I have followed the steps in https://www.servuo.com/threads/tutorial-free-aws-ec2-shard-on-debian.14340/

After hosting a copy of ServUO in a Ubuntu EC2 instance, I started seeing that GMs doing teleports would get their clients frozen.

This doesn't happen if I host the same copy locally, either in Windows or in a Ubuntu local machine.

These are the ways I can reproduce this issue:

  1. GM teleporting fast moving around the world.
    1. Sometimes it takes only one teleport to happen, even if the player is not GM.
  2. Sailing in general.
  3. Having a macro that tries to move items too fast. This may happen by accident and it kills the client.

Does anyone have a rough idea of why this may happen? Or info on how can I debug this behaviour to find out?

I have the feeling that the EC2 instance is somehow failing to route all the traffic, but I have absolutely no idea and I don't know where to beging.
 
Not sure, I can just say that back then. I was helping with a shard that was running a server on ovh, same issue as you mentioned.

Then I pinged the server for a period of time and I always had a few dropped. The host company said "thats normal", after that we moved to another server (kinda the same performance since that is not the issue) and no more dropped pings / icmp and voila no more client freezes
 
Thank you for the reply, this is good to know.

I was thinking to try to host it on OVH actually. However you are making me reconsider. What hosting would you suggest instead?
 
I am not sure what good US hosters there are. I personally am running a server on netcup but thats in germany / eu.
You can sometimes get a nice sale there.

For US hosts you could search the forum. That question came up a lot
 
Back