VM Terminal Servers displaying periodic overload
Posted on 2011-05-02
I am in need of some serious guidance here..
First a bit of background, we have two main applications, one being a financial package and secondly a web based clinical system. Previously they were both served from a single physical server. This server crashed and as part of a recovery plan we built two VM terminal servers to separate out the applications/load and provide a bit of redundancy. Both these VM are accessed via RDP, over our WAN and from external connections. WAN is a microwave linked series of VPNs
We are now frequently finding that over the course of the day, but particularly between 2:30 & 3:30pm one or both of these VMs are being seriously hit in the CPU stakes causing very poor response times at the user interface. (typically 40 users on one box, about a dozen on the other))
From the VM host perspective, looking at the resources for the particular VM - nothing seems to be over the top and each VM is on a separate host. (VM hosts are IBM Blades with SAS external storage, ds3200, Database and web server are also separate blades - but physical)
A collegue of mine is suggesting the network, I'm thinking the VMs themselves as we never had this problem when it was a physical device, we're looking to monitor all network devices, but would love any additional thoughts anyone might have as to where to look for the bottleneck!
thanks in advance for your advices