IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 14 May 2012 (all times are UTC)
| 00:51 | raven has left IRC (raven!~raven@114.79.16.121, Quit: Leaving) | |
| 01:38 | Phantomas has left IRC (Phantomas!~Phantomas@unaffiliated/phantomas) | |
| 01:43 | Billy_Way has left IRC (Billy_Way!billy@nat/redhat/x-jgfpqwqjldfnyppn, Quit: Good night ^_^) | |
| 01:45 | Billy_Way has joined IRC (Billy_Way!billy@nat/redhat/x-jhoopddhzyqmaale) | |
| 02:15 | vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Quit: leaving) | |
| 02:35 | [GuS] has left IRC ([GuS]!~gustavo@unaffiliated/gus/x-663402, Quit: Konversation terminated!) | |
| 02:42 | Parker955 is now known as Parker955_Away | |
| 03:10 | adrianorg has left IRC (adrianorg!~adrianorg@187.115.104.79, Ping timeout: 240 seconds) | |
| 03:49 | bieb has left IRC (bieb!~Me@99-2-101-23.lightspeed.tukrga.sbcglobal.net) | |
| 04:49 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 04:59 | andygraybeal_ has left IRC (andygraybeal_!~andy@h174.212.22.98.dynamic.ip.windstream.net, Read error: Operation timed out) | |
| 05:22 | cyberorg has left IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg, Ping timeout: 260 seconds) | |
| 05:43 | cyberorg has joined IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg) | |
| 07:10 | <alkisg> Good morning everyone
| |
| 07:11 | knipwim: may I commit changes to existing files, or will that put overhead with your merge, and I should wait till you fiinish?
| |
| 07:26 | <knipwim> alkisg: i can commit also what i've done yesterday for the server
| |
| 07:26 | almost finished
| |
| 07:26 | <alkisg> knipwim: ok, np, I can wait for today, please ping me if you're done
| |
| 07:28 | bobby_C has joined IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at) | |
| 07:37 | <Hyperbyte> Good morning alkisg!
| |
| 07:39 | ayo572 has joined IRC (ayo572!~ayo@41.177.59.120) | |
| 07:39 | <ayo572> hello everyone
| |
| 07:39 | i need help with running a fat client
| |
| 07:40 | i have configured and its working alright but the only problem is with files that are aved while the fat client is running diont get saved to the server
| |
| 07:40 | instead once i log out i loose all files
| |
| 07:40 | <alkisg> Distro/version?
| |
| 07:40 | <ayo572> i would greatly appreciate any help
| |
| 07:41 | ubuntu 10.04
| |
| 07:42 | @alkisg ....ubuntu 10.04
| |
| 07:45 | <alkisg> ayo572: did you specify anything related to NFS?
| |
| 07:45 | Or you're using the default SSHFS?
| |
| 07:46 | Also, you're using LDM, not GDM, right?
| |
| 07:46 | <ayo572> yes s
| |
| 07:47 | <alkisg> And you didn't create users in the chroot... correct?
| |
| 07:47 | <ayo572> i specified the export directory for nfs
| |
| 07:47 | i did create a user in chroot
| |
| 07:47 | i did the server side config for nfs
| |
| 07:47 | and its working alrite on the server side
| |
| 07:48 | <alkisg> cat /opt/ltsp/i386/etc/X11/default-display-manager
| |
| 07:48 | What's the output?
| |
| 07:49 | asmok has joined IRC (asmok!~asmok@cs78145191.pp.htv.fi) | |
| 07:50 | <ayo572> /usr/sbin/ldm
| |
| 07:51 | <alkisg> Remove any references for NFS from your lts.conf and see if it works that way, and use a server user, not a chroot user.
| |
| 07:52 | Also, don't add users in the chroot, as that might break the ltsp users handling
| |
| 07:52 | <ayo572> i should use a server user ?
| |
| 07:52 | also ncan server users log in to ltsp on the clients
| |
| 07:53 | <alkisg> LDM uses ssh to authenticate the server users on the clients
| |
| 07:53 | So you should *only* be using server users, not chroot users
| |
| 07:53 | <ayo572> ok
| |
| 07:54 | <asmok> alkisg: I have CuBox - http://www.solid-run.com/phpbb/viewtopic.php?f=4&p=2368#p2368 - just for learning experience, i can run my own kernel and 12.04 armhf in Cubox
| |
| 07:54 | <ayo572> so how do i allow userson d server access to ltsp enviroment
| |
| 07:55 | <asmok> alkisg: there is this old project page - http://projects.powerdeveloper.org/project/imx515/799 - stgraber: "The project, would be to test and fix the use of the existing ARM port of Ubuntu on thin clients."
| |
| 07:56 | <alkisg> ayo572: you don't need to allow anything, all server users are allowed by default
| |
| 07:56 | <ayo572> ok yes
| |
| 07:56 | just discovered that
| |
| 07:56 | <alkisg> asmok: cool, but we don't have any arm hardware at all in greek schools, so I haven't yet looked into any arm-related stuff :) vagrantc is your man. :)
| |
| 07:56 | <asmok> alkisg: is that project live?
| |
| 07:57 | alkisg: ok
| |
| 07:59 | <alkisg> asmok: no idea, if stgraber commented there, you may ask him
| |
| 07:59 | <asmok> alkisg: maybe i just try to do something first and then drop in #ubuntu-arm
| |
| 07:59 | <alkisg> Really, no arm experience whatsoever
| |
| 07:59 | <ayo572> still not working
| |
| 08:00 | <alkisg> ayo572: tail /var/log/auth.log => did you see the user login there?
| |
| 08:01 | <ayo572> yes
| |
| 08:01 | <asmok> alkisg: we did have that ltsp workshop with ltsp-pnp - after it is officially out i will use ltsp-pnp in my Mexico project - 512/P4/nVidia Vanta
| |
| 08:02 | <ayo572> its there
| |
| 08:03 | <asmok> alkisg: people was really happy when they understood how easy ltsp system is, thanks for ltsp-pnp and support for fat clients ;-)
| |
| 08:03 | <alkisg> asmok: nice, although there won't be an "ltsp-pnp" package, but an "ltsp-server-dnsmasq" package an a --cleanup option to ltsp-update-image
| |
| 08:03 | I'll post a wiki page when it's ready, in a couple of weeks
| |
| 08:04 | ayo572: did you remove NFS references from your lts.conf?
| |
| 08:04 | <asmok> alkisg,: really great news
| |
| 08:04 | <ayo572> yes i removed it
| |
| 08:05 | <asmok> have a good day, have to go, bye
| |
| 08:05 | dobber has joined IRC (dobber!~dobber@213.169.45.222) | |
| 08:05 | <alkisg> ayo572: the client was able to login, right? can you run cat /proc/mounts ?
| |
| 08:05 | bb asmok
| |
| 08:05 | asmok has left IRC (asmok!~asmok@cs78145191.pp.htv.fi, Quit: Lähdössä) | |
| 08:05 | <alkisg> ayo572: and see there if there are any nfs or sshfs references
| |
| 08:15 | <ayo572> yes client looged in
| |
| 08:16 | ok checking nfs
| |
| 08:16 | i cant run cat /proc/mounts
| |
| 08:17 | <alkisg> Why?
| |
| 08:17 | <ayo572> its say no succh directory
| |
| 08:17 | <alkisg> ... cat /proc/self/mounts ?
| |
| 08:18 | If you don't have /proc mounted, that's a serious problem
| |
| 08:19 | <ayo572> no such file or directory ok
| |
| 08:20 | its showing
| |
| 08:20 | <alkisg> Did you open a gnome-terminal?
| |
| 08:20 | <ayo572> yes
| |
| 08:20 | <alkisg> ls /proc
| |
| 08:20 | Is anything there?
| |
| 08:21 | <ayo572> yes
| |
| 08:21 | <alkisg> ls /proc/mounts
| |
| 08:21 | And this isn't there?
| |
| 08:21 | <ayo572> its is now
| |
| 08:22 | /proc/mounts thats wats there
| |
| 08:22 | <alkisg> You were probably running the wrong command
| |
| 08:22 | cat /proc/mounts
| |
| 08:22 | Does that display the mounts now?
| |
| 08:22 | <ayo572> yes i figured that was my mistake
| |
| 08:22 | yyes
| |
| 08:22 | <alkisg> Do you see nfs or sshfs there?
| |
| 08:22 | <ayo572> it shows /proc/mounts
| |
| 08:22 | yes it sjhoows sshfts
| |
| 08:23 | <alkisg> Does it say /home/<username> ... ?
| |
| 08:23 | <ayo572> yes
| |
| 08:23 | <alkisg> OK, create a file in your desktop, it should appear on the server instantaneously
| |
| 08:23 | <ayo572> home/user home/user fuse.sshfs
| |
| 08:24 | <alkisg> I.e. it should be working now, if you create a file on your desktop and logoff/logon, it should be there again
| |
| 08:24 | <ayo572> yes
| |
| 08:24 | <alkisg> OK so problem solved
| |
| 08:24 | <ayo572> ok restarting now to check
| |
| 08:29 | <knipwim> alkisg: http://bazaar.launchpad.net/~wimmuskee/ltsp/ltsp-restructure-the-tree/revision/2196
| |
| 08:29 | server done, and going into a meeting now
| |
| 08:29 | probably speak to you tonight
| |
| 08:30 | <alkisg> knipwim: push it, we can make further changes later on
| |
| 08:30 | (if needed)
| |
| 08:30 | <ayo572> thank you soooooo much man
| |
| 08:30 | <knipwim> how do i merge these changes from this branch to the ltsp-trunk?
| |
| 08:30 | <alkisg> ayo572: you're welcome
| |
| 08:30 | <ayo572> alkisg u the
| |
| 08:30 | u the man
| |
| 08:31 | <alkisg> knipwim: not sure, maybe a brz push <upstream-url> would do it
| |
| 08:31 | As it's not a diverted branch
| |
| 08:31 | knipwim: try: bzr push bzr+ssh://bazaar.launchpad.net/~ltsp-upstream/ltsp/ltsp-trunk/
| |
| 08:32 | <knipwim> check, looks like it worked :)
| |
| 08:32 | <alkisg> knipwim: ty, talk to you tonight
| |
| 08:35 | <ayo572> ttyl guys
| |
| 08:35 | ayo572 has left IRC (ayo572!~ayo@41.177.59.120, Quit: irc2go) | |
| 08:58 | Steve_The_Pirate has joined IRC (Steve_The_Pirate!~Gary@82-68-255-22.dsl.in-addr.zen.co.uk) | |
| 09:10 | Steve_The_Pirate has left IRC (Steve_The_Pirate!~Gary@82-68-255-22.dsl.in-addr.zen.co.uk, Ping timeout: 272 seconds) | |
| 09:23 | Steve_The_Pirate has joined IRC (Steve_The_Pirate!~Gary@82-68-255-22.dsl.in-addr.zen.co.uk) | |
| 09:45 | dobber has left IRC (dobber!~dobber@213.169.45.222, Remote host closed the connection) | |
| 09:51 | dobber has joined IRC (dobber!~dobber@213.169.45.222) | |
| 09:52 | Trixboxer has joined IRC (Trixboxer!~Trixboxer@115.124.115.71) | |
| 09:55 | CousinLarry has joined IRC (CousinLarry!~CousinLar@112.134.71.228) | |
| 09:55 | <CousinLarry> hi
| |
| 09:57 | got a problem, i got a LTSP setup on ubuntu with a server and about 8 machines, there is a issue with the server and i had to give it for repair and now obviously none of the clients can work, how could i have avoided such a situation ?
| |
| 10:07 | <alkisg> It's the same problem with regular machines or other servers, not LTSP related... You could have a second server, or if the server disk was working, you could just put that in another machine till the original one was repaired
| |
| 10:19 | <CousinLarry> true, but is there such a way 2 servers caz run concurrently perthaps where if one fails operations would continue as normal ?
| |
| 10:22 | <alkisg> Without downtime at all? Sounds difficult... For the hard disks part, there's hardware RAID with NFS capabilities...
| |
| 10:22 | If you knew beforehand that your server is to go down, you could do VM teleporting, but with a hard crash, it sounds difficult
| |
| 10:24 | <CousinLarry> the problem was that the server started to randomly reboot
| |
| 10:25 | <alkisg> The easiest implementation is to have RAID, in case the disks fail (some even support hotswapping), and a second machine where you can place the RAID if other server hardware fails
| |
| 10:25 | This way you only have a couple of minutes downtime, I'm not sure if something more is worth the trouble.
| |
| 10:25 | <CousinLarry> the disk(s) are fine, ive had to send it back to shop for the second time now and since the clients have no disks, nothing can be done
| |
| 10:25 | <alkisg> Right, so you only needed to swap the disks to another machine
| |
| 10:26 | <CousinLarry> :) unfortunately there is only 1 server, is there any documentation to read up on for situations like this with LTSP ?
| |
| 10:26 | beside what we just talked about
| |
| 10:35 | <alkisg> I'm not sure what you're searching about, if you only have 1 server and the hardware has problems, how would the clients boot?
| |
| 10:38 | Do you mean to utilize their local hard disks, if they have any?
| |
| 10:39 | <CousinLarry> no there are none, im just trying to find more information on how this sort of thing can be mitigated
| |
| 10:39 | <alkisg> Without a second server? I can't imaging what you're searching for.
| |
| 10:40 | <CousinLarry> thats what everybody told indiana jones but that never stopped him
| |
| 10:40 | <alkisg> If you can netboot clients without a server (as it'd be sent for repairs), I'm sure many vendors would be interested in hearing about it, as it would be they wouldn't need to ship any hardware then :)
| |
| 10:41 | *wouldn't be any need...
| |
| 10:42 | <CousinLarry> "A single-server LTSP set-up creates a single point of failure scenario. With a LTSP Cluster set of servers, one or more servers can fail and users would be able to log in and continue with work as usual." https://www.ltsp-cluster.org/tour
| |
| 10:42 | what is this
| |
| 10:42 | <alkisg> It needs 2 or more servers
| |
| 10:43 | You're saying you only have 1, and you're asking how can you boot the clients if it breaks, i.e. without a server
| |
| 10:43 | It's just impossible, unless if you use local hard disks or usb sticks or whatever to boot them locally
| |
| 10:43 | <CousinLarry> no, thats what has happened, im trying to find a way so i can prevent this from happening without users being affected
| |
| 10:43 | <alkisg> (01:25:05 μμ) alkisg: The easiest implementation is to have RAID, in case the disks fail (some even support hotswapping), and a second machine where you can place the RAID if other server hardware fails
| |
| 11:03 | leio_ has joined IRC (leio_!~leio@gentoo/developer/leio) | |
| 11:04 | leio has left IRC (leio!~leio@gentoo/developer/leio, Ping timeout: 272 seconds) | |
| 11:21 | andygraybeal_ has joined IRC (andygraybeal_!~andy@h174.212.22.98.dynamic.ip.windstream.net) | |
| 11:46 | Parker955_Away is now known as Parker955 | |
| 11:47 | CousinLarry has left IRC (CousinLarry!~CousinLar@112.134.71.228) | |
| 11:51 | awilliams has left IRC (awilliams!mistik1@unaffiliated/mistik1, Ping timeout: 250 seconds) | |
| 11:51 | awilliams has joined IRC (awilliams!mistik1@unaffiliated/mistik1) | |
| 11:58 | awilliams has left IRC (awilliams!mistik1@unaffiliated/mistik1, Ping timeout: 265 seconds) | |
| 11:58 | awilliams has joined IRC (awilliams!mistik1@unaffiliated/mistik1) | |
| 12:07 | bobby_C has left IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at, Ping timeout: 245 seconds) | |
| 12:10 | [GuS] has joined IRC ([GuS]!~MysT@unaffiliated/gus/x-663402) | |
| 12:29 | muppis has left IRC (muppis!muppis@viuhka.fi, Read error: Operation timed out) | |
| 12:30 | muppis has joined IRC (muppis!muppis@viuhka.fi) | |
| 12:31 | awilliams has left IRC (awilliams!mistik1@unaffiliated/mistik1, Read error: Connection reset by peer) | |
| 12:31 | awilliams has joined IRC (awilliams!mistik1@unaffiliated/mistik1) | |
| 12:35 | <alkisg> stgraber: currently we're using nbdroot=:ltsp_i386 in the command line, because that's the name of our export in /etc/nbd-server/conf.d/ltsp_i386.conf. But nbdroot=, similarly to nfsroot=, can be defined in the root-path option of DHCP.
| |
| 12:35 | In order to avoid having to change the DHCP configuration, I propose that we rename the export name of our files to [/opt/ltsp/i386], so that the same root path works for NFS and NBD too (sadly that's what I originally proposed but vagrantc thought :ltsp_i386 was better).
| |
| 12:35 | This will break compatibility with existing nbd-server configuration, we'll have to fix them on postinst (of 12.04 => 12.10+ upgrades). Is that OK?
| |
| 12:37 | awilliams has left IRC (awilliams!mistik1@unaffiliated/mistik1, Ping timeout: 244 seconds) | |
| 12:37 | sideffect has left IRC (sideffect!sideffect@gateway/shell/bshellz.net/x-mwvmfmqjclwlohml, Ping timeout: 245 seconds) | |
| 12:38 | awilliams has joined IRC (awilliams!mistik1@unaffiliated/mistik1) | |
| 12:38 | sideffect has joined IRC (sideffect!sideffect@gateway/shell/bshellz.net/x-lemgslyxeizjksks) | |
| 12:39 | <alkisg> Instead of putting the upgrade code in postinst, I can also put it in ltsp-update-image, if you think it's more appropriate
| |
| 12:41 | dsorf has left IRC (dsorf!~Miranda@ip-89-176-9-128.net.upcbroadband.cz, Ping timeout: 265 seconds) | |
| 12:45 | awilliams has left IRC (awilliams!mistik1@unaffiliated/mistik1, Ping timeout: 240 seconds) | |
| 12:45 | awilliams has joined IRC (awilliams!mistik1@unaffiliated/mistik1) | |
| 12:46 | vmlintu__ has left IRC (vmlintu__!~vmlintu@nblzone-240-143.nblnetworks.fi, Remote host closed the connection) | |
| 12:59 | <eddytv> Looking for suggestions on how to track down why /tmp on my "fat" LTSP clients ends up with mode 755 instead of 1777.
| |
| 13:00 | bengoa has joined IRC (bengoa!~bengoa@2001:1291:229:2:216:cbff:feab:6cc9) | |
| 13:00 | <eddytv> This is preventing X from starting. If I ssh in to the client and chmod /tmp, X starts.
| |
| 13:08 | <alkisg> eddytv: grep tmp /proc/mounts on the client, is there something there?
| |
| 13:08 | grep /tmp /proc/mounts, better
| |
| 13:11 | <eddytv> Nope
| |
| 13:12 | There are other 'tmpfs' mounts though, like /run and /cow
| |
| 13:13 | <alkisg> eddytv: put break=50-update-motd in the kernel command line (pxelinux.cfg/default)
| |
| 13:14 | At that point, check /tmp, if it has the correct attributes
| |
| 13:17 | <eddytv> Is there a way to know if that "break" really happened?
| |
| 13:18 | <alkisg> You'll get a shell. Ah, and remove quiet splash plymouth etc if you have them
| |
| 13:18 | You put that in the append line, right?
| |
| 13:18 | <eddytv> Right. I added it, ssh'd in, and /tmp is still unmounted / 755
| |
| 13:19 | <alkisg> Errr no you should get a local shell, before sshd is started
| |
| 13:19 | Pastebin your current pxelinux.cfg/default
| |
| 13:19 | That's ubuntu 12.04, right?
| |
| 13:20 | /var/lib/tftpboot/ltsp/<arch>/pxelinux.cfg/default
| |
| 13:21 | <eddytv> http://pastebin.com/4kDW3trF
| |
| 13:22 | <alkisg> remove vt.handoff too
| |
| 13:22 | And try break=50-fstab instead
| |
| 13:23 | <eddytv> Hmm, it doesn't seem to be honoring the break
| |
| 13:23 | It's still trying to start X
| |
| 13:23 | <alkisg> That's weird, maybe you also have i386/pxelinux.cfg/default in your tftp?
| |
| 13:24 | If so, try changing that one
| |
| 13:24 | <eddytv> No, everything is 64-bit
| |
| 13:25 | I added an "S00test" in rc2.d that did an 'ls -ld /tmp", and /var/log/boot.log shows it is 755 at that point
| |
| 13:26 | <Hyperbyte> So we did this research under our clients, to see how satisfied they are (or aren't) with our products and services
| |
| 13:26 | Was a form with about 40, 50 questions
| |
| 13:26 | Looking over them now...
| |
| 13:27 | One person left the entire form blank, didn't fill out any info - except "What do you think we don't do well?"
| |
| 13:27 | They answered 'Communication'
| |
| 13:27 | The rest of the form is completely empty...
| |
| 13:27 | <eddytv> hahaha.
| |
| 13:28 | <alkisg> Hehe :D
| |
| 13:28 | <eddytv> alkisg: since you brought up /tmp not being mounted, it made me realize that I still have CONFIGURE_FSTAB=false in my lts.conf from 11.04...
| |
| 13:29 | and if I cat /etc/fstab, it just says "# UNCONFIGURED FSTAB FOR BASE SYSTEM"
| |
| 13:30 | <Hyperbyte> eddytv, that's correct
| |
| 13:31 | <eddytv> OK. I just removed that option, and /tmp is still 755 instead of 1777 :(
| |
| 13:31 | <alkisg> eddytv: you can also try: RCFILE_01=ls -ld /tmp > /tmp/status, to see what goes on at the point when rc.local runs
| |
| 13:32 | eddytv: it doesn't sound ltsp related, I would be searching evil services you (or mythbuntu) installed if I were you
| |
| 13:32 | <eddytv> FWIW, now /etc/fstab has an entry for /proc (and that’s all)
| |
| 13:32 | <alkisg> Yup that's normal unless you use NFS home etc
| |
| 13:33 | <eddytv> I did find's | xargs grep chmod for a while trying to find something that could be messing up /tmp
| |
| 13:33 | Came up empty.
| |
| 13:35 | Also, I haven't installed anything else since I initially ran "ltsp-build-client --fat-client-desktop mythbuntu-desktop --mount-package-cache --keep-packages --copy-sourceslist".
| |
| 13:35 | <alkisg> Btw did the splash screen go away when you removed the quiet splash part from pxelinux.cfg/default ?
| |
| 13:36 | <eddytv> Yes
| |
| 13:36 | I saw the "typical" linux boot output
| |
| 13:36 | adrianorg has joined IRC (adrianorg!~adrianorg@187.115.104.79) | |
| 13:37 | <eddytv> I added the RCFILE_01 entry to lts.conf and it shows /tmp with mode 755
| |
| 13:38 | <alkisg> And with ssh you see that it happens before login?
| |
| 13:39 | So, after rc.local, and before login? There's only a few services that run between them...
| |
| 13:39 | Ouch sorry you said 755
| |
| 13:39 | OK
| |
| 13:39 | <eddytv> Yeah, I was confused there for a sec
| |
| 13:39 | But the 'break' doesn't seem to be working
| |
| 13:40 | <alkisg> If you saw it with rc.d, never mind then, the break would be even earlier
| |
| 13:41 | <eddytv> so where does /tmp normally get mounted as a 'tmpfs'?
| |
| 13:41 | <alkisg> I assume that with SCREEN_02=shell you also get 755, right?
| |
| 13:42 | It doesn't, I was just wondering if mythbuntu did that for power management issues
| |
| 13:42 | <eddytv> ah, ok
| |
| 13:43 | I do see that there is no /tmp mount on the running 12.04 system...
| |
| 13:44 | dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174) | |
| 13:46 | <eddytv> Where does the /tmp dir get "configured" for clients?
| |
| 13:53 | Oh geez
| |
| 13:53 | I just looked in /opt/ltsp/amd64 and /tmp has mode 755 there!
| |
| 13:54 | <||cw> whoops
| |
| 13:55 | <eddytv> Ugh. I could've sworn I checked that before and made sure it was correct
| |
| 13:55 | But looking through my zsh history, I see that I was cd'd in to /opt/ltsp/amd64 and ran 'ls -ld /tmp' instead of 'ls -ld tmp' (probably out of habit).
| |
| 13:56 | * eddytv bangs head on desk | |
| 13:59 | vmlintu has joined IRC (vmlintu!~vmlintu@nblzone-240-143.nblnetworks.fi) | |
| 14:00 | <eddytv> alkisg: thanks for helping me eventually find the solution to this issue. I chmod'd the tmp dir in /opt/ltsp/amd64 and rebuild the image, and now the client autologins and X starts, so I'm starting to make progress
| |
| 14:01 | <alkisg> eddytv: nice, although you didn't find what did change the chroot /tmp/ mode in the first place
| |
| 14:03 | <eddytv> True. But there were also files in that tmp dir, so it's possible I was cd'd in to /opt/ltsp/amd64 and I ran something that trounced on /tmp
| |
| 14:04 | Once I get this working, my plan is to mv /opt/ltsp to /opt/ltsp.save and "start fresh", following my notes, to make sure I can re-create a "clean" LTSP environment
| |
| 14:05 | <alkisg> For any notes that belong upstream, do file bugs :)
| |
| 14:05 | <eddytv> If the ltsp tmp dir gets messed up again, then I'll know there's an issue, and if not, then I'll just know it was the side effect of something I was doing trying to get the mythtv diskless clients working in 12.04
| |
| 14:06 | For now, I'm going to forge ahead and see if I can get things like 'lircd', sound, etc. working
| |
| 14:08 | <alkisg> The lircd part should be implemented from /usr/share/ltsp/init-ltsp.d/50-lircd
| |
| 14:08 | See similar scripts in that dir
| |
| 14:12 | <eddytv> OK, will do. Out of curiosity and for my understanding of the ltsp architecture, is the /cow mountpoint a temporary storage point where files in the root filesystem that are changed are written to (like stuff in /var/log and /var/cache) that is "lost" on each reboot?
| |
| 14:12 | <alkisg> Yes
| |
| 14:13 | mmetzger has left IRC (mmetzger!~mmetzger@99-71-214-196.lightspeed.mdldtx.sbcglobal.net, Ping timeout: 248 seconds) | |
| 14:14 | dsorf has joined IRC (dsorf!~Miranda@89-24-91-2.i4g.tmcz.cz) | |
| 14:15 | <eddytv> OK. I see something about support for persistent user home directories. Right now, every time X starts, XFCE (the default window manager for mythbuntu since you really never see it) asks "Welcome to Panel. Would like to use the default Panel configuration?" because there are no ~mythtv/.config/xfce4/xfconf/... files.
| |
| 14:15 | mmetzger has joined IRC (mmetzger!~mmetzger@99-71-214-196.lightspeed.mdldtx.sbcglobal.net) | |
| 14:17 | <eddytv> My guess is that if I set up persistent home directories, files written to ~mythtv would get saved across reboots, and alleviate that issue... but I'm also guessing that would present a problem if multiple clients were all powered up simultaneously and they all were using the 'mythtv' userid...
| |
| 14:17 | <alkisg> Fat clients have persistent home directories by default
| |
| 14:18 | Each user gets his own directory
| |
| 14:18 | <eddytv> How is that set up on the "backend"?
| |
| 14:18 | <alkisg> ssh authentication, sshfs mounting of /home/username
| |
| 14:19 | <eddytv> is configuration of that automatic? that doesn't seem to be happening
| |
| 14:19 | <Hyperbyte> It is default.
| |
| 14:19 | It happens after a user logs in via ldm.
| |
| 14:19 | <eddytv> including "autologin"?
| |
| 14:20 | <Hyperbyte> eddytv, I believe there's no difference. But feel free to turn off autologin and test.
| |
| 14:21 | Few things to check for
| |
| 14:21 | - You did create the user on the LTSP server?
| |
| 14:21 | - The user has a homedir specified in their finger output which resides in /home/ ?
| |
| 14:21 | - The home directory exists on the server?
| |
| 14:22 | <eddytv> so I should see an 'sshfs' in /proc/mounts when this is working?
| |
| 14:22 | <alkisg> After login, yes
| |
| 14:22 | <eddytv> Yes, the 'mythtv' user exists on the LTSP server with a valid home directory.
| |
| 14:23 | <alkisg> Don't use the "mythtv" user for all accounts, create separate ones for each user
| |
| 14:23 | <eddytv> for each frontend you mean?
| |
| 14:23 | <alkisg> For each client (pc), or member of the family, etc
| |
| 14:24 | Check also that you're using LDM instead of GDM or LightDM
| |
| 14:24 | <eddytv> LDM is definitely running on the client.
| |
| 14:25 | And I want autologin, so the mythtv clients can just be "turned on" and work, like an appliance (especially since they don't have a keyboard)
| |
| 14:25 | <alkisg> LDM_AUTOLOGIN=True, LDM_USERNAME/password etc
| |
| 14:25 | bbiab
| |
| 14:25 | <eddytv> Right, and I have the autologin working, but no sshfs home stuff.
| |
| 14:29 | <Hyperbyte> eddytv, pastebin the output of 'finger <user>' and 'ls -al /home'
| |
| 14:29 | Where <user> is the user you're logging in with on LTSP
| |
| 14:31 | <eddytv> Is 'finger' actually required? The 'finger' pkg is not installed.
| |
| 14:35 | <Hyperbyte> Just install it, little program for fetching userinfo
| |
| 14:36 | <eddytv> So the ltsp stuff calls 'finger'?
| |
| 14:37 | Just wondering if that could explain why the sshfs stuff isn't automatically working
| |
| 14:38 | <Hyperbyte> No
| |
| 14:38 | I just want to see the entire user info
| |
| 14:39 | Can you provide me with the output or not? :)
| |
| 14:40 | <eddytv> Yup, one sec
| |
| 14:40 | <Hyperbyte> And those commands should be ran on the server by the way
| |
| 14:40 | Not the client.
| |
| 14:41 | <eddytv> http://pastebin.com/kA97ELsm
| |
| 14:41 | leio_ is now known as leio | |
| 14:43 | <Hyperbyte> There is a home dir on the client, right?
| |
| 14:43 | /opt/ltsp/<arch>/home/
| |
| 14:44 | <eddytv> Yes
| |
| 14:44 | <Hyperbyte> I don't mean a 'mythtv' subdir, but the actual /home dir exists on the client?
| |
| 14:44 | Okay... pastebin your lts.conf as well. :)
| |
| 14:44 | <alkisg> eddytv: does an /opt/ltsp/amd64/home/mythtv (non-empty) directory exist?
| |
| 14:45 | The sshfs options that LTSP uses don't work with existing dirs with contents
| |
| 14:45 | <eddytv> Ah, that must be it
| |
| 14:45 | ltsp/amd64/home/mythth exists and has a .mythtv dir
| |
| 14:46 | <alkisg> Try with another user then, and also remove that mythtv folder
| |
| 14:46 | You shouldn't have any user accounts in the chroot
| |
| 14:47 | <eddytv> ok, let me clear it out and rebuild the image
| |
| 14:49 | are user accounts between the ltsp server and thick clients somehow linked?
| |
| 14:50 | in other words, if I create 'mythtv1', 'mythtv2', 'mythtv3', etc. users on the ltsp server, will I be able to "autologin" to those accounts automatically on a client?
| |
| 14:51 | <Hyperbyte> eddytv, yes.
| |
| 14:51 | But accounts aren't linked
| |
| 14:52 | LTSP uses SSH to authenticate
| |
| 14:52 | So if you log in on an LTSP client, you're logging in via SSH on the server
| |
| 14:52 | If succesful, the user account is transferred to the client (for this session only) and the home dir is mounted
| |
| 14:53 | That's why creating user accounts or home dirs on LTSP clients is not very useful... it only disturbs the workings of LTSP, if anything.
| |
| 14:57 | As you have just witnessed. :)
| |
| 15:03 | <eddytv> OK, let me give me this a shot. I'll report back in a minute.
| |
| 15:11 | Steve_The_Pirate has left IRC (Steve_The_Pirate!~Gary@82-68-255-22.dsl.in-addr.zen.co.uk, Quit: Leaving) | |
| 15:13 | komunista has joined IRC (komunista!~slavko@adsl-195-168-232-196.dynamic.nextra.sk) | |
| 15:14 | <knipwim> alkisg: on r2199, good catch, i can't believe i missed that one
| |
| 15:15 | <alkisg> np :)
| |
| 15:19 | <eddytv> What causes /opt/ltsp/amd64/etc/ktsp/update-kernels.conf to get re-written?
| |
| 15:20 | Trying to understand how to set the default kernel options for pxelinux.cfg/default (so I don't have to keep manually editing the file after every ltsp-update-image to remove 'quiet' and 'splash')
| |
| 15:23 | <alkisg> Yeah it's a mess, I'm just removing that part in ltsp-update-image that re-creates that file all the time
| |
| 15:23 | <eddytv> Ah, OK. Nice.
| |
| 15:23 | Hyperbyte: I now see /home/mythtv1 mounted via fuse.sshfs
| |
| 15:24 | <alkisg> You can put the options in the server /etc/ltsp/ltsp-update-image.conf, so that they persist
| |
| 15:24 | (in 12.04, later on it'll be simplified)
| |
| 15:24 | e.g. BOOTPROMPT_OPTIONS="root=/dev/nbd0 init=/sbin/init-ltsp quiet splash plymouth:force-splash vt.handoff=7"
| |
| 15:25 | <eddytv> Gotcha. Thanks!
| |
| 15:25 | dsorf has left IRC (dsorf!~Miranda@89-24-91-2.i4g.tmcz.cz, Ping timeout: 252 seconds) | |
| 15:27 | <eddytv> FWIW, it might be nice if 'ltsp-update-image' issued a warning if /opt/ltsp/<arch>/tmp did not have perms 1777, and along the same lines, a warning about sshfs home dirs not working if /opt/ltsp/<arch>/home/* has files :D
| |
| 15:32 | <alkisg> For the first... nah :D Make a request for that in your distro initscripts, to check if /tmp has the correct permissions
| |
| 15:32 | For the second, there's a bug report filed, it'll work with non-empty dirs too in the future
| |
| 15:40 | dobber has left IRC (dobber!~dobber@213.169.45.222, Remote host closed the connection) | |
| 15:40 | staffencasa has joined IRC (staffencasa!~staffenca@128.193.8.220) | |
| 15:41 | <knipwim> alkisg: cool, you can now see how much extension code each distro has ;)
| |
| 15:54 | Phantomas has joined IRC (Phantomas!~Phantomas@unaffiliated/phantomas) | |
| 15:55 | <knipwim> alkisg: also, i think i can move the /server/Debian/share/plugins/ltsp-build-client/functions to /server/Debian/share/ltsp-build-client-functions
| |
| 15:56 | and remove the VENDOR include part in ltsp-build-client altogether
| |
| 15:56 | * alkisg checks... | |
| 15:58 | <alkisg> knipwim: you'd need to include ltsp-server-functions then in ltsp-build-client
| |
| 15:59 | <knipwim> these are included already
| |
| 15:59 | for the vendor_detect()
| |
| 16:00 | <alkisg> Ah sorry I was looking too low :D
| |
| 16:00 | Yup, sounds good
| |
| 16:00 | <knipwim> and while were at it, lose the common dir
| |
| 16:01 | override the packaged common plugins with the distro specific ones in /usr/share/ltsp/plugins/ltsp-build-client/.
| |
| 16:01 | <alkisg> knipwim: I'm not sure if that would be accepted by all
| |
| 16:01 | I think there are thoughts that "one can build a gentoo chroot on an ubuntu server"
| |
| 16:01 | <knipwim> kk, waiting for that one
| |
| 16:02 | <alkisg> Not sure if that would ever be possible, but it needs some chatting before doing it
| |
| 16:02 | Should server/share/plugins/functions be renamed to server/share/ltsp-build-client-functions ?
| |
| 16:02 | <knipwim> would be a nice challenge though
| |
| 16:03 | <alkisg> Also, I think ltsp-build-client.d would be a more common name than plugins/ltsp-build-client/
| |
| 16:03 | <knipwim> hmm, it would get overwritten by the Debian one
| |
| 16:04 | <alkisg> It's not like we use that dir for anything else...
| |
| 16:04 | The debian one would go to ltsp-build-client-vendor-functions
| |
| 16:04 | So both would automatically be sourced by ltsp-server-functions
| |
| 16:04 | As it sources $tool-functions and $tool-vendor-functions
| |
| 16:04 | <knipwim> we include <command>-vendor-functions?
| |
| 16:04 | <alkisg> I think so, let me look...
| |
| 16:05 | Nope
| |
| 16:06 | Ah yeah we said that:
| |
| 16:06 | common functions go inside the tool,
| |
| 16:06 | overridable functions go in -functions
| |
| 16:06 | So functions should be merged into ltsp-build-client, and the debian one renamed into -functions, right, but lets see if the function merging makes sense...
| |
| 16:06 | <knipwim> hmm, it doesn't apply to ltsp-build-client apparently
| |
| 16:07 | <alkisg> Why not?
| |
| 16:07 | <knipwim> because the Debian ltsp-build-client-functions also doesn't override anything
| |
| 16:07 | <alkisg> It doesn't matter
| |
| 16:07 | It can provide new functions if it needs to
| |
| 16:07 | I'd say go for it
| |
| 16:07 | Merge functions into ltsp-build-client, rename the debian one into ltsp-build-client-functions
| |
| 16:08 | It's only 3 functions
| |
| 16:08 | <knipwim> will do that tonight, first arrange some diner
| |
| 16:09 | are you going to merge a lot tonight?
| |
| 16:10 | bieb has joined IRC (bieb!~hbieber@173.226.16.130) | |
| 16:10 | <alkisg> I'm not doing any more moves, just pushing new code
| |
| 16:10 | So I don't think we'll have overlaps
| |
| 16:18 | Ghidorah has joined IRC (Ghidorah!4a7e0f78@gateway/web/freenode/ip.74.126.15.120) | |
| 16:19 | alexqwesa has left IRC (alexqwesa!~alex@109.172.12.47, Remote host closed the connection) | |
| 16:20 | alexqwesa has joined IRC (alexqwesa!~alex@109.172.12.47) | |
| 16:31 | alexqwesa has left IRC (alexqwesa!~alex@109.172.12.47, Remote host closed the connection) | |
| 16:35 | Steve_The_Pirate has joined IRC (Steve_The_Pirate!~Gary@cpc16-haye17-2-0-cust169.haye.cable.virginmedia.com) | |
| 16:59 | toscalix has joined IRC (toscalix!~toscalix@174.165.219.87.dynamic.jazztel.es) | |
| 17:07 | alexqwesa has joined IRC (alexqwesa!~alex@109.172.12.47) | |
| 17:19 | vagrantc has joined IRC (vagrantc!~vagrant@c-76-115-60-19.hsd1.or.comcast.net) | |
| 17:19 | vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc) | |
| 17:20 | alexqwesa has left IRC (alexqwesa!~alex@109.172.12.47, Remote host closed the connection) | |
| 17:21 | alexqwesa has joined IRC (alexqwesa!~alex@109.172.12.47) | |
| 17:22 | <alkisg> vagrantc: I'm about to push my last changes (ltsp-cleanup etc), so the next thing in my TODO list is the packaging... first, the debian ltsp-server package needs to be updated to reflect the trunk changes, then the ltsp-server-dnsmasq package needs to be written, and after debian is done, the ubuntu packaging should be synced as much as possible...
| |
| 17:22 | Will you have any time today for any of that?
| |
| 17:25 | <vagrantc> looks like NEW queue processing has really slowed lately... might be hard to get ltsp-server-dnsmasq through
| |
| 17:26 | looks like i'll have time today to look at things
| |
| 17:26 | * vagrantc isn't so sure the changes to the server/plugins/ltsp-build-client are less work | |
| 17:27 | <vagrantc> again, it's got the symlink stupidity
| |
| 17:27 | only lots of them
| |
| 17:28 | it looks fine for debian, but i'd hate to be the Ubuntu maintainer
| |
| 17:29 | <knipwim> it should be better for debian, there were some symlinks in the debian dir also, pointing to ubuntu
| |
| 17:30 | <vagrantc> the old way, symlinks were fine, the new way, symlinks require manual handling of each symlink.
| |
| 17:32 | alexqwesa has left IRC (alexqwesa!~alex@109.172.12.47, Remote host closed the connection) | |
| 17:33 | <vagrantc> or you can set up some dead links
| |
| 17:33 | but that doesn't exactly make much sense either
| |
| 17:33 | alexqwesa has joined IRC (alexqwesa!~alex@109.172.12.47) | |
| 17:34 | alexqwesa has left IRC (alexqwesa!~alex@109.172.12.47, Remote host closed the connection) | |
| 17:34 | <knipwim> or a "Debian-related" dir, with all Debian and Ubuntu common plugins in them
| |
| 17:35 | alexqwesa has joined IRC (alexqwesa!~alex@109.172.12.47) | |
| 17:36 | <vagrantc> or i guess ubuntu can include another symlink in ../Debian to trick the otherwise dead links
| |
| 17:42 | so that's the one case that looks like it's actually more complicated than before
| |
| 17:45 | <knipwim> only in packaging apparently
| |
| 17:45 | for ubuntu
| |
| 17:46 | <vagrantc> or for any other potential derived distro
| |
| 17:46 | <knipwim> gentoo doesn't have a problem with symlinks
| |
| 17:46 | <vagrantc> wait a minute ...
| |
| 17:46 | looks like i can't install ubuntu from debiann and vice-versa with this setup anymore...
| |
| 17:47 | <alkisg> vagrantc: we didn't touch the ltsp-build-client part, as someone mentioned ^ this
| |
| 17:47 | So it's still the same, both for packaging and cross distro installs
| |
| 17:48 | <vagrantc> ah, then server/Debian/share/plugins/ltsp-build-client/* should be moved to server/Debain/share/plugins/ltsp-build-client/Debian/
| |
| 17:48 | <alkisg> (08:28:07 μμ) vagrantc: it looks fine for debian, but i'd hate to be the Ubuntu maintainer => the ubuntu just copies Ubuntu/share/* over Debian/share/* over client/*, i.e. it's 1 line longer than the debian's packaging, which would be 20 lines smaller than it is now :)
| |
| 17:49 | (08:48:21 μμ) vagrantc: ah, then server/Debian/share/plugins/ltsp-build-client/* should be moved to server/Debain/share/plugins/ltsp-build-client/Debian/ => true
| |
| 17:49 | <knipwim> server/Debian/share/plugins/ltsp-build-client/* packages to /usr/share/plugins/ltsp-build-client/Debian/.
| |
| 17:49 | <vagrantc> alkisg: and then you'll end up with a bunch of broken symlinks
| |
| 17:49 | <knipwim> right?
| |
| 17:49 | <vagrantc> if the files don't end up in the actual structure layout in the package, what's the point of the restrcture?
| |
| 17:50 | <alkisg> vagrantc: why would I have broken symlinks?
| |
| 17:50 | vagrantc: it's the same as it is now for init-ltsp.d, we don't have broken symlinks in Ubuntu
| |
| 17:50 | <vagrantc> alkisg: because you've got a messload of broken symlinks in your plugin dir, which will be installed on top of the files.
| |
| 17:50 | <alkisg> vagrantc: we didn't (or shouldn't) touch the ltsp-build-client plugin dir
| |
| 17:51 | <vagrantc> alkisg: well, the Ubuntu plugin dir has lots of symlinks currently
| |
| 17:51 | <alkisg> They won't be touched by the restructuring, so I don't see what will change there
| |
| 17:51 | If they were, we should put them back
| |
| 17:52 | <vagrantc> this is the same problem we had with init-ltsp.d and why we removed the symlinks there.
| |
| 17:52 | <alkisg> And now init-ltsp.d works fine
| |
| 17:52 | And we can see the Ubuntu diff from Debian, and the Debian diff from common very easily
| |
| 17:52 | And it's much easier for packaging
| |
| 17:52 | <vagrantc> the ltsp-build-client plugin dirs definitely got restructured.
| |
| 17:52 | <alkisg> We should undo that part
| |
| 17:53 | <knipwim> vagrantc: how did it get restructured, being not functional anymore
| |
| 17:53 | the dirs just got moved
| |
| 17:53 | after packaging the same structure can be recreated
| |
| 17:54 | <vagrantc> knipwim: and the symlinks now point to ../../../../Debian/...
| |
| 17:54 | in Ubuntu's plugin dir.
| |
| 17:54 | the symlinks need to point to ../Debian/
| |
| 17:54 | <alkisg> knipwim: the server/$distro dirs would only appear in specific distro packages, while the ltsp-build-client plugins should appear in all distros, so they don't belong in server/$distro, but back where they were
| |
| 17:54 | <knipwim> why?
| |
| 17:55 | <vagrantc> knipwim: because that's how they need to be installed on the server side.
| |
| 17:55 | knipwim: in the package.
| |
| 17:55 | knipwim: also, it allows for the potential for cross-distro installs.
| |
| 17:55 | * knipwim processing | |
| 17:56 | <knipwim> check, you guys need the symlinks
| |
| 17:56 | and the symlinks to be the same as upstream
| |
| 17:56 | <alkisg> I don't know though if it's possible to build ltsp cross-distro, has anyone succeeded in that?
| |
| 17:57 | <vagrantc> alkisg: i've built Debian chroots on ubuntu with a one-line change in ltsp-build-clienyt
| |
| 17:57 | <knipwim> hhmmokay, i'll move them back
| |
| 17:58 | <alkisg> vagrantc: but that's not really really cross distro... how would you run debootstrap on gentoo?
| |
| 17:58 | <vagrantc> and now that the plugins do less mangling of the chroot, it will only get easier
| |
| 17:58 | <knipwim> crap, a lot of symlinks to move back :(
| |
| 17:58 | <vagrantc> alkisg: debootstrap is only a shell script
| |
| 17:58 | <alkisg> vagrantc: but it doesn't exist on gentoo or on non-debian distros anyway
| |
| 17:58 | <vagrantc> alkisg: it doesn't guarantee that all distros will have all the tools to do cross-distro installs, but it leaves the possibility open
| |
| 17:58 | <knipwim> you would need a portage and emerge as a bare minimum
| |
| 17:59 | <vagrantc> for those that do have the tools.
| |
| 17:59 | <alkisg> I wouldn't want to have to care about ltsp-build-client ubuntu scripts being runnable in gentoo
| |
| 17:59 | <knipwim> however, such an exercise could make the ltsp code more generic
| |
| 18:00 | <vagrantc> alkisg: don't break that bank over it, but at least Debian/Ubuntu is trivial, and we shouldn't break that :)
| |
| 18:00 | * alkisg doesn't mind either way as he won't be using ltsp-build-client anymore :P :D | |
| 18:00 | <vagrantc> hah
| |
| 18:00 | <alkisg> But yeah sure let's put those back
| |
| 18:00 | Anything else?
| |
| 18:01 | <vagrantc> how feasible is it to just revert the commit? is it co-mingled with other commits?
| |
| 18:01 | <knipwim> the plugin commit was the whole server tree
| |
| 18:02 | a new commit would be less complicated
| |
| 18:02 | <alkisg> My root is 18 Gb, `ltsp-build-client --cleanup /` generated an /opt/ltsp/i386 of 1.7 Gb, let's see it if boots...
| |
| 18:02 | <knipwim> alkisg: what does --cleanup do?
| |
| 18:03 | <alkisg> Creates tmpfs/cow etc of /, removes user accounts, logs, regenerates ssh keys etc
| |
| 18:03 | So that one can use his / instead of a chroot
| |
| 18:03 | <knipwim> cool
| |
| 18:03 | <alkisg> (or a chroot in a VM, with users, desktop enviroment and all for managing it)
| |
| 18:04 | <vagrantc> alkisg: what about a setup that actually leaves users?
| |
| 18:04 | <alkisg> vagrantc: I'll put code in the future to specify which "plugins" one doesn't want to execute
| |
| 18:05 | So one may keep his server ssh keys, another might keep the users etc
| |
| 18:05 | <vagrantc> it could be a really useful kiosk to leave the user ... (or autogenerate the user)
| |
| 18:05 | and then have that user configured to a known good state
| |
| 18:06 | knipwim: did you want one of us to revert the ltsp-build-client plugins? i could do it .. since you've done so much on it soo far.
| |
| 18:11 | the ltsp-build-client plugins actually had a somewhat sane structure from the get-go :)
| |
| 18:11 | in that it was actually planned ... it wasn't perfect, but at least it was consistant
| |
| 18:12 | knipwim: it might be easier to revert the whole commit, and then re-add the other bits.
| |
| 18:14 | Gremble has joined IRC (Gremble!~Ben@cpc10-aztw24-2-0-cust114.aztw.cable.virginmedia.com) | |
| 18:15 | <knipwim> vagrantc: no, it's allright
| |
| 18:16 | only the Ubuntu symlinks will take some concentration
| |
| 18:21 | <vagrantc> the nbd-proxy stuff isn't really debian specific
| |
| 18:22 | it's mostly an Ubuntu thing, although in theory it could be used by anyone
| |
| 18:22 | <alkisg> You ship it in debian, that's why we put it there
| |
| 18:22 | Packaging ==> common => debian => ubuntu
| |
| 18:22 | So stuff common in debian and ubuntu should go in debian, if it's shipped there, even if their development and main usage is in ubuntu.... right?
| |
| 18:23 | If it's not shipped by debian, ok, it would only go in ubuntu
| |
| 18:23 | <vagrantc> i guess so. although it seems as much an upstream feature as a debian one
| |
| 18:23 | <alkisg> I think it relies on debian specific bits about how initramfs tools work
| |
| 18:23 | And personally I think it should go in a separate package...
| |
| 18:24 | <vagrantc> i don't think client/Debian/nbd-proxy contains anything debian/ubuntu specific
| |
| 18:25 | but yes, it really ought to go in it's own source package.
| |
| 18:26 | <alkisg> We can put it in client/nbd-proxy till then if it sounds better
| |
| 18:26 | <vagrantc> ltsp-localapps is actually server-side ... that doesn't map well to the new model.
| |
| 18:28 | <alkisg> Should we put them back in ltsp-trunk/ then?
| |
| 18:29 | Or only put that specific script in the server dir?
| |
| 18:29 | <vagrantc> dunno.
| |
| 18:29 | <alkisg> ...and its man page...
| |
| 18:29 | and the examples/ltsp*.desktop files...
| |
| 18:30 | and iptables.sh (which shouldn't be in localapps anyway, fat clients can also use that)
| |
| 18:30 | <vagrantc> there are a number of tricks
| |
| 18:31 | so the share dirs in theory get installed to /usr/share/ltsp ?
| |
| 18:31 | <alkisg> Yes, and each distro just overlaps its own dir
| |
| 18:32 | <vagrantc> ok, so client/Debian/initramfs shouln't really get moved to client/Debian/share
| |
| 18:33 | dbp has joined IRC (dbp!~5minmajor@adsl-072-156-187-139.sip.bhm.bellsouth.net) | |
| 18:34 | <alkisg> Yeah it should go to client/Debian/initramfs
| |
| 18:38 | <knipwim> ok, finished the plugins
| |
| 18:38 | moving the initramfs also
| |
| 18:38 | bobby_C has joined IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at) | |
| 18:38 | <knipwim> and the README-DEVELOPMENT-POLICY to common/doc
| |
| 18:39 | <bieb> Hello! Long time no talk.. I am getting ready to totally rebuild my LTSP server.. currently it is ubuntu 10.04 (but up to date on patches), running thin clients, Firefox on the local clients.. and crossover office to give students MS Office. I have read that I can change the 12.04 desktop from the new shiny unity to the old look.. is that what everyone seems to be doing? Are there any other gotchas I should be checking in going to 12.04?
| |
| 18:40 | <knipwim> erm, the initramfs is in Debian/initramfs already
| |
| 18:40 | <vagrantc> knipwim: moving initramds?
| |
| 18:40 | <highvoltage> bieb: on LTSP, yes, it's widely done
| |
| 18:40 | <vagrantc> knipwim: yeah, it's fine where it is.
| |
| 18:40 | <highvoltage> bieb: you can install the gnome-session-fallback package, it installs everything you need
| |
| 18:40 | <knipwim> nothing else to move?
| |
| 18:40 | nbd-proxy ok?
| |
| 18:41 | <highvoltage> bieb: then you just need to set it as default in your lts.conf and you're good to go
| |
| 18:41 | <bieb> highvoltage: any pros/cons to keeping the 12.04 desktop?
| |
| 18:41 | <knipwim> then i'm committing
| |
| 18:41 | <vagrantc> knipwim: i'd commit one change at a time.
| |
| 18:41 | <highvoltage> bieb: unity /can/ potentially use network bandwidth more efficiently, since it uses compiz as window manager by default. it is more cpu intensive on your server though, and compiz can use lots of memory over time.
| |
| 18:42 | <vagrantc> knipwim: so moving the plugins back is one change.
| |
| 18:42 | <bieb> highvoltage: gotcha
| |
| 18:42 | <knipwim> vagrantc: touche
| |
| 18:43 | <bieb> highvoltage: I know 12.04 lts is recently released.. have there been any issues with LTSP reported that you know of?
| |
| 18:44 | <highvoltage> bieb: not that I know of. I've done a lot of testing (only with a few pcs at a time though) with Edubuntu 12.04 LTS with Gnome Fallback (with metacity compositing enabled and disabled), Unity 3D, Unity 2D and Xfce
| |
| 18:44 | bieb: I can honestly say it's the best LTSP release on Ubuntu in years
| |
| 18:44 | (actually, imho the best)
| |
| 18:44 | <vagrantc> knipwim: does gentoo run the autogen scripts as part of it's build process?
| |
| 18:44 | <bieb> highvoltage: good to know
| |
| 18:45 | <knipwim> not for ltsp-client
| |
| 18:45 | <bieb> highvoltage: we only have 13 thin clients in our lab so that should work well for us.
| |
| 18:45 | <knipwim> i believe only for ldm and ltspfs
| |
| 18:45 | <highvoltage> bieb: great.
| |
| 18:46 | <bieb> highvoltage: back a couple years ago Scott B from in here was working on Sabayon to manage multiple profiles.. I saw sabayan as a distro.. is that the same? or did Scott get it working as a package?
| |
| 18:47 | <highvoltage> bieb: yes there's both sabayon the program and the distro, but they are unrelated. Scott got it working, but gnome changed everything underneath so it stopped working
| |
| 18:48 | (I think someone got it working again recently, but I'm not sure)
| |
| 18:48 | <bieb> highvoltage: damn! what is good way to manage multiple profiles? we have 2 stand alone linux "kiosks" in the library that I would like to replace with thin clients
| |
| 18:49 | <highvoltage> bieb: well, ltsp has a kiosk plugin that might work for you, I haven't used it myself yet but if you hang around here someone should be able to help :)
| |
| 18:50 | <bieb> highvoltage: thanks..
| |
| 18:51 | highvoltage: you know of anyone that uses Crossover Office to provide MS office on their LTSP server? we have used it and it works well.. except for recognizing the USB that a student plugs into the thin client
| |
| 18:52 | <Gremble> just upgraded a server from the last LTS to the new one
| |
| 18:52 | nightmare upgrade on the server but I got there
| |
| 18:53 | <highvoltage> Gremble: please share! what went wrong?
| |
| 18:53 | <dbp> hi guys...ltsp novice here. I was wondering if someone could help guide me through diagnosing my inability to complete build-client on 10.04 LTS server after upgrade from 8.04 LTS
| |
| 18:53 | <Gremble> honestly highvoltage, if I went through the ordeal in my mind again, I'd end up in some kind of therapy
| |
| 18:54 | upgrade resulted in a whole heep of dependancy problems
| |
| 18:54 | <highvoltage> Gremble: oops, ok. I thought that this was the LTSP therapy channel :)
| |
| 18:54 | <Gremble> the system was without a libc6 for a large period of time
| |
| 18:54 | the multi-arch changes make it tricky
| |
| 18:54 | <highvoltage> did you use "apt-get dist-upgrade" to upgrade?
| |
| 18:55 | <Gremble> sometimes aptitude, sometimes apt-get upgrade sometimes apt-get dist-upgrade
| |
| 18:55 | sometime dpkg --force-all -i
| |
| 18:55 | <highvoltage> in that case I'm sorry to break it to you that it was your fault that it broke.
| |
| 18:55 | <vagrantc> dbp: tell us more ... what happened?
| |
| 18:55 | <Gremble> with downloaded packages
| |
| 18:55 | <highvoltage> dist-upgrade isn't supported on ubuntu upgrades, you have to use update-manager
| |
| 18:56 | <vagrantc> really? wow.
| |
| 18:56 | <highvoltage> update-manager is aware of things like multi-arch changes (and other big transitions)
| |
| 18:56 | <alkisg> Well it's mostly working except for a few glitches with --cleanup, I'll troubleshoot that after the packaging, knipwim may I push code changes?
| |
| 18:56 | <vagrantc> debian supports using apt-get dist-upgrade for multi-arch ...
| |
| 18:56 | <Gremble> I did ask a couple of times for advice and peoples experiences on this channel, but got nothing back
| |
| 18:56 | <highvoltage> you /can/ use dist-upgrade if you're willing to fix all the things that break manually, but for a supported Ubuntu upgrade you have to use update-manager.
| |
| 18:57 | <Gremble> vagrantc, yep no trouble with multiarch stuff on Debian
| |
| 18:57 | cyberplopo has joined IRC (cyberplopo!b29c559c@gateway/web/freenode/ip.178.156.85.156) | |
| 18:57 | <highvoltage> vagrantc: yes, debian supports dist-upgrade between releases, Ubuntu doesn't
| |
| 18:57 | <cyberplopo> hello
| |
| 18:57 | <vagrantc> anything that wouldn't properly upgrade with dist-upgrade would be seen as a major bug.
| |
| 18:57 | <cyberplopo> i have a problem with ltsp
| |
| 18:57 | <Gremble> highvoltage, so updating from one LTS to the next is fully supported?
| |
| 18:58 | <vagrantc> ah well.
| |
| 18:58 | <cyberplopo> when i up eth1 i can't navigate
| |
| 18:58 | if I down eth1 I can
| |
| 18:58 | someone can help me?
| |
| 18:58 | <Gremble> how are you managing you interfaces cyberplopo ?
| |
| 18:58 | <highvoltage> Gremble: yes, as long as you use update manager and not apt-get (or aptitude) dist-upgrade
| |
| 18:58 | <alkisg> knipwim: pushed, remember to pull before pushing
| |
| 18:59 | <vagrantc> alkisg, knipwim: any objection to moving nbd-proxy and localapps back to the top of the tree? or at least move client/Debian/nbd-proxy to client/nbd-proxy?
| |
| 18:59 | <Gremble> highvoltage, do you know anyone who's done that for the current and previous LTS releases without trouble?
| |
| 18:59 | <dbp> after upgrade when I try to build-client I get the failed getting release file error. wget retrieves the file from the repository.
| |
| 18:59 | <knipwim> alkisg: sorry
| |
| 18:59 | <highvoltage> Gremble: you can't skip an LTS either, like some people try to upgrade from 8.04 to 12.04. you have to do the jumps through the releases inbetween. so from 8.04 it's 8.04 -> 10.04 -> 12.04
| |
| 19:00 | <Gremble> right, I went from 10.04 to 12.04
| |
| 19:00 | anyway, all good now
| |
| 19:00 | <cyberplopo> <Gremble> I managing with /etc/network/interfaces
| |
| 19:00 | <highvoltage> Gremble: yep
| |
| 19:00 | <Gremble> the chroot upgraded easily
| |
| 19:00 | * knipwim eating first | |
| 19:00 | <Gremble> but I'm rebuilding it now as there was some chaff
| |
| 19:00 | <alkisg> vagrantc: nope, no objections at all
| |
| 19:00 | <Gremble> I must say, the new build scripts are top banana!
| |
| 19:00 | <alkisg> knipwim: /me goes to eat too :D
| |
| 19:01 | <Gremble> cyberplopo, no network-manager at all?
| |
| 19:01 | <cyberplopo> network-manager disapeared
| |
| 19:01 | <Gremble> yoiks
| |
| 19:02 | <vagrantc> dbp: so you're running ubuntu 10.04, and when running ltsp-build-client it gets a release file error?
| |
| 19:02 | <dbp> that is correct
| |
| 19:02 | <cyberplopo> when network-manager goes it says me "eth1 is not managed"
| |
| 19:02 | <vagrantc> dbp: if you could cut and paste the error that would be more helpful. if it's many lines, please use a pastebin
| |
| 19:02 | !pastebin
| |
| 19:02 | <ltsp`> vagrantc: pastebin: the LTSP pastebin is at http://ltsp.pastebin.com. Please paste all text longer than a line or two to the pastebin, as it helps to reduce traffic in the channel. Don't forget to paste the URL of the text here.
| |
| 19:03 | <vagrantc> dbp: any reason not to upgrade to 12.04, also?
| |
| 19:03 | komunista has left IRC (komunista!~slavko@adsl-195-168-232-196.dynamic.nextra.sk, Ping timeout: 244 seconds) | |
| 19:03 | <Gremble> cyberplopo, you could pastebin your /etc/network/interfaces file if you liked, and it had no compromising info in it
| |
| 19:04 | <cyberplopo> well, i'm going to conect from that pc
| |
| 19:04 | wait me for a moment, please
| |
| 19:05 | <vagrantc> alkisg, knipwim: or maybe move the server-specific parts of localapps into a server-side dir ... and leave nbd-proxy in client/nbd-proxy ... ?
| |
| 19:06 | cyberpelopo has joined IRC (cyberpelopo!b29c559c@gateway/web/freenode/ip.178.156.85.156) | |
| 19:06 | <cyberpelopo> hello Gremble
| |
| 19:06 | I'm here
| |
| 19:06 | <Gremble> hi
| |
| 19:06 | <alkisg> vagrantc: that sounds OK too
| |
| 19:07 | <cyberpelopo> auto lo iface lo inet loopback auto eth0 iface eth0 inet dhcp auto eth1 iface eth1 inet static address 192.168.0.254 netmask 255.255.255.0 gateway 192.168.0.254
| |
| 19:07 | this is my /etc/network/interfaces
| |
| 19:07 | <bieb> Anyone here use Crossover Office on their LTSP setup?
| |
| 19:07 | cyberplopo has left IRC (cyberplopo!b29c559c@gateway/web/freenode/ip.178.156.85.156, Quit: Page closed) | |
| 19:09 | <cyberpelopo> if I up eth1 I can boot thin client, but have no internet
| |
| 19:09 | obviusly, if I down eth1 I can't boot thin clients
| |
| 19:10 | <vagrantc> alkisg, knipwim: i'm wondering if we shouldn't have client/share/ltsp and client/share/ldm ... i'd like to move one of the localapps/ldm-rc.d scripts outside of localapps, as it really doesn't belong there, but it gets installed to /usr/share/ldm/rc.d ...
| |
| 19:10 | <Gremble> down the interface eth1 remove the line that says gateway 192.168.0.254
| |
| 19:10 | up the interface
| |
| 19:11 | cyberpelopo, ^
| |
| 19:11 | <alkisg> vagrantc: right now "share" actually means "share/ltsp", so it isn't a very good name, so yeah either renaming "share" to "ltsp" or doing it as you say sound ok...
| |
| 19:11 | <cyberpelopo> well, i'm going to do this. if I disconect from chat is for this
| |
| 19:12 | <alkisg> vagrantc: if you rename "share" to "ltsp", then you'd add an "ldm" directory alongside "ltsp"
| |
| 19:13 | <cyberpelopo> well, i'm still here
| |
| 19:13 | <vagrantc> alkisg: that might make more sense...
| |
| 19:13 | <cyberpelopo> eth1 is up
| |
| 19:13 | i'm going to boot a thin client
| |
| 19:13 | <highvoltage> alkisg: are you still up for working on disk image cloning in LTSP? is it something we can give you an action item for? :)
| |
| 19:14 | <alkisg> highvoltage: it's in the long term plans... for 13.xsomething probably
| |
| 19:14 | <highvoltage> alkisg: ok
| |
| 19:14 | <cyberpelopo> eth1 is up, i have internet in the server but can't boot thinclient
| |
| 19:14 | <vagrantc> gah. this is so much shuffling of code.
| |
| 19:14 | this is why we've put it off so long.
| |
| 19:15 | <Gremble> <cyberpelopo> eth1 is up, i have internet in the server but can't boot thinclient
| |
| 19:15 | * alkisg bets vagrantc will like having many more less lines in his debian/install though :) | |
| 19:15 | <alkisg> Bah "more less"
| |
| 19:15 | <Gremble> cyberpelopo, what is the IP of your eth0?
| |
| 19:15 | <cyberpelopo> 192.168.0.102
| |
| 19:16 | <Gremble> then you can't use that range for the eth1
| |
| 19:16 | dbp has left IRC (dbp!~5minmajor@adsl-072-156-187-139.sip.bhm.bellsouth.net, ) | |
| 19:16 | <Gremble> you'll need to use 192.168.<something else>.102
| |
| 19:16 | toscalix has left IRC (toscalix!~toscalix@174.165.219.87.dynamic.jazztel.es, Remote host closed the connection) | |
| 19:16 | <Gremble> 192.168.1.254 for example
| |
| 19:17 | <cyberpelopo> mmmm
| |
| 19:17 | yujuuuuu
| |
| 19:18 | I have internet in server and thin client is ...
| |
| 19:18 | oooohh
| |
| 19:18 | no
| |
| 19:18 | <Gremble> :)
| |
| 19:18 | <cyberpelopo> thin crash
| |
| 19:18 | <Gremble> you might need to re-jig some config files
| |
| 19:18 | <cyberpelopo> re-jig??
| |
| 19:18 | <Gremble> dhcpd.conf will need to know to give out 192.168.1.x addresses
| |
| 19:19 | and not 192.168.0.x
| |
| 19:19 | which it is probably currently set to
| |
| 19:19 | komunista has joined IRC (komunista!~slavko@adsl-195-098-014-163.dynamic.nextra.sk) | |
| 19:19 | <cyberpelopo> how can I fix this?
| |
| 19:19 | <Gremble> edit the file for dhcpd.conf
| |
| 19:21 | /etc/dhcp/dhcpd.conf
| |
| 19:21 | though if you've added include "/etc/ltsp/dhcpd.conf";
| |
| 19:21 | to the bottom of that file, you'll actually need to edit /etc/ltsp/dhcpd.conf
| |
| 19:22 | <cyberpelopo> I can't restart dhcp3-server
| |
| 19:23 | <Gremble> what error message?
| |
| 19:23 | <cyberpelopo> bad range, address 192.168.0.20 not in subnet 192.168.1.0 netmask 255.255.255.0
| |
| 19:23 | <Gremble> change 192.168.0.20 to 192.168.1.20
| |
| 19:24 | you should probably read up on networking
| |
| 19:24 | <cyberpelopo> check syslog for diagnostics.
| |
| 19:24 | can't restart dhcp3
| |
| 19:25 | <Gremble> did you change 192.168.0.20 to 192.168.1.20 ?
| |
| 19:25 | <cyberpelopo> yes, range 192.168.1.20 192.168.1.250;
| |
| 19:25 | <Gremble> odd
| |
| 19:26 | <eddytv> quick package question: if I want to remove pacakges from my LTSP image, I can just run "LTSP_HANDLE_DAEMONS=false ltsp-chroot -c -p aptitude", right?
| |
| 19:27 | <cyberpelopo> what do you suggest
| |
| 19:27 | ?
| |
| 19:27 | <Gremble> what distro are you running?
| |
| 19:27 | what package are you using for dhcpd?
| |
| 19:27 | if you are running isc-dhcp-server then have a look at /etc/default/isc-dhcp-server
| |
| 19:27 | <cyberpelopo> don't know
| |
| 19:28 | i installed edubuntu 12.04 and 10.04 (current)
| |
| 19:29 | <Gremble> isn't that two different things?
| |
| 19:29 | anyway, have a look at /etc/default/isc-dhcp-server
| |
| 19:29 | if you have one
| |
| 19:29 | <cyberpelopo> isc-dhcp-server not exists
| |
| 19:29 | <Gremble> /etc/default/dhcpd3
| |
| 19:29 | or something?
| |
| 19:30 | /etc/default/dhcpd
| |
| 19:30 | ?
| |
| 19:30 | <cyberpelopo> whith dhcp only exists dhcp3-server
| |
| 19:30 | <Gremble> that's the one
| |
| 19:30 | <cyberpelopo> yes
| |
| 19:31 | <Gremble> does it specify any interfaces
| |
| 19:31 | ?
| |
| 19:31 | INTERFACES="eth1" I think is what you want
| |
| 19:31 | <cyberpelopo> eth1
| |
| 19:31 | yes
| |
| 19:32 | <Gremble> what's the current error?
| |
| 19:32 | <cyberpelopo> don't boot thin client
| |
| 19:32 | thin client detects isolinux from net bau can't boot
| |
| 19:33 | <Gremble> is there still an error with dhcp3-server?
| |
| 19:33 | cyberpelopo, that could be a million things
| |
| 19:33 | <cyberpelopo> INTERFACES="eth1"
| |
| 19:33 | I followed the tutorial to install edubuntu
| |
| 19:34 | <Gremble> where's that?
| |
| 19:34 | <cyberpelopo> the instalation is clean
| |
| 19:34 | thousand of manuals, howto's
| |
| 19:34 | are all the same
| |
| 19:35 | <Gremble> you need to do what all sysadmins do, check the logs and google for the errors
| |
| 19:35 | <cyberpelopo> introduce DVD Ubuntu 12.04, install, select eth1 to ltsp(because eth0 goes to internet)
| |
| 19:35 | <alkisg> eddytv: no need for LTSP_HANDLE_DAEMONS, ltsp-chroot does that automatically, and also aptitude isn't preinstalled in 12.04, you might want to use apt-get purge
| |
| 19:36 | <cyberpelopo> when finish, I can boot thin clients but whitout internet
| |
| 19:37 | I need to install this tomorrow in the school
| |
| 19:38 | some suggest?
| |
| 19:38 | <vagrantc> alkisg: i just removed the "documentation" symlink for ltsp-client-common-functions from the Debian dir ... i'm thinking that's better served as a readme or a Makefile in common/
| |
| 19:39 | <Gremble> cyberpelopo, pastebin your errors to us
| |
| 19:39 | <eddytv> alkisg: Thanks, as always.
| |
| 19:40 | <Gremble> http://ltsp.pastebin.com is the place to put your logs cyberpelopo
| |
| 19:41 | <cyberpelopo> what do you need to paste?
| |
| 19:42 | <Gremble> logs of the errors with dhcp
| |
| 19:42 | <cyberpelopo> where can I find this?
| |
| 19:42 | <Gremble> /var/log/dhcp3.log or something like that
| |
| 19:42 | have a look around /var/log
| |
| 19:43 | also see /var/log/syslog for what happened when the clients tried to boot
| |
| 19:43 | its useful to run tail -f /var/log/syslog
| |
| 19:43 | while a client is booting
| |
| 19:43 | to see what's going on
| |
| 19:48 | <cyberpelopo> I'd paste that in pastebin
| |
| 19:48 | syslog from cyberpelopo
| |
| 19:50 | Phantomas has left IRC (Phantomas!~Phantomas@unaffiliated/phantomas) | |
| 19:50 | <alkisg> vagrantc: I think bzr supports hardlinks, not sure if the whole toolchain is good enough to use them
| |
| 19:52 | <Gremble> cyberpelopo, have you done that? what is the pastebin page?
| |
| 19:52 | you need to post the link
| |
| 19:52 | <cyberpelopo> http://pastebin.com/6RLVECpC
| |
| 19:52 | <vagrantc> alkisg: i'm not so sure of that.
| |
| 19:52 | <cyberpelopo> sorry, did'n know
| |
| 19:53 | <eddytv> alkisg: yikes, I ran 'ltsp-chroot -c -p apt-get purge pulseaudio' and it wants to remove ltsp-client.
| |
| 19:53 | <Gremble> cyberpelopo, now pastebin your dhcpd.conf, something is wrong there
| |
| 19:55 | <cyberpelopo> http://pastebin.com/rMTZNzy3
| |
| 19:55 | <alkisg> eddytv: yeah ltsp-client depends on pulseaudio, maybe it should recommend it instead, as it's not necessary on fat clients nor on xfreerdp clients...
| |
| 19:56 | <eddytv> Bummer. For a MythTV frontend, you don't want pulseaudio... I guess I can just disable it to avoid breaking the dependency?
| |
| 19:57 | <alkisg> eddytv: just set sound=false in lts.conf
| |
| 19:57 | (caps, SOUND=false)
| |
| 19:57 | eddytv: btw why, doesn't mythtv play well with pulseaudio?
| |
| 19:58 | <Gremble> cyberpelopo, see private messages
| |
| 19:58 | <cyberpelopo> ??
| |
| 19:59 | how can i see private messages?
| |
| 19:59 | <Gremble> sorry, /query window, I sent them to your other login
| |
| 19:59 | copied to your current one now
| |
| 19:59 | <eddytv> Well, if mythtv sees PulseAudio running, it somehow "suspends it" while MythTV is running and then re-actives it when you quit. Not sure of the technical details as to why. But since MythTV doesn't want it/use it, I'd rather not have it running.
| |
| 20:00 | [GuS] has left IRC ([GuS]!~MysT@unaffiliated/gus/x-663402, Quit: Konversation terminated!) | |
| 20:01 | <Gremble> are you getting the PMs cyberpelopo ?
| |
| 20:08 | bieb has left IRC (bieb!~hbieber@173.226.16.130) | |
| 20:11 | <vagrantc> alkisg: here's a question ... should localapps maintain it's own ldm/rc.d hook section, or should that be moved to share/ldm/rc.d ?
| |
| 20:11 | <alkisg> vagrantc: moved to ldm/rc.d ?
| |
| 20:12 | I think it's better this way, fat clients also use some parts of their code
| |
| 20:12 | <vagrantc> alkisg: no, ltsp-trunk/share/ldm/rc.d
| |
| 20:12 | er, ltsp-trunk/client/share/ldm/rc.d
| |
| 20:12 | <alkisg> Ah I thought you were going to rename share to ltsp
| |
| 20:13 | <vagrantc> it definitely belongs in ltsp, as it's a feature that ltsp-client* provides by hooking into ldm's rc.d structure
| |
| 20:13 | <alkisg> Wait you have both share and ldm?
| |
| 20:13 | <vagrantc> ldm could run without it.
| |
| 20:13 | <alkisg> I thought it would be ltsp-trunk/client/ltsp and ltsp-trunk/client/ldm
| |
| 20:13 | <vagrantc> alkisg: there's now ltsp-trunk/client/share/ltsp and ltsp-trunk/client/share/ldm
| |
| 20:14 | <alkisg> OK
| |
| 20:14 | <vagrantc> alkisg: you had proposed that as one of the options, i went with subdirs in share
| |
| 20:14 | <alkisg> I meant without subdirs in share, directly under client/, but anyway :)
| |
| 20:14 | <vagrantc> yes, i know.
| |
| 20:14 | :)
| |
| 20:14 | <alkisg> Ah :)
| |
| 20:15 | We may also do the same with an etc/ dir :D
| |
| 20:15 | <vagrantc> i think it's clearer where they actually go that way...
| |
| 20:15 | oh fun.
| |
| 20:15 | <alkisg> /etc/ltsp, /etc/xinetd, /etc/nbd-server/, etc etc
| |
| 20:15 | <vagrantc> should we just ... mirror where everything goes?
| |
| 20:15 | <alkisg> Hehe... only when it's really easier for packaging, not for things that only need a couple of lines in install
| |
| 20:16 | <vagrantc> for things that have build dirs (nbd-proxy, localapps, getltscfg) it doesn't quite work, but for everything else
| |
| 20:18 | did we kill off ltsp-keys?
| |
| 20:18 | the one that went in if-up.d ?
| |
| 20:19 | and what happened to server//plugins/functionns
| |
| 20:19 | and what happened to server//plugins/functions
| |
| 20:19 | oh, that got merged into ltsp-build-client ... we had dreams of the plugin system being used for more than ltsp-build-clienty
| |
| 20:21 | <alkisg> (11:18:15 μμ) vagrantc: did we kill off ltsp-keys? => dunno, but we shouldn't need it anymore now that we use hostnames, right?
| |
| 20:23 | <vagrantc> haven't really tested that code much :)
| |
| 20:23 | cyberpelopo has left IRC (cyberpelopo!b29c559c@gateway/web/freenode/ip.178.156.85.156, Ping timeout: 245 seconds) | |
| 20:23 | * alkisg is waiting for ltsp-server-dnsmasq to start the heavy testing | |
| 20:23 | cyberpepolo has joined IRC (cyberpepolo!b29c559c@gateway/web/freenode/ip.178.156.85.156) | |
| 20:24 | <knipwim> well, i'm off to bed, see you guys next week
| |
| 20:25 | and what's left of the tree :)
| |
| 20:26 | cyberpelopo has joined IRC (cyberpelopo!b29c559c@gateway/web/freenode/ip.178.156.85.156) | |
| 20:28 | Gremble has left IRC (Gremble!~Ben@cpc10-aztw24-2-0-cust114.aztw.cable.virginmedia.com, Quit: I Leave) | |
| 20:28 | <vagrantc> we've done a bit of pruning and grafting today, yes.
| |
| 20:28 | a frankenstein tree.
| |
| 20:31 | cyberpelopo has left IRC (cyberpelopo!b29c559c@gateway/web/freenode/ip.178.156.85.156, Client Quit) | |
| 20:39 | <eddytv> any reason LTSP depends on so many foreign-language fonts?
| |
| 20:40 | <alkisg> 'night all
| |
| 20:40 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Quit: Leaving.) | |
| 20:46 | andygraybeal_ has left IRC (andygraybeal_!~andy@h174.212.22.98.dynamic.ip.windstream.net, Read error: Operation timed out) | |
| 20:47 | Trixboxer has left IRC (Trixboxer!~Trixboxer@115.124.115.71, Quit: "Achievement is not the end, its the beginning of new journey !!!") | |
| 20:50 | <vagrantc> well, first go of a package build failed ... but it's probably easier to just keep rebuilding till it suceeds.
| |
| 20:53 | staffencasa_ has joined IRC (staffencasa_!~staffenca@128.193.8.220) | |
| 20:53 | staffencasa__ has joined IRC (staffencasa__!~staffenca@128.193.8.220) | |
| 20:55 | staffencasa has left IRC (staffencasa!~staffenca@128.193.8.220, Ping timeout: 244 seconds) | |
| 20:59 | andygraybeal_ has joined IRC (andygraybeal_!~andy@h216.88.31.71.dynamic.ip.windstream.net) | |
| 21:06 | Steve_The_Pirate has left IRC (Steve_The_Pirate!~Gary@cpc16-haye17-2-0-cust169.haye.cable.virginmedia.com, Quit: Leaving) | |
| 21:10 | komunista has left IRC (komunista!~slavko@adsl-195-098-014-163.dynamic.nextra.sk, Ping timeout: 252 seconds) | |
| 21:26 | stgraber has left IRC (stgraber!~stgraber@ubuntu/member/stgraber, Ping timeout: 248 seconds) | |
| 21:45 | <vagrantc> hrm. merging the plugins for ltsp-build-client directly into ltsp-build-client broke any possibility of cross-distro installs.
| |
| 21:45 | well, if you expect certain functions to be present.
| |
| 21:46 | or rather, only debian ships some functions.
| |
| 21:46 | stgraber has joined IRC (stgraber!~stgraber@ubuntu/member/stgraber) | |
| 21:48 | <vagrantc> i guess i can just add in ltsp_tool-vendor-functions
| |
| 21:48 | hmm...
| |
| 21:57 | dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Quit: Leaving...) | |
| 22:14 | staffencasa__ has left IRC (staffencasa__!~staffenca@128.193.8.220, Quit: Leaving) | |
| 22:14 | staffencasa_ has left IRC (staffencasa_!~staffenca@128.193.8.220, Quit: Leaving) | |
| 22:14 | staffencasa has joined IRC (staffencasa!~staffenca@128.193.8.220) | |
| 22:21 | <eddytv> Any idea what starts 'pulseaudio' on an LTSP client? I set "SOUND=false" in lts.conf and it still started, then I ran "update-rc.d" to remove the startup scripts from /etc/rc*.d, but it's still getting started on the clients.
| |
| 22:23 | <vagrantc> SOUND=false in lts.conf should disable the pulseaudio calls from LTSP
| |
| 22:24 | if it's actually reading your lts.conf
| |
| 22:24 | though something else (i.e. mythtv?) might start it...
| |
| 22:24 | <eddytv> OK. That's what I thought. Let me double check.
| |
| 22:25 | Is lts.conf case-sensitive?
| |
| 22:26 | FWIW, I grep'd in /opt/ltsp/*/use/share/ltsp/init-ltsp.d for 'sound' and didn't see any references in there
| |
| 22:26 | Are there other places that process lts.conf directives?
| |
| 22:27 | <vagrantc> not case sensitive, no.
| |
| 22:28 | lots
| |
| 22:33 | <eddytv> any idea where the SOUND stuff is parsed / pulseaudio is started?
| |
| 22:34 | <vagrantc> with 12.04?
| |
| 22:35 | we just restructure the entire source tree today, i don't remember where anything is :)
| |
| 22:42 | <eddytv> LOL
| |
| 22:45 | <vagrantc> eddytv: most everything ltsp does is in /opt/ltsp/i386/usr/share/ltsp/
| |
| 22:45 | eddytv: grep should get you far
| |
| 22:47 | well, the latest and greatest ltsp on debian now kernel panics at init ... old version works fine.
| |
| 22:47 | hrm.
| |
| 22:48 | bobby_C has left IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at, Ping timeout: 256 seconds) | |
| 22:52 | <vagrantc> or that was just a freak install.
| |
| 22:52 | seems fine now.
| |
| 22:53 | * vagrantc retests... | |
| 22:56 | JesseC has left IRC (JesseC!JesseCW@wsip-98-175-20-126.br.br.cox.net, ) | |
| 23:08 | <vagrantc> for reasons i cannot determine, the new ltsp init-ltsp.d scripts are more fragile than ltsp 5.3.7
| |
| 23:09 | init-ltsp completely dies if something in any script fails
| |
| 23:18 | enyc has left IRC (enyc!~enyc@muddle.enyc.org.uk, Ping timeout: 272 seconds) | |
| 23:19 | <vagrantc> seems related to NBD swap
| |
| 23:41 | what i can't figure out is what's changed...
| |