Hello dear Life Is Feudal Community!
We have 3 LifeIsFeudal Servers running on one root machine.
Server 1 was created before the last game update so it's from an older version and it ran on another machine.
Later we created Server 2 and 3 on a newer machine and then also moved Server 1 onto the same one (DB dumped, copied and imported - this task was completed without any problems and every player still had their stuff properly).
The World IDs are perfectly in order:
Server 1 = ID 1, DB: LIF_1;
Server 2 = ID 2, DB: LIF_2;
Server 3 = ID 3, DB: LIF_3;
The DB is running on the same machine as the game servers.
The System is Windows Server 2008 x64
CPU x12 (Intel Xeon) (unrestricted performance/speed/amount)
16 GB RAM for that specific VM where those 3 Servers are running.
System load is low ~60% RAM
2-20 % CPU load
Now to the problem:Server 1 is crashing a lot, Server 2 and 3 are not even nearly crashing as much as Server 1.
Also it's important to tell that the affected Server 1 is freezing as soon as it reaches the 4 GB RAM mark and needs to be restarted MANUALLY - which creates a huge problem as we can't be there 24/7 to do that. It doesn't shut down by itself as it can't react when it's frozen.
We think that the reason could be a lot of memory-trashing.
For example by objects not being deleted properly and/or too many objects in the world that should be deleted (specific post to that topic
here)
Another (side-)issue:Also, we have FireDaemon running to make sure the processes get restarted properly if they ever crash. Weirdly enough Server 1 (which is running on the machine for about 2 days already) made it out of Firedaemon's control, Firedaemon lost it "out of sight" and assumed it crashed, trying to restart it all the time, causing plenty of logs being created and the process stopping because it detected that the instance is still running. The only thing we could do about it was shutting down the proccess manually and restarting it with Firedaemon afterwards.
But mostly another problem is the case FireDaemon can't help any more: The process freezes as stated above.
Please ask for further informations if neccessary.
We will provide the Log Files per PM as soon as we get a response by any staff and if you need a proccess dump, we can provide that too.
Thanks for reading.
PS:
Yes we already applied this fix.
Also we deactivated WerFault Service (it causes a system CPU-load of 100% everytime one of the server crashes).