Last month we had a failover cluster that we had to upgrade from 2012 R2 to 2016. Ever since the upgrade the storage performance (connected to a SAN via iSCSI) had tanked and backups were basically causing the whole house of cards to tumble.

Eventually found that the MPIO policy on the cluster disk had been reset by the in-place upgrade (bad practice to upgrade in place I know but we are a small shop and I don’t exactly have the time to rebuild another cluster from scratch)

But that week after the upgrade was a total shitshow