Thread Exception on start of jitsi-videobridge service

When I try to start the jitsi-videobridge task ( systemctl start jitsi-videobridge ), systemctl returns me an error.
I have tried to inspect some logs (journalctl -xn, /var/logs/jitsi/jvb.log) and those are the errors I get :

Result of journalctl -xn
Content of jvb.log

How can I get more information on what is going wrong ?

(That’s a report of the issue I just opened on github, on the wrong thread !)

@damencho asked me the available memory on the machine, are you talking about the available ram or the preconfigured ram allowed to the JVM ?

And Yes, Jicofo is running on the same server.

The service used to work (for 24 hours, I’d say), then at 0 pm there was a error : SocketException, too many files open (translated from french, so I’m not sure), with message “(Error creating socket)”. which is too in ice4j.

Then at 5 pm the error changed (can’t say if both are linked or not.)

Yes, I saw Out of memory error that’s why I asked about the available RAM.

Too many open files is another story. We had pushed to latest jvb (latest from unstable repo) a change with custom systemd unit file. So just doing service jvb restart will set the appropriate settings and the limit of the tasks and files (as described in the quick-install doc) should no longer be an issue.

The server has 32 Go Ram.

I’m not sure about why but service jvb restart doesn’t do anything (unit not found). I’m assuming you’re talking about jitsi-videobridge, in which case it throws an error.

But Maybe I’m mistaken, given that I haven’t installed the unstable, only stable. Is that a mistake ?

I followed https://github.com/jitsi/jitsi-meet/blob/master/doc/quick-install.md (without advanced configuration)

I’m not sure why you got the out of memory error with 32 GB of RAM.

The settings I was mentioning and the systemd is available only from unstable.

32GB is probably an overkill; for instance our own instances run on 8GB and scale horizontally, instead of vertically. I remember we were plagued by this exception a few months ago when we upgraded to xenial and the fix was, as @damencho correctly pointed out, to adjust the available memory in the systemd file.

It’s strange that you get a “unit not found” error. What if you do systemctl restart jitsi-videobridge? Can you make sure that the jitsi-videobridge systemd init file is actually there?

systemctl restart jitsi-videobridge fails, and asks me to check journalctl, which is described up here, but it does fail as in “service unknown”. Therefore, I assume there is a systemd configuration.

# systemctl is-active jitsi-videobridge
failed
# systemctl is-active jicofo
active

@damencho I can’t find the systemd file in the repo. Have we released it yet or does it only live in the infra scripts?