Closed condector closed 5 years ago
It's not something we've ever seen in production. I guess one thing that may have happened is that the LXD daemon restarted, closing the connection between lxd-demo-server and lxd itself?
I installed the lxd from snap:
Client version: 3.8
Server version: 3.8
And, because lxd-demo-server
not have the Unix socket path to the LXD, I create a symbolic link to fulfill that:
lrwxrwxrwx 1 root root 36 dez 20 08:46 /var/lib/lxd/unix.socket -> /var/snap/lxd/common/lxd/unix.socket
The lxd-demo-server
create the container, but not start it... but when I kill it and run the server again, it start to work normally for a while.
First, I try to start a new container (I did not create a new one for a day). As you could see
on the terminal running the lxd-demo-server
, no error's:
I accept the rules and start to see the wheel:
On the server, I could see the container that was created (so lxc-demo-server
are connected to the LXD normally):
but the container not start and the wheel continues running on the browser. So, I reload the webpage and it creates a new container:
Now, I stopped the lxd-demo-server
and start it again. As you can see, now it creates
the container and start it normally:
and, on the browser, I could see the terminal window:
Could I set any parameter on the lxd-demo-server
to debug it better?
I will close it because I imagine that you have a lot of problems to solve other things (much more important). Just remember that this problem is occurring.
No point in closing an issue that hasn't been resolved
Ok, no problem. Thanks.
I've been trying to replicate this issue by restarting LXD, making it crash, ... with no success.
The snap is a bit outdated though so I'll at least refresh it to use a newer build of the client but if there's a particular case where the event handling gets into a bad state as the above suggests, I've not yet found a way to replicate it.
@stgraber do the test here in another situation.
I copy the build lxd-server-demo
from my old server to a new one installed from scratch (Ubuntu 18.04 LTS). Not use btrfs
as disk, only use dir
this time, and the issue has not appeared again. Maybe something related to the go become unstable on the other server.
Now I will close the "bug" because the problem was in my old environment. I'm testing the new one and the problem does not occur again.
I saw a similar question but running on your own server. Here, my
lxd-demo-server
is running and create normally the container when the "rules" were accepted.But, if I not create any container for some hours and try to create a new one, the server creates the container itself (but it remains stopped), but the container was not started, so the interface did not open the terminal.
Not saw any logs on the terminal on the running
lxd-demo-server
process.I think that, probably, some process was not starting the container after that change related to the alias container image.
If I kill the
lxd-demo-server
process, it start to create and remove the containers again, but keep one container on stopped state and I need to remove it manually.Only to clarify, I create this environment here to allow my computer science students to play with Linux. Thanks a lot for LXD/LXD-demo-server environment created.