I am running ESXi on a Dell PowerEdge r310. I currently have 2x 250GB hard drives installed in a RAID1 for the Hypervisor and 2x 2TB drives setup in RAID1 for the VM datastore. I figured keeping them separate would be nice in case there were issues with the Hypervisor array. This has been running for approximately 2 days now and last night I went ahead and moved our production VM over to the machine. Today around peak time I receive this message:
Device naa.600508e000000000a54c6f66b1cbaf06
performance has deteriorated. I/O latency
increased from average value of 12049
microseconds to 242737 microseconds.
warning
At about the same time my production VM locked up. I shut it off and started it back up and has been running fine since.
I know that the 2x 2TB drives in RAID1 is less than ideal. My question is this. Am I just wasting my time keeping the Hypervisor separate from the virtual machines? Would I be wise to install the Hypervisor on a single drive (zero redundancy) and then use 3 drives in RAID5 for the vmstore? Should I just use all 4 drives in one array in RAID10?
Any advice would be appreciated.