Hi guys, apologies if I have posted this in the wrong section, I have registered hoping you can help with our VPS.
We have a VPS running Windows Server 2008r2, with 2 cpu cores on an i7, 1gb ram, and 100mb connection. We use it to run a small game server, a small website/forum with IIS, and a filezilla FTP server.
Recently we have been having problems where the connection drops completely to everything, including RDP, for 10-20 seconds a time. During this time, all packets are lost. Sometimes the game server reports packets were modfied.
I have attached a picture of performance monitor when this problem occurs. Although the graph shows bytes total, when I change that to two lines, it is bytes received that shoots up beyond the total bandwidth.
Which virtualizaton technology is used on the VPS?
I believe it is Xen based.
Originally Posted by AvailHosting-Jeff
How often does this connection drop occur? Does it occur in any sort of a pattern (e.g., every 30 minutes, only during high traffic, etc)?
The pattern seems completely random. Sometimes it can run for a few hours under heavy traffic, and then have these connection drops happen every few minutes, or it can do it with almost no traffic at all.
Originally Posted by Lincxu
the picture above does not tell too much about the problem, I would ask the hoster to moniter your server form some time and see what happened when problem occurs.
The hoster said they could not find any problems and were unsure of the cause, so we are currently trying our best to monitor it ourselves to try and find a cause.
After manually adding the setting, the missing gateway was fixed, however that has turned out to be a seperate problem, not the original one. I managed to install screen capture software on the VPS to capture when the problem occurs, and CPU usage never goes above 40-45% when the problem occurs, it is still just the total bytes received shooting up that is the only indicator of the problem.
Apologies for the double post, please ignore my last post. After looking at the graph again, it appears the CPU usage also shoots up at the same time as the total bytes received. It is hard to tell however due to the gap. I have attached an image of the graph, you can see the problem occuring twice.
How can we track down the rogue process? We never see it as the server seems to lock up when it occurs. Could that because of the total bytes received causing the CPU to shoot up?
The red line is CPU usage, the yellow line is Total Bytes received/sec.