Is anyone else observing that vMotion between servers with Fling 2.1 seems to be much longer than previously? I applied Fling 2.1 as an upgrade on the older fling series, all networking is the same config, with 10GbE connections used for the vmotion kernel adapter. All being done using a shared iSCSI VMFS filesystem, so no data movement should be necessary, just a computer resource migration.
I don't normally use vMotion a lot, but I'm also dealing with unkillable/restartable VMs after CPU soft-lockups, and a 160 core, 1TB Ampere box with 28 VMs takes about an hour for all of the VMs to migrate off. No single VM is more than 20core/128GB. Most are <64GB RAM.
-------------------------------------------