| 21:34 | alexqwesa has quit IRC | |
| 21:34 | MorningSon has quit IRC | |
| 22:15 | Androidfan616 has joined #ltsp | |
| 22:16 | Androidfan616 has quit IRC | |
| 22:52 | alkisg has joined #ltsp | |
| 23:35 | leio has quit IRC | |
| 23:41 | leio has joined #ltsp | |