Hi - I have an 8 node cluster running on HP G6 BL490c blades which is currently about 90% committed on RAM. They are running ESXi 4.1 (build 348481)
Currently I am seeing very little TPS activity and quite a lot of swapping and balooning which concerns me. My understanding was that TPS kicked in later on this hardware due to the preference for using large pages but I would have expected it to trigger before seeing swapping began.
There are no reservations or limits set on VMs or resource pools & no advanced settings have been changed from defaults. MMU is set to automatic on all running VMs. DRS is automatic at 3 stars.
Is there some vital setting I have overlooked to get TPS working?