Benchmark and possibly optimize longhorn (moc-infra)

[naved001]

I was expecting RWO storage provided by longhorn to be faster than RWX but that wasn't the case for the very brief FIO tests I did. I am opening this issue to remember to do it at some point.

• Benchmark the storage on the host
• Benchmark pods with RWO and RWX PVCs in both block and fs mode.
• See if there's any parameters we can tune in longhorn.