Crashing after character select (Fixed)

[kedra]

I wanted to recount my recent experience troubleshooting an eqemu server I was trying to host on a linux virtual machine. I spent an embarrassingly long time troubleshooting this problem which turned out to have a very easy fix. Hopefully this will help someone else if they run into a similar issue.

I had an existing, functioning eqemu server running on debian 10 that I tried booting up after a long period of disuse and found that all of my existing characters would crash to the server select screen after the character select screen.

Since this was a test server with nothing valuable on it, I ended up reinstalling the vm with debian 11. I followed the linux install guide on the wiki and used Akkadius install script which executed nicely and all binaries compiled successfully. I modified my eqemu_config.json, and double checked the port forwarding on my firewall.

Everything looked good but after starting the server and trying to log in I crashed after the character select screen again. The world log appeared to show a successful spin up of a zone instance, but the character never entered the zone. Zone logs simply showed the instance spin up, then go idle again.

I double checked config files and ports and scoured the wiki and discord for clues. Most of the threads I saw on the forums with similar symptoms turned out to be problems with the config file or with port forwarding. I tried reinstalling with ubuntu, moving the VM to a new host, and everything else I could think of.

I found a post on discord that described the problem exactly:

cobaltblue — 08/14/2022

I've got a hyper-v debian 11 vm hosting the server (configured w/ external switch). Both the VM and my machine have separate static IPs on the router and the VM has ports forwarded. Trying to run the client on my machine. I successfully authenticate to the VMs private login server. I successfully created a character. no luck on <Enter World> however...the server is not responding. -

client log.. dbg.txt
Zone addr [192.168.1.x:22555] received...
ERROR: World did not drop the connection.
Failed to connect to zoneserver (192.168.1.x, port 7000), result = 5

server log.. world_679.log
(Nostalgia) [Zoning from character select] [felwitheb] (Zone ID [62]: Instance ID: [0])
Broadcasting a world time update
Setting zone process to Zone: Southern Felwithe (felwitheb) ID: 62
Local zone address was [127.0.0.1], setting local address to: [192.168.1.x]
Sending client to zone [felwitheb] ([62]:[0]) at [192.168.1.x]:[7000]
[Client Login] ClientListEntry::SetOnline for [myUsername] (1) = [Online] (4)
Client disconnected (not active in process)
Removing client from [192.168.1.y]:[3289]

appreciate the guidance. I know @tibbs3d had a similar issue several months back. (edited)

Unfortunately there weren’t any clues in the follow up discussion.

I’m happy to report that my particular problem was fixed by reconfiguring the virtual machine to use 8 virtual processors instead of 4. The only other notable spec is 8GB RAM which remains unchanged. I have no idea why 4 processors were apparently insufficient but my best guess is that it was resulting in a delay and timeout when passing the client from the world process to zone...

So there you have it. Hopefully someone finds that useful.