TrueOsiris
Restarting container on server crash
Server was running fine all day until I got an unhandled exception. The server crashed, but the container didn't restart. Is there a way to restart the container when the server encounters an exception?
Here's a snip of the logs when it crashed:
wine: Unhandled exception 0x80000003 in thread 24 at address 0000000180EE7265 (thread 0024), starting debugger...
vrising | 032c:fixme:event:wait_for_withdrawn_state window 0x10056/1000001 wait timed out
vrising | 00f4:fixme:file:CancelSynchronousIo (000000000000006C): stub
vrising | 00f4:fixme:file:CancelSynchronousIo (000000000000006C): stub
vrising | 00f4:fixme:file:CancelSynchronousIo (000000000000006C): stub
vrising | 00f4:fixme:file:CancelSynchronousIo (000000000000006C): stub
vrising | 00f4:fixme:file:CancelSynchronousIo (000000000000006C): stub
vrising | 00f4:fixme:file:CancelSynchronousIo (000000000000006C): stub
vrising | 00f4:fixme:file:CancelSynchronousIo (000000000000006C): stub
vrising | 00f4:fixme:file:CancelSynchronousIo (000000000000006C): stub
vrising | 00f4:fixme:file:CancelSynchronousIo (000000000000006C): stub
vrising | 00f4:fixme:file:CancelSynchronousIo (000000000000006C): stub
vrising | 00f4:fixme:file:CancelSynchronousIo (000000000000006C): stub
vrising | 00f4:fixme:file:CancelSynchronousIo (000000000000006C): stub
Mine has been running for 3 weeks without 1 crash. How often do you encounter this? Seems like a disk IO issue on the dockerhost.
It happened to me about 3 times in 2 weeks. I'm running the container on an AWS EC2 instance. I wrote a script that parses the logs for that specific message and restarts it, so it's all good now.
I am experiencing a similar issue, where the server kicks everyone out and the logs have that same 0x80000003 error with a CancelSynchronousIo loop afterwards. I'm running the docker image on my synology box. When trying to restart the container through synology's UI, the server only lasts for about 10 minutes before happening again. But if I shut down the whole thing and boot it up again later, I am able to get a couple days out of it.
I saw https://github.com/TrueOsiris/docker-vrising/issues/16, but I'm not seeing a 0 byte sized folder the same way that user was. I'll keep playing around with things to see if there are any other consistent patterns. I'm not sure if there's anything I can do to improve stability, or if this is just part of playing an early access game.
It seems to me that this is a server-running-behing issue, as memory/gamefiles are growing to large for the server to handle.