our cluster is starting to have high memory usage and since the number of VMs isn't decreasing anytime soon so I need to start to think about increasing the memory.
I'm wondering can I just add more memory to the current servers or are there any other bottle necks. We aren't running any big workloads, mostly test servers, AD and web-servers. Email and database are on dedicated HW.
I've attached screenshots of the server details, so it's easier to see what we are running.
All data is stored in a HP P6300 SAN. It has 50 10k RPM disks and 4Gb connectivity.
One 1Gb NIC is dediated to vSphere traffic
Two 1Gb NICs are in our LAN
One 1Gb NIC is in our DMZ
None of the performance monitors show high usage on disk, network or CPU. But somehow I have the feeling that the performance is not optimal.
Those server specs look okay - it does look like too much RAM is being used though - the CPU brief graphic looks good, if you lost a host you certainly wouldn't have failover capacity as the resources don't add up.
You have two options really
1) Reduce the RAM requirements on the VM's to allow failover, just switching off the test servers not in use would be enough.