00:20 | Faithful has quit IRC | |
00:33 | Faithful has joined #ltsp | |
01:04 | alkisg has joined #ltsp | |
01:16 | BWMerlin has joined #ltsp | |
01:26 | Faithful has quit IRC | |
01:38 | ofirk has joined #ltsp | |
01:39 | Faithful has joined #ltsp | |
01:49 | <ofirk> Hi to all
| |
01:49 | I want to deploy a LTSP server/client on my network
| |
01:50 | I have a router with a DHCP service and a computer that gets STATIC IP address from the router based on its MAC address
| |
01:50 | gnunux has joined #ltsp | |
01:51 | <ofirk> Since I have more computers on the network, I want to keep the router as is, and only install software on the server
| |
01:51 | The setup should look something like this: http://img267.imageshack.us/img267/4448/networksetup.png
| |
01:52 | <alkisg> ofirk: https://help.ubuntu.com/community/UbuntuLTSP/ProxyDHCP
| |
01:53 | (or, if the router is good enough to provide a boot filename, you can even avoid that)
| |
01:53 | Good morning to all
| |
01:55 | <ofirk> thanks
| |
01:55 | alkisg: it seems really useful
| |
01:56 | alkisg: I am still having hard time to understand the entire process
| |
01:56 | First I need to follow https://help.ubuntu.com/community/UbuntuLTSP/LTSPQuickInstall and then https://help.ubuntu.com/community/UbuntuLTSP/ProxyDHCP ?
| |
01:57 | What about the client? I just need to set it to boot from network?
| |
01:59 | <alkisg> Yes
| |
01:59 | <ofirk> alkisg: to all questions?
| |
01:59 | <alkisg> To all :)
| |
02:00 | <ofirk> Thanks :)
| |
02:00 | <alkisg> (sorry in a hurry, but don't worry that's the way to go - bb!)
| |
02:00 | <ofirk> I will start deploying it right away
| |
02:00 | alkisg has quit IRC | |
02:17 | dobber has joined #ltsp | |
02:28 | mikkel has joined #ltsp | |
03:00 | Da-Geek has joined #ltsp | |
03:09 | gu has left #ltsp | |
03:27 | Mixologic_ has joined #ltsp | |
03:28 | Mixologic has quit IRC | |
03:28 | Mixologic_ is now known as Mixologic | |
03:40 | alkisg has joined #ltsp | |
04:07 | sweetpi has quit IRC | |
04:25 | alkisg has quit IRC | |
04:39 | gnunux has quit IRC | |
05:17 | Faithful has quit IRC | |
05:30 | Faithful has joined #ltsp | |
06:19 | pmatulis has joined #ltsp | |
06:27 | mikkel has quit IRC | |
06:31 | Faithful has quit IRC | |
06:33 | ogra_cmpc has quit IRC | |
06:53 | MorningSon has joined #ltsp | |
07:10 | mikkel has joined #ltsp | |
07:15 | Blinny has joined #ltsp | |
07:38 | BWMerlin has quit IRC | |
07:56 | ofirk has quit IRC | |
08:14 | wwx has quit IRC | |
08:14 | wwx has joined #ltsp | |
08:22 | Faithful has joined #ltsp | |
08:28 | methril_work has joined #ltsp | |
08:30 | <methril_work> morning
| |
08:31 | hawkI has joined #ltsp | |
08:37 | ofirk has joined #ltsp | |
08:37 | ofirk has joined #ltsp | |
08:37 | <ofirk> hi again :)
| |
08:37 | I followed the instructions at https://help.ubuntu.com/community/UbuntuLTSP/LTSPQuickInstall
| |
08:37 | and then at https://help.ubuntu.com/community/UbuntuLTSP/ProxyDHCP
| |
08:38 | to setup a LTSP server on my Kubuntu desktop with a router acting as a DHCP server
| |
08:38 | Faithful has quit IRC | |
08:38 | <ofirk> however, I don't understand the following line at the config file /etc/dnsmasq.d/ltsp.conf :
| |
08:38 | # This range(s) is for the public interface, where dnsmasq functions
| |
08:38 | # as a proxy DHCP server providing boot information but no IP leases.
| |
08:38 | # Any ip in the subnet will do, so you may just put your server NIC ip here.
| |
08:38 | dhcp-range=10.160.31.10,proxy
| |
08:39 | what is my server NIC ip?
| |
08:39 | if say my server as a static IP 192.168.0.150 then it would be it?
| |
08:51 | Da-Geek has quit IRC | |
08:52 | Faithful has joined #ltsp | |
08:57 | Faithful has quit IRC | |
09:08 | Mixologic has quit IRC | |
09:12 | jammcq has joined #ltsp | |
09:12 | <jammcq> good morning friends
| |
09:12 | cliebow: ping
| |
09:12 | Faithful has joined #ltsp | |
09:15 | ofirk has quit IRC | |
09:41 | sutula_gone is now known as sutula | |
09:44 | Guerdal82 has joined #ltsp | |
09:45 | <methril_work> i'm seing a slow login with ldm, but with gdm is faster
| |
09:45 | do you have any idea where to look for?
| |
09:45 | Mixologic has joined #ltsp | |
09:45 | <methril_work> it still applies the policikit don't allowing to make connections with ssh -X?
| |
09:46 | i'm using ubuntu 10.04
| |
09:48 | shogunx has quit IRC | |
09:48 | shogunx has joined #ltsp | |
09:58 | [GuS] has joined #ltsp | |
10:01 | jappiecr has joined #ltsp | |
10:01 | <jappiecr> -join
| |
10:03 | I have my ltps server installed and I can connect tc's with different usernames and such! However, when I start the app Thin Client manager, no tc's are shown and so there is no control over the tc's...any ideas?!
| |
10:15 | andrew_ubuntu has joined #ltsp | |
10:16 | <andrew_ubuntu> Hello everyone. Is there documentation on how to setup a scanner either on the server that the clients can use, or on a client for Ubuntu LTSP 9.10? Thanks!
| |
10:16 | <jammcq> andrew_ubuntu: I don't know if there's docs, but I have lots of thin clients with scanners attached
| |
10:17 | you need to install saned and micro-inetd on the thin client chroot
| |
10:17 | <andrew_ubuntu> ok, easy enough :).
| |
10:18 | <Kyle__> jammcq: saned==scanner-server?
| |
10:19 | <jammcq> yeah
| |
10:19 | <Kyle__> Neat.
| |
10:20 | Nokio has joined #ltsp | |
10:20 | * Kyle__ reads | |
10:21 | <Nokio> Hi all, When i do an update of my system and it want to reboot because of a new kernel version for example all my dummy has a window opening saying do you want to reboot now. Is this normal ? in my opinion dummy should not be able to reboot my server.
| |
10:23 | * Kyle__ blinks | |
10:23 | <jammcq> andrew_ubuntu: you'll need a line in your lts.conf file that says:
| |
10:24 | RCFILE_01 = /etc/init.d/scanner.sh
| |
10:24 | * andrew_ubuntu takes note | |
10:24 | <Kyle__> Nokio: When you go to reboot your dumb-terminals, your linux server reboots?
| |
10:24 | * Kyle__ blinks | |
10:25 | <jammcq> andrew_ubuntu: and that scanner.sh script should contain:
| |
10:25 | #!/bin/sh
| |
10:25 | SANE_PORT=$(grep sane-port /etc/services | awk '{print $2}')
| |
10:25 | start-stop-daemon -q -b --start --exec /usr/bin/micro-inetd ${SANE_PORT%%/tcp} /usr/sbin/saned
| |
10:25 | that last line probably wrapped in your IRC client
| |
10:25 | but it's basically 3 lines of code
| |
10:26 | and then, you need to configure the saned config files
| |
10:26 | enable network access
| |
10:26 | <Nokio> kyle_, right now i didnt click on the dummy to reboot so im not quite sure yet and i will only be able to try during lunch. Do you know if it will reboot the server, will it reboot that dummy only or will it reboot all dummy?
| |
10:27 | jappiecr has quit IRC | |
10:27 | <andrew_ubuntu> jammcq: where does the scanner.sh file live?... in the chroot?
| |
10:27 | <jammcq> in /etc/init.d
| |
10:27 | as pointed to by the lts.conf entry
| |
10:28 | <andrew_ubuntu> just on the server then?
| |
10:29 | <Kyle__> Nokio: I guess I'm not quite following what you are saying.
| |
10:29 | Nokio: If you reboot a dumb-terminal, only the dumb-terminal should reboot. If the command is somehow sent to the server, it will reboot if, and only if, the user who sent it has the right to reboot.
| |
10:30 | Nokio: Usually, only root has the right, and you're not logging in as root on the dumb-term, right?
| |
10:31 | <jammcq> andrew_ubuntu: no, in the chroot. so more like: /opt/ltsp/i386/etc/init.d/scanner.sh
| |
10:33 | <andrew_ubuntu> ooh, ok... sorry. I was pretty sure it was the chroot, but thought I should clarify.
| |
10:33 | <Nokio> Kyle_ no im logued in as a normal user in the dummy. BTW im sorry for my bad english ! So what you say is that is i click on the windows that says reboot on the dummy, only the dummy should reboot and not the server.
| |
10:36 | <Kyle__> Nokio: Yes. I believe so, but I'm used to fat clients, not dumb-terminals. HOldon.
| |
10:37 | Can anyone else verify that if you click reboot on a thin-client, that only the thin-client will be rebooted?
| |
10:39 | <andrew_ubuntu> jammcq: just to clarify, this is what I have done so far: 1) installed libsane and micro-inetd in the chroot 2) made the /etc/init.d/scanner.d file in the chroot and chmod 755 3) ran ltsp-update-image 4) put the RCFILE_01 line in lts.conf
| |
10:40 | <jammcq> andrew_ubuntu: you might like to install sane-utils in the chroot as well, to aid in troubleshooting
| |
10:41 | also, on the chroot, edit your /etc/sane.d/dll.conf file to exclude all of the drivers that you don't need. It'll speed up scanner detection quite a bit
| |
10:42 | <andrew_ubuntu> oh good ideas... thanks a bunch btw for all your help.
| |
10:42 | <jammcq> and edit the /etc/sane.d/saned.conf file and add a '+' to the end of the file, or add an entry specifically for the machine that will be running the client application
| |
10:42 | that's also in the chroot
| |
10:46 | <andrew_ubuntu> done :)
| |
10:48 | now to test.... it'll be a bit before I can do this, they are using the scanner :(
| |
10:48 | <jammcq> what kind of scanner do you have?
| |
10:48 | <andrew_ubuntu> epson perfection .... eh... umm... 1660 photo I believe
| |
10:49 | or it might be the 2400... either way I know it works with linux and I have it working in LTSP 4.2 also.
| |
10:49 | <jammcq> ah, should be a piece of cake then
| |
10:50 | <andrew_ubuntu> ya, should be; especially with your help. I tried a week or so ago to do it, but was trying to get it to work on the server (thought it would be easier)... wasn't.
| |
10:51 | <jammcq> oh, there was something I needed to do with udev to grant the correct permissions to the usb device nodes
| |
10:52 | in the chroot, I added:
| |
10:52 | /lib/udev/rules.d/99-sane-group.rules
| |
10:52 | the contents are:
| |
10:52 | ENV{DEVTYPE}!="usb_device", GOTO="libsane_rules_end"
| |
10:52 | ENV{libsane_matched}=="yes", GROUP="saned"
| |
10:52 | LABEL="libsane_rules_end"
| |
10:53 | that sets the group on the device node to 'saned' so that the saned program can access it
| |
10:53 | <Kyle__> Actually, I've got a fat-client question: how come the clients don't start all the normal services on boot? The one that sticks out for me is cups.
| |
10:54 | <Nokio> kyle_ what runlevel are you in ? i had a similar problem
| |
10:57 | jappiecr has joined #ltsp | |
10:58 | <Kyle__> Nokio: I'm using ubuntu, and there aren't really run-levels like in normal linucies.
| |
10:58 | <jappiecr> Anybody out there with some Thin Client Manager experience...my TC manager does not see my thin clients.
| |
10:58 | <jammcq> jappiecr: sorry, i'm not familiar with it
| |
10:59 | <Kyle__> Nokio: I can start it, if I use a launch it using an RCFILE_XX in lts.conf, but it seems slightly hackish.
| |
11:01 | <Nokio> kyle__ same here im with ubuntu but when you use the command runlevel you should still get 2 as an answer. i was getting unknown for some reason and that was why cups and ssh in my case were not starting
| |
11:02 | kyle__ if you get unknown its a known bug ill try to get the link again let me know
| |
11:02 | <Kyle__> Nokio: Oh. You know I didn't check that on my clients.
| |
11:03 | Guerdal82_ has joined #ltsp | |
11:04 | <Kyle__> Nokio: THey show runlevel 2.
| |
11:04 | Hum.
| |
11:04 | dobber has quit IRC | |
11:12 | <andrew_ubuntu> jammcq: getting closer to testing... wish me luck ;).
| |
11:20 | Faithful has quit IRC | |
11:27 | <andrew_ubuntu> jammcq: xsane says no scanner detected :(
| |
11:33 | Faithful has joined #ltsp | |
11:33 | <jappiecr> I am setting up a compu lab for a school in Costa Rica. I (kind of) have the server up and running, that is my thin clients boot and can I can log in using different usernames on different thin clients and all works beautiful!
| |
11:33 | <jappiecr> <jappiecr> Now the challanges that I run in to: 1) The Thin Client Manager does not show my thin clients.. I followed the manual and have somewhat set it. Made profile for Students and teachers (+ one for adminsitrator), assigned different user to to those groups in Profile Manager and here I am...NO CLUE where to go from here!! even with the GCompris suite I setup the adm. section and say that clients need to login but when I log in from a T
| |
11:33 | C as a student, there is no login required...
| |
11:35 | Faithful has quit IRC | |
11:35 | <jappiecr> ps...I tried installing ltsp-manager and ltsp-aulacontrol but after restart my system crashed and had to uninstall both for the server to work again!
| |
11:36 | Faithful has joined #ltsp | |
11:40 | <johnny> ltsp-manager is only for ltsp 4
| |
11:40 | it shouldn't be in your repos
| |
11:41 | never heard of aulacontrol, can't comment
| |
11:41 | thin client manager i don't think it works anymore, tcm-ng might be better
| |
11:46 | <andrew_ubuntu> jammcq: btw, I had them login to the console on the terminal the scanner is hooked to and run 'scanimage -L' and it sees the scanner, so that part is working. I just don't know why it's not showing up in gnome.... must have missed a step.
| |
11:49 | <methril_work> hi, someone has a slow login with ldm, but works fast with old startx?
| |
11:49 | i'm using Ubuntu10.04
| |
11:49 | it could be just the wm?
| |
11:49 | but the same wm (gnome) work fine in gdm
| |
11:49 | <jappiecr> johnny: thanks! I tried looking up tcm-ng in synaptic but it only shows "tcm" and says that "phyton tcm" is installed
| |
11:49 | <Kyle__> methril_work: what's the load on the server look like when they log in using LDM?
| |
11:50 | <methril_work> lower that 1
| |
11:51 | Kyle__, it takes about 14 seconds since ldm pass the login prompt until the desktop get the gnome desktop environment configured
| |
11:51 | <Kyle__> methril_work: Wow. Thin clients or fat?
| |
11:51 | <methril_work> Kyle__, with some users (with old settings) it could take 24 secs
| |
11:51 | thin clients
| |
11:51 | but the same thin client starts in 2/3 seconds with startx
| |
11:52 | <Kyle__> methril_work: I've got fat clients on ubuntu 10.04, and I found using LDM_DIRECTX=True in lts.conf sped the login up a bit. Maybe by 5-10 seconds when the server is loaded down.
| |
11:52 | <methril_work> i know is not the same, but the performance could be the same, isn't it?
| |
11:52 | * Kyle__ nods | |
11:52 | <methril_work> it's enabled
| |
11:52 | <Kyle__> methril_work: Err... I thought LDM_DIRECTX worked fairly analogously to STARTX....
| |
11:52 | <methril_work> i've LDM_DIRECTX
| |
11:53 | it has differences
| |
11:53 | <Kyle__> Right, but I thought they both used X, instead of X tunned over ssh.
| |
11:54 | <methril_work> yes, they both use X
| |
11:54 | Faithful has quit IRC | |
11:54 | <Kyle__> What's the difference in login speed if they go into an xterm session?
| |
11:55 | <methril_work> it breaks with xterm
| |
11:55 | <Kyle__> You can't log in with xterm?
| |
11:55 | <methril_work> i cant
| |
11:57 | <Kyle__> I wonder if something else is wrong.
| |
11:57 | An xterm session should always work right?
| |
11:57 | <methril_work> i set the gconf performance settings in mandatory
| |
11:58 | ok
| |
11:58 | it's working now
| |
11:58 | let me check the start time
| |
11:58 | * Kyle__ nods | |
11:59 | <methril_work> 9 secs
| |
11:59 | it's there a way to see what ltsp settings are enabled in a thin client?
| |
11:59 | like environment/export for ltsp vars?
| |
12:01 | <Kyle__> I really don't know. I think sometime should setup some thin clients just so I can understand them better, but right now I"m using all fat clients.
| |
12:01 | * Kyle__ wishes his fat clients were a good deal fatter. A P4 is not exactly blazing nowdays. | |
12:01 | <methril_work> well, thank you for the suggestions
| |
12:02 | <Kyle__> No problem. Whish I could help more.
| |
12:02 | * methril_work get envy due to Kyle__'s fat clients | |
12:03 | <Kyle__> methril_work: Heh, don't be too envious. I recently found out the guy who used to be in my job was lazier than I imagined.
| |
12:03 | <johnny> ram and network is key..
| |
12:03 | ram more so for fat clients
| |
12:03 | <methril_work> i've upgrading the server's RAM
| |
12:03 | <johnny> but stil, it beats cpu
| |
12:03 | <Kyle__> He opted to buy out of the department budget, instead of lease out of the university's budget...because he didn't want to have to reconfigure new machines every three years....
| |
12:04 | <methril_work> i was thinking to change to NFS, but i see a boot speed degradation
| |
12:04 | <Kyle__> So I'm stuck with 5 year old PCs, when I could have brand new ones....
| |
12:04 | <methril_work> Kyle__, well, it's the IT fun ;)
| |
12:04 | <Kyle__> methril_work: NFS home for fat clients is a huge speed boost, but I'm still using nbd for /
| |
12:04 | methril_work: Hehe.
| |
12:05 | <johnny> nbd is faster
| |
12:05 | <Kyle__> The man was a menace to anything with a keyboard.
| |
12:05 | <johnny> that's why we use it..
| |
12:05 | <methril_work> i see nbd faster
| |
12:05 | * Kyle__ nods | |
12:05 | <johnny> but it's not writable in the way necessary, so nfs is used for that
| |
12:05 | <methril_work> Kyle__, dam windows users! (that call they own admins)
| |
12:06 | we don't need NFS home
| |
12:06 | <Kyle__> johnny: NFS is amazing in many situations though. There are comanies who opt to run their databases over NFS rather than iSCSI.
| |
12:06 | <methril_work> it's located on the ltsp server
| |
12:06 | Faithful has joined #ltsp | |
12:06 | <johnny> well, nfs is better for home , since it is easy to write to..
| |
12:06 | <Kyle__> methril_work: Ahh the advantage of thin clients.... I'd do it if I could get the budget for a big enough server...but I only give users windows via virtualization, so it would need to be a _big_ server.
| |
12:07 | <johnny> hmm.. isn't the licensing more expensive that way?
| |
12:07 | <methril_work> really big one!!
| |
12:07 | <johnny> vs rdp connections to a regular windows servers?
| |
12:08 | <methril_work> how is the minimum amount of RAM recommended ?
| |
12:08 | <johnny> depends on if they need administrator privileges or not
| |
12:08 | <Kyle__> johnny: Not really. I'm at a school, and we have an MSDN liscence. Each user gets their very own windows VM, to modify, tweak, break, infect, etc.... and when they ask, or I notice it's too bad, I just restore it to a clean state.
| |
12:08 | <methril_work> i read somewhere that nbd for thin clients does not scale so well, if they don't have so much RAM
| |
12:08 | <johnny> for thin clients i it's 192 = (128 x clients)
| |
12:08 | <Kyle__> johnny: Oh yea, windows thin clients are expensive. YOu need an RDP CAL for each thin client. It gets pricey.
| |
12:09 | <johnny> err 192 + (128 x clients)
| |
12:09 | <methril_work> johnny, in the server
| |
12:09 | <johnny> yes
| |
12:09 | <methril_work> but what about the thin clients?
| |
12:09 | <johnny> 128
| |
12:09 | <Kyle__> I'm guessing the system mounting ndb stores all changes in RAM, right?
| |
12:10 | <johnny> sure.. but what changes?
| |
12:10 | <methril_work> if i've less., i works, butt....
| |
12:10 | i was thinking in that case the overhead of NFS is smaller than nbd
| |
12:10 | <Kyle__> johnny: anything written to a log, /tmp if you don't have tmpfs setup, etc.
| |
12:10 | <johnny> few things are stored on the client
| |
12:10 | methril_work, i doubt it
| |
12:10 | * Kyle__ nods | |
12:10 | <johnny> Kyle__, those logs should be syslogged remotely
| |
12:10 | <jappiecr> my network connections does not show up (anymore!) in my panel, how do i activate this again?
| |
12:10 | <johnny> ltsp has support for that built in, you just have to enable it on the server and set the lts.conf var
| |
12:11 | <Kyle__> johnny: I agree, but some (poorly behaving) software may not.
| |
12:11 | * Kyle__ blinks | |
12:11 | <methril_work> uff!! trying to do optimizataions in a production server is a nightmare
| |
12:11 | <johnny> hmm?
| |
12:11 | that poorlly bheaving software is most likely on the server, not on the client
| |
12:11 | jimjimovich has joined #ltsp | |
12:11 | <Kyle__> there's an lts.conf var for syslog? Damn... I edited the rsyslogd.conf on my image to do that.
| |
12:11 | <methril_work> SYSLOGD_HOST
| |
12:11 | <johnny> Kyle__, i don't know how it works with fat clients at all
| |
12:11 | probably does..
| |
12:11 | but i do't really know
| |
12:12 | <Kyle__> johnny: For my fat clients they all log to /var/log, unless changed.
| |
12:12 | <johnny> sure
| |
12:12 | <jimjimovich> what's the best way to do remote support (something like VNC with Remote Desktop Viewer) with LTSP? I want to be able to connect and help my staff with problems (both they and I are on the same LTSP server)
| |
12:12 | <johnny> same with thin clients, but if you have ram issues where that is a big deal, then you would want to setup remote logging
| |
12:12 | jimjimovich, italc maybe?
| |
12:12 | <methril_work> johnny, how about AoE?
| |
12:12 | <johnny> no idea
| |
12:13 | <Kyle__> johnny: The clients all have 2GB of ram, so it shouldn't be a big issue.
| |
12:13 | <johnny> it doesn't support AoE
| |
12:13 | nobody has wanted it bad enough to add it
| |
12:13 | <Kyle__> methril_work: AoE will only work on a local LAN segment, so that would require the server to be on the very same switch.
| |
12:13 | <methril_work> OpenSuse are doing something with KIWI
| |
12:13 | <Nokio> kyle__ then im sorry i dont know what this can be!
| |
12:13 | * Kyle__ really wants to play with AoE anyway, it seems cool :) | |
12:13 | <johnny> nfs is just easier most of the time..
| |
12:14 | <Kyle__> Nokio: That's ok. Thanks anyway. I can manually start cups.
| |
12:14 | johnny: Yea.
| |
12:14 | <methril_work> i see better performance with LDM than startx option
| |
12:14 | due to gnome-watchdog enabled
| |
12:15 | it allow to keep the server with clean process
| |
12:15 | <Kyle__> methril_work: Maybe the startx option is using more bandwidth, and thus constraining yoru network that way?
| |
12:15 | Eh, nahh.
| |
12:15 | <methril_work> i doubt it
| |
12:15 | maybe is getting more process on the server, but not on the client
| |
12:16 | <Nokio> kyle__ i know you probably already checked but is cups configured to start on boot in /etc/rc2.d?
| |
12:16 | * Kyle__ nods | |
12:16 | <methril_work> the problem is that if the people see that the login takes more than 10 secs, they though the machine is not responding and they force a switch off
| |
12:17 | * Kyle__ nods | |
12:18 | <jimjimovich> Am I right to assume that I can't use VNC with Ubuntu LTSP?
| |
12:18 | <Kyle__> methril_work: I have that problem myself. But for me I ended up being more disk-bandwidth constrained (no raid). I ended up seperating users out to different physical drives, and it helped.
| |
12:18 | jimjimovich: YOu should be able to use VNC with ltsp just fine.
| |
12:19 | <jimjimovich> Kyle__: how do you distinguish which session/person you're connecting to?
| |
12:19 | <Kyle__> jimjimovich: You may have to configure a few things first. You could use the vnc module that's built into XOrg so you could VNC into any session, or setup vinarge to start by default on all desktops once they're logged in.
| |
12:19 | jimjimovich: On the server you can see who's logged in from which client.
| |
12:20 | <Nokio> kyle__ do you have something in your /etc/network/interfaces ?
| |
12:20 | <Kyle__> Nokio: Sorry? What do you mean? I have my primary interface setup there, staticlly.
| |
12:21 | <Nokio> kyle__ https://bugs.launchpad.net/ubuntu/+source/upstart/+bug/500520 the bug i faced with my cups not starting was related to booting in an unknown runlevel and the fact that my lo interface was not defined in that file
| |
12:28 | <Kyle__> Nokio: Humm. Interesting. I'll check that.
| |
12:29 | Nokio: it's set to "auto lo" and ifconfig looks correct for lo, but I wonder if that's the crux the problem.
| |
12:32 | gecko_zo6 has joined #ltsp | |
12:33 | <gecko_zo6> I have some errors and my ltsp client do not start
| |
12:33 | the error is tftp file not found
| |
12:33 | but my tftp config seems to be ok
| |
12:34 | could someone help???
| |
12:37 | <Kyle__> What starts nbd-server? I mean, what's _supposed_ to start it on an ubuntu 10.04 ltsp system?
| |
12:38 | <johnny> openbsd-inetd
| |
12:39 | <Kyle__> OK.
| |
12:39 | * Kyle__ digs | |
12:41 | Nokio has quit IRC | |
12:42 | Lns has joined #ltsp | |
12:42 | Lns has joined #ltsp | |
12:44 | gecko_zo6 has quit IRC | |
12:46 | rpm_ has joined #ltsp | |
12:48 | <Kyle__> Ahh. Weird. It updated inetd.conf to include two backup images I had, but not my main image. I mustve done somethign strange.
| |
12:50 | jappiecr has quit IRC | |
12:52 | al34n1x has joined #ltsp | |
12:54 | jappiecr has joined #ltsp | |
12:54 | <al34n1x> Hi guys, i just got the LTSP-Cluster over OpenVZ link and i get excited to test it in our labs..my questions is...do you know if the implementation over Debian is just that straightforward?
| |
13:07 | dobber_ has joined #ltsp | |
13:20 | <Kyle__> al34n1x: I use a debian-based virtualization distro called proxmox-ve for my servers.
| |
13:20 | It runs OpenVZ just fine.
| |
13:20 | al34n1x: Admittedly, there are issues running an LTSP server inside of an OpenVZ container, but hey, thems the brakes.
| |
13:25 | Faithful has quit IRC | |
13:27 | <jappiecr> I a, running edubuntu 10.04 with ltps server setup and 2 nics. eth0= for Internet and eth1 = for TC network (with ip range: 10.0.30.) . In my "Network Connections" it now shows the following wired connections: Ethernet automatic, ifupdown (eth1), ifupdown (eth0) and Auto eth0 ....does this sound right?!
| |
13:27 | jimjimovich has left #ltsp | |
13:28 | <abeehc-_> anyone know why rdeskop runs on the client side now when you define something like SCREEN_07 = rdesktop servername
| |
13:30 | <Kyle__> abeehc-_: Errr, for my fat clients that's how I always define their sessions.
| |
13:31 | abeehc-_: I have SCREEN_03=comodore64, which launches a script that runs a C64 basic emulator on vt 3.
| |
13:31 | And it always runs on the client.
| |
13:32 | * Kyle__ sighs | |
13:32 | <Kyle__> Sadly, not a single student has noticed it yet.
| |
13:36 | rpm_ has quit IRC | |
13:38 | ogra_ac has joined #ltsp | |
13:39 | Faithful has joined #ltsp | |
13:43 | Faithful has quit IRC | |
13:44 | <Lns> Kyle__: =)
| |
13:44 | I did something similar as just a test a while back - started a Zork game on tty6.
| |
13:48 | bobby_C has joined #ltsp | |
13:49 | <abeehc-_> hehe
| |
13:49 | my clients are starved for resources though and from 9.04 to 9.10 rdesktop started running on the client and acting strange
| |
13:50 | ogra_ac has quit IRC | |
13:54 | <Lns> abeehc-_: do you mean rdesktop is being started as a localapp?
| |
13:54 | <Kyle__> Lns: Did any of the kids see it?
| |
13:54 | <Lns> Kyle__: no it was just a test at my own office =)
| |
13:55 | <Kyle__> Ahh I see.
| |
13:55 | jpg has joined #ltsp | |
13:55 | <Kyle__> I really wanted to put apple basic on there, but couldn't find the c-source code anywhere anymore.
| |
13:55 | <Lns> though I have always wanted to put old-school text based games/emulators in the 'games' list..i KNOW they'd find them there ;)
| |
13:55 | * Kyle__ nods | |
13:55 | <Kyle__> Lns: My labs are all used for CS classes, so I figure anything that allows them to program is fair game. Hell, I even put ICON on all of them for one of my dear old dinosaur professors.
| |
13:56 | jpg has left #ltsp | |
13:57 | mgariepy has joined #ltsp | |
13:57 | mgariepy has joined #ltsp | |
13:57 | Faithful has joined #ltsp | |
13:59 | <abeehc-_> Lns: yeah
| |
13:59 | it appears to run as a localapp but I've not build the image for that purpose, and no localapps are defined in lts.conf
| |
14:00 | <Lns> abeehc-_: have you verified by looking at the local process list on the tc?
| |
14:00 | sounds strange - you usually have to configure an app to run as a localapp ;)
| |
14:01 | [GuS] has quit IRC | |
14:02 | Faithful has quit IRC | |
14:04 | <abeehc-_> yeah i'm wathcing the proc on the client via top
| |
14:04 | ogra_ac has joined #ltsp | |
14:04 | <abeehc-_> the behaviour changed once i migrated to 9.10/ppa from 9.04
| |
14:04 | ubuntu that is
| |
14:05 | jeez i just messed up update-alternatives ldm-theme somehow
| |
14:17 | <al34n1x> Kyle_:Yeap i saw promox that is based on debian, i will try to use that as base for this LTSP cluster
| |
14:18 | <Kyle__> al34n1x: Well, maybe use it as a model more than base your cluster on it :)
| |
14:24 | <al34n1x> yeap, u right
| |
14:25 | Another question...do you know why the LTSP clients get stucks if there are two different DHCP across the environment?
| |
14:32 | <Kyle__> al34n1x: Two DHCP servers in the same environment?
| |
14:32 | s/environment/subnet
| |
14:33 | al34n1x has quit IRC | |
15:04 | sweetpi has joined #ltsp | |
15:20 | pmatulis has quit IRC | |
15:27 | <Kyle__> Is there anythign special in the LTSP kernels to mount the nbd root?
| |
15:29 | mikkel has quit IRC | |
15:32 | Blinny has quit IRC | |
15:32 | <sweetpi> the nbd block driver in the ramdisk
| |
15:34 | <Kyle__> When you install a new kernel in ubuntu, is the initrd.img rebuilt, or is it one out of the package? (I'm assuming that's the ramdisk you're referring to).
| |
15:35 | <johnny> rebuilt probably
| |
15:35 | <sweetpi> I dont use ubuntu, but on debian its rebuilt
| |
15:35 | <johnny> via initramfs-tools
| |
15:35 | just like debian
| |
15:35 | <Kyle__> Humm.
| |
15:35 | OK this is odd. Cause I loaded a mainline kernel, and it books, initrd, runs, then complains it can't find the init
| |
15:36 | <sweetpi> what are the kernel args?
| |
15:37 | <Kyle__> I didn't change them from what I had before. In pxelinux.cfg/default: append ro initrd=initrd.img quiet splash nbdport=2000
| |
15:38 | <sweetpi> try netboot=nbd
| |
15:38 | <Kyle__> Hu. I just walked over to one of them. the nbd image is mounted at /root/rofs, so it's there.
| |
15:38 | OK I'll add that.
| |
15:39 | <sweetpi> otherwise, get rid of quiet splash and see whats going on
| |
15:39 | mgariepy has quit IRC | |
15:40 | <Kyle__> Yea same sitch.... Trying again.
| |
15:42 | Removing quiet and splash didn't help. Busybox takes over and you cant scroll back. But it looks like anormal boot...just doesn't have an init.
| |
15:43 | When I run gzip -dc initrd.img |strings|grep /sbin/init
| |
15:43 | <sweetpi> this initrd works fine with the ltsp kernel?
| |
15:43 | <Kyle__> I see it's looking in /sbin /etc and /bin for an init.
| |
15:43 | * Kyle__ nods | |
15:45 | <Kyle__> Ooh... with the new kernel it's mounting it as /root/rofs.... but in the init script rootmount is set to /root
| |
15:46 | Would that have anything to do with it, or no?
| |
15:46 | * Kyle__ hasn't dug through initrds much. | |
15:47 | <johnny> initramfs has a plugin architecture, if you don't build the initramfs usingthe right command (or pointing to the right directory) it won't add the nbd support
| |
15:48 | nbd root isn't in initramfs by default, you have to use the plugin
| |
15:48 | <Kyle__> johnny: It's got the nbd support, that doesn't seem to be the problem. I know, because it's mounting it at /root/rofs
| |
15:48 | * Kyle__ frowns. | |
15:49 | jappiecr has quit IRC | |
15:57 | <sweetpi> Kyle__: not sure whats wrong, netboot should have worked. I guess try boot=nbd
| |
15:58 | other than that, try break=premount and see if you can figure out what is going on
| |
15:58 | maybe the nbd driver wont load on that specific kernel
| |
16:00 | <Kyle__> sweetpi: The driver is loading, becuase it is mounting, just not finding init where it expects it.
| |
16:00 | I'll try the boot=nbd now
| |
16:01 | <sweetpi> so its just mounting in the wrong place, cant remember how to change that
| |
16:02 | unless its just root=
| |
16:03 | al34n1x has joined #ltsp | |
16:03 | <al34n1x> my connection dropped off...did anyone see the issue about the dhcp servers?
| |
16:03 | i got two servers on the same network
| |
16:03 | one belong to the school and i'm not able to manage
| |
16:04 | <Kyle__> al34n1x: Yea I saw.
| |
16:04 | <al34n1x> the other one is under my scope
| |
16:04 | <Kyle__> al34n1x: You mean both respond to all requests,and you don't manage one of them?
| |
16:05 | <al34n1x> i set up the DHCP file to server only to 10 computers...no matter what if the other Servers is providing IPs my computers get stucks
| |
16:05 | and don't boot
| |
16:05 | <sweetpi> al34n1x: are you using proxydhcp?
| |
16:06 | <Kyle__> root= or root=nbd?
| |
16:06 | <al34n1x> nop
| |
16:06 | <sweetpi> al34n1x: then its not always going to work right
| |
16:06 | <Kyle__> al34n1x: I suppose you could set them up to boot locally, instead of via DHCP+pxe.
| |
16:07 | <sweetpi> Kyle__: give me a min I'll see what I can dig up
| |
16:07 | <Kyle__> al34n1x: Do you have a managed switch in the room your clients are in?
| |
16:07 | sweetpi: Thanks.
| |
16:07 | * Kyle__ is trying ALL these suggestions one by one, in ohpes of one working, buecase kernel bugs are not fun to live with. | |
16:07 | <al34n1x> nop, i don't have access to the switch
| |
16:07 | yeap
| |
16:08 | <Kyle__> al34n1x: Damn. Are these real thin clients, old desktops being used as thin clients, or fat clients?
| |
16:08 | <sweetpi> al34n1x: I doubt this will fix _everything_ but you can specify the full path to init with init=
| |
16:08 | <al34n1x> a mix of them
| |
16:09 | how i could do that with init?
| |
16:09 | <Kyle__> al34n1x: I think that one was meant for me :)
| |
16:09 | <sweetpi> init=/root/rofs/sbin/init or something like that
| |
16:10 | but I think the scripts are still going to need it mounted on /root
| |
16:10 | <Kyle__> al34n1x: For the old-pcs as thin clients, and the fat clients, you can boot locally, grabbing the nbd image and whatever else you need. No need to rely on DHCP.
| |
16:11 | <al34n1x> ok, on the thin client where i can store that, because i don't have any chance of store anything due to they are diskless
| |
16:12 | <Kyle__> al34n1x: Ugh. Damn.
| |
16:12 | al34n1x: Maybe someone who knows more about thin clinets needs to help :) I only use the fat ones.
| |
16:13 | sweetpi: It still fails, but in much more intersting ways now :)
| |
16:14 | <sweetpi> yeah I figured :)
| |
16:15 | <Kyle__> sweetpi: What's confusing is that it's not seeing the built-in init in initramfs that it should be. Correct me if I'm wrong, but there _should_ be one in initram that sets it all up, then passes control along to the one on the real root image.
| |
16:17 | <al34n1x> Another question...i noted that if all the session is encripted the response from the clients is slow if you compare it against a normal session with only the login encripted
| |
16:17 | my question is..how many clients the LTSP server support
| |
16:17 | ?
| |
16:18 | i could get until 11 more or less, i'm not sure if it's a problem on my network...perhaps
| |
16:18 | * Kyle__ rolls back to the older kernel. Less than 2 hours left in the day, I won't risk leaving late tongiht. | |
16:19 | <sweetpi> Kyle__: init in in /init
| |
16:19 | Kyle__: so try passing init=/init since its looking for /sbin/init
| |
16:19 | * Kyle__ blinks | |
16:19 | <Kyle__> Humm.
| |
16:20 | K, bouncing two clients now to test.
| |
16:21 | <andrew_ubuntu> anyone know if it is possible to ssh to a clients console? I can do this in LTSP4.2, but not sure how to in LTSP5 (probably need to install sshd into the chroot for starters...).
| |
16:21 | <Kyle__> Drat. Didn't help. Similar situation, but not identical to passing in /root/rofs/sbin/init.
| |
16:22 | <sweetpi> andrew_ubuntu: correct, install sshd
| |
16:23 | <andrew_ubuntu> sweetpi: thanks...do I have to put anything in lts.conf to get it to start on boot?
| |
16:23 | <sweetpi> Kyle__: im out of ideas, only thing I can say is break=premount and try to do everything step by step. or wait for one of the more knowledgeable people here
| |
16:23 | andrew_ubuntu: not that I remember
| |
16:24 | <Kyle__> Thanks for all the help sweetpi. All of it _should_ have worked. It really just could be something funky with how ubuntu makes their mainline kernels.
| |
16:24 | <andrew_ubuntu> sweetpi: ok cool
| |
16:25 | <sweetpi> Kyle__: np I'm sure someone else here can help when they wake up :)
| |
16:27 | <Kyle__> I hope so. I'm going to try again tomorrow, I just don't want to leave my lab in an un-usable state.
| |
16:28 | <sweetpi> Kyle__: look for alkisg or vagrantc, they might know
| |
16:29 | M4gic5t0rM has joined #ltsp | |
16:30 | <sweetpi> Kyle__: you have a working kernel for the lab right?
| |
16:41 | <Kyle__> sweetpi: Yes I do. Sorry Iwas walking around hard-rebooting machines that had locked up due to that kernel bug I've got
| |
16:41 | * Kyle__ sighs. | |
16:42 | <Kyle__> Of all the old hardware to have, a lab full of P4 dells that lockup under high load.
| |
16:45 | pmatulis has joined #ltsp | |
16:46 | bobby_C has quit IRC | |
16:46 | M4gic5t0rM has quit IRC | |
16:47 | jammcq has quit IRC | |
16:48 | cliebow_ has joined #ltsp | |
16:48 | cliebow__ has joined #ltsp | |
16:49 | alexqwesa has quit IRC | |
16:50 | alexqwesa has joined #ltsp | |
16:51 | cliebow_ has quit IRC | |
16:55 | Lns has quit IRC | |
17:07 | ogra_ac has quit IRC | |
17:08 | ogra_ac has joined #ltsp | |
17:14 | <andrew_ubuntu> sweetpi: Thanks for you help on the SSH... it works great!
| |
17:14 | now if I can just get my scanner working.....
| |
17:15 | but that will have to wait until tomorrow.... night guys!!!
| |
17:15 | andrew_ubuntu has quit IRC | |
17:20 | al34n1x has quit IRC | |
17:41 | hawkI has quit IRC | |
17:49 | alkisg has joined #ltsp | |
17:51 | johnny has left #ltsp | |
17:55 | dobber_ has quit IRC | |
18:19 | johnny has joined #ltsp | |
18:23 | cliebow__ has quit IRC | |
18:24 | ogra_ac_ has joined #ltsp | |
18:24 | ogra_ac has quit IRC | |
18:24 | ogra_ac_ is now known as ogra_ac | |
18:55 | jhutchins_lt has quit IRC | |
19:02 | alkisg has quit IRC | |
19:03 | ogra_ac has quit IRC | |
19:03 | ogra_ac has joined #ltsp | |
19:13 | sweetpi has quit IRC | |
19:43 | Mixologic has quit IRC | |
20:05 | Guerdal82 has quit IRC | |
20:27 | alexqwesa has quit IRC | |
20:31 | cliebow_ has joined #ltsp | |
20:32 | johnny has left #ltsp | |
20:38 | sweetpi has joined #ltsp | |
21:06 | johnny has joined #ltsp | |
21:32 | ogra has quit IRC | |
21:33 | ogra has joined #ltsp | |
22:05 | pmatulis has quit IRC | |
22:12 | Guerdal82 has joined #ltsp | |
22:16 | cliebow_ has quit IRC | |
23:16 | sweetpi has quit IRC | |
23:46 | try2free has joined #ltsp | |
23:56 | try2free has left #ltsp | |