All,
I currently run an R610 server (X5550 procs) and a newly acquired R620 unit (e5-2650) unit in an vSphere 5 cluster. ESXi running latest version and latest patches. Dell servers up-to-date with regards to firmware of all components. The R610 unit uses internal 15k SAS drives to host the ESXi hypervisor, while the R620 uses mirrored 2GB SD cards for the ESXi hypervisor.
When I migrated some workloads over to the new cluster member (r620), I saw execution times for some of our processes increase dramatically. The processes ran on a MS SQL 2008 r2 server. When I migrated the VM back over to the original cluster member (r610), execution times dropped back to normal times.
I then proceeded to run a series of benchmarks against a test VM that was migrated back and forth between each cluster member and saw reproducible results that showed the r620 was SLOWER than the R610 in terms of CPU and Memory benchmarks (using Passmark v7.0). Cluster was using EVC mode Nahalem. I also ran tests by removing the EVC mode and was able to get similar results with the R610 being the faster performer.
I've verified all cores and hyperthreading is enabled. Glancing at the Best Practices guide for ESXi 5.0 yielded nothing different in my setup..
Has anyone seen or heard of these issues with the R620 unit, or the E5-2650 processors?
Thanks,
Elvis
