IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 12 July 2010 (all times are UTC)
| 00:36 | ltspbot has joined #ltsp | |
| 00:37 | ltspbot` has quit IRC | |
| 00:42 | alexqwesa has quit IRC | |
| 00:42 | vmlintu_ has quit IRC | |
| 00:46 | gentgeen__ has quit IRC | |
| 00:46 | highvoltage has quit IRC | |
| 00:46 | atkuepker has quit IRC | |
| 00:46 | Mobe has quit IRC | |
| 00:46 | loathing has quit IRC | |
| 01:17 | shogunx has joined #ltsp | |
| 01:50 | alkisg has quit IRC | |
| 02:33 | vagrantc has quit IRC | |
| 02:43 | alexqwesa has joined #ltsp | |
| 02:43 | vmlintu_ has joined #ltsp | |
| 02:48 | gentgeen__ has joined #ltsp | |
| 02:48 | highvoltage has joined #ltsp | |
| 02:48 | atkuepker has joined #ltsp | |
| 02:48 | Mobe has joined #ltsp | |
| 02:48 | loathing has joined #ltsp | |
| 03:03 | alkisg has joined #ltsp | |
| 03:13 | ogra_cmpc has quit IRC | |
| 03:25 | ogra_cmpc has joined #ltsp | |
| 03:32 | ogra has quit IRC | |
| 03:32 | ogra has joined #ltsp | |
| 03:45 | alkisg has quit IRC | |
| 04:05 | vmlintu_ has quit IRC | |
| 04:15 | Da-Geek has joined #ltsp | |
| 04:27 | Selveste1_ has joined #ltsp | |
| 04:28 | mrcarrot has joined #ltsp | |
| 04:29 | <mrcarrot> inside of an lts client chroot apt-get -t lenny-backports upgrade is breaking the environment
| |
| 04:29 | it can not mount rofs on nbd0 after that
| |
| 04:29 | has anyone of you encountered this with debian
| |
| 04:44 | dobber has joined #ltsp | |
| 04:46 | <mrcarrot> out of curiousity... what distro do you all other use for ltsp. i am running debian in one school and ubuntu in another. both are having troubles
| |
| 04:46 | both needs 3:rd part packages to be useful
| |
| 04:46 | for flash and other things
| |
| 04:47 | it would be good to not need 3:rd part packages and backported packages
| |
| 05:03 | ogra_cmpc has quit IRC | |
| 05:08 | alkisg has joined #ltsp | |
| 05:25 | alkisg has quit IRC | |
| 05:33 | alkisg has joined #ltsp | |
| 06:19 | F-GT has quit IRC | |
| 06:22 | alkisg has quit IRC | |
| 06:26 | pmatulis has joined #ltsp | |
| 06:36 | F-GT has joined #ltsp | |
| 06:46 | peter_tm has joined #ltsp | |
| 07:04 | Blinny has joined #ltsp | |
| 07:06 | <Blinny> Morning. Anyone know why my Intel D510M0 boards (built-in r5169 GB NICs) would freeze halfway through a PXE boot but not when BIOS set to gPXE?
| |
| 07:08 | <Appiah> gPXE is not in the bios?
| |
| 07:09 | ogra_cmpc has joined #ltsp | |
| 07:09 | <Appiah> you could always try a bios update
| |
| 07:11 | <Blinny> It is an option in the BIOS. And that one works. But regular PXE boot locks with a NIC message just after doing the USB stuff.
| |
| 07:12 | And truth be told, the gPXE looks exactly like a PXE boot, and not at all like an etherboot boot.
| |
| 07:12 | I'd prefer not to use gPXE because the user must press some keys to get it to boot gPXE.
| |
| 07:12 | BTW, this is on a Hardy Heron installation.
| |
| 07:12 | mrcarrot has left #ltsp | |
| 07:13 | <Appiah> didnt know mainboards shipped with gpxe
| |
| 07:14 | can you get the tftpserver to dump some info about the freeze?
| |
| 07:14 | maybe a timeout?
| |
| 07:15 | hersonls has joined #ltsp | |
| 07:15 | alkisg has joined #ltsp | |
| 07:15 | <Blinny> I did some wireshark sniffs and everything looks kosher. TFTP data packets, and ACKs all the way to the freeze.
| |
| 07:16 | Any suggestions of how to get the tftp server to give some details?
| |
| 07:17 | <alkisg> Blinny: is that after pxelinux gets loaded? Or before?
| |
| 07:18 | It goes like this: dhcp hands out an ip address, pxelinux.0 is loaded with tftp, and then pxelinux.cfg/default and the kernel/initrd are loaded again with tftp.
| |
| 07:19 | (in other words, could it be a pxelinux problem? - I've seen it fail lots of times...)
| |
| 07:21 | <Blinny> No, I get tons of kernel messages.
| |
| 07:22 | (If I remove the quiet/splash directives.
| |
| 07:22 | sweetpi had me put 'break=preload' or some such in the kernel arguments, and that got me to a busybox prompt, but I couldn't figure out what to do from there.
| |
| 07:42 | <alkisg> Blinny: If you're getting kernel messages, then it isn't a tftp problem
| |
| 07:43 | And you're saying that it works with gpxe and not with gpxe?! That's very strange, because when the kernel loads those programs have completed their tasks sucessfully.
| |
| 07:44 | Try to remove quiet/splash and see what's the last message before the freeze
| |
| 07:46 | *working with gpxe and not with pxe, i meant...
| |
| 07:54 | <Blinny> The last message is almost always a r5169 message about link being up or down
| |
| 07:55 | Thanks alkisg. I'm on the offender testing it out so let me switch and I'll get you specific details. Appreciate it mate.
| |
| 07:55 | BRB
| |
| 07:55 | Blinny has quit IRC | |
| 07:58 | Egyptian[Home] has quit IRC | |
| 08:01 | Blinny has joined #ltsp | |
| 08:01 | <Blinny> OK!
| |
| 08:03 | alkisg: The last message is, this time, a hid-core.c USB HID core driver message.
| |
| 08:03 | peter_tm has quit IRC | |
| 08:04 | <Blinny> alkisg: However, a few lines up there are 'IP-Config: eth0 hardware address blabh mtu 1500 DHCP RARP' then 'r8169: eth0: link down'
| |
| 08:04 | On some boots, that 'eth0: link down' comes as 'eth0: link up'
| |
| 08:04 | I've plugged into a 100M switch to make sure it's not a cabling issue too.
| |
| 08:05 | strangely, when I unplugged my cat5 cable, and then plugged into my local 100M switch I get two 'eth0: link down' messages, though link lights are good.
| |
| 08:06 | Selveste1_ has quit IRC | |
| 08:06 | F-GT has quit IRC | |
| 08:07 | <Blinny> Same thing in the 100M switch.
| |
| 08:08 | cyberorg has quit IRC | |
| 08:10 | <Blinny> One strange thing. The 'IP Config: eth0 hardware address 70:71:... 1500 DHCP RARP' does not come with a '[microtime stamp]' like all the other messages. And, if I take out 'quiet splash' then it shows up on ctrl+alt+f1 just after the initrd.img loading, also without any kernel timestamp.
| |
| 08:12 | It's similar to the output in this post: http://ubuntuforums.org/archive/index.php/t-874568.html
| |
| 08:12 | though I only have one NIC
| |
| 08:12 | Egyptian[Home] has joined #ltsp | |
| 08:12 | <alkisg> Blinny: ok let's check which stage it reaches
| |
| 08:13 | Open /var/lib/tftpboot/ltsp/i386/pxelinux.cfg/default
| |
| 08:13 | And put "break=init" in the place were "quiet splash" was
| |
| 08:13 | cyberorg has joined #ltsp | |
| 08:13 | <alkisg> Then reboot the client. If you get a busybox shell, then it passes the ipconfig stage.
| |
| 08:14 | (better yet, put "break=mount" there instead - that will give you a shell right after the networking is configured)
| |
| 08:16 | <Blinny> with break=mount then I get the same thing - EXCEPT there is no 'link down' or 'ip-config: eth0' message
| |
| 08:16 | ohwait
| |
| 08:16 | if I hit enter then i get (initramfs)
| |
| 08:17 | peter_tm has joined #ltsp | |
| 08:18 | <alkisg> Blinny: ok, there, if you run `ifconfig` do you see eth0 configured properly?
| |
| 08:18 | <Blinny> No. Nothing returned from 'ifconfig'
| |
| 08:19 | <alkisg> OK, try running : ipconfig eth0
| |
| 08:19 | ...and see what that tells you
| |
| 08:19 | <Blinny> That gives 'ip-config: eth0 hardware address blahblah mtu dhcp rarp'
| |
| 08:20 | then a kernel message 'r8169: eth0: link up' then the 'NET: Registered protocol family 17'
| |
| 08:20 | but I'm NOT back to a promp
| |
| 08:20 | t
| |
| 08:20 | <alkisg> Then ipconfig can't initialize your card properly, either because of a dhcp problem or a module problem.
| |
| 08:21 | <Blinny> In the server's logs I get a request, offer, request, ack
| |
| 08:21 | <alkisg> But, did you say it's working with gpxe? That's too strange... :)
| |
| 08:21 | <Blinny> It is working with gPXE
| |
| 08:21 | <alkisg> OK, try a different approach.
| |
| 08:21 | <Blinny> Though I found a forum post that says intel's gPXE isn't really gPXE, but it falls back to PXE. I don't know if I believe that.
| |
| 08:21 | <alkisg> In pxelinux.cfg/default, remove "break=mount". Then, in a NEW line at the end of the file, write "ipappend 3"
| |
| 08:22 | Reboot the client and see if it boots.
| |
| 08:23 | F-GT has joined #ltsp | |
| 08:23 | <Blinny> Yah I got an address. Stopped after rootserver: 192.168.1.1 then rootpath: then on another line filename :
| |
| 08:23 | but no boody
| |
| 08:24 | er, booty
| |
| 08:25 | I get a 'tftp: client does not accept options' in the server's logs also.. have seen this message before though
| |
| 08:27 | <alkisg> Do you have any weird dhcp configuration?
| |
| 08:27 | Put break=mount back (and leave ipappend 3 as well). Paste your pxelinux.cfg/default to pastebin. Then reboot the client and see if `ifconfig` tells you that all are ok.
| |
| 08:28 | <Blinny> I have dynamic DNS
| |
| 08:29 | http://pastebin.com/bP376PEJ
| |
| 08:30 | F-GT has quit IRC | |
| 08:30 | <Blinny> ifconfig eth0 gives me info about eth0, but doesn't grab an address
| |
| 08:30 | <alkisg> What do you mean "dynamic dns"?
| |
| 08:30 | <Blinny> Hrm. The server says it ACK'd the request
| |
| 08:31 | I mean creating bind records for the hosts that get dhcp addresses
| |
| 08:31 | For example, 192.168.1.23 is ws1-023.mydomain.fun
| |
| 08:31 | http://en.wikipedia.org/wiki/Dynamic_DNS
| |
| 08:32 | <alkisg> I don't think ipconfig supports that. Is the client supposed to notify the dns server or dhcp3-server itself?
| |
| 08:33 | <Blinny> dhcp3-server notifies DNS
| |
| 08:33 | <alkisg> Anyway, is dhcp3 server on the same machine as the ltsp server?
| |
| 08:33 | <Blinny> Yes.
| |
| 08:33 | <alkisg> Could you post your /etc/ltsp/dhcpd.conf as well?
| |
| 08:33 | <Blinny> And the presence/absence of a nameserver to notify doesn't (isn't supposed to) prevent booting like this.
| |
| 08:34 | Is there a safe/secure pastebin that allows file upload
| |
| 08:34 | <alkisg> PM me if you like...
| |
| 08:36 | <Blinny> Gonna be a long PM
| |
| 08:36 | <alkisg> np
| |
| 08:36 | <Blinny> OK trying a file transfer
| |
| 08:37 | ..or my firewall is blocking it (;
| |
| 08:37 | Blinny has quit IRC | |
| 08:38 | <alkisg> Ooops
| |
| 08:38 | Blinny has joined #ltsp | |
| 08:38 | <Blinny> Oops.
| |
| 08:38 | <alkisg> :)
| |
| 08:38 | That doesn't seem to work :D
| |
| 08:38 | peter_tm has left #ltsp | |
| 08:39 | <alkisg> Well anyway... since ipappend 3 didn't solve the problem, then it isn't a dhcp problem
| |
| 08:39 | <Blinny> Here. http://pastebin.com/43yqLzpn
| |
| 08:40 | <alkisg> It's probably some module problem - maybe gpxe leaves the card in a better state than pxe
| |
| 08:40 | <Blinny> Is there another module I could fall back to with an r8169?
| |
| 08:41 | <alkisg> Are you going to update to 10.04 any time soon? If so, the problem may go by itself..
| |
| 08:42 | <Blinny> I've got four servers and am in the middle of switching telephones to SIP trunks. It'll be a few months.
| |
| 08:43 | <alkisg> Do try replacing pxelinux.0 with the one from debian, it helps in some similar cases
| |
| 08:43 | http://people.ubuntu.com/~alkisg/pxelinux.0
| |
| 08:43 | (the ubuntu one is quite old)
| |
| 08:45 | <Blinny> WIll that pxelinux.0 get overwritten when I do security updates?
| |
| 08:45 | <alkisg> No - it'll be overwritten when you run ltsp-update-kernels
| |
| 08:45 | ...but if you also put it in the chroot, then it won't.
| |
| 08:45 | <Blinny> Ah. Dig
| |
| 08:46 | F-GT has joined #ltsp | |
| 08:47 | <Blinny> Yeah the new pxelinux.0 didn't change matters.
| |
| 08:50 | <alkisg> You could try using udhcpc (like karmic+) do instead of ipconfig....
| |
| 08:50 | sudo chroot /opt/ltsp/i386
| |
| 08:51 | <Blinny> That sounds scary (;
| |
| 08:51 | <alkisg> Ah no it's not that easy you'll need some manual file copying too
| |
| 08:52 | <Blinny> Well hell I'm up for it.
| |
| 08:53 | Does udhcp require your pxelinux.0 ? I reverted when I noticed no change.
| |
| 08:53 | <alkisg> Nope
| |
| 08:53 | OK, those steps should do it:
| |
| 08:53 | 1) install udhcpc in the chroot
| |
| 08:54 | 2) Save http://pastebin.com/xDhFviM3 as /opt/ltsp/i386/usr/share/initramfs-tools/hooks/udhcp
| |
| 08:54 | 3) Save http://pastebin.com/JHTpRNX8 as /opt/ltsp/i386/usr/share/initramfs-tools/scripts/init-premount/udhcp
| |
| 08:55 | 4) make them both executable (chmod +x)
| |
| 08:55 | 5) in the chroot, run: update-initramfs -u
| |
| 08:55 | 6) out of the chroot, run ltsp-update-kernels
| |
| 08:56 | That should do it, udhcpc will be used instead of ipconfig (if no compatibility issues are there which I don't know of)
| |
| 08:56 | <Blinny> OK. Do I need any changes on the server / non-chroot side?
| |
| 08:57 | <alkisg> No, except for ltsp-update-kernels
| |
| 08:59 | <Blinny> OK here goes.
| |
| 09:01 | Yuck.
| |
| 09:01 | <alkisg> What?
| |
| 09:01 | <Blinny> udhcpc Sending discover... No lease, failing.
| |
| 09:01 | <zeitsofa> hello, what can be wrong if a user on a client can't umount a usbstick?
| |
| 09:01 | <Appiah> is it only a specific client?
| |
| 09:01 | <zeitsofa> on all clients
| |
| 09:01 | <Appiah> only one user?
| |
| 09:01 | others can?
| |
| 09:01 | <zeitsofa> all users on all clients can't umount
| |
| 09:02 | <Appiah> Dist and version?
| |
| 09:02 | <alkisg> Blinny: hmmm maybe you should try with a minimal dhcpd.conf to see if it changes anything? /usr/share/doc/ltsp-server/examples/dhcpd.conf
| |
| 09:02 | <Blinny> The server does ACK
| |
| 09:02 | <alkisg> Is that for pxe or for udhcpc?
| |
| 09:02 | <Blinny> OK. I'll try that.
| |
| 09:02 | udhcpc
| |
| 09:03 | through PXE
| |
| 09:03 | <alkisg> No, I mean that the server should send 2 acks for each client
| |
| 09:03 | One for when it requests an address at the pxe stage, and one for the udhcpc or ipconfig stage
| |
| 09:03 | <Blinny> An offer and an ACK
| |
| 09:03 | <alkisg> No, two offers and two acks for each client
| |
| 09:03 | It goes like this:
| |
| 09:03 | pxe requests address, dhcp offers / acks etc
| |
| 09:04 | then the kernel loads, but it needs a new address
| |
| 09:04 | So again it sends a request (through ipconfig or udhcpc) and the dhcp server again offers/acks etc
| |
| 09:04 | <Blinny> Yah so I'm only getting one offer/ack with udhcpc
| |
| 09:04 | <alkisg> So udhcp doesn't get any offers.
| |
| 09:05 | Then no, that doesn't sound like a dhcp problem at all. The module is just not working right.
| |
| 09:05 | To revert the udhcp changes, just uninstall udhcpc from the chroot and run update-initramfs / ltsp-update-kernels (the last steps above)
| |
| 09:06 | So I'm not seeing any easy solutions. Either a kernel upgrade or a lucid chroot.
| |
| 09:06 | (or keep using gpxe :))
| |
| 09:06 | <Blinny> I had a backup i386 chroot from a few days ago, I just moved that one back. Thank you.
| |
| 09:06 | heh
| |
| 09:06 | Weird that gPXE works though eh?
| |
| 09:06 | <zeitsofa> Appiah: ubuntu 10.04 and version Version: 5.2.1
| |
| 09:07 | <alkisg> Blinny: If you run through the module history, I bet you'll see something about initialization issues...
| |
| 09:08 | <Appiah> zeitsofa: localdev enabled in lts.conf?
| |
| 09:08 | <Blinny> alkisg: I read about WOL issues when dual-booting but I'm not.
| |
| 09:08 | <zeitsofa> yes it'i enabled
| |
| 09:11 | <Blinny> alkisg: And modules are built against kernels right, so I couldn't just grab an updated module.
| |
| 09:12 | <Appiah> zeitsofa: https://wiki.ubuntu.com/DebugLocalDev checked this?
| |
| 09:13 | <alkisg> Blinny: no, you'd have to manually compile it against your kernel headers
| |
| 09:13 | zeitsofa: you're not supposed to unmount usb sticks on the clients - you just pull them off after 2 seconds of inactivity
| |
| 09:13 | ltspfsd flushes its buffers every 2 secs
| |
| 09:25 | Selveste1_ has joined #ltsp | |
| 09:36 | <Blinny> And it probably wouldn't build against the older kernel.
| |
| 09:56 | <alkisg> I think it would.
| |
| 10:00 | alkisg has quit IRC | |
| 10:04 | Blinny has quit IRC | |
| 10:05 | jbrett has joined #ltsp | |
| 10:08 | Blinny has joined #ltsp | |
| 10:10 | Selveste1_ has quit IRC | |
| 10:16 | staffencasa has joined #ltsp | |
| 10:20 | dobber has quit IRC | |
| 10:23 | <Blinny> What packages do I need in the chroot to build a kernel module? build-essentials, -headers and -modules ?
| |
| 10:24 | I guess I don't need -modules for building
| |
| 10:25 | F-GT has quit IRC | |
| 10:42 | vagrantc has joined #ltsp | |
| 10:42 | F-GT has joined #ltsp | |
| 10:47 | Blinny has quit IRC | |
| 10:54 | alkisg has joined #ltsp | |
| 10:59 | The_Code has joined #ltsp | |
| 11:05 | alkisg has quit IRC | |
| 11:10 | pmatulis has quit IRC | |
| 11:11 | pmatulis has joined #ltsp | |
| 11:38 | Da-Geek has quit IRC | |
| 11:45 | alkisg has joined #ltsp | |
| 11:53 | makghosh has joined #ltsp | |
| 11:59 | <johnny> alkisg, not sure if you care,but maybe you might like this
| |
| 11:59 | http://blogs.gnome.org/xclaesse/2010/07/12/ssh-your-im-contacts/
| |
| 11:59 | <alkisg> Hey johnny
| |
| 12:01 | Heh, nice! In this cases I'm using bash sockets though, `bash 0<>/dev/tcp/alkisg.dyndns.org/25547` and done :)
| |
| 12:01 | No need to install anything at all, nor forward any ports...
| |
| 12:03 | <johnny> i still think you'd think it was cool, especially if it ends up being a package that you could just tell somebody to install via the software manager
| |
| 12:04 | vs telling them to download and mark a script executable
| |
| 12:04 | <alkisg> Sure, I do think it's cool
| |
| 12:05 | I'd also like a shared terminal within the chat session, if possible :P :D
| |
| 12:06 | But for remote help / connections reverse bash + reverse vnc solved my problem - now if only screen worked over tcp, so that both sides saw the session...
| |
| 12:07 | (btw the ssh-im code is really small - nice example for using telepathy tubes...)
| |
| 12:07 | <johnny> using it with a concept that people like sbalnaev might understand :)
| |
| 12:07 | hah :)
| |
| 12:08 | <alkisg> !seen sbalneav
| |
| 12:08 | <ltspbot> alkisg: sbalneav was last seen in #ltsp 6 weeks, 5 days, 0 hours, 27 minutes, and 26 seconds ago: <sbalneav> win 1
| |
| 12:08 | <vagrantc> ltsptelepathytubefs
| |
| 12:08 | ouch!
| |
| 12:08 | * alkisg wonders if sbalneav is learning telepathy and that's why he's been away... :P | |
| 12:08 | <vagrantc> !seen gadi
| |
| 12:08 | <ltspbot> vagrantc: gadi was last seen in #ltsp 2 weeks, 4 days, 23 hours, 35 minutes, and 7 seconds ago: <Gadi> that's two
| |
| 12:08 | <vagrantc> ouch!
| |
| 12:08 | <alkisg> My god ltsp is falling apart... :P
| |
| 12:08 | <vagrantc> !seen stgraber
| |
| 12:08 | <ltspbot> vagrantc: stgraber was last seen in #ltsp 1 week, 1 day, 21 hours, 36 minutes, and 59 seconds ago: <stgraber> cool
| |
| 12:09 | <vagrantc> alkisg: it's you and me :)
| |
| 12:09 | <johnny> !seen vagrantc
| |
| 12:09 | <ltspbot> johnny: vagrantc was last seen in #ltsp 7 seconds ago: <vagrantc> alkisg: it's you and me :)
| |
| 12:09 | <alkisg> Nah, I was in Turkey for a week and now I'll be gone for the next 2 years for my phd...
| |
| 12:09 | <johnny> hmm?
| |
| 12:09 | <alkisg> So LTSP == vagrantc for now :D
| |
| 12:09 | <johnny> really alkisg ?
| |
| 12:09 | congratulations
| |
| 12:09 | <vagrantc> alkisg: eeyk!
| |
| 12:09 | <alkisg> Well I'll be around, but I'll mostly work with delphi :( for the next 2 years
| |
| 12:10 | <vagrantc> johnny: i've been seeing a few gentoo-ish commits... you still without a gentoo box?
| |
| 12:10 | <johnny> yes
| |
| 12:11 | seems like i'm closer to the dream of buying a new laptop than ever tho
| |
| 12:12 | * vagrantc wonders what an appropriate update of /topic would look like | |
| 12:12 | The_Code has quit IRC | |
| 12:12 | otavio has quit IRC | |
| 12:13 | otavio has joined #ltsp | |
| 12:13 | <monteslu> Any chance Gavin Spurgeon hangs out in here?
| |
| 12:14 | <johnny> who's that?
| |
| 12:15 | <alkisg> monteslu: he showed up two weeks ago but I haven't seen him since
| |
| 12:15 | <vagrantc> alkisg: next year's debian conference will be in your neck of the woods... http://wiki.debconf.org/wiki/DebConf11/BanjaLuka
| |
| 12:16 | no idea how realistically close that is, but looks pretty close from over here
| |
| 12:17 | <alkisg> Hmmm it should be doable within 1 day car drive...
| |
| 12:17 | * alkisg reads more about this... | |
| 12:20 | <alkisg> http://maps.google.com/maps?f=d&source=s_d&saddr=Io%C3%A1nnina,+Greece&daddr=Bosnia+and+Herzegovina&hl=el&geocode=&mra=ls&sll=39.674495,20.840141&sspn=0.129218,0.308647&ie=UTF8&t=h&z=6
| |
| 12:23 | chupacabra has quit IRC | |
| 12:24 | chupacabra has joined #ltsp | |
| 12:24 | <monteslu> alkisg, thanks. He's going to be doing redhat related LTSP stuff, I think
| |
| 12:24 | <alkisg> monteslu: did you read the related mails in ltsp-discuss ?
| |
| 12:25 | <monteslu> yeah, and on the k12osn list
| |
| 12:25 | <dberkholz> johnny: gentoo/dracut work seems to be going fairly well
| |
| 12:25 | <johnny> whoo..
| |
| 12:25 | <monteslu> I think he's waiting for an ok from Warren, but I was hoping he would just go ahead without waiting for an ok
| |
| 12:26 | <johnny> dberkholz, it's gonna be great to remove the last major barrier of entry..
| |
| 12:27 | i guess at that point.. we'll have to try to convince ltsp peple to put up some tarballs
| |
| 12:27 | we use sourceforge.. why can't snaps be hosted there??? :(
| |
| 12:27 | chupacabra has quit IRC | |
| 12:27 | <johnny> altho there is some sort of .deb unpacker.. we could just unpack vagrantc 's code :)
| |
| 12:27 | or fedora 13's rpms..
| |
| 12:27 | <alkisg> johnny: you can just run bzr branch and get the code?
| |
| 12:28 | <johnny> that's what we do now
| |
| 12:28 | <dberkholz> that's not really suitable for distro packages, though
| |
| 12:28 | <johnny> butwe don't want to add a dep of zr
| |
| 12:28 | bzr
| |
| 12:28 | <dberkholz> rpms can be dealt with, if warren happens to include all the stuff we need
| |
| 12:28 | <alkisg> johnny: "getting the code" is part of some dependency?!
| |
| 12:28 | <johnny> alkisg, right now, if you emerge ltsp, it will pull in bzr as a dep
| |
| 12:28 | <alkisg> Ouch
| |
| 12:28 | <johnny> and then check out a specified rev or trunk
| |
| 12:29 | alkisg, ebuilds are just recipes for how to get and build something
| |
| 12:29 | it's not a binary, it's just a text file
| |
| 12:29 | so.. if ltsp could host some cronned snaps of the code at tagged release times
| |
| 12:29 | that would be pretty cool
| |
| 12:30 | daily would be nice too.. but not required..
| |
| 12:32 | chupacabra has joined #ltsp | |
| 12:32 | <vagrantc> johnny: both debian and ubuntu's process includes tarballs you could just grab.
| |
| 12:34 | no need to extract from an rpm; debian-based distros ship the tarball as it's own file.
| |
| 12:34 | the "upstream" tarball, that is
| |
| 12:34 | but yeah, it would be nice to have an "official" upstream.
| |
| 12:35 | <johnny> so, who do we have to convince?
| |
| 12:35 | <vagrantc> someone just needs to do the work...
| |
| 12:35 | and host it somewhere.
| |
| 12:35 | <johnny> yeah.. somebody who has access
| |
| 12:35 | sourceforge
| |
| 12:35 | we hae sourceforge
| |
| 12:35 | use that
| |
| 12:35 | it's free
| |
| 12:36 | <vagrantc> sourceforge is painful
| |
| 12:36 | <johnny> didn't they maek it easier for handle files recently?
| |
| 12:36 | and there's also a command line program
| |
| 12:36 | <vagrantc> didn't know about the commandline program... and the web interface is not completely obfuscating anymore ... but it's still kind of painful.
| |
| 12:38 | since the VCS is hosted on launchpad, it might be feasible to have launchpad generate a tarball as a post-push hook or some such
| |
| 12:39 | otherwise, you just have a cron job running somewhere that pulls the repository regularly, and if a new tag is found, generates a tarball from that tag...
| |
| 12:40 | i could do a proof of concept pretty easily...
| |
| 12:49 | Blinny has joined #ltsp | |
| 12:52 | alkisg has quit IRC | |
| 12:53 | Lns has joined #ltsp | |
| 12:53 | Blinny has quit IRC | |
| 12:54 | makghosh is now known as makghosh|afk | |
| 13:04 | wurzzero has joined #ltsp | |
| 13:08 | <wurzzero> Hi... i have a problem in my LTSP server (Ubuntu 10.04). I asked Alkis and he told me to come here ... maybe someone could help me...
| |
| 13:08 | <Appiah> maybe...
| |
| 13:08 | <wurzzero> good
| |
| 13:08 | so ..
| |
| 13:08 | <johnny> !ask
| |
| 13:08 | <ltspbot> johnny: "ask" :: Don't ask to ask a question, simply ask it, and if someone knows the answer, they'll respond. Please hang around for at least 15 minutes after asking a question, as not everybody constantly monitors the channel.
| |
| 13:09 | <wurzzero> there is a bug in the LTSP (https://bugs.launchpad.net/ubuntu/+source/gnome-session/+bug/491940) and because this problem i restarted my server without properly log off the client..
| |
| 13:10 | now when i boot the server it stays about 8 secs in the gdm screen and then shutdown
| |
| 13:10 | don't know what to do
| |
| 13:11 | <Appiah> so you patched it and after that you get that shutdown problem?
| |
| 13:12 | <wurzzero> i didn't patch because i found that after the problem
| |
| 13:13 | i forgot to say that i configured the client to auto login and bug make a loop
| |
| 13:13 | <Appiah> what does the /var/log/messages on the server say?
| |
| 13:13 | or /var/log/auth
| |
| 13:14 | <wurzzero> i'll paste the last lines..
| |
| 13:14 | svc: failed to register lockdv1 RPC service (errno 97).
| |
| 13:14 | NFSD: Using /var/lib/nfs/v4recovery as the NFSv4 state recovery directory
| |
| 13:14 | NFSD: starting 90-second grace period
| |
| 13:15 | Creating dsa-hostkey for LucidLynx
| |
| 13:15 | Creating rsa-hostkey for LucidLynx
| |
| 13:15 | ppdev: user-space parallel port driver
| |
| 13:15 | Jul 8 17:48:46 localhost pulseaudio[1803]: lock-autospawn.c: Cannot access autospawn lock.
| |
| 13:15 | man
| |
| 13:15 | monteslu has quit IRC | |
| 13:16 | <wurzzero> is there a way to upload the file in irc?
| |
| 13:20 | monteslu has joined #ltsp | |
| 13:20 | <wurzzero> i chmod -x /etc/init.d/ltsp-sirvecole and the server do not shutdown, i'd like to know how can i look for wath it loads
| |
| 13:36 | <johnny> sirvecole ?
| |
| 13:36 | and why are you using nfs?
| |
| 13:36 | nfs is not default for ubuntu
| |
| 13:39 | <wurzzero> i only install the ltsp-server-standalone and build the client
| |
| 13:42 | is there something like a 'cache' that stores the client configuration? so i could reset, delete.....
| |
| 13:42 | <johnny> the client configuration is only one file /var/lib/tftpboot/ltsp/i386/lts.conf
| |
| 13:43 | but on a fresh install you won't have one
| |
| 13:43 | did you mess with any files in /opt/ltsp/i386?
| |
| 13:43 | if so.. you might want to start over by deleting it and generating a new client
| |
| 13:43 | ltsp-build-client --arch i386
| |
| 13:45 | <wurzzero> i modified the /var/lib/tftpboot/ltsp/i386/lts.conf to auto login, but i removed this configuration and its empty again, but the problem persists
| |
| 13:46 | i'll do that, build a new client...
| |
| 13:49 | tanks! i'll be back in some minutes ...
| |
| 13:49 | dobber_ has joined #ltsp | |
| 13:50 | wurzzero has quit IRC | |
| 13:53 | wurzzero has joined #ltsp | |
| 13:55 | <wurzzero> man, didn't work, i deleted the /opt/ltsp/i386 and it shutdown again.....btw, since the problem i'm using mandriva 2010.0 tryng to solve the problem....
| |
| 13:55 | triple boot
| |
| 14:00 | now /var/log/messages have a new message Jul 12 15:54:25 localhost ltsp: No client chroots found, please run ltsp-build-client
| |
| 14:00 | but i can't creat cause it shutdown
| |
| 14:01 | <vagrantc> it just keeps shutting down on boot?
| |
| 14:02 | <wurzzero> it loads everything, it stays about 8 secs in the GDM screen and shutdown
| |
| 14:03 | i logged in, but it shutdown too
| |
| 14:03 | <vagrantc> you could boot into single-user mode and disable gdm...
| |
| 14:03 | or boot off a live CD or some other OS
| |
| 14:04 | <wurzzero> i'm in other OS, Mandriva 2010, a make the changes here and then reboot to test, but its not working
| |
| 14:04 | you mean boot in init3?
| |
| 14:06 | i'll try to boot in single-user, but i need to find out how to....
| |
| 14:07 | i Mandriva i usually modify the /etc/inittab, but in ubuntu there isn't...
| |
| 14:07 | <vagrantc> i don't think ubuntu implements runlevel distinctions
| |
| 14:07 | though upstart may be different than debian here...
| |
| 14:07 | look at /boot/grub/menu.list and see if there's options to disable the hidden menu, and if there's boot options for single user mode...
| |
| 14:08 | <wurzzero> all right...
| |
| 14:08 | <vagrantc> actually, you could just disable gdm from there...
| |
| 14:09 | or uninstall whatever version of gdm you have...
| |
| 14:09 | and re-install the un-patched one from ubuntu...
| |
| 14:09 | if that really caused your problem...
| |
| 14:14 | hersonls has quit IRC | |
| 14:15 | <wurzzero> but, i use gdm ... i can not do that, i'll boot in single-user to test ....i'll be back in some minutes...
| |
| 14:16 | wurzzero has quit IRC | |
| 14:41 | <Lns> stgraber: Has anyone used ltsp-cluster on Debian ?
| |
| 14:45 | johnny has left #ltsp | |
| 14:54 | pmatulis has quit IRC | |
| 14:57 | wurzzero has joined #ltsp | |
| 14:58 | <wurzzero> man, didn't work
| |
| 14:59 | after disable ltsp-sirvecole, i booted in the 10.04, deleted the /opt/ltsp/i386 and the ltsp-build-client, but it shutdown again
| |
| 15:02 | Jul 12 16:58:49 localhost kernel: [ 19.650879] ppdev: user-space parallel port driver Jul 12 16:58:50 localhost pulseaudio[1796]: lock-autospawn.c: Cannot access autospawn lock.
| |
| 15:05 | <vagrantc> wurzzero: what is this ltsp-sirvecole ?
| |
| 15:05 | wurzzero: that's not part of LTSP as far as i know...
| |
| 15:06 | Lns: it's long been on my todo list to try the ltsp-cluster stuff
| |
| 15:07 | but i haven't
| |
| 15:08 | <_UsUrPeR_> atkuepker: ping?
| |
| 15:09 | <wurzzero> man, i do not know what it is, i've installed ltsp-server-standalone........
| |
| 15:10 | <Lns> vagrantc: same here. Let me know how I can help at all, I'm going to set up some VMs to test an install out on
| |
| 15:10 | <wurzzero> i'll install that package in my virtualbox to see what it installs...
| |
| 15:11 | <vagrantc> wurzzero: on a hunch, it sounds spanish...
| |
| 15:12 | it's part of ltsp-controlaula ... which is developed by a big LTSP deployment in extremadura, spain
| |
| 15:12 | <_UsUrPeR_> can anybody give me some pointers on setting up a mknbi etherboot image instead of elf in ubuntu 10.04?
| |
| 15:13 | <vagrantc> _UsUrPeR_: why do you need to?
| |
| 15:13 | <_UsUrPeR_> because some of our older units do not support elf or gpxe
| |
| 15:13 | <wurzzero> man, thats it...
| |
| 15:14 | <vagrantc> wurzzero: you uninstalled it?
| |
| 15:14 | <wurzzero> i messed everytihng here, i think the best is to uninstall everything and the do a new install.....
| |
| 15:15 | <vagrantc> _UsUrPeR_: install the mknbi package in the chroot ...
| |
| 15:15 | wurzzero: or re-install from scratch
| |
| 15:16 | _UsUrPeR_: and then run: ltsp-chroot /usr/share/ltsp/update-kernels ; ltsp-chroot update-initramfs -u ; ltsp-update-kernels
| |
| 15:16 | <wurzzero> i was testing new tools, and installed the 'ltsp-controaula', thought it would help me...
| |
| 15:16 | <vagrantc> _UsUrPeR_: that will generate legacy-nbi.img files, that you'll need to configure DHCP to tell those clients to load.
| |
| 15:16 | no idea if it works.
| |
| 15:17 | <_UsUrPeR_> vagrantc: thanks :)
| |
| 15:17 | I'll do this now
| |
| 15:17 | <vagrantc> _UsUrPeR_: this is presuming ubuntu or debian
| |
| 15:17 | <_UsUrPeR_> yeah, ubuntu 10.04
| |
| 15:17 | <wurzzero> all right, tanks man!
| |
| 15:17 | i'll be back tomorow ...bye
| |
| 15:21 | wurzzero has quit IRC | |
| 15:34 | alkisg has joined #ltsp | |
| 15:37 | irule has joined #ltsp | |
| 15:48 | Selveste1 has joined #ltsp | |
| 15:52 | <alkisg> vagrantc: will you be attending debconf 2011?
| |
| 15:53 | <vagrantc> alkisg: hope to. it sounds like a really excellent opportunity!
| |
| 15:53 | <alkisg> I'll try to attend it as well... If you want to visit Greece too, I can give you a ride & a room to stay :)
| |
| 15:54 | Is there a date set?
| |
| 15:54 | <vagrantc> no date yet
| |
| 15:54 | <alkisg> k
| |
| 15:57 | <vagrantc> well, i'm hopeful something will work out :)
| |
| 16:03 | jhutchins_lt has quit IRC | |
| 16:09 | Lns has quit IRC | |
| 16:11 | jhutchins_lt has joined #ltsp | |
| 16:13 | <nutron> ok can one differentiate between 32 bit and 64 bit client chroots? if so, how do I map the macs to use specific architecture locations?
| |
| 16:13 | if there's a doc on this, just point me to that, google (or myself) is failing me
| |
| 16:14 | <alkisg> You can easily do it manually (with hardcoded mac addresses in dhcpd.conf), but there's also some preliminary support for autodetection in latest versions of pxelinux, which method are you interested in?
| |
| 16:15 | (and, you do know that there won't be much difference for regular ltsp clients, as they only boot a minimal system and the rest of the programs run on the server, right?)
| |
| 16:15 | <nutron> alkisg: whichever you suggest. I was just reading the man page for ltsp-build-client, seems that I can only choose i386 or powerpc architectures. Does this mean that native amd64 is not possible for clients?
| |
| 16:15 | <alkisg> Sure it is, --arch amd64
| |
| 16:16 | (or if the server is already amd64, then that's the default)
| |
| 16:16 | <nutron> --arch i386 | powerpc Builds an i386 chroot on an amd64 or a powerpc chroot on a ppc64. An error will be presented if the arch chosen is unsupported.
| |
| 16:16 | oh I see.
| |
| 16:17 | I guess the debian wiki is more unclear. Not ltsp's fault.
| |
| 16:17 | <alkisg> But it's usually not worth it to have 2 chroots just to get a login screen on the clients. Unless you have very good clients and want to run them as fat clients or something like that..
| |
| 16:18 | <nutron> alkisg: I think I'll just manually configure the dhcp config to select the correct arch, that seems straightforward enough.
| |
| 16:18 | alkisg: I think you're right, I was contemplating just using the lowest common denominator. I have 20+ dual core amd64 machines, as clients (going from diskfull workstations to thin clients)
| |
| 16:19 | <alkisg> Another way to do it is with pxelinux.cfg/01-mac-address files, which point to different ports on the ltsp server.
| |
| 16:19 | <nutron> But I don't have a compelling reason to go amd64.
| |
| 16:19 | oh I've never heard of that, I'll check into it.
| |
| 16:19 | <alkisg> Then I'd suggest you just use an i386 chroot, I don't think it's worth it to maintain 2 thin chroots.
| |
| 16:20 | And if the clients have more than 4 GB RAM, just use a -pae kernel.
| |
| 16:20 | <nutron> alkisg: You're probably right, just seems like a waste, though I am removing some of the client ram sticks, to dumb them down and re-distribute the wealth so to speak
| |
| 16:20 | alkisg: after reading the documentation, it seems like I can get away with a gig or two.
| |
| 16:20 | no?
| |
| 16:21 | <alkisg> For thin clients? With no localapps? 256 MB should be more than enough...
| |
| 16:21 | <nutron> alkisg: ... seriously?
| |
| 16:21 | Nice
| |
| 16:21 | <alkisg> Sure, they're "thin" clients ;)
| |
| 16:21 | But well... in my installations, any PCs that have more than 512 MB RAM, I use them as fat clients instead.
| |
| 16:22 | <nutron> Ok, figured the extra ram could be used for caching or what have you, but if it's not needed, I'm gonna yoink it.
| |
| 16:22 | <alkisg> (fat == all apps run locally)
| |
| 16:22 | Or if you use localapps (e.g. local firefox + flash), then the ram will be used there too.
| |
| 16:22 | <nutron> alkisg: oh, is that an easy "setting"? or would it run like trying to maintain two chroots?
| |
| 16:23 | I remember something in the docs about it, but I'm unclear on how it's configured. I'll go re-read that section
| |
| 16:23 | <alkisg> It's easy enough but I think it's only tested in ubuntu for the time being. One fat chroot can serve both fat + thin clients.
| |
| 16:23 | <nutron> ahh, well I'm using debian :/
| |
| 16:24 | <alkisg> So localapps probably is the best way to utilize your good clients.
| |
| 16:24 | <nutron> ok, arch doesn't sound like the answer then, fat clients does
| |
| 16:24 | <alkisg> Nope, fat ==> ubuntu, so for debian ==> go for localapps
| |
| 16:24 | <vagrantc> i've done some testing with fat clients on debian
| |
| 16:25 | <alkisg> How did it work out?
| |
| 16:25 | <nutron> alkisg: can I ask what your "general" setup is? ie, what number of clients do you run off of a single server?
| |
| 16:25 | <vagrantc> it basically works, though anything that requires authentication (namely, locking screensavers) is a problem.
| |
| 16:25 | since they don't have proper pam authentication
| |
| 16:25 | <nutron> alkisg: oh did I miss what you meant about localapps? Local to the server then? Not the client?
| |
| 16:25 | <alkisg> I think we have a gconf setting that disables screensavers in the last versions...
| |
| 16:26 | <nutron> Uhh :) I think you lost me. Let me re-read
| |
| 16:26 | :P
| |
| 16:26 | <alkisg> nutron: yeah read about localapps, this way you'll e.g. run firefox locally on the client. That greatly reduces your server requirements
| |
| 16:26 | <vagrantc> there are approximately 4 possibilities ...
| |
| 16:27 | thin clients, thin clients + local apps, fat clients, fat clients + remote apps
| |
| 16:27 | <nutron> localapps is defined in the pdf manual?
| |
| 16:27 | vagrantc: ahh I see.
| |
| 16:27 | So I just tag which apps I want to run local to the client?
| |
| 16:27 | <vagrantc> local apps run on the thin client hardware, remote apps run on the server.
| |
| 16:28 | and install them in the chroot, yes.
| |
| 16:28 | localapps also have the problem with anything that requires authentication
| |
| 16:28 | <nutron> Oh! I see, I already looked at that, my terminology is shoddy as I'm new to this. Already tried that... good stuff.
| |
| 16:29 | vagrantc: but apps such as icedove don't require pam, so it must be good to go right?
| |
| 16:29 | <vagrantc> possibly
| |
| 16:29 | <nutron> Ok, I'll try it.
| |
| 16:30 | <vagrantc> nutron: the versions of ltsp related packages in debian lenny/stable don't support that stuff, but it is supported with backports.org
| |
| 16:30 | http://wiki.debian.org/LTSP/Howto/Lenny-With-Backports
| |
| 16:31 | or, debian squeeze/testing
| |
| 16:31 | <nutron> vagrantc: oh crappy, ok yeah I read that.
| |
| 16:31 | <vagrantc> of course
| |
| 16:31 | <nutron> vagrantc: do you know if sabayon is an "easy" backport? as it's not available in backports, or lenny
| |
| 16:31 | <vagrantc> we were busy developing a lot of those features right when lenny froze.
| |
| 16:31 | <alkisg> Btw, for those who haven't seen it, toy story 3 is a great movie (and now Debian has more names to choose from :D)
| |
| 16:32 | <vagrantc> don't know much about sabayon
| |
| 16:32 | heh
| |
| 16:32 | alkisg: i don't think it's even dipped into toy story 2 for names...
| |
| 16:32 | <nutron> Debian is stuck to toy story 1 though.
| |
| 16:32 | as bruce worked on it
| |
| 16:32 | <alkisg> Ah, I didn't know that
| |
| 16:33 | <nutron> that's why debian uses toy story 1 names. It's going to run out soon enough though.
| |
| 16:33 | Mip5 has joined #ltsp | |
| 16:33 | <nutron> Unless they start using inanimate names like "dresser" or "floor"
| |
| 16:33 | <alkisg> I think sbalneav managed to backport sabayon to hardy. Is that similar to Lenny?
| |
| 16:33 | (ubuntu 8.04)
| |
| 16:34 | <nutron> Hmm, I'm not sure, though because I've never used it, if it's a standalone program, I can make use of it somehow, but because it depends on Python, that could mean one of a billion different possibilities.
| |
| 16:34 | <Mip5> Hey Gang - I'm having trouble getting clients to be able to login in a new install of ltsp lucid. Anyone up for helping out?
| |
| 16:34 | <vagrantc> lenny is newer on some things, older on other things than hardy
| |
| 16:34 | <alkisg> \ask
| |
| 16:34 | !ask
| |
| 16:34 | <ltspbot> alkisg: "ask" :: Don't ask to ask a question, simply ask it, and if someone knows the answer, they'll respond. Please hang around for at least 15 minutes after asking a question, as not everybody constantly monitors the channel.
| |
| 16:35 | <Mip5> The clients boot fine, present a login screen, appear to login to a desktop, but then quickly jump back to the login prompt.
| |
| 16:35 | <alkisg> nutron: here's the hardy sabayon ppa link, if it helps: https://launchpad.net/~sbalneav/+archive/ppa?field.series_filter=hardy
| |
| 16:36 | !compiz
| |
| 16:36 | <ltspbot> alkisg: "compiz" :: if compiz is giving you problems, one way to disable it for all users is: sudo gconftool-2 --direct --config-source xml:readwrite:/etc/gconf/gconf.xml.mandatory --type string --set /desktop/gnome/session/required_components/windowmanager metacity
| |
| 16:36 | <nutron> alkisg: thanks.
| |
| 16:36 | <alkisg> Mip5: try this ^^
| |
| 16:37 | <Mip5> alkisg - I'll try it.
| |
| 16:39 | <nutron> Simple question based on opinion/preference. Is it not suggested that chroots and other services provided from a server reside under /srv? I guess /opt has been around forever and /srv is new ... so I guess that pretty much answers it.
| |
| 16:39 | <Mip5> alkisg: Worked!!
| |
| 16:39 | <alkisg> Mip5: do file a bug in compiz about it, it's giving pain to lots of people (I don't have any hardware that's affected to file it myself...)
| |
| 16:39 | <vagrantc> nutron: right. /srv would be preferable, but /opt/ltsp was established before /srv was ever even thought of
| |
| 16:41 | <Mip5> alkisg: okay - I'm sort of a nube about bug filing. Is there a way I can easily send the relevant logs or info to help?
| |
| 16:41 | <alkisg> Mip5: write as much info as you can under https://bugs.launchpad.net/ubuntu/+source/compiz
| |
| 16:41 | (click on file bug etc)
| |
| 16:42 | <Mip5> alkisg: okay - will do. thanks again for the snappy help. I spent about 45 googling before I jumped on IRC
| |
| 16:42 | <alkisg> lspci on the client will be required, I guess
| |
| 16:42 | (to submit the pci id of your client graphics card)
| |
| 16:43 | <nutron> vagrantc: Hmm. ltsp-build-client failed due to dependencies, this a common bug in lenny? or did I miss something?
| |
| 16:43 | <vagrantc> nutron: failing with lenny seems really unlikely.
| |
| 16:43 | nutron: were you using the backports?
| |
| 16:43 | <nutron> vagrantc: yeah, it's why I asked
| |
| 16:44 | vagrantc: no, plain 'ol lenny
| |
| 16:44 | <alkisg> !learn restore_compiz if you disabled compiz by running the command suggested by the !compiz factoid, and you want to revert the change, run: sudo gconftool-2 --direct --config-source xml:readwrite:/etc/gconf/gconf.xml.mandatory --recursive-unset /desktop/gnome/session/required_components
| |
| 16:44 | <ltspbot> alkisg: (learn [<channel>] <key> as <value>) -- Associates <key> with <value>. <channel> is only necessary if the message isn't sent on the channel itself. The word 'as' is necessary to separate the key from the value. It can be changed to another word via the learnSeparator registry value.
| |
| 16:44 | <vagrantc> nutron: it's been a while since i've tried a plain install...
| |
| 16:44 | <alkisg> !learn restore_compiz as if you disabled compiz by running the command suggested by the !compiz factoid, and you want to revert the change, run: sudo gconftool-2 --direct --config-source xml:readwrite:/etc/gconf/gconf.xml.mandatory --recursive-unset /desktop/gnome/session/required_components
| |
| 16:44 | <ltspbot> alkisg: The operation succeeded.
| |
| 16:45 | <nutron> ok I'll try backports
| |
| 16:45 | irule has quit IRC | |
| 16:46 | <vagrantc> nutron: either way, it seems unlikely.
| |
| 16:47 | nutron: without a full log, my best guess would be transient network failures
| |
| 16:47 | irule has joined #ltsp | |
| 16:47 | <nutron> vagrantc: it worked on a subsequent retry... odd
| |
| 16:47 | pmatulis has joined #ltsp | |
| 16:48 | <nutron> net/sources issues I'm sure
| |
| 16:53 | yanu has quit IRC | |
| 16:53 | <vagrantc> i tend to use approx to cache downloads, which minimizes the likelihood of network failures
| |
| 16:55 | * alkisg loves ltsp-build-client --mount-package-cache :) | |
| 16:58 | yanu has joined #ltsp | |
| 16:58 | yanu has joined #ltsp | |
| 17:05 | <vagrantc> it's useful, though it stalls several apt operations on the server for as long as the build is running.
| |
| 17:23 | <nutron> seems that on my second attempt I failed to add the --arch i386 option, so it built an amd64 chroot with the correct boot files under /var/lib/tftpboot ... I'm retrying the i386 option .. I'm waiting to see how it works
| |
| 17:27 | question: I changed the tftproot directory when I installed a different tftp server. With ltsp-build-client can I specify where the kernel installs ought to be?
| |
| 17:28 | so instead of /var/lib/tftproot it's /srv/tftproot
| |
| 17:28 | vagrantc has quit IRC | |
| 17:28 | <alkisg> I think if you move that to /srv/tftp instead, it'll be populated automatically
| |
| 17:29 | grep srv `which ltsp-update-kernels`
| |
| 17:29 | <nutron> if I just move the directory?
| |
| 17:30 | <alkisg> If you mkdir -p /srv/tftp
| |
| 17:30 | BenG has joined #ltsp | |
| 17:30 | <nutron> oh, ok. So I should just move my current contents of /srv/tftproot to /srv/tftp? and then I'm good to go?
| |
| 17:31 | <alkisg> Yes, but you'll need to update your tftpd as well
| |
| 17:31 | <nutron> indeed
| |
| 17:31 | <alkisg> When you run ltsp-build-client, at the end ltsp-update-kernels is called,
| |
| 17:31 | and that copies the kernels from the chroot to all of those dirs, if they exist: /var/lib/tftpboot /tftpboot /srv/tftp
| |
| 17:32 | <nutron> aye, I saw that .. I think that's when the i386 version failed, waiting to see if it'll work this time
| |
| 17:32 | I see
| |
| 17:32 | <alkisg> (at least in the ltsp trunk version, not sure about lenny - run the grep command I wrote above to be sure)
| |
| 17:32 | <nutron> perfect
| |
| 17:32 | it returns nothing
| |
| 17:32 | (the grep command)
| |
| 17:32 | <alkisg> Hmm can you paste your ltsp-update-kernels to pastebin?
| |
| 17:32 | !pastebot
| |
| 17:32 | <ltspbot> alkisg: "pastebot" :: The LTSP pastebot is at http://ltsp.pastebin.com. Please paste all text longer than a line or two to the pastebot, as it helps to reduce traffic in the channel. Don't forget to paste the URL of the text here.
| |
| 17:32 | <nutron> k
| |
| 17:34 | http://ltsp.pastebin.com/Aa0YBMHA
| |
| 17:35 | <alkisg> Hmmm yeah that's a little old. You'll need to create a /etc/ltsp/ltsp-update-kernels.conf file,
| |
| 17:35 | and put TFTPDIRS=yourdir there
| |
| 17:35 | If you do that before running ltsp-build-client, it should then work automatically.
| |
| 17:35 | <nutron> oh nice
| |
| 17:35 | alkisg: thanks very much
| |
| 17:38 | of course, if I'd read that myself, I would've figured it out heh :/ I hate beeing a noob
| |
| 17:38 | s/beeing/being/
| |
| 17:40 | dobber_ has quit IRC | |
| 17:40 | BenG has quit IRC | |
| 17:40 | <nutron> are there many admitted users in business using ltsp? I wonder if this will fix our upgrade and maintenance hell
| |
| 17:42 | <alkisg> There are installations with thousands of ltsp clients, e.g. 5000 of them...
| |
| 17:42 | I'm using ltsp in schools though with just a dozen of clients each, so I can't tell you about big installations
| |
| 17:42 | <nutron> holy ... ok seems like my ~80 workstation setup is pretty dinky in comparison
| |
| 17:43 | <alkisg> It _did_ solve our maintainance problems though :)
| |
| 17:43 | <nutron> good to hear
| |
| 17:43 | I'm pretty excited about all of this
| |
| 17:50 | hmm... am I missing something? tftpd is saying that ltsp/i386/pxelinux.0 is not found... but.. it's there...
| |
| 17:51 | <alkisg> Did you restart tftpd?
| |
| 17:51 | (btw which tftpd server are you using?)
| |
| 17:57 | <nutron> nevermind I got it (I'm using atftpd) and yeah I restarted it
| |
| 17:57 | for some reason the command line was screwed up and used "5/" as the root directory... heh I think it's a typo, I'm trying to fix it atm
| |
| 17:58 | alkisg has quit IRC | |
| 17:59 | <nutron> Yeah .. I forgot the --verbose=5 (the equals) and it used 5 as the root dir... ouch .. lame typo
| |
| 18:00 | ok, so now it's on to configure it. Out of the box I should be able to log in as I'm a user already on the server correct?
| |
| 18:00 | as right now it just goes back to the login screen
| |
| 18:00 | must be something else
| |
| 18:12 | can you clarify something rather simple that I've missed along the way? Do I add a user in the chroot for him/her to exist in there? or does ltsp actually run from the users inside the server?
| |
| 18:15 | Mip5 has quit IRC | |
| 18:22 | UtrinqueParatus has joined #ltsp | |
| 18:23 | <UtrinqueParatus> what are the cheapest terminals you can get for use with ltsp
| |
| 18:24 | <nutron> a pentium 3 with 256mb ram? $50?
| |
| 18:41 | staffencasa has quit IRC | |
| 18:51 | <nutron> how do I look at the ldm log? I'm trying to login and it's telling me it's incorrect (on VT1 and just returns to login in ldm)
| |
| 18:56 | <UtrinqueParatus> how about a mini terminal aesthetically pleasing
| |
| 18:59 | Selveste1 has quit IRC | |
| 19:11 | vagrantc has joined #ltsp | |
| 19:17 | <map7> UtrinqueParatus, Shuttle X27D seems pretty good and cheap
| |
| 19:19 | It has dual screen as well, for $233AU
| |
| 19:19 | +Ram
| |
| 19:20 | <vagrantc> nutron: you can configure the tftproot, though by default it will copy to /var/lib/tftpboot, /tftpboot, /srv/tftp
| |
| 19:21 | <UtrinqueParatus> thanks map7
| |
| 19:21 | <vagrantc> nutron: you can set TFTPBOOTDIR in /etc/ltsp/ltsp-update-kernels.conf
| |
| 19:22 | but TFTPBOOTDIR should be the same as the root of your tftp server, otherwise it might have the paths in the generated pxelinux config files wrong
| |
| 19:22 | <map7> UtrinqueParatus, You could also put together your own, these are often cheaper. Something like a Intel D945GSEJT Motherboard and a M350 case, check out www.mini-box.com
| |
| 19:23 | These can be built for $222AU + Ram
| |
| 19:23 | <UtrinqueParatus> for the sake of $11 for a deployment of 500 probably not worth it
| |
| 19:24 | <map7> no not really but mini-box can put configurations together for you for $22 and they have some cool boards and stuff
| |
| 19:24 | might be worth checking out.
| |
| 19:25 | <UtrinqueParatus> yeah will do
| |
| 19:52 | daynov has joined #ltsp | |
| 19:57 | daynov has left #ltsp | |
| 20:06 | irule has quit IRC | |
| 20:10 | UtrinqueParatus has quit IRC | |
| 20:26 | <nutron> ok... one more... the rest I've figured out
| |
| 20:26 | how do allow only a single login ... in the subnet?
| |
| 20:26 | <vagrantc> nutron: not sure i get what you mean... ?
| |
| 20:27 | <nutron> vagrantc: well. I can login as nutron on client1, and client2, and client3 etc etc etc ad infinitum. gconf and others get confused...
| |
| 20:27 | <vagrantc> ah.
| |
| 20:27 | there's some lts.conf value for that...
| |
| 20:27 | !lts.conf
| |
| 20:27 | <ltspbot> vagrantc: "lts.conf" :: http://manpages.ubuntu.com/lts.conf
| |
| 20:28 | <nutron> crap... I looked already, must've missed it. Sorry.
| |
| 20:28 | <vagrantc> might not be documented there...
| |
| 20:28 | hrm. manpage wasn't found...
| |
| 20:29 | LDM_LIMIT_ONE_SESSION=true
| |
| 20:29 | <nutron> oh! .. where can I see that documented?
| |
| 20:30 | <vagrantc> in the source code, apparently...
| |
| 20:30 | <nutron> :/ I don't have that :P
| |
| 20:31 | <vagrantc> there's also LDM_LIMIT_ONE_SESSION_PROMPT, which will propmpt the user if it detects an old session before blindly killing the old processes
| |
| 20:31 | nutron: /opt/ltsp/i386/usr/share/ldm/rc.d/S15-userLoginCheck
| |
| 20:32 | <nutron> perfect, thanks vagrantc
| |
| 20:32 | * vagrantc adds to the docs... | |
| 20:35 | <nutron> hmm doesn't seem to do much in 5.1... think I'll have to go to that backports version I was intending to in the first place
| |
| 20:35 | <vagrantc> yes, you'll need the backports version
| |
| 20:38 | <nutron> ugh crap...
| |
| 20:38 | dependency problems going to backports...
| |
| 20:39 | <vagrantc> i'll try it myself and see if i have the same issue
| |
| 20:39 | nutron: you following the instructions specified on the wiki i referenced earlier?
| |
| 20:39 | <nutron> vagrantc: yessir
| |
| 20:39 | aptitude -t lenny-backports install ltsp-server-standalone
| |
| 20:40 | <vagrantc> ltsp-info?
| |
| 20:40 | !pastebot
| |
| 20:40 | <ltspbot> vagrantc: "pastebot" :: The LTSP pastebot is at http://ltsp.pastebin.com. Please paste all text longer than a line or two to the pastebot, as it helps to reduce traffic in the channel. Don't forget to paste the URL of the text here.
| |
| 20:41 | <vagrantc> nutron: could you paste the results of "ltsp-info" ?
| |
| 20:41 | <nutron> I don't have an ltsp-info
| |
| 20:41 | I'll run a find
| |
| 20:41 | <vagrantc> dpkg -l 'ltsp*'
| |
| 20:42 | <nutron> is it in ltsp-utils?
| |
| 20:42 | <vagrantc> no
| |
| 20:42 | it's in recent versions of ltsp-server
| |
| 20:42 | <nutron> http://ltsp.pastebin.com/x88FACiH
| |
| 20:43 | <vagrantc> nutron: that's clearly not the version from backports.org
| |
| 20:43 | <nutron> yeah
| |
| 20:43 | :) I have conflicts going to backports
| |
| 20:43 | <vagrantc> paste the results?
| |
| 20:43 | <nutron> I'm trying to paste the results
| |
| 20:43 | sec
| |
| 20:44 | http://ltsp.pastebin.com/bqEbtUSH
| |
| 20:44 | <vagrantc> i thought you meant conflicts when running ltsp-build-client ...
| |
| 20:45 | nutron: ah.
| |
| 20:45 | <nutron> :o no. I don't understand why ldm server is being kept back
| |
| 20:45 | <vagrantc> nutron: try with apt-get
| |
| 20:45 | <nutron> k
| |
| 20:45 | that worked... wth... :/
| |
| 20:46 | * vagrantc is not a fan of aptitude | |
| 20:46 | <vagrantc> nutron: you weren't following the instructions after all :P
| |
| 20:46 | <nutron> I wasn't?
| |
| 20:46 | What'd I do wrong?
| |
| 20:46 | oh lol
| |
| 20:46 | aptitude
| |
| 20:46 | heh
| |
| 20:46 | <vagrantc> aptitude vs. apt-get
| |
| 20:46 | <nutron> ok
| |
| 20:46 | now i have to restart the clients
| |
| 20:47 | and re-do the build
| |
| 20:47 | <vagrantc> nutron: if you said "n" at "Accept this solution?" it would have tried another way to resolve the issue
| |
| 20:47 | <nutron> probably, but I always feel that aptitude is on crack
| |
| 20:47 | it's been forced on me
| |
| 20:47 | <vagrantc> actually, answering "y" should've worked too
| |
| 20:49 | actually, my instructions don't suggest which version to even install..
| |
| 20:49 | nutron: how did you configure apt pinning?
| |
| 20:53 | makghosh|afk has quit IRC | |
| 20:54 | <nutron> vagrantc: I didn't specifically, just told apt the default distro is lenny
| |
| 20:54 | vagrantc: I should set backports packages to be preferred
| |
| 20:54 | then I wouldn't have started with the version in lenny
| |
| 20:54 | as I'm trying to keep it "clean and safe"
| |
| 20:55 | I see that ldm has themes... is there a doc to walk me through creating a custom one for the company? :/
| |
| 20:57 | Ok I'll figure that one out, google gave me some hope there
| |
| 20:58 | <vagrantc> just copy the directory from one of the existing themes and mangle it till you have something you like.
| |
| 20:58 | <nutron> Errm, I set the hostname in dhcp, does ltsp not ... assign it? I'd like it to read the hostname (for example .. workstation55)
| |
| 20:58 | vagrantc: Yeah I saw that on a mailing list archive :P, I'll do just that
| |
| 20:59 | <vagrantc> it's not totally clear to me how early hostname stuff works...
| |
| 20:59 | <nutron> k it's not a big deal
| |
| 20:59 | Maybe I can get it figured out. I just want to be able to have the managers assign a seat, then ... ie workstation55 is where john sits etc...
| |
| 20:59 | <vagrantc> you can configure the autogenerated hostname with HOSTNAME variables, also just recently added to lts.conf
| |
| 21:00 | (i.e. not yet released)
| |
| 21:00 | <nutron> :o crap... heh
| |
| 21:00 | <vagrantc> the code supports it, it's just not yet in a released version of the docs
| |
| 21:01 | <nutron> oh oh I see. Is there a cvs/svn code viewer that I can read to understand how to use it?
| |
| 21:02 | <vagrantc> https://code.launchpad.net/~ltsp-docwriters/ltsp/ltsp-docs-trunk
| |
| 21:02 | http://bazaar.launchpad.net/~ltsp-docwriters/ltsp/ltsp-docs-trunk/files
| |
| 21:02 | <nutron> wow, great, thanks
| |
| 21:02 | <vagrantc> it's in reasonably human-readable xml
| |
| 21:03 | just added LDM_LIMIT_ONE_SESSION*
| |
| 21:06 | <nutron> :P
| |
| 21:06 | I'm still waiting for ltsp-build-client to finish.
| |
| 21:07 | <vagrantc> may take some time, yes.
| |
| 21:07 | nutron: you did run it with the --backports-mirror option?
| |
| 21:10 | pmatulis has quit IRC | |
| 21:17 | <nutron> vagrantc: yes i did
| |
| 21:17 | sorry I went afk there for a few mins
| |
| 21:20 | ok I'm going to go check to see if the limit one session setting worked
| |
| 21:22 | hmm didn't work
| |
| 21:23 | I used LDM_LIMIT_ONE_SESSION_PROMPT .. do I need to use it in conjunction with LDM LIMIT ONE SESSION?
| |
| 21:24 | <vagrantc> as the docs state, yes. :)
| |
| 21:25 | the just written 10 minutes ago docs :)
| |
| 21:25 | <nutron> oh :)
| |
| 21:25 | thanks
| |
| 21:26 | <vagrantc> LTSP feels like it's finally stabalized a bit, so whenever i run across some undocumented feature i take a couple seconds to document it.
| |
| 21:26 | <nutron> question... when the client now boots, tar complains that certain files are in the future. is this a timezone thing now? I have ntp set, so.. hmm I'll go check timezone
| |
| 21:26 | You seem to be very good for the project, I appreciate the time you're taking with me.
| |
| 21:27 | oh fantastic, it works now
| |
| 21:27 | <vagrantc> well, i'm the primary maintainer for the debian side of things... so i do what i can :)
| |
| 21:28 | most LTSP users appear inclined towards ubuntu, and while i appreciate ubuntu, my heart is definitely with debian.
| |
| 21:28 | <nutron> This is so great, I have no idea why I discounted it the last 28 times I looked at it. I think maybe because of the website...
| |
| 21:29 | Yeah, I have some serious angst against ubuntu, but I'm sure that's a discussion for another time/channel :P
| |
| 21:29 | <vagrantc> we're not always the greatest with presentation
| |
| 21:29 | <nutron> nah, it wasn't presentation
| |
| 21:30 | last week when I weighed installs between fai and ltsp, I chose fai, because the website sends me to commercial offerings. Not that I'm against it, but the docs links sent me to distro specific locations... hard to explain but it didn't feel like one cohesive project
| |
| 21:30 | <vagrantc> we've had a pretty good cross-distro project with ltsp5...
| |
| 21:30 | <nutron> so I figured it was some sort of shell .. presented by hp or the dw.com site etc.
| |
| 21:31 | <vagrantc> the fai folks are very much community driven
| |
| 21:31 | <nutron> but that was purely based on first impression, I'm not knocking it, just explaining my thought process.
| |
| 21:31 | <vagrantc> but it's also focused on Installing... the thin-client stuff is sort of a byproduct of the real goal of the project
| |
| 21:31 | <nutron> err I meant the ltsp site sent me to commercial offerings
| |
| 21:31 | <vagrantc> oh!
| |
| 21:32 | <nutron> well I do like it, i was very concerned about maintainability. we have 12 routers here that boot off of nfs, using pxe etc... and they're a total pain in the butt to maintain. I'm sure glad I went with ltsp
| |
| 21:32 | and you were on the channel to help :)
| |
| 21:33 | <vagrantc> i seem to be in the mood for it today... sometimes i'm grumpier :)
| |
| 21:33 | <nutron> :P that's the usual mood for debian guys
| |
| 21:34 | I know, I've been using deb for ... ages
| |
| 21:34 | and I still feel retarded
| |
| 21:36 | <vagrantc> whoah... the default file download on the sourceforge site is an ltsp 4.1 .iso
| |
| 21:39 | <nutron> there was someone talking about that before on here today, about having recent snapshots available.
| |
| 21:41 | <vagrantc> yeah... i should just go ahead and set something up...
| |
| 21:54 | artista_frustrad has joined #ltsp | |
| 22:36 | alexqwesa has quit IRC | |
| 22:38 | alexqwesa has joined #ltsp | |
| 22:47 | try2free has joined #ltsp | |
| 23:23 | <nutron> Ok I'm having a hard time with the thin client hostname. I can reverse lookup the ip, forward, and its set in dhcp, yet I still get ltsp*** hostnames. Do I really need to set HOSTNAME in lts.conf for each machine?
| |
| 23:30 | <vagrantc> there's HOSTNAME_BASE
| |
| 23:30 | and HOSTNAME_EXTRA
| |
| 23:31 | <nutron> hmm yes I saw those
| |
| 23:31 | <vagrantc> so HOSTNAME_BASE defaults to "ltsp"
| |
| 23:31 | <nutron> but I was hoping to use the DNS entries to name them, from what I understand, that should work .. no?
| |
| 23:31 | <vagrantc> and HOSTNAME_EXTRA defaults to something based on IP address.
| |
| 23:32 | <nutron> oh..
| |
| 23:32 | <vagrantc> should, i guess.
| |
| 23:32 | the HOSTNAME_BASE and HOSTNAME_EXTRA should only trigger if it doesn't get the hostname from dhcp/dns ... but maybe that doesn't consistantly work.
| |
| 23:33 | <nutron> i just setup root login in the chroot
| |
| 23:33 | and I'm poking around on the client. ... from what I can tell it works ok.
| |
| 23:33 | <vagrantc> for debugging, SCREEN_08=shell is the way to go...
| |
| 23:33 | <nutron> do you know which script pulls the info?
| |
| 23:34 | <vagrantc> it's done in the initramfs
| |
| 23:34 | <nutron> k I'll go set that up
| |
| 23:34 | <vagrantc> see /opt/ltsp/i386/usr/share/initramfs-tools/scripts/nfs-bottom/ltsp
| |
| 23:35 | <nutron> on boot it says warning: /etc/syslog.conf doesn't exist...
| |
| 23:35 | <vagrantc> that's fine
| |
| 23:35 | <nutron> do I care (this is on the client)
| |
| 23:35 | <vagrantc> default is rsyslog now
| |
| 23:36 | in fact, those warning messages should probably be changed to something less alarming ...
| |
| 23:36 | <nutron> k, oh SCREEN_08=shell kills the X VT.. :o
| |
| 23:36 | <vagrantc> yeah... SCREEN_07=ldm, SCREEN_08=shell
| |
| 23:36 | include both
| |
| 23:37 | it only sets a default if no SCREEN_NN are defined.
| |
| 23:37 | <nutron> ahh, makes sense
| |
| 23:38 | oh and ipconfig says it's using nbi.img... odd
| |
| 23:38 | <vagrantc> that's not used at that point
| |
| 23:38 | <nutron> on boot again, and it also states that SLEEP 0.1 is not a valid number
| |
| 23:38 | <vagrantc> yeah...
| |
| 23:38 | there's a bit of more-or-less useless noise
| |
| 23:39 | <nutron> just thought I'd bring it up as you seem to have the power to change things :)
| |
| 23:39 | <vagrantc> some
| |
| 23:39 | <nutron> I've been ignoring them merrily
| |
| 23:49 | is a simple chroot /opt/... enough to install dnsutils into the chroot? or ... ehh what was I asking... heh
| |
| 23:49 | <vagrantc> ltsp-chroot sets some variables that prevent daemons from starting on the server
| |
| 23:50 | it defaults to using /opt/ltsp/$(dpkg --print-architecture)
| |
| 23:51 | so if you're using i386 chroot on an amd64 server, you'd first want to: export ROOT=/opt/ltsp/i386
| |
| 23:51 | and alkisg rewrote it so it takes commandline options
| |
| 23:51 | try2free has left #ltsp | |
| 23:52 | <nutron> yeah cool I like it
| |
| 23:53 | <vagrantc> the ltsp-chroot rewrite is still just in sid/unstable
| |
| 23:54 | <nutron> oh..
| |
| 23:56 | ok that doesn't work, reverse lookups are fine
| |
| 23:56 | yet still get the pre and post hostname
| |
| 23:56 | I'm gonna have to dig into that initramfs stuff
| |
| 23:59 | Huh, seems that maybe my dhcp was borked
| |
| 23:59 | get-lease-hostnames didn't have true after it
| |