I have a few virtual servers running on a VMWare-ESXi server. One of those is a mailserver running Windows server 2003, Has been running quite well for a while now. About 4 days ago, the mailserver runs beautifully for about 5-8 hours at which point it slows down to an unusable crawl.
When I try to VM into the server it most likely tells me that it' can't connect and I can't see what's going on.
When I try to ping the server, it usually drops 2-4 packets, ping 2 seconds later a different 2-4 packets with the pingback time ranging from <1 to 200+ wtih a few request timeouts and somethimes a destination unreachable. In fact on some pings I get destination unreachable, request time out, and success at <1ms pingback. And not in that order.
To remedy the situation I do an F12 reboot of the physical machine and start up the virtuals inside one by one, Then I get a few hours before it starts crawling again.
But before I start up ANY virtual machines I get the same unstable ping behaviour when pining the server's management adress. If I ping the virtual servers when everything runs nicely I get no packet loss. When I VMWare to manage the server because I need to boot up the virtuals I will frequently see it lose the console and get it back.
This server is 1 of 4 twins in a rack the other twins have no such problem, very solid, very fast, ping anything <1ms no loss.
Has anyone encoutered a similar problem or does the behaviour ring a bell.
Am I looking at hardware failure or bad config/software?