01:12 | vagrantc has left IRC (vagrantc!~vagrant@unaffiliated/vagrantc, Quit: leaving) | |
01:42 | GodFather has joined IRC (GodFather!~rcc@wsip-66-210-242-210.ph.ph.cox.net) | |
05:37 | quinox has left IRC (quinox!~quinox@ghost.qtea.nl, Quit: WeeChat 2.9) | |
05:40 | quinox has joined IRC (quinox!~quinox@ghost.qtea.nl) | |
07:33 | ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz) | |
07:45 | woernie has joined IRC (woernie!~werner@pd9e8b5cc.dip0.t-ipconnect.de) | |
07:46 | gvy_ has joined IRC (gvy_!~mike@193.43.10.250) | |
07:47 | gvy_ has left IRC (gvy_!~mike@193.43.10.250, Remote host closed the connection) | |
07:57 | GodFather has left IRC (GodFather!~rcc@wsip-66-210-242-210.ph.ph.cox.net, Ping timeout: 256 seconds) | |
08:12 | bcg has left IRC (bcg!~b@dg4ybwyyyyyyyyyyyyyyt-3.rev.dnainternet.fi, *.net *.split) | |
08:23 | bcg has joined IRC (bcg!~b@dg4ybwyyyyyyyyyyyyyyt-3.rev.dnainternet.fi) | |
08:26 | uumas has left IRC (uumas!uumaskapsi@gateway/shell/matrix.org/x-vgzjrxjxzjjbaerd, Ping timeout: 246 seconds) | |
09:17 | woernie has left IRC (woernie!~werner@pd9e8b5cc.dip0.t-ipconnect.de, Remote host closed the connection) | |
09:21 | uumas has joined IRC (uumas!uumaskapsi@gateway/shell/matrix.org/x-naistafqnaqnwuki) | |
10:01 | woernie has joined IRC (woernie!~werner@p50867cae.dip0.t-ipconnect.de) | |
10:05 | bobby8 has joined IRC (bobby8!6dc048e9@HSI-KBW-109-192-072-233.hsi6.kabel-badenwuerttemberg.de) | |
10:05 | woernie_ has joined IRC (woernie_!~werner@p50867e5b.dip0.t-ipconnect.de) | |
10:05 | woernie has left IRC (woernie!~werner@p50867cae.dip0.t-ipconnect.de, Ping timeout: 264 seconds) | |
10:09 | <bobby8> Hi there. There is a strange issue with my LTSP instance. Each time I update the image, the first client that logs in will have its configuration broken. I.e. on the next login, some of the desktop settings are gone and the desktop has to be recreated or replaced from a backup. This is only the first client that connects after the image has been
| |
10:09 | updated (no matter who it is).
| |
10:09 | Any idea what's going on? Which log should I check?
| |
10:10 | <alkisg> bobby8: nothing comes to mind that would cause that
| |
10:10 | Is this the new ltsp?
| |
10:10 | <bobby8> Yes
| |
10:10 | <alkisg> The output of `mount` at that time, would be useful
| |
10:11 | I could also see via reverse-vnc if you'd like
| |
10:11 | <bobby8> mount at what time?
| |
10:11 | <alkisg> At the first broken login
| |
10:11 | On the client itself
| |
10:12 | <bobby8> I usually find the broken login when logging on on the server itself.
| |
10:13 | <alkisg> I didn't get that
| |
10:13 | So you log on in a client as user1
| |
10:13 | Then you logout, and you login as administrator on the server
| |
10:13 | You run ltsp image, then you logout
| |
10:13 | <bobby8> My procedure is as follows: I update the image, then I walk to one of the client machines to see if everything is fine. I log on to one of the accounts on the client, shut it down again. When I log on to that account on the server, the desktop is gone.
| |
10:14 | <alkisg> Then you login as user1 on the server, and the settings are broken?
| |
10:15 | In some cases I've seen people run `sudo app` and thus change ~/.config/dconf, all the settings, to be owned by root
| |
10:15 | This then causes the settings to be inaccessible by the user
| |
10:15 | And in similar cases, corrupted dconf instead of wrong permissions
| |
10:15 | That's all that comes to mind. If you do have a broken profile, we could test that theory.
| |
10:15 | <bobby8> As one would expect... but I was not even aware of that command.
| |
10:16 | <alkisg> You've never used sudo?
| |
10:16 | E.g. sudo gedit, could cause that problem
| |
10:16 | <bobby8> you wrote "app", I expected that to be a command I don't know
| |
10:16 | <alkisg> app means "any application"
| |
10:17 | Sorry I was being brief
| |
10:17 | <bobby8> When this occurs, I just log in on the client and shut it down again. I don't do anything else.
| |
10:17 | <alkisg> Are you using the default sshfs for mounting /home?
| |
10:18 | <bobby8> The first couple of times this happened, I copied the desktop configuration back from a backup. I thing I would have noticed if some of the files would have been owned by root...
| |
10:18 | Everything is plain LTSP, so yes, this should be sshfs.
| |
10:19 | <alkisg> Which distro/version/flavor?
| |
10:19 | <bobby8> Kubuntu 20.04
| |
10:19 | <alkisg> Ah, so no dconf-based but kde-settings-something-based (not using KDE, although it seems nice)
| |
10:20 | I've heard of some caching problems of kde in the past, that it's putting things in /var/cache instead of in ~/.cache
| |
10:20 | And that some people were disabling cache to optimize things
| |
10:20 | I don't have enough experience to advice though, I'd need to check via vnc when the problem occurs
| |
10:21 | *enough experience with kde
| |
10:21 | <bobby8> Are you saying that disabling cache might cause it or fix it?
| |
10:21 | <alkisg> It's an idea, yeah
| |
10:22 | <bobby8> What strikes me most is that it only happens on the first login after updating the image.
| |
10:22 | I have never seen this on any other occasion.
| |
10:22 | <alkisg> Maybe kde stores a "cache counter" at /var/cache/something, and it goes in the image, and then kde again tries to match ~/.cache with /var/cache/something
| |
10:23 | So after it updates ~/.cache once, everything is good again
| |
10:23 | It's too much guesswork, it needs hands-on work
| |
10:23 | <bobby8> I see
| |
10:25 | Hold on... I'll be trying something
| |
10:30 | On the other hand... if it is a caching problem, why doesn't it happen on each account once after updating the image?
| |
10:32 | Is it suficcient to run "ltsp kernel"?
| |
10:32 | sorry
| |
10:32 | ltsp image
| |
10:39 | <alkisg> bobby8: it only happens on the first account that you connect, where, to EACH pc (so many accounts) or to ANY pc (so one account)?
| |
10:40 | <bobby8> I have a server with three accounts so anyone can work on the server or any of the clients.
| |
10:41 | It does neither depend on the machine that is used nor on the account that logs on.
| |
10:41 | I tried to force it to happen right now, but it didn't happen... :-|
| |
10:41 | <alkisg> So if user1 logs in to pc1, it's broken. If then user2 logs in to pc2, it's not broken?
| |
10:41 | <bobby8> Right
| |
10:42 | <alkisg> Yeah that doesn't make any sense at all
| |
10:42 | As users data don't communicate with each other
| |
10:42 | <bobby8> Or even if user2 log in to pc1 after user1 has logged off.
| |
10:42 | <alkisg> That could be explained, but what I said above, can't be
| |
10:45 | <bobby8> There is one thing that is different on a local and a remote shutdown:
| |
10:46 | When I close a local session, the log in xsession-errors goes:
| |
10:46 | kdeinit5: PID ... terminated
| |
10:46 | ...
| |
10:46 | kdeinit5: terminate KDE
| |
10:46 | klauncher: Exiting on signal 1
| |
10:47 | Closing the remot session goes like this:
| |
10:47 | kdeinit5: PID ... terminated
| |
10:47 | ...
| |
10:47 | klauncher: Exiting on signal 15
| |
10:47 | X connection to :0 broken (explicit kill or server shutdoen).
| |
10:47 | The X11 connection broke (error 1). Did the X11 server die?
| |
10:50 | <alkisg> That sounds like an abrupt logoff; it could explain /home/username getting corrupted; but not "only the first user"
| |
10:52 | <bobby8> Well, "Only the first user" is just my impression from experience. I have no proof that there is a real causality.
| |
10:52 | <alkisg> If it's a race/random condition, then yeah many events could explain it
| |
10:54 | <bobby8> There's another difference in handling: On the client, I usually shut the machine down from KDE once I'm done. I would not do that when I'm sitting in front of the server; I just log out in that case.
| |
10:56 | <alkisg> I don't think `ltsp image` is related then
| |
10:56 | Try this on the server and client: sudo -i; echo 3 > /proc/sys/vm/drop_caches
| |
10:56 | This drops caches (same thing that ltsp image does), and the next time the server needs to read from disk, it won't have it in cache/ram, so it'll be a bit slower, so it *might* cause the race condition again
| |
10:57 | (ltsp image loads all the disk and image, so previous cache contents are replaced)
| |
10:57 | It might be some race condition in KDE, that gets more intense when /home is on the network
| |
10:58 | In which case, maybe switching to nfs home could help
| |
10:58 | (nfs is more stable than sshfs, yet insecure in its default setup)
| |
11:08 | <bobby8> Well, it didn't happen... But I have the "The X11 connection broke" three times in the log, plus:
| |
11:08 | kdeinit5: Fatal IO error: client killed
| |
11:08 | kdeinit5: sending SIGHUP to children
| |
11:08 | klauncher: Exiting on signal 1
| |
11:08 | epoptes-client got signal: 15, relaunching...
| |
11:14 | <alkisg> Anyway try to log off and THEN shutdown for a few days, is if it doesn't happen that way
| |
11:14 | If it doesn't, then it's some race condition with kde poweroff
| |
11:14 | *see if it doesn't happen...
| |
11:15 | <bobby8> I will watch this carefully and come back...
| |
11:15 | Thanks for now.
| |
11:18 | bobby8 has left IRC (bobby8!6dc048e9@HSI-KBW-109-192-072-233.hsi6.kabel-badenwuerttemberg.de, Quit: Connection closed) | |
11:25 | woernie_ has left IRC (woernie_!~werner@p50867e5b.dip0.t-ipconnect.de, Ping timeout: 264 seconds) | |
11:30 | woernie has joined IRC (woernie!~werner@p50867e5b.dip0.t-ipconnect.de) | |
12:31 | woernie has left IRC (woernie!~werner@p50867e5b.dip0.t-ipconnect.de, Remote host closed the connection) | |
14:17 | ricotz_ has joined IRC (ricotz_!~ricotz@ubuntu/member/ricotz) | |
14:18 | ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Ping timeout: 260 seconds) | |
14:23 | ricotz_ is now known as ricotz | |
16:01 | GodFather has joined IRC (GodFather!~rcc@wsip-66-210-242-210.ph.ph.cox.net) | |
16:10 | ltspbot has joined IRC (ltspbot!~supybot@devs.ts.sch.gr) | |
16:27 | GodFather has left IRC (GodFather!~rcc@wsip-66-210-242-210.ph.ph.cox.net, Remote host closed the connection) | |
16:30 | GodFather has joined IRC (GodFather!~rcc@wsip-66-210-242-210.ph.ph.cox.net) | |
16:33 | lucascastro has left IRC (lucascastro!~lucascast@177-185-133-183.dynamic.isotelco.net.br, Ping timeout: 272 seconds) | |
16:49 | lucascastro has joined IRC (lucascastro!~lucascast@45-167-143-6.netfacil.inf.br) | |
16:57 | GodFather has left IRC (GodFather!~rcc@wsip-66-210-242-210.ph.ph.cox.net, Ping timeout: 260 seconds) | |
17:18 | mwalters has left IRC (mwalters!~ubox@204.111.136.22, *.net *.split) | |
17:21 | mwalters has joined IRC (mwalters!~ubox@204.111.136.22) | |
17:31 | GodFather has joined IRC (GodFather!~rcc@wsip-66-210-242-210.ph.ph.cox.net) | |
17:33 | lucascastro has left IRC (lucascastro!~lucascast@45-167-143-6.netfacil.inf.br, Ping timeout: 240 seconds) | |
18:02 | lucascastro has joined IRC (lucascastro!~lucascast@45-167-143-6.netfacil.inf.br) | |
18:13 | GodFather has left IRC (GodFather!~rcc@wsip-66-210-242-210.ph.ph.cox.net, Remote host closed the connection) | |
18:14 | vagrantc has joined IRC (vagrantc!~vagrant@unaffiliated/vagrantc) | |
18:15 | GodFather has joined IRC (GodFather!~rcc@wsip-66-210-242-210.ph.ph.cox.net) | |
18:22 | GodFather has left IRC (GodFather!~rcc@wsip-66-210-242-210.ph.ph.cox.net, Remote host closed the connection) | |
18:23 | GodFather has joined IRC (GodFather!~rcc@wsip-66-210-242-210.ph.ph.cox.net) | |
18:39 | woernie has joined IRC (woernie!~werner@p50867e5b.dip0.t-ipconnect.de) | |
18:44 | woernie_ has joined IRC (woernie_!~werner@p50867e5b.dip0.t-ipconnect.de) | |
18:44 | woernie has left IRC (woernie!~werner@p50867e5b.dip0.t-ipconnect.de, Ping timeout: 272 seconds) | |
18:49 | woernie_ has left IRC (woernie_!~werner@p50867e5b.dip0.t-ipconnect.de, Ping timeout: 256 seconds) | |
18:54 | woernie has joined IRC (woernie!~werner@p54bec8f3.dip0.t-ipconnect.de) | |
18:59 | woernie has left IRC (woernie!~werner@p54bec8f3.dip0.t-ipconnect.de, Ping timeout: 246 seconds) | |
19:32 | jgee98 has joined IRC (jgee98!~jgee@190.159.118.121) | |
19:32 | bluejaypop has left IRC (bluejaypop!~7f000001@unaffiliated/josefig) | |
19:36 | jgee9 has left IRC (jgee9!~jgee@190.159.118.121, *.net *.split) | |
19:36 | highvoltage has left IRC (highvoltage!~highvolta@ubuntu/member/highvoltage, *.net *.split) | |
19:38 | ltspbot has joined IRC (ltspbot!~supybot@devs.ts.sch.gr) | |
19:40 | highvoltage has joined IRC (highvoltage!~highvolta@ubuntu/member/highvoltage) | |
20:36 | lucascastro has left IRC (lucascastro!~lucascast@45-167-143-6.netfacil.inf.br, Ping timeout: 265 seconds) | |
20:41 | lucascastro has joined IRC (lucascastro!~lucascast@45-167-143-6.netfacil.inf.br) | |
20:46 | lucascastro has left IRC (lucascastro!~lucascast@45-167-143-6.netfacil.inf.br, Ping timeout: 272 seconds) | |
21:58 | adrianorg has left IRC (adrianorg!~adrianorg@177.18.51.212, Ping timeout: 265 seconds) | |
22:13 | ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Quit: Leaving) | |
22:29 | adrianorg has joined IRC (adrianorg!~adrianorg@187.113.250.197) | |