Web UI Sometimes Very Slow to Become Available and Server Key Disappeared

Xibo 1.8.2 - Docker
Ubuntu 16.04.2 LTS VM
No custom ports.
RAM: 2300MB Starting - Dynamic

This has spiraled out a bit over time.

We came in today to find 99% of our displays had frozen on one slide out of their layouts.

Logged into the web UI and found that the last display contact date was the 29th. Given it was essentialy all the displays etc, we rebooted the Xibo server.

Now when trying to access the web ui, Chrome gave and Err_Empty_Response msg.

I tried docker-compose down followed by docker-compose up -d but the error remains.

docker ps -a lists all containers as up for 9 minutes but still no response from the web ui.

I then started writing up this post and generated docker-compose logs results which I can PM. used >> xibo-docker-logs.txt, if there is a better way to export them more cleanly let me know. It has long blocks of near jibberish that looks like its being spat out from the database, not sure if that is supposed to happen.

Just before hitting submit on this I checked again and the web ui was responsive, checked on the display entries and they still hadn’t updated. Then checked on one of the displays and found that it was generating an error that the server key it had didn’t match the key the server was giving. Checked the key in web ui and found that it was now blank. Copied the key in and things seem better, however, I also looked in the logs and audit trail in hopes of finding clues and found that the logs are empty and the trail is very nearly so.

So in light of it taking over 9 minutes for the web ui to become available, and that wasn’t just 1 time, there were a few other reboots where we went over the 5 minute mark on the containers with no web ui and I ended up rebooting again, our key getting wiped out, and the potential log issues, are there any known issues that we need to go in and fix?

Basically its all running now but I’m concerned that its still broken in some subtle way(s) and would appreciate feedback on if I should be concerned and/or if there is anything I can do.

Thank you for your time.

I’m happy to look at the output of your docker-compose log - you can send it to me via private message.

It could be also good to check your server stats with top and free -h commands to see how the memory, cpu usage etc looks like.

That’s rather odd, obviously it should not be removed automatically from the CMS settings.

Perhaps the docker log you’ve collected when it was ‘broken’ would tell us more.