Ok, I tried WinMTR tests. The ping times were always pretty low. Every now and then there would be a lost packet (like 1 in over a thousand packets). I don’t think that minimal amount of loss would cause what I’m seeing. Also, the packet loss didn’t coincide with the occurrences of the freezes.
I logged a customer support ticket 2 weeks ago. The first GM gave the following instructions:
- Uncheck the “Optimize Network for Speed” option in the in-game Network settings.
- Perform an ipconfig /flushdns, ipconfig /release, ipconfig /renew.
- Prioritize the game in Task Manager to High Priority.
- Perform a “Scan and Repair” in the Battle.net client.
- Shutdown computer. Unplug router and modem. Wait 60 seconds. Plug modem back in and wait for it to come up. Plug router back in and wait until it is ready. Turn computer back on.
I did all of the above. The problem seemed to be fixed–for two days. And then the freezes started happening again. I tried repeating all the same steps, but this time the freezes continued to happen.
I re-opened the ticket, and the next GM gave these instructions:
- If you’re using a wireless connection (which I never was), try wired.
- Reset your user interface.
- Reduce the size of the game packets (i.e., the first step that the first GM said to do).
So, that wasn’t helpful, since I’m wired, I had previously tried the interface reset, and had already turned off the setting.
In a dungeon last night when I experienced the freeze, over voice coms I told my group that I was frozen. Whenever I freeze, I will continue to move my character around with the hope that I won’t die to targeted ground mechanics, and will try to keep casting my rotation (though on my side, I never see any casts complete). Another guy in the group told me that even though I said I was frozen, he saw my character moving around and casting spells.
So, This is definitely an issue related to packets from the server. It seems like when the issue occurs, all of the incoming packets from the server are batched together, and eventually get processed all at once. I don’t know if the server is batching them, I don’t know if Windows is batching them, and I don’t know if the client is just delaying pulling them off its message queue (while it continues to be able to send messages to the server).
A third GM recommended reaching out to the Technical Support Forums for further assistance, providing my observations.
Anyone have any other ideas of other diagnostics or logs I can look for to figure out where the packets are getting delayed or if they are getting batched or anything else?