Interesting that the smart-path setting helped to reduce that latency. If it was that configuration even after you changed out HW/parts, then I would be looking at the firmware.
As to why it happened - that would definitely be a casefile to send to HP. It reminds me of some problems I used to have with HW that had an unknown software bug that allowed for a buffer overwrite to fill up an SQL server overnight. We had to get the HW vendor involved to do a firmware update. It was not a common bug or occurrence, it just happened when we had a certain configuration applied. We just did a pentest without realizing we were doing a pentest.
I'm glad you got some resolution though. It's up to you if you want to open a case with HP. Or VMware for that matter (just to cover all bases)
