IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 29 December 2013 (all times are UTC)
| 00:00 | <gdi2k> I'm also seeing a lot of corrupted login graphics - after login things are fine, but the graphics on the login screen is corrupted on many clients here
| |
| 00:00 | gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Ping timeout: 260 seconds) | |
| 00:00 | <gdi2k> (it's not a big deal, they still work, but it does look ugly)
| |
| 00:02 | xet7 has left IRC (xet7!~xet7@80-186-81-85.elisa-mobile.fi, Quit: Lähdössä) | |
| 00:18 | gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com) | |
| 00:37 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 245 seconds) | |
| 00:58 | markit has left IRC (markit!~marco@88-149-177-66.v4.ngi.it, ) | |
| 01:18 | gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Remote host closed the connection) | |
| 01:18 | Enslaver has left IRC (Enslaver!~Enslaver@fedora/Enslaver, Read error: Connection reset by peer) | |
| 01:20 | <gdi2k> just wanted to provide some feedback on my audio issue on fat clients (only "Dummy output" shows up in Pulse Audio Volume Control. Maybe obvious in hindsight, but it turns out users on fat clients must be added to the audio group for them to be able to access audio devices on the fat client. This is not true for thin clients, and also not for regular local users (such as from a live cd). Issue resolved :)
| |
| 01:22 | Enslaver has joined IRC (Enslaver!~Enslaver@fedora/Enslaver) | |
| 01:36 | adrianorg has left IRC (adrianorg!~adrianorg@177.156.58.138, Ping timeout: 252 seconds) | |
| 01:37 | adrianorg has joined IRC (adrianorg!~adrianorg@177.156.58.138) | |
| 01:50 | gdi2k has left IRC (gdi2k!~gdi2k@222.127.254.113, Remote host closed the connection) | |
| 01:59 | gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com) | |
| 02:00 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 02:01 | <alkisg> vagrantc: hi! two packaging changes: https://code.launchpad.net/~ts.sch.gr/sch-scripts/ltsp-debian-packaging
| |
| 02:05 | gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Ping timeout: 252 seconds) | |
| 02:10 | <alkisg> OK, reports...
| |
| 02:10 | First, xprop -spy works fine, so localapps are fine
| |
| 02:10 | ltsp-localapps xterm was actually more snappier now that there's no delay+polling involved
| |
| 02:14 | vagrantc: cdrom without cdpinger didn't work for me... /me searches why..
| |
| 02:18 | alexqwesa has left IRC (alexqwesa!~alex@109.172.12.47, Ping timeout: 252 seconds) | |
| 02:21 | <alkisg> Some rules are missing...
| |
| 02:21 | ...Finally, pxelinux.cfg/default wasn't upgraded with ltsp-update-image etc, I'll test that tomorrow too
| |
| 02:26 | <vagrantc> alkisg: i had it not install extra rules on ubuntu
| |
| 02:26 | alkisg: it will only upgrade it if the chroot has pxelinux.cfg/ltsp
| |
| 02:26 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Ping timeout: 245 seconds) | |
| 02:27 | alexqwesa has joined IRC (alexqwesa!~alex@109.172.12.47) | |
| 03:23 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 03:30 | alexqwesa has left IRC (alexqwesa!~alex@109.172.12.47, Ping timeout: 246 seconds) | |
| 04:14 | gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com) | |
| 04:19 | gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Ping timeout: 245 seconds) | |
| 04:39 | |Paradox| has left IRC (|Paradox|!iamparadox@c-71-206-130-134.hsd1.va.comcast.net, Excess Flood) | |
| 04:41 | |Paradox| has joined IRC (|Paradox|!~iamparado@c-71-206-130-134.hsd1.va.comcast.net) | |
| 04:44 | |Paradox| has left IRC (|Paradox|!~iamparado@c-71-206-130-134.hsd1.va.comcast.net, Excess Flood) | |
| 04:44 | |Paradox| has joined IRC (|Paradox|!iamparadox@c-71-206-130-134.hsd1.va.comcast.net) | |
| 05:13 | cyberorg has left IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg, Ping timeout: 246 seconds) | |
| 05:24 | F-GT has left IRC (F-GT!~phantom@ppp59-167-136-109.static.internode.on.net, Read error: Connection reset by peer) | |
| 05:29 | F-GT has joined IRC (F-GT!~phantom@ppp59-167-136-109.static.internode.on.net) | |
| 05:47 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 272 seconds) | |
| 05:49 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 06:13 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Quit: Leaving.) | |
| 06:23 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 06:26 | * vagrantc waves to alkisg | |
| 06:26 | <alkisg> Hi vagrantc, hi all!
| |
| 06:27 | vagrantc: so pxelinux updating isn't supposed to work in ltsp-pnp?
| |
| 06:27 | <vagrantc> alkisg: should work fine...
| |
| 06:27 | alkisg: you'll have to run /usr/share/ltsp/update-kernels first, though
| |
| 06:28 | if you're upgrading
| |
| 06:28 | should work fine on a fresh install
| |
| 06:28 | presuming you've got updated version of both update-kernels (ltsp-client-core) and ltsp-update-kernels (ltsp-server)
| |
| 06:28 | <alkisg> vagrantc: ah, cool, so, update-kernels deleted pxelinux.cfg/default and created /ltsp,
| |
| 06:29 | * vagrantc nods | |
| 06:29 | <alkisg> so now ltsp-update-image -c / && ltsp-update-kernels should create it in tftp only?
| |
| 06:29 | <vagrantc> that was how i figured out how to make it backwards compatible
| |
| 06:29 | i.e. if default is present in the chroot, ltsp-update-kernels doesn't do anything
| |
| 06:30 | <alkisg> The vmlinuz symlink etc was replaced by vmlinuz-generic...
| |
| 06:30 | <vagrantc> alkisg: yup.
| |
| 06:30 | shouldn't have deleted any symlinks...
| |
| 06:30 | <alkisg> Ah no actually that symlink wasn't there in /boot, only in tftp
| |
| 06:30 | <vagrantc> should have just created more symlinks
| |
| 06:30 | <alkisg> Now the symlinks are in /boot
| |
| 06:30 | <vagrantc> and should get copied?
| |
| 06:31 | <alkisg> ltsp-update-image will take a long time.... lets see... /me runs it
| |
| 06:31 | <vagrantc> alkisg: regarding ltspfs ... i had it not create any extra udev rules, as you said those weren't needed on ubuntu
| |
| 06:32 | <alkisg> vagrantc: now it doesn't have any rules to call ltspfsmounter at all
| |
| 06:32 | <vagrantc> no rules at all?
| |
| 06:32 | <alkisg> I think when I complained, it had rules for both add and change
| |
| 06:32 | <vagrantc> yeah.
| |
| 06:32 | <alkisg> Now I don't think it matches CD drives at all
| |
| 06:32 | <vagrantc> it should work if you have media inserted, at the very least
| |
| 06:33 | i.e. boot with media inserted
| |
| 06:37 | <alkisg> Ah, this one:
| |
| 06:37 | # other drives:
| |
| 06:37 | ACTION=="add", SUBSYSTEM=="block", ENV{ID_TYPE}!="floppy", RUN+="ltspfs_entry add %k"
| |
| 06:38 | <vagrantc> basically, if you need extra rules, i kinda implemented an ugly sed hack at build time.
| |
| 06:47 | <alkisg> root@ltsp241:~# cat /var/run/ltspfs_fstab (inserted on boot, it worked fine)
| |
| 06:47 | /dev/sr0 /var/run/drives/cdrom iso9660 defaults,utf8 0 0
| |
| 06:47 | root@ltsp241:~# cat /var/run/ltspfs_fstab (inserted after login, didn't work)
| |
| 06:47 | /dev/sr0 /var/run/drives/cdrom-sr0 iso9660,udf defaults 0 0
| |
| 06:48 | ...after a few more insertions (vbox)...
| |
| 06:48 | root@ltsp241:~# cat /var/run/ltspfs_fstab
| |
| 06:48 | /dev/sr0 /var/run/drives/cdrom-sr0 iso9660,udf defaults 0 0
| |
| 06:48 | /dev/sr0 /var/run/drives/cdrom iso9660,udf defaults 0 0
| |
| 06:48 | /dev/sr0 /var/run/drives/cdrom-sr0 iso9660,udf defaults 0 0
| |
| 06:48 | <vagrantc> alkisg: oh, i tried AoE ... it's awesome!
| |
| 06:49 | alkisg: although i had to tell it to bring up networking!
| |
| 06:49 | * alkisg likes the "networking doesn't need to be configured" part | |
| 06:49 | <alkisg> vagrantc: did you put the script that I pastebin'ed, that does exactly that?
| |
| 06:50 | $ pastebinit /usr/share/initramfs-tools/scripts/local-top/aoe_ltsp
| |
| 06:50 | http://pastie.org/8584378
| |
| 06:51 | <vagrantc> alkisg: i dropped that in local-top, whipped up a CMDLINE_AOE bit, and used IPAPPEND=3 to configure networking
| |
| 06:51 | gdi2k has joined IRC (gdi2k!~gdi2k@222.127.254.113) | |
| 06:51 | <alkisg> It shouldn't need IPAPPEND...
| |
| 06:51 | The older version would only bring up eth0 though, so if the client has multiple nics use the latest ^ pastebin
| |
| 06:51 | <vagrantc> well, LTSP kind of assumes you have networking
| |
| 06:51 | maybe i've got your older script
| |
| 06:52 | <alkisg> Actually the udhcp script always assigns an IP, so I never tested booting without an IP...
| |
| 06:52 | <vagrantc> i did.
| |
| 06:53 | it was slightlly mind-blowing
| |
| 06:53 | so, the quickest, easiest way i set it up to use networking was with IPAPPEND 3, and then LTSP worked better
| |
| 06:54 | <alkisg> I think it also behaves much better than nbd when there's lost connectivity for a while
| |
| 06:54 | <vagrantc> wonder if i can simulate that in libvirt easily
| |
| 06:55 | guess i could ifdown the server's interface
| |
| 06:55 | <alkisg> With vbox I had to use bridged networking, it wouldn't work at all with host-only networking
| |
| 06:55 | ltspfs /media/root/cdrom-sr0 fuse.ltspfs rw,nosuid,nodev,relatime,user_id=0,group_id=0 0 0
| |
| 06:55 | ltspfs /media/angelos/cdrom-sr0 fuse.ltspfs rw,nosuid,nodev,relatime,user_id=1014,group_id=1014 0 0
| |
| 06:55 | ltspfs /media/root/cdrom fuse.ltspfs rw,nosuid,nodev,relatime,user_id=0,group_id=0 0 0
| |
| 06:55 | ltspfs /media/angelos/cdrom fuse.ltspfs rw,nosuid,nodev,relatime,user_id=1014,group_id=1014 0 0
| |
| 06:55 | ltspfs /tmp/.root-ltspfs/cdrom-sr0 fuse.ltspfs rw,nosuid,nodev,relatime,user_id=0,group_id=0 0 0
| |
| 06:55 | ltspfs /tmp/.angelos-ltspfs/cdrom-sr0 fuse.ltspfs rw,nosuid,nodev,relatime,user_id=1014,group_id=1014 0 0
| |
| 06:55 | That's my `cat /proc/mounts`, I assume I'm missing some remove rule?
| |
| 06:56 | <vagrantc> it should auto-remove the entry on failure
| |
| 06:56 | * alkisg is using the ltspfsd.rules that debian uses now... | |
| 06:56 | <alkisg> I should probably try it in 2 steps though from vbox, first remove, then insert another
| |
| 06:56 | <vagrantc> it's probably triggering on both add and change...
| |
| 07:03 | * alkisg can't unmount those even manually: | |
| 07:03 | <alkisg> ltspfs /media/angelos/cdrom-sr0 fuse.ltspfs rw,nosuid,nodev,relatime,user_id=1014,group_id=1014 0 0
| |
| 07:03 | ltspfs /media/root/cdrom fuse.ltspfs rw,nosuid,nodev,relatime,user_id=0,group_id=0 0 0
| |
| 07:03 | ltspfs /media/angelos/cdrom fuse.ltspfs rw,nosuid,nodev,relatime,user_id=1014,group_id=1014 0 0
| |
| 07:06 | vagrantc: pxelinux.cfg/default now is supposed to be a symlink?
| |
| 07:06 | If so, it worked fine :)
| |
| 07:08 | vagrantc: err, except I didn't have any "default" entries
| |
| 07:08 | So I got stuck at the "boot:" prompt
| |
| 07:12 | <vagrantc> alkisg: hrm.
| |
| 07:12 | wonder why it didn't have default entries
| |
| 07:13 | alkisg: so yeah, pxelinux.cfg/default should be a symlink pointing to pxelinux.cfg/ltsp
| |
| 07:13 | (well, "ltsp", but yeah)
| |
| 07:14 | alkisg: so maybe you need the "change" remove events, but not the add events?
| |
| 07:14 | <alkisg> Maybe... let me try
| |
| 07:19 | ltspfs_entry: # if user's home directory is mounted via sshfs, assume we want a local
| |
| 07:19 | # mount for localapps
| |
| 07:19 | gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com) | |
| 07:19 | <alkisg> ...the code there is a bit broken, it's seeing /home/Shared as a user...
| |
| 07:21 | cyberorg has joined IRC (cyberorg!~cyberorg@opensuse/member/Cyberorg) | |
| 07:23 | gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Ping timeout: 245 seconds) | |
| 07:29 | <highvoltage> /win 14
| |
| 07:35 | <alkisg> ltspfs testing is a pain on vbox, I'll try it again at work
| |
| 07:35 | <vagrantc> heh
| |
| 07:35 | <alkisg> vagrantc: do you have any stuff that's not committed on -trunk yet?
| |
| 07:35 | * vagrantc should set up a virtual ubuntu server | |
| 07:36 | <vagrantc> alkisg: i don't think so
| |
| 07:36 | alkisg: so what's broken with the pxelinux.cfg/default symlink?
| |
| 07:36 | <alkisg> It lacks those:
| |
| 07:36 | default ltsp-NBD
| |
| 07:36 | ontimeout ltsp-NBD
| |
| 07:37 | <vagrantc> it has no entries for default or ontimeout?
| |
| 07:39 | <alkisg> Right
| |
| 07:39 | <vagrantc> hah.
| |
| 07:39 | it's because we remove it after generating that header...
| |
| 07:40 | fix on it's way to ltsp-trunk
| |
| 07:42 | clearly i didn't test that too well...
| |
| 07:42 | i think that's why i hesitated to commit it, actually
| |
| 07:43 | alkisg: fixed in trunk.
| |
| 07:46 | <alkisg> vagrantc: nice, it works fine now
| |
| 07:46 | vagrantc: what's left for me to test, except for cdpingerless ltspfs?
| |
| 07:47 | <vagrantc> alkisg: i think that's it for testing
| |
| 07:47 | <alkisg> Hrm. I need more :P
| |
| 07:47 | OK, what's left to do before branching ltsp6?
| |
| 07:47 | <vagrantc> alkisg: the only other question i had was using a pxelinux.cfg/default file with an include instead of the symlink.
| |
| 07:48 | <alkisg> vagrantc: nah symlinking is fine, it avoid the second tftp request
| |
| 07:48 | <vagrantc> alkisg: it would be a little easier for users to simply edit it than remove the symlink and create a pxelinux.cfg/default with an include line
| |
| 07:48 | <alkisg> We can do the include stuff when we merge everything under /pxelinux.cfg
| |
| 07:49 | <vagrantc> alkisg: yeah, that's why i went with symlink.
| |
| 07:49 | alkisg: it's still possible for the admin to edit it, but it's still only one tftp request
| |
| 07:50 | <alkisg> I like the idea, maybe we can do that too in ltsp6
| |
| 07:50 | <vagrantc> converting it from a symlink with a template with ltsp-config might be interesting
| |
| 07:50 | that could have the localized comments and all that in the templates
| |
| 07:51 | <alkisg> ltsp-config pxe --menu --theme=xxx :P
| |
| 07:51 | Yeah, that sounds good
| |
| 07:53 | * alkisg wonders if vagrantc would agree in dropping all of our older-versions-compatibility-code for ltsp 6... | |
| 07:55 | <alkisg> E.g. "alkisg: let's change the output format of ldminfod! vagrantc: ok!" :P
| |
| 07:57 | <vagrantc> hah!
| |
| 08:00 | staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu) | |
| 08:07 | staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Ping timeout: 240 seconds) | |
| 08:08 | alkisg1 has joined IRC (alkisg1!~alkisg@ubuntu/member/alkisg) | |
| 08:09 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Ping timeout: 260 seconds) | |
| 08:11 | alkisg1 is now known as alkisg | |
| 08:21 | gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com) | |
| 08:26 | gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Ping timeout: 252 seconds) | |
| 08:34 | vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Quit: leaving) | |
| 08:38 | <gdi2k> is there a way of allowing users of fat clients to reboot / shutdown their machines? Currently they prompted for root's password when they want to reboot etc. from the DE. If they log out and reboot from LDM it works fine, but users are not understanding
| |
| 08:44 | <alkisg> gdi2k: which de? which distro/version?
| |
| 08:45 | <gdi2k> Xubuntu 13.10
| |
| 08:45 | XFCE
| |
| 08:46 | <alkisg> In normal xubuntu (non ltsp), are users not in the "sudo" group able to shutdown the system?
| |
| 08:48 | Btw gdi2k I saw somewhere (in the irc logs? in the mailing list?) that you were having some other issues with fat clients that you shouldn't have, and I'm wondering if your setup is a bit broken...
| |
| 08:49 | Ah, right, users in the audio group. They shouldn't need to be in the audio group.
| |
| 08:52 | <gdi2k> for shutdown question: yes, users not a member of sudo can shutdown the machine from within XFCE provided no-one else is logged into the machine. I have a feeling when on the fat client, the DE things there are others logged in - maybe it is trying to shut down the server, not the fat client?
| |
| 08:54 | for the audio issue: I spent quite a bit of time on that, and adding users to "audio" definitely solves it. I read somewhere in some arch documents that if audio devices are accessed "remotely", users need to be a member of the audio group, otherwise not. Was wondering if, because the way fat clients work, this would be considered remote
| |
| 08:56 | <alkisg> About 2 users currently logged in: what's the output of `w` and of `ck-list-sessions` ?
| |
| 08:57 | About the audio, no, only services like pulseaudio should be allowed direct access to audio devices
| |
| 09:00 | bennabiy has left IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com, Read error: Connection reset by peer) | |
| 09:02 | <gdi2k> checking... now I have this issue (for the 3rd time already) where clients don't boot and end up at a busybox prompt with "Target filesystem doesn't have requested /sbin/init-ltsp. restarting nbd server fixes it, but it's a pain. happens every few server reboots
| |
| 09:02 | wondering if my install is borked
| |
| 09:04 | bennabiy has joined IRC (bennabiy!~Thunderbi@24-181-55-79.dhcp.gnvl.sc.charter.com) | |
| 09:05 | <gdi2k> w just shows the user actually using that machine, nothing else. have to install consolekit in chroot to get ck-list-sessions
| |
| 09:05 | <alkisg> No, don't
| |
| 09:06 | <gdi2k> ok
| |
| 09:06 | <alkisg> It might be using logind instead...
| |
| 09:06 | Well, the workaround that I put for fat client users to be considered active was for consolekit
| |
| 09:07 | So if you see systemd/logind in the list of processes, then it's probably not in effect
| |
| 09:07 | And the user might be considered as "remote" in logind, and that would need some fixing in ltsp
| |
| 09:08 | <gdi2k> ok got it, will check on that
| |
| 09:09 | <alkisg> gdi2k: loginctl list-sessions
| |
| 09:10 | And then loginctl show-session <number>
| |
| 09:13 | <gdi2k> http://pastebin.ubuntu.com/6655961/
| |
| 09:16 | <alkisg> gdi2k: if you insert a usb stick, is it automounted?
| |
| 09:23 | gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com) | |
| 09:27 | gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Ping timeout: 240 seconds) | |
| 09:29 | <gdi2k> no, not automountedf
| |
| 09:31 | <alkisg> Yeah then it's a logind issue, it needs a similar workaround like that one with consolekit
| |
| 09:33 | <gdi2k> ok - thanks for the info - would it help to switch to console kit? I think I saw some steps for that on the arch wiki
| |
| 09:33 | (which may or may not apply to Ubuntu)
| |
| 09:41 | <alkisg> It would help to switch to 12.04 :D
| |
| 09:41 | LTSP users should rarely use non-LTS Ubuntu versions...
| |
| 09:42 | They're really untested, the ubuntu ltsp maintainers don't care much about them
| |
| 09:45 | <gdi2k> yeh, I should have considered that - the issue with 12.04 is the outdated versions of libreoffice etc.
| |
| 09:46 | will live with this for a few months, then switch to 14.04 LTS
| |
| 09:47 | (that was the original plan, but we had a server issue, so I thought I'd take the opportunity to update things - maybe not such a great idea!)
| |
| 09:55 | * alkisg doesn't see much different in libreoffice in 12.04 vs 14.04... | |
| 09:56 | gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com) | |
| 09:57 | <gdi2k> 65k max rows in calc gives us a headache, but I think there are PPA with more recent versions
| |
| 10:04 | alkisg, all sorts of stuff isn't working here, so heading to the 12.04 download page :(
| |
| 10:18 | <Lumiere> gdi2k: on a positive note, 12.04 -> 14.04 should be fairly painless when the release the upgrader
| |
| 10:20 | <gdi2k> hope so - maybe I'll leave that step until 2015 lol
| |
| 11:07 | gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Remote host closed the connection) | |
| 11:44 | bybjbx has joined IRC (bybjbx!iamparadox@c-71-206-130-134.hsd1.va.comcast.net) | |
| 11:44 | |Paradox| has left IRC (|Paradox|!iamparadox@c-71-206-130-134.hsd1.va.comcast.net, Read error: Connection reset by peer) | |
| 11:44 | bybjbx is now known as |Paradox| | |
| 12:00 | gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com) | |
| 12:06 | gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Ping timeout: 252 seconds) | |
| 12:18 | khildin has joined IRC (khildin!~khildin@ip-213-49-87-104.dsl.scarlet.be) | |
| 12:40 | alexqwesa has joined IRC (alexqwesa!~alex@109.172.12.47) | |
| 13:02 | gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com) | |
| 13:07 | gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Ping timeout: 245 seconds) | |
| 13:49 | gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com) | |
| 13:58 | bitcheker has joined IRC (bitcheker!~bitchecke@dynamic-adsl-84-221-30-117.clienti.tiscali.it) | |
| 15:06 | PhoenixSTF has joined IRC (PhoenixSTF!~rudi@78.29.154.124) | |
| 15:08 | <gdi2k> alkisg, thanks for your help. 12.04 is up and running now and things are much better behaved. no random errors, sound doesn't require uses to be part of the audio group and all the shutdown / reboot stuff is working. Bad side effect: All users' XFCE panels are a mess due to downgrading XFCE versions
| |
| 15:09 | <alkisg> gdi2k: sudo rm /home/*/.config/lxde/whatever*
| |
| 15:14 | <gdi2k> thanks!
| |
| 15:20 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Ping timeout: 246 seconds) | |
| 15:21 | mattcen has left IRC (mattcen!~mattcen@c110-22-201-130.sunsh4.vic.optusnet.com.au, Ping timeout: 240 seconds) | |
| 15:24 | mattcen has joined IRC (mattcen!~mattcen@c110-22-201-130.sunsh4.vic.optusnet.com.au) | |
| 15:25 | freedomrun has joined IRC (freedomrun!~freedomru@unaffiliated/freedomrun) | |
| 15:28 | lee_ has joined IRC (lee_!~lee@loathe.ms) | |
| 15:33 | PhoenixSTF has left IRC (PhoenixSTF!~rudi@78.29.154.124, *.net *.split) | |
| 15:33 | dsugar100 has left IRC (dsugar100!~dsugar@columbia.tresys.com, *.net *.split) | |
| 15:33 | lee has left IRC (lee!~lee@loathe.ms, *.net *.split) | |
| 15:40 | PhoenixSTF has joined IRC (PhoenixSTF!~rudi@78.29.154.124) | |
| 15:41 | gdi2k has left IRC (gdi2k!~gdi2k@222.127.254.113, Ping timeout: 252 seconds) | |
| 15:57 | staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu) | |
| 16:03 | staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Ping timeout: 246 seconds) | |
| 16:22 | |Paradox| has left IRC (|Paradox|!iamparadox@c-71-206-130-134.hsd1.va.comcast.net, Read error: Connection reset by peer) | |
| 16:22 | |Paradox| has joined IRC (|Paradox|!iamparadox@c-71-206-130-134.hsd1.va.comcast.net) | |
| 16:26 | vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc) | |
| 16:42 | lifeboy has joined IRC (lifeboy!~roland@105-236-159-32.access.mtnbusiness.co.za) | |
| 16:46 | <lifeboy> Hi all, anyone actually around here on a Sunday? I have a strange problem in a debian chroot. I can't set the locales because dpkg-configure gives me a multicoloured screen and the keyboard doesn't work as expected: eg <down arrow> gives me ^[[B and so forth. Any ideas on how to bypass/fix this?
| |
| 16:48 | All the info in can find on the web wrt to this required that I use dpkg-configure :-( which I can't until I've fixed this problem!
| |
| 16:48 | s/info in can find/info I can find/
| |
| 16:50 | When I run the actual chroot as a client, I don't have this problem, it's only in the chroot environment
| |
| 16:56 | <vagrantc> try the command "reset"
| |
| 16:56 | and see if that clears up your terminal
| |
| 16:59 | <lifeboy> It clears up the terminal, but when I run e.g. dpkg-reconfigure locales, I get the same behaviour again.
| |
| 16:59 | <vagrantc> alternately, use a different frontend: sudo ltsp-chroot dpkg-reconfigure --frontend=readline locales
| |
| 17:00 | lifeboy: what's the size of your screen?
| |
| 17:00 | i mean, the window or terminal you're running dpkg-reconfigure from?
| |
| 17:01 | <lifeboy> Yes, I have tried that, it gives another error: Use of uninitialized value $_[1] in join or string at /usr/share/perl5/Debconf/DbDriver/Stack.pm line 111, <FIN> line 2. So it doesn't give me the option to enter a value, it just continues and fails
| |
| 17:01 | I mean when use -freadline
| |
| 17:03 | I connect via ssh, but the window is not too large, I'd say about 800x600.
| |
| 17:04 | <vagrantc> echo $LINES $COLUMNS
| |
| 17:05 | <lifeboy> I've made it smaller, no difference: Currently 21 x 78
| |
| 17:07 | I just tried with terminal 80 x 24 and it's still the same.
| |
| 17:07 | same behaviour.
| |
| 17:08 | <vagrantc> what's $TERM set to?
| |
| 17:08 | <lifeboy> xterm
| |
| 17:08 | <vagrantc> ltsp-chroot dpkg -l | egrep -v ^ii
| |
| 17:09 | i.e. are any packages not fully installed?
| |
| 17:10 | <lifeboy> iU console-common 0.7.87
| |
| 17:10 | iF console-data 2:1.12-2
| |
| 17:10 | <vagrantc> you've got a bunk install
| |
| 17:10 | ltsp-chroot apt-get -f install
| |
| 17:10 | lifeboy: is this a new install?
| |
| 17:11 | <lifeboy> When I do that, the same dialog is shown and it wants to configure the incomplete packages.
| |
| 17:11 | <vagrantc> could you reinstall from scratch? sounds like the initial install may have failed
| |
| 17:12 | <lifeboy> No, it's one that alkisg has been helping me with, but the locales were never properly configured. So I'm trying to fix that.
| |
| 17:13 | zamba has left IRC (zamba!marius@flage.org, *.net *.split) | |
| 17:13 | <lifeboy> I could theoretically create it again, but I'd need to build a debian-vm first at the site. Could I just force remove the incomplete packages?
| |
| 17:14 | <vagrantc> well, those broken packages are a likely indicator that you've got many quirks ahead of you...
| |
| 17:14 | <lifeboy> I have another site where there is a debian-vm that I could use to generate a new chroot and then transfer it across to this site?
| |
| 17:15 | <vagrantc> what were you attempting to do that wasn't part of a typical installation?
| |
| 17:15 | <lifeboy> If I use a chroot that work perfectly at the other site, it should also work, right? Same hardware for thin clients...
| |
| 17:15 | zamba has joined IRC (zamba!marius@flage.org) | |
| 17:15 | freedomrun has left IRC (freedomrun!~freedomru@unaffiliated/freedomrun, Quit: So long and thanks for all the fish) | |
| 17:16 | <vagrantc> sure, you could transfer the chroot, as long as it's done correctly :)
| |
| 17:16 | <lifeboy> It's a debian chroot (wheezy) on Ubuntu 12.04 server
| |
| 17:16 | <vagrantc> was it working, and then it stopped working?
| |
| 17:16 | i.e. did it ever work?
| |
| 17:17 | <lifeboy> I tar'ed the directory, copied it with sftp and untared it. Yes, it was working, ie it boot, but the users can't log in. Some I'm trying to troubleshoot it, but need to have sshd running for instance,
| |
| 17:18 | and wanted to add some other tools
| |
| 17:18 | pretty straight forward stuff, nothing weird :-)
| |
| 17:18 | It's still working.
| |
| 17:18 | <vagrantc> uid/gid mismatches when you untarred it might have caused problems
| |
| 17:19 | <lifeboy> hmm...
| |
| 17:20 | I can check, but I think they where both root uid 1000
| |
| 17:20 | on both servers
| |
| 17:20 | <vagrantc> yes, but the various system users and groups within the chroot are likely different from the server
| |
| 17:21 | lifeboy: how did you tar it up, and how did you untar it?
| |
| 17:21 | <lifeboy> I only have one user (myself) in the chroot and all's the same
| |
| 17:21 | let me check the history...
| |
| 17:21 | <vagrantc> system users, not login accounts
| |
| 17:21 | and system groups
| |
| 17:22 | gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Remote host closed the connection) | |
| 17:23 | <vagrantc> i.e. root is (pretty much) always uid 0, but other system groups, such as "fuse" may have different numeric IDs depending on the system.
| |
| 17:23 | <lifeboy> Are you refering to the server or the chroot?
| |
| 17:24 | <vagrantc> yes.
| |
| 17:24 | :P
| |
| 17:24 | <lifeboy> Sorry, I mistyped above I see, root and uid 1000
| |
| 17:25 | <vagrantc> lifeboy: what question are you answering?
| |
| 17:26 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 17:26 | * vagrantc waves to alkisg | |
| 17:26 | <lifeboy> That was in reference to my comment: "<lifeboy> hmm...
| |
| 17:26 | I can check, but I think they where both root uid 1000
| |
| 17:26 | on both servers"
| |
| 17:26 | <vagrantc> alkisg: seems like you should commit aoe_ltsp and then we call it ltsp 5.4.7 ?
| |
| 17:27 | lifeboy: i was referring to all the other uid/gid
| |
| 17:27 | <lifeboy> regarding the tar command: I used sudo tar -xf /home/roland/i386-wheezy.tar
| |
| 17:28 | from /opt/ltsp/i386
| |
| 17:28 | <vagrantc> lifeboy: i don't think that preserves permissions, does it?
| |
| 17:28 | <alkisg> Hi vagrantc, hi all
| |
| 17:28 | vagrantc: I'm still testing the cdpingerless stuff
| |
| 17:28 | <lifeboy> nope, should I have preserved them.
| |
| 17:28 | <vagrantc> alkisg: but that's not dependent on ltsp
| |
| 17:28 | lifeboy: how did you create the tarball?
| |
| 17:28 | <alkisg> Ah, sure, I was just taking them one by one
| |
| 17:28 | * lifeboy waves at alkisg | |
| 17:28 | <alkisg> Hi lifeboy
| |
| 17:29 | <vagrantc> lifeboy's got a badly broken ltsp install
| |
| 17:29 | <alkisg> vagrantc: I'd also like to revisit udhcp, so I'm leaving those for the next couple of days... cdpinger comes first
| |
| 17:29 | <vagrantc> alkisg: just some bugfix testing, or major rewrite?
| |
| 17:29 | <lifeboy> I created the tarbal using sudo tar -xf /home/roland/i386-wheezy.tar from /opt/ltsp/i386
| |
| 17:30 | <vagrantc> lifeboy: that would extract the tarball, wouldn't it?
| |
| 17:30 | <alkisg> vagrantc: no not major rewrite, just a small rewrite :P to adhere to our coding standards, to stop using sleep and use wait_for_udev, to not use an external script etc
| |
| 17:30 | <lifeboy> duh, I copied the wrong line...
| |
| 17:31 | * vagrantc wonders when wait_for_udev was introduced | |
| 17:31 | <alkisg> vagrantc: also, I'd like to test aoe with real clients (to check for flow control issues) before committing it, in case we need to disable flow control on the server NICs too
| |
| 17:31 | It's there in 12.04, don't know about before
| |
| 17:32 | * alkisg would like to limit his backports-caring factor to stable-1 :D | |
| 17:32 | <vagrantc> seems to be present in wheezy, so that's good.
| |
| 17:32 | <alkisg> Soo.... I don't get "add" events in 14.04 non-ltsp clients
| |
| 17:33 | <vagrantc> alkisg: i typically only really care about the current debian stable and testing, but i like to at least *know* if it works on oldstable
| |
| 17:33 | <alkisg> I wonder if that's due to 14.04 being more in sync with debian, or if I was using the wrong method while testing, possibly with cdpinger still running...
| |
| 17:33 | <vagrantc> alkisg: and if it's easy to support on oldstable, then may as well.
| |
| 17:34 | <alkisg> vagrantc: so you have a squeeze chroot around for such tests?
| |
| 17:34 | <lifeboy> vagrantc: It seems I used the tarbal I kept as a backup and just extracted that. I can make a fresh one, transfer it to the other site and set it up again. Let's see if that works then
| |
| 17:34 | <vagrantc> alkisg: i've mostly given up on squeeze for now :)
| |
| 17:34 | * alkisg could keep some installations around for tests... 14.04, 12.04, 10.04, and jessie, wheezy, squeeze, and maybe even some fedora + opensuse VM | |
| 17:34 | <vagrantc> lifeboy: you'll probably want: tar xpf
| |
| 17:35 | <lifeboy> vagrantc: Should I make sure the permissions are kept?
| |
| 17:35 | when I make the tarbal?
| |
| 17:35 | <vagrantc> alkisg: since squeeze is end-of-support in just 5 months, i don't really bother at this point
| |
| 17:35 | <alkisg> Ah, cool
| |
| 17:35 | <vagrantc> lifeboy: that would make it a *meaningful* backup :)
| |
| 17:36 | alkisg: i.e. i only bother backporting testing/unstable to stable, and stable to oldstable...
| |
| 17:37 | alkisg: but i do like to keep the versioned dependencies correct, which is why i ask about some things
| |
| 17:37 | <lifeboy> I've never considered that it may be a problem, since everything was running as root anyway when the thin client boots up to the point where the user logon occurs, but that is on the server, not the thin client, not so?
| |
| 17:37 | <vagrantc> lifeboy: not really.
| |
| 17:37 | <alkisg> vagrantc: sure, but "does wait_for_udev exist in squeeze" is roughly equivalent do "look it up in a squeeze chroot" :)
| |
| 17:37 | So I'll keep some chroots around to be sure...
| |
| 17:38 | <vagrantc> alkisg: you're too kind :)
| |
| 17:38 | <lifeboy> I'll just make sure permissions are kept :-)
| |
| 17:39 | <alkisg> vagrantc: we don't have anywhere a "how to mount localdevs manually", right?
| |
| 17:39 | ...in order to better debug the problems I see...
| |
| 17:39 | <vagrantc> alkisg: the only place i remember is that broken wiki page
| |
| 17:39 | well, outdated
| |
| 17:39 | <alkisg> Yeah that was too broken, nm
| |
| 17:47 | <vagrantc> alkisg: so, if you add the udev rules that work for me on debian, it generates too many entries and they don't get removed from ltspfs_fstab ?
| |
| 17:48 | * alkisg is testing exactly now... trying to note things as best as I can | |
| 17:58 | <alkisg> When I insert a disk, it gets mounted locally but I don't see it on the server... checking why...
| |
| 17:58 | <lifeboy> vagrantc: I see the default for making a tarbal as root is to preserve permissions, so I think my tarbal is ok. I also purged console-common, console-data and locales from the chroot and now there are no more complaints it seems
| |
| 18:01 | <vagrantc> lifeboy: are they in a partially configured state on the original chroot?
| |
| 18:01 | <lifeboy> vagrantc: No
| |
| 18:02 | <alkisg> vagrantc: I'm looking at ltspfs_entry and it looks a bit broken... e.g. I have a problem with the label, I'm always getting plain "cdrom" because of broken code, do you get that too?
| |
| 18:03 | <vagrantc> alkisg: i think we intentionally set it to "cdrom" in order for some desktop icons to automatically show up with the "right" icon.
| |
| 18:04 | <alkisg> case $link in cdrom*) LABEL="$link"
| |
| 18:04 | <lifeboy> vagrantc: Actually the locales package is installed in the original chroot and it works fine, also dpkg-reconfigure works fine.
| |
| 18:05 | <vagrantc> lifeboy: so how did it get broken?
| |
| 18:05 | <alkisg> Dunno it looks like we're going into a whole lot of trouble, calling udevinfo etc just to create a static label?
| |
| 18:05 | This is a bit broken too, it just provides 1 alternative filename, it doesn't generate a semi-random one... # Let's change the label LABEL="${LABEL}-${DEVICENAME}"
| |
| 18:06 | <lifeboy> vagrantc: I'm not sure. I'll try installing it again and see what happens.
| |
| 18:06 | <alkisg> So I ended up having 2 plain "cdrom" and 3 "cdrom-sr0" :D
| |
| 18:07 | <vagrantc> alkisg: dunno. works fine for me...
| |
| 18:07 | even tried with multiple CDROMs
| |
| 18:07 | <alkisg> whoah ltspfsmounter is in python...
| |
| 18:07 | <vagrantc> even did some testing with real USB CDROM drives using usb-passthrough on a virtual thinclient :)
| |
| 18:08 | <alkisg> Hehe
| |
| 18:08 | highvoltage has left IRC (highvoltage!~highvolta@abathur.jonathancarter.org, Ping timeout: 252 seconds) | |
| 18:09 | <vagrantc> alkisg: i wonder what's so different with your setup that it's triggering such different issues
| |
| 18:10 | highvoltage has joined IRC (highvoltage!~highvolta@ubuntu/member/highvoltage) | |
| 18:10 | <alkisg> ...will know in a while :)
| |
| 18:14 | <vagrantc> alkisg: well, if your setup is triggering two events, that could create race conditions...
| |
| 18:14 | <alkisg> vagrantc: currently, the part that's missing seems to be the ltspfsmounter call on the server
| |
| 18:15 | <vagrantc> alkisg: that would happen if there was no ID_FS_TYPE or FS_TYPE or whatever it's called
| |
| 18:15 | i think
| |
| 18:15 | <alkisg> There's iso9660,udf
| |
| 18:15 | Maybe the comma is breaking things?
| |
| 18:16 | <vagrantc> shouldn't
| |
| 18:16 | "mount -t iso9660,udf /dev/cdrom /mnt" works for me
| |
| 18:17 | <alkisg> vagrantc: can you fill in the missing bits for me?
| |
| 18:17 | udev calls ltspfs_entry
| |
| 18:17 | <vagrantc> interestingly enough, maybe we shouldn't hard-code that with CDROMs ... that may have only been needed for cdpinger...
| |
| 18:17 | alkisg: you are testing with all the changes in place?
| |
| 18:18 | <alkisg> Yup
| |
| 18:18 | At least, if our -trunk and -packaging are correct...
| |
| 18:18 | <vagrantc> alkisg: i'd start putting some echo debugging in ltspfs_entry
| |
| 18:18 | <alkisg> I've put some
| |
| 18:18 | So, it mounts things locally, as the user and as root
| |
| 18:19 | And then it's supposed to call ltspfsmounter on the server
| |
| 18:19 | Next, I tried to call ltspfsmounter manually from gnome-terminal
| |
| 18:19 | Like, ltspfsmounter /media/alkisg/cdrom add
| |
| 18:20 | ...and it errored with "/media/alkisg not mounted"
| |
| 18:20 | <vagrantc> this is on 14.04 or 12.04 ?
| |
| 18:20 | <alkisg> 14.04
| |
| 18:20 | * alkisg is missing some bits there, what does ltspfsmounter exactly do? Why does it need access to $DISPLAY? | |
| 18:21 | <vagrantc> it's been so long since i've really had to debug this stuff ... it's "just worked" for the most part.
| |
| 18:22 | alkisg: it uses the DISPLAY to get the ltspfs token
| |
| 18:22 | ltspfsd will refuse connections that don't have the correct token
| |
| 18:22 | it's a weak security measure
| |
| 18:23 | the ltspfs_token should be in an xatom/xproperty
| |
| 18:25 | <alkisg> Ah so I need SSH_CONNECTION to find out the IP... it doesn't use LTSP_CLIENT
| |
| 18:26 | <vagrantc> it does, at least currently, rely on the ssh tunnel
| |
| 18:29 | <alkisg> vagrantc: does "ltspfsmounter /media/alkisg/cdrom add" sound correct?
| |
| 18:30 | It says "wrote $COOKIE size 32, waiting"
| |
| 18:30 | the mount point appears, but I cannot access it, d???????? ? ? ? cdrom
| |
| 18:35 | <vagrantc> !ltspfs
| |
| 18:35 | <ltsp> I do not know about 'ltspfs', but I do know about these similar topics: 'ltspfs-source'
| |
| 18:37 | <alkisg> Ah I was using the wrong path, it should be /var/run/drives/cdrom
| |
| 18:38 | <vagrantc> aha
| |
| 18:38 | i was trying to research that
| |
| 18:38 | markit has joined IRC (markit!~marco@88-149-177-66.v4.ngi.it) | |
| 18:39 | * vagrantc would like to put all the /var/run/* into /var/run/ltspfs/* | |
| 18:39 | <alkisg> Hrm. Running it manually made it work
| |
| 18:39 | <vagrantc> problem solved, uploading!
| |
| 18:39 | <alkisg> Haha!
| |
| 18:40 | <markit> hi alkisg :) how are you? Working hard in some interesting ltsp improvements?
| |
| 18:40 | <vagrantc> alkisg: so, think you'll meet your end-of-year deadline with these new ltsp improvements?
| |
| 18:40 | <alkisg> markit: hi! yup, vagrantc and I are trying to wrap up ltsp5 and move on to ltsp6...
| |
| 18:41 | vagrantc: not sure... I may end up moving the deadline for a few weeks or even months :D
| |
| 18:41 | <markit> any "roadmap" or "planned features" online I can read and calm my curiosity? :) Already testing ubuntu 14.04?
| |
| 18:41 | <vagrantc> alkisg: i'd like to at least get what's in -trunk uploaded by end-of-year
| |
| 18:44 | <alkisg> vagrantc: I have a few hours now, and I'll have a few more tomorrow, so yes if you're just expecting my testing (and maybe aoe, not udhcp) I'll have them ready...
| |
| 18:44 | markit: nah, nothing that targets users currently
| |
| 18:44 | (no documentation, I mean)
| |
| 18:45 | <vagrantc> markit: we haven't gotten too far into ltsp6 implementation ... it's largely ideas at this point with a little bit of proof-of-concept
| |
| 18:46 | <markit> thanks vagrantc :)
| |
| 18:47 | my only problem is "will kde still work with ltsp?" :) I've decided to bug kde devs again about it but no one has ltsp direct experience, we will see
| |
| 18:48 | <alkisg> Both thins and fats have issues with KDE?
| |
| 18:49 | <markit> alkisg: with my workarounds (nothing special just my ignorance) they both work well, only problem is that
| |
| 18:49 | <alkisg> vagrantc: I think the issue is that it doesn't have access to DISPLAY
| |
| 18:49 | <markit> they have a higher I/O compared with other DE, like double
| |
| 18:49 | that impacts a lot boot time, expecially at first login (when there are tons of writes)
| |
| 18:50 | in short, with the same hw you can fire up double the fat clients as me
| |
| 18:50 | <alkisg> Thin client login shouldn't be affected by IO...
| |
| 18:50 | Fats, ok...
| |
| 18:50 | <markit> in fact thin work fine, expecially at first login
| |
| 18:50 | <vagrantc> alkisg: why doesn't it?
| |
| 18:50 | <markit> since in any case a lot of data is wrote in the /var/tmp
| |
| 18:51 | also I had to use nfs with "async" that is a risk
| |
| 18:51 | without it clients does not run the DE, it hangs
| |
| 18:52 | gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com) | |
| 18:52 | <vagrantc> alkisg: why doesn't it have access to DISPLAY?
| |
| 18:52 | <alkisg> vagrantc: I put an "env > file" and I don't see DISPLAY there
| |
| 18:52 | ...for ltspfsmounter
| |
| 18:55 | * alkisg doesn't get the "call_ltspfsmounter" code... | |
| 18:55 | <alkisg> ..example content for $DISPLAY_INFO?
| |
| 18:57 | <vagrantc> alkisg: are you running with LDM_DIRECTX ?
| |
| 18:57 | <alkisg> vagrantc: yes
| |
| 18:57 | Ah, that could be our difference..
| |
| 18:57 | I don't see how the ldm_directx code would ever work
| |
| 18:58 | <vagrantc> i've definitely tested it in the past
| |
| 18:58 | oh!
| |
| 18:58 | pgrep -l -f has changed in jessie
| |
| 18:58 | <alkisg> Aaah
| |
| 18:58 | <vagrantc> probably 14.04 too
| |
| 18:58 | meh.
| |
| 18:58 | <alkisg> Was it supposed to return the full command line?
| |
| 18:58 | <vagrantc> that's going to be hard to maintain backwards compatibility on
| |
| 18:59 | we need to audit all of our code for that.
| |
| 18:59 | i knew that was going to burn us someday
| |
| 19:00 | it sure annoys me from the commandline
| |
| 19:00 | i think it's a bug in pgrep, actually...
| |
| 19:03 | <alkisg> vagrantc: there's a new "-a" parameter
| |
| 19:04 | <vagrantc> yup
| |
| 19:04 | <alkisg> It's documented in the man page etc...
| |
| 19:04 | <vagrantc> i bet
| |
| 19:04 | <alkisg> So we'd have to do it like "nc -q"
| |
| 19:04 | <vagrantc> right
| |
| 19:04 | set a variable, or maybe a function or something
| |
| 19:07 | gah. and pgrep returns 0 even when you've passed an invalid argument
| |
| 19:08 | <alkisg> ./client/Fedora/initscripts/ltsp-client-launch: SERVER=$(pgrep -f -l $DEVICE | awk '{print $3}')
| |
| 19:08 | ./client/Redhat/initscripts/ltsp-client-launch: SERVER=$(pgrep -f -l $DEVICE | awk '{print $3}')
| |
| 19:08 | ./client/Redhat/share/ltsp/ltsp-client-launch: SERVER=$(pgrep -f -l $DEVICE | awk '{print $3}')
| |
| 19:08 | ltspfs-trunk/scripts/ltspfs_entry: IS_DIRECTX=$(pgrep -f -l ${LDM_SOCKET}.*DISPLAY)
| |
| 19:08 | I think those are all...
| |
| 19:09 | <vagrantc> for all i know, redhat doesn't include the backwards-incompatible pgrep
| |
| 19:09 | <alkisg> vagrantc: in ltspfs, why do we care about directx?
| |
| 19:09 | That's only to get the cookie, right?
| |
| 19:09 | * vagrantc doesn't remember why we care | |
| 19:09 | <alkisg> We could just use ssh -X in all cases...
| |
| 19:10 | The communication to the client's ltspfsd is always unencrypted and unrelated to ssh and ldm_directx, right?
| |
| 19:10 | <vagrantc> sadly, yes.
| |
| 19:10 | <alkisg> And we always use `ssh -Y` in our master connection, so we do have access to the X display
| |
| 19:11 | We could do ssh port fowarding if we wanted encryption... but anyway that's another topic
| |
| 19:12 | NTBlade has joined IRC (NTBlade!81d7a0c5@gateway/web/freenode/ip.129.215.160.197) | |
| 19:13 | <NTBlade> Hi, anyone out there this evening?
| |
| 19:13 | <alkisg> vagrantc: yup, adding `-a` fixed the issue
| |
| 19:13 | !ask | echo NTBlade:
| |
| 19:13 | <ltsp> NTBlade: 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 a full hour after asking a question, as not everybody constantly monitors the channel.
| |
| 19:14 | <vagrantc> alkisg: so the quick fix is conditionally add "pgrep -f -l -a"
| |
| 19:15 | freedomrun has joined IRC (freedomrun!~freedomru@unaffiliated/freedomrun) | |
| 19:15 | <vagrantc> alkisg: a better fix might be to drop the need for pgrep?
| |
| 19:15 | <alkisg> vagrantc: I'm thinking of completely ignoring ldm_directx
| |
| 19:15 | Btw, the cookie could just go in $LTSP_SESSION somewhere, no need for X properties...
| |
| 19:16 | <vagrantc> we liked a cookie better than a variable for some reason, but it's been years since we thought about it
| |
| 19:17 | alkisg: if the variable was set in an environment variable, all users on the server would be able to steal the other users cookies
| |
| 19:17 | can't you?
| |
| 19:17 | well, setting the environment variable from the commandline...
| |
| 19:17 | <alkisg> vagrantc: in my thinking, LTSP_SESSION is a user-owned (and readable) directory
| |
| 19:18 | Inside it we have files for xdg, for environment, for cookies...
| |
| 19:18 | <vagrantc> ah, for ltsp6
| |
| 19:18 | <alkisg> Yeah
| |
| 19:18 | But no, other users can't see our environment
| |
| 19:22 | vagrantc: so, with this change, and by commenting out 2 of the debian/ubuntu-specific rules (the 2 last with the ## in front of them), ltspfs work fine without cdpinger
| |
| 19:23 | <vagrantc> you need the polling rule?
| |
| 19:23 | alkisg: which change, btw?
| |
| 19:24 | <alkisg> vagrantc: http://pastie.org/8585792
| |
| 19:24 | I don't need the polling rule, but do you want me to see if it harms anything, and if not, to have it in both debian and ubuntu?
| |
| 19:27 | <vagrantc> alkisg: it would be a lot easier if we could use the same exact rules, but if i need the change events anyways, then may as well not add anything to ubuntu that doesn't need to be there
| |
| 19:27 | <alkisg> OK, let's keep it as a build-time diff then
| |
| 19:28 | vagrantc: what else? only aoe (and at some later release, udhcp)?
| |
| 19:29 | * alkisg waits for an "ok" to commit the ltspfs_entry fix... | |
| 19:31 | <vagrantc> alkisg: i'm testing your change to ltspfs_entry now
| |
| 19:35 | <lifeboy> alkisg: Should epoptes "see" a thin-client before someone has actually logged on?
| |
| 19:38 | <alkisg> lifeboy: yes
| |
| 19:38 | (if it's installed in the chroot too)
| |
| 19:40 | <vagrantc> alkisg: works like a charm!
| |
| 19:41 | <alkisg> Nice, pushing...
| |
| 19:41 | <vagrantc> simpler more functional code!
| |
| 19:43 | alkisg: probably need to change it also for scripts/ldm/X10-delayed-mounter
| |
| 19:43 | well, it seems to work fine as is...
| |
| 19:43 | but i wonder if it would work without all that
| |
| 19:44 | <alkisg> It should... /me changes that too
| |
| 19:45 | <vagrantc> alkisg: slight preference for ssh -X -S over ssh -XS
| |
| 19:45 | * alkisg obeys to the DD even though he slightly prefers the second form himself... :P | |
| 19:46 | <alkisg> vagrantc: any idea what are the last lines X10-delayed-mounter for?
| |
| 19:47 | Actually... now that we're always creating the local user, maybe we no longer need the root ltspfs mount anymore?
| |
| 19:47 | I see 2 mounts, one for root, one for user
| |
| 19:47 | norrie has joined IRC (norrie!~norrie@129.215.160.197) | |
| 19:48 | <vagrantc> alkisg: it just becomes easier to search in the manpage ... makes it clearer it's two options, rather than one strangely formatted longhand option
| |
| 19:48 | it's also clearer that ssh -X -S <socket> couldn't be written as ssh -SX <socket>
| |
| 19:49 | at least, that's my opinion
| |
| 19:49 | <alkisg> Balance, as always... shell newbies :D are supposed to learn about short options, and others want them to remember parameter combinations more easily, like e.g. netstat -nap, it would be more difficult for me to remember it as -n -a -p
| |
| 19:49 | <vagrantc> alkisg: works in delayed_mounter as well
| |
| 19:49 | <norrie> Can someone help fix a problem with LTSP on Linux mint please?
| |
| 19:50 | <alkisg> OK, I'll commit all that
| |
| 19:50 | norrie: I think bennabiy worked on Linux Mint support
| |
| 19:50 | Ah, vagrantc, about the root mount?
| |
| 19:50 | <norrie> Thin client images build fine but I want to use fat. The fat image builds but has the wrong desktop (Ubuntu instead of MATE)
| |
| 19:51 | <alkisg> vagrantc, sbalneav: /media/root/cdrom and /media/user/cdrom
| |
| 19:51 | ....do we need the root one?
| |
| 19:51 | (we always have a local user now)
| |
| 19:51 | <vagrantc> alkisg: it's also used by rdesktop/xfreerdp screen scripts
| |
| 19:52 | <alkisg> vagrantc: hrm, well those should have a special user for that
| |
| 19:52 | <vagrantc> alkisg: so we don't always have a local user...
| |
| 19:52 | * vagrantc nods | |
| 19:52 | <alkisg> Like the kiosk user
| |
| 19:52 | OK I'll leave that for now then
| |
| 19:53 | <vagrantc> yes, i think we could fix it up later, but for now...
| |
| 19:53 | NTBlade has left IRC (NTBlade!81d7a0c5@gateway/web/freenode/ip.129.215.160.197, Ping timeout: 272 seconds) | |
| 19:53 | <vagrantc> norrie: what version of ltsp are you using? linuxmint support was only recently added.
| |
| 19:54 | <norrie> How do I check that please?
| |
| 19:54 | <vagrantc> ltsp-info
| |
| 19:54 | <norrie> two secs...
| |
| 19:55 | <vagrantc> !pastebin | echo norrie
| |
| 19:55 | <ltsp> norrie 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.
| |
| 19:55 | <vagrantc> and the linuxmint support still is a little rough around the edges
| |
| 19:55 | <norrie> ltsp-server 5.4.3-0ubuntu6
| |
| 19:56 | <vagrantc> it was added in ltsp 5.4.6 ...
| |
| 19:56 | you could probably manually fix it up to do the right thing.
| |
| 19:56 | there also may be a ppa for it/
| |
| 19:56 | ?
| |
| 19:57 | <norrie> Sorry, I'm still here and new to IRC
| |
| 19:57 | <lifeboy> Alkisg: Yes, epoptest is install on the client. I thought it should show, but it doesn't although the client starts and reports that it connects to the server on port 789, but it doesn't show up in the epoptes console. Are there other ports I can check?
| |
| 19:57 | <alkisg> There's ltsp daily builds, it does have packages for saucy
| |
| 19:57 | <vagrantc> bennabiy: i don't think you ever got the artwork/fatclient stuff into the code.
| |
| 19:57 | <alkisg> lifeboy: get a local shell on the client and run epoptes-client from there
| |
| 19:57 | !screen_02
| |
| 19:57 | <ltsp> screen_02: To get a root shell on an Ubuntu thin client: https://help.ubuntu.com/community/UbuntuLTSP/ClientTroubleshooting#Using_a_shell_SCREEN
| |
| 19:58 | <alkisg> !screen_08
| |
| 19:58 | <ltsp> screen_08: To get a root shell on a Debian thin client, put SCREEN_07=ldm, SCREEN_08=shell and SCREEN_DEFAULT=07 to lts.conf.
| |
| 19:58 | <alkisg> (don't remember what chroot you have)
| |
| 19:59 | <vagrantc> wheezy makes 02-06 available, i think...
| |
| 20:00 | <alkisg> vagrantc: did you get any mails about failed builds yesterday, or did I succeed in me being the only one that gets themm?
| |
| 20:00 | <vagrantc> alkisg: your successes were in other things. :P
| |
| 20:01 | <alkisg> Ouch
| |
| 20:01 | Then I think I'll change the owner of ltsp-daily-builds ppa to me, and restore it to ltsp-upstream when someone wants to do something with it...
| |
| 20:02 | I could just use my own ppa, but it just doesn't make sense to have the ltsp-daily-builds ppa then
| |
| 20:03 | <vagrantc> if we didn't build for each and every ubuntu release, it'd be less annoying when it fails
| |
| 20:03 | it is good to know when the packaging is broken
| |
| 20:03 | but following debian's packaging doesn't necessarily follow -trunk
| |
| 20:05 | if we reduced it down to one arch and one release, and maybe made a separate packaging branch that we try to keep in sync with -trunk ... ?
| |
| 20:05 | alkisg: so ltspfs-trunk ready for upload? looks good to me.
| |
| 20:06 | <alkisg> I can tell it to only build trusty, which I'm currently testing...
| |
| 20:06 | Let me check 1 last time, but yeah it should be ready
| |
| 20:07 | <vagrantc> 5 files changed, 6 insertions(+), 59 deletions(-)
| |
| 20:07 | maintaining full feature-compatibility, i'd say that's a good day.
| |
| 20:08 | it might even build on kfreebsd or hurd now :)
| |
| 20:08 | * vagrantc kind of cringes at the thought | |
| 20:09 | <alkisg> Haha
| |
| 20:09 | vagrantc: how about uncommenting the 2 last actions in ltspfsd.rules?
| |
| 20:09 | (statically instead of at build time...)
| |
| 20:10 | I believe other distros would probably want those too?
| |
| 20:13 | And another thing, probably for lstp6, would be to allow unmounting from nautilus (helper=lbmount as a mount parameter)
| |
| 20:16 | <vagrantc> alkisg: the change events?
| |
| 20:16 | <alkisg> Yup
| |
| 20:16 | <vagrantc> alkisg: i thought you didn't need them on ubuntu?
| |
| 20:17 | <alkisg> vagrantc: I did
| |
| 20:17 | (07:32:58 μμ) alkisg: Soo.... I don't get "add" events in 14.04 non-ltsp clients
| |
| 20:17 | (07:33:30 μμ) alkisg: I wonder if that's due to 14.04 being more in sync with debian, or if I was using the wrong method while testing, possibly with cdpinger still running...
| |
| 20:18 | vagrantc: let's do the release and I'll test in 12.04 when it builds in the ppa
| |
| 20:18 | * vagrantc is assuming other distros are more like ubuntu | |
| 20:18 | <lifeboy> alkisg: I ran "/usr/sbin/epoptes-client -c server 789" on the client and it retrieved the cert. Thereafter, running the client without any parameters works. How can I make this "stick" so it will do so automatically in future?
| |
| 20:18 | <vagrantc> but every distro will have to decide
| |
| 20:19 | which is part of the fragile nature of this that makes me nervous
| |
| 20:19 | <alkisg> vagrantc: basically I think that the debian rules file would work in all distros...
| |
| 20:19 | It's just that 1 line is already present on ubuntu, but I don't think it would be a problem if it was defined 2 times
| |
| 20:19 | <vagrantc> that's why i left them upstream as comments, but dynamically uncomment them for debian.
| |
| 20:20 | <alkisg> lifeboy: see epoptes.org/installation
| |
| 20:20 | Read the part about ltsp chroots
| |
| 20:21 | <lifeboy> alkisg: thanks, will do.
| |
| 20:22 | <vagrantc> alkisg: ah, well, that would be simpler.
| |
| 20:23 | and cleaner looking
| |
| 20:27 | <lifeboy> alkisg: when I do "epoptes-client -c" in the chroot, I get "epoptes-client ERROR: Failed to fetch certificate from localhost:789", although the localhost is listening on 789?
| |
| 20:28 | <alkisg> lifeboy: what's the output of this command outside of the chroot? ls -l /etc/epoptes/ /opt/ltsp/i386/etc/epoptes
| |
| 20:29 | <lifeboy> -rw-r--r-- 1 root root 875 Dec 20 22:49 server.crt
| |
| 20:29 | -rw------- 1 root root 916 Dec 20 22:49 server.key
| |
| 20:29 | /opt/ltsp/i386/etc/epoptes:
| |
| 20:29 | total 0
| |
| 20:29 | -rw-r--r-- 1 root root 0 Dec 29 22:27 server.crt
| |
| 20:29 | <alkisg> vagrantc: about pgrep -a: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=526355
| |
| 20:30 | <lifeboy> should the cert have 600 permissions
| |
| 20:30 | ?
| |
| 20:30 | <alkisg> No it should be readable
| |
| 20:30 | But the chroot one shouldn't be zero
| |
| 20:30 | <lifeboy> Hmm... yes
| |
| 20:31 | <alkisg> In the chroot, does this produce some code as output? socat - ssl:localhost:789,verify=0
| |
| 20:32 | <lifeboy> 2013/12/29 22:32:33 socat[11746] E SSL_connect(): Success
| |
| 20:33 | <alkisg> That doesn't sound right
| |
| 20:34 | <lifeboy> If it's is able to get the cert from the booted thin-client, is it re-generating the cert when it boots then?
| |
| 20:34 | <alkisg> Maybe you typed something wrong
| |
| 20:34 | Anyway, the quick way around it, just copy the .crt from the server to the chroot with the cp command
| |
| 20:34 | <lifeboy> ok
| |
| 20:35 | <alkisg> sudo cp /etc/epoptes/server.crt /opt/ltsp/i386/etc/epoptes/
| |
| 20:35 | Then run ltsp-update-image if you're using nbd
| |
| 20:38 | <lifeboy> Yes, now the thin-client shows in epoptes! Great! Thanks, alkisg for the help!
| |
| 20:38 | <alkisg> np
| |
| 20:43 | <vagrantc> alkisg: alright, well, i'll uncomment them upstream
| |
| 20:44 | alkisg: er, uncomment the udev rules upstream... makes it clear what's actually needed to work, but distros may need to comment them out individually depending on their udev configuration
| |
| 20:46 | gbaman_ has joined IRC (gbaman_!~gbaman@host81-130-35-88.in-addr.btopenworld.com) | |
| 20:46 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Ping timeout: 272 seconds) | |
| 20:48 | <vagrantc> scared em off!
| |
| 20:49 | gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Ping timeout: 246 seconds) | |
| 20:53 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 20:53 | <alkisg> @(#*&$# net connectivity... aoe behaves much better :P
| |
| 20:54 | vagrantc: want me to test for sideeffects of the first rule that's not needed on ubuntu?
| |
| 20:59 | vagrantc: it appears to work fine even with it
| |
| 20:59 | So just uncomment all 3 of them ...
| |
| 21:05 | <vagrantc> ok!
| |
| 21:07 | norrie has left IRC (norrie!~norrie@129.215.160.197, Quit: Leaving) | |
| 21:11 | <lifeboy> How is it possible to have a directory in a chroot (/home/roland) doing ltsp-update-image, booting the image and then not having that directory on the thin client?!
| |
| 21:13 | <vagrantc> because ltsp-update-image purges some directories, including /home, i think.
| |
| 21:13 | markit has left IRC (markit!~marco@88-149-177-66.v4.ngi.it, ) | |
| 21:14 | <vagrantc> you can edit/create /etc/ltsp/ltsp-update-image.excludes if you want to change that behavior
| |
| 21:14 | <lifeboy> Ah! Was this added in more recent versions?
| |
| 21:15 | Would that included the certificate files for openssh_server too?
| |
| 21:15 | <alkisg> Yup
| |
| 21:15 | <lifeboy> s/included/include/
| |
| 21:15 | Aaargh! I have been hacking away at this for hours :-(
| |
| 21:16 | What is the rationale to do this? I found it useful to be able to ssh into the thin client from time to time.
| |
| 21:17 | <alkisg> It wasn't implemented just for ssh
| |
| 21:17 | The ssh keys are included there for security reasons
| |
| 21:17 | E.g. if someone uses ltsp-pnp, he doesn't want the server keys to be world-readable
| |
| 21:18 | <lifeboy> So it's basically to have a more secure enviroment for ltsp?
| |
| 21:18 | <alkisg> You can regenerate them with the -c (cleanup) parameter so that they're new, unrelated to your server, and then remove them from ltsp-update-image.excludes
| |
| 21:19 | If we didn't do that, people that knew stuff would be able to steal your data
| |
| 21:19 | Passwords, login to your server, etc etc
| |
| 21:19 | <lifeboy> I'm looking that the .excludes file now and I'm starting to get the picture.
| |
| 21:19 | looking at the .excludes file...
| |
| 21:20 | alkisg: where do I use the -c parameter to regen ssh-server keys?
| |
| 21:20 | <alkisg> E.g. ltsp-update-image -c /opt/ltsp/i386, BUT
| |
| 21:21 | hmm ok np
| |
| 21:21 | <lifeboy> And then I comment out that line in the .excludes file, or must I delete it?
| |
| 21:22 | <alkisg> I think comments work too, but I'm not sure, check the nbd manpages
| |
| 21:25 | <lifeboy> It seems -c only works if my clients and server are the same architecture?
| |
| 21:29 | <alkisg> Yeah if you have a different arch ignore it
| |
| 21:30 | vagrantc: I pushed the aoe root code, untested, but I think it should work...
| |
| 21:31 | <vagrantc> alkisg: i'll test it a bit.
| |
| 21:31 | alkisg: working on the ltspfs upload right now
| |
| 21:31 | <alkisg> Nice
| |
| 21:32 | vagrantc: meh I forgot to put CMDLINE_AOE in debian's update-kernel.conf
| |
| 21:33 | <vagrantc> alkisg: i can always add it :)
| |
| 21:33 | <alkisg> :)
| |
| 21:33 | For ltsp6 let's ship the same update-kernels.conf file
| |
| 21:33 | *in debian/ubuntu
| |
| 21:34 | <vagrantc> the only difference really was the 64/pae/32 ?
| |
| 21:34 | <alkisg> And the boot order, which should better be server-depended
| |
| 21:35 | <vagrantc> oh, and NBD
| |
| 21:35 | <alkisg> And the plymouth parameters which are gone now :D
| |
| 21:38 | 'night vagrantc, night all :)
| |
| 21:39 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Remote host closed the connection) | |
| 21:52 | <lifeboy> vagrantc: I'm still trying to come to terms with the usage of keys on the thin-client chroot. To be able to make a connection to the server with ssh, the thin client needs a correct host key in /root/.ssh/known_hosts. How do I do that?
| |
| 21:52 | normally /root/* would be .excluded, so the client shouldn't be able to connect?
| |
| 21:53 | <vagrantc> lifeboy: configure /etc/ssh/sshd_config to accept some keys from different places
| |
| 21:53 | <lifeboy> sshd_config? Isn't that for the ssh server on the client?
| |
| 21:54 | <vagrantc> AuthorizedKeysFile %h/.ssh/authorized_keys, /var/lib/custom/%u
| |
| 21:54 | do you want to ssh to the cleint or the server?
| |
| 21:55 | lifeboy: ok, /opt/ltsp/<arch>/etc/ssh/sshd_config
| |
| 21:55 | lifeboy: and then create /var/lib/custom/<username> files for ssh keys to log in with
| |
| 21:55 | <lifeboy> My client can't log in using the normal ltsp gui, due to the server keys not being correct in the known hosts file
| |
| 21:56 | <vagrantc> ltsp-update-sshkeys ; ltsp-update-image
| |
| 21:56 | or wait, you're using nfs?
| |
| 21:56 | <lifeboy> The problem of connecting from the elsewhere to the thin client has been resolved
| |
| 21:56 | <vagrantc> then just ltsp-update-sshkeys
| |
| 21:56 | <lifeboy> No, using nbd
| |
| 21:57 | I did, but let me just check it all again
| |
| 21:57 | <vagrantc> is this on the server where you generated the image elsewhere?
| |
| 21:57 | <lifeboy> Yes
| |
| 21:58 | I went through all the issues step by step and this is the last one that I need to fix
| |
| 21:58 | <vagrantc> the last one you know of :P
| |
| 21:59 | lifeboy: what OS are you running on the serveer?
| |
| 21:59 | <lifeboy> hehe... I hope so. Alkisg did the basic config so I hope it will now be fine.
| |
| 22:00 | Server: Ubuntu 12.04 64bit
| |
| 22:01 | I didn't have much choice but to create a debian VM and create a chroot from there and copy it across. Worked fine for site 1, but site 2 gave me more issues. Furtunately I'm learning in th eprocess.
| |
| 22:02 | * vagrantc forgets how ltsp-update-kernels worked on 12.04 | |
| 22:03 | <vagrantc> we did some fixes shortly after 12.04 was released
| |
| 22:06 | lifeboy: what's in /opt/ltsp/<arch>/etc/ssh/ssh_known_hosts ?
| |
| 22:06 | <lifeboy> I did ltsp-update-sshkeys again, but it makes no difference. It still complains that I don't have the correct key(s) in known_hosts
| |
| 22:06 | when I attempt to ssh from the client to the server
| |
| 22:06 | <vagrantc> lifeboy: and does that match what's in the image's /etc/ssh/ssh_known_hosts
| |
| 22:06 | what's the exact error message?
| |
| 22:07 | ssh from the console?
| |
| 22:07 | you're trying to troubleshoot ldm failing to connect?
| |
| 22:07 | gbaman_ has left IRC (gbaman_!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Remote host closed the connection) | |
| 22:08 | <lifeboy> Yes, the clients cannot logon. So I'm checking what happens when I ssh from a terminal to the server
| |
| 22:08 | gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com) | |
| 22:08 | <lifeboy> The ECDSA host key for server has changed,
| |
| 22:08 | and the key for the corresponding IP address 192.168.8.100
| |
| 22:08 | is unknown. This could either mean that
| |
| 22:08 | etc...
| |
| 22:08 | <vagrantc> both servers have the same ip address?
| |
| 22:08 | <lifeboy> no
| |
| 22:09 | <vagrantc> does "md5sum /etc/ssh/ssh_known_hosts" on the client match what's in /opt/ltsp/<arch>/etc/ssh/ssh_known_hosts ?
| |
| 22:10 | or rather, md5sum /opt/ltsp ...
| |
| 22:12 | basically, are they the same file
| |
| 22:12 | also, do either of the files have ip addresses in them?
| |
| 22:13 | you might have better luck getting an updated version of ltsp for your ubuntu server.
| |
| 22:13 | <lifeboy> no, they don't have ip addresses in them, the have the hostnames.
| |
| 22:13 | <vagrantc> btw, why are you going through all this trouble? what's wrong with an ubuntu chroot?
| |
| 22:14 | lifeboy: is there an /opt/ltsp/<arch>/root/.ssh/known_hosts ?
| |
| 22:14 | <lifeboy> We have just brought them all to 12.04, since it's the LTS version! The client hardware has the cmov issue
| |
| 22:14 | <vagrantc> i prefer debian, but it seems like a lot of hassle to support multiple distros
| |
| 22:14 | ah, cmov.
| |
| 22:15 | <lifeboy> So alkisg suggested I use a wheezy chroot
| |
| 22:17 | /opt/ltsp/i386/etc/ssh/ssh_known_hosts exists
| |
| 22:17 | <vagrantc> lifeboy: what about /opt/ltsp/i386/root/.ssh/known_hosts ?
| |
| 22:17 | lifeboy: and does that contain ip addresses in it? or other information?
| |
| 22:18 | <lifeboy> 174462055312cb75cb744edf7e600614 /opt/ltsp/i386/etc/ssh/ssh_known_hosts
| |
| 22:18 | 174462055312cb75cb744edf7e600614 /etc/ssh/ssh_known_hosts
| |
| 22:19 | The files in the chroot and the client are identical
| |
| 22:19 | <vagrantc> lifeboy: what about /opt/ltsp/i386/root/.ssh/known_hosts ?
| |
| 22:19 | <lifeboy> which is what I would expect.
| |
| 22:19 | <vagrantc> lifeboy: and does that contain ip addresses in it? or other information?
| |
| 22:19 | <lifeboy> no ip's, but root@parow, which is the server
| |
| 22:20 | <vagrantc> lifeboy: i would delete that file and run ltsp-update-image again
| |
| 22:20 | it may have conflictting content
| |
| 22:20 | <lifeboy> no known_hosts file in root
| |
| 22:21 | * vagrantc feels like we're out of sync and isn't sure which answers are to which questions | |
| 22:23 | <lifeboy> sorry, vagrantc, I'm responding slowly since I'm checking the files this side. I have no deleted /opt/ltsp/i386/etc/ssh/ssh_known_hosts, ran ltsp-update-sshkeys and then lts-update-image.
| |
| 22:24 | ltsp-update-image
| |
| 22:24 | let's see...
| |
| 22:24 | <vagrantc> i'd be surprised if you got anything different
| |
| 22:25 | do you have any customizations in lts.conf ?
| |
| 22:25 | <lifeboy> Ok, deleting the file first did the trick!
| |
| 22:26 | No users can log in, I can see them in epoptes and they don't boot Unity by default.
| |
| 22:26 | * vagrantc is really confused now | |
| 22:26 | <lifeboy> Now user can log in! Sorry! Typo
| |
| 22:26 | <vagrantc> heh.
| |
| 22:26 | telex has left IRC (telex!~telex@freeshell.de, Remote host closed the connection) | |
| 22:26 | <vagrantc> still confused how that fixed anything
| |
| 22:26 | <lifeboy> It's late! :-)
| |
| 22:27 | <vagrantc> but ok!
| |
| 22:27 | <lifeboy> I only did ltsp-update-sshkeys before, but somehow it didn't fix the key. Deleting the ssh_know_hosts file first, seems to have worked. Dunno why
| |
| 22:28 | <vagrantc> oh, a wheezy chroot should be able to handle ssh_known_hosts from ubuntu... the other way around likely to have issues
| |
| 22:28 | telex has joined IRC (telex!~telex@freeshell.de) | |
| 22:28 | <vagrantc> lifeboy: it is a mystery to us all.
| |
| 22:28 | ltsp-update-sshkeys should blindly overwrite what's in ssh_known_hosts.
| |
| 22:29 | <lifeboy> I'll just noting this side what we did and hopefully it will help me in future. They still have two years on this hardware, so it has to work the way it is now.
| |
| 22:30 | <vagrantc> lifeboy: if there are no ip addresses in the shs_known_hosts files, you should be fine.
| |
| 22:30 | ssh_known_hosts
| |
| 22:30 | <lifeboy> I did rebuild the image after the ltsp-update -sshkeys, which I hadn't done before
| |
| 22:31 | ltsp-update-sshkeys that is...
| |
| 22:31 | Anyway, thanks for the help! I'm off now...
| |
| 22:32 | * vagrantc grumbles | |
| 22:33 | <vagrantc> could've saved us 20 minutes of troubleshooting by having done that when i asked :P
| |
| 22:39 | oh well, ltspfs didn't build on kfreebsd-amd64
| |
| 22:39 | * vagrantc is kind of relieved | |
| 22:53 | khildin has left IRC (khildin!~khildin@ip-213-49-87-104.dsl.scarlet.be, Quit: I'm gone, bye bye) | |
| 23:07 | gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Remote host closed the connection) | |
| 23:38 | gbaman has joined IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com) | |
| 23:45 | gbaman has left IRC (gbaman!~gbaman@host81-130-35-88.in-addr.btopenworld.com, Ping timeout: 260 seconds) | |
| 23:48 | PhoenixSTF has left IRC (PhoenixSTF!~rudi@78.29.154.124, Quit: Leaving) | |
| 23:49 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |