This sounds like an HPE issue and you should be lodging a support case with them.
We are facing excatly the same problem (2x HPE Proliant G9).
Case at HPE is ongoing.
Current Actionplan from Level 3 Support:
1. Update all VMware hosts that connecto the the volumes to MEM version 6.7/
2. Uninstall MEM on all hosts.
What a suprise, this worked. I replied and ask if they dont recommend to install the plugin, even with the released version 6.7 ...
Keep you up to date here!
Thanks for answering. Glad to hear I'm not the only one with this problem.
So HPE advice is to uninstall the MEM-driver on all hosts and revert to default VMWare round-robin?
I'm sure that works but then Multipathing to the SAN is gone. We need multipathing to work, so this is not an acceptable solution for us.
Please keep me up to date on your progress.
Any more information on this one ?
We also have the PSOD's on reboot/shutdown with the MEM module, and after reboot it happens not all datastores come back online. The devices get status "Not consumed".
Removing MEM brings back stability.
Running on Gen9 Synergy blades, VMware 6.7U2 and U3, with the 6.7-version of hpe-lh-mem extension.
Today we tried to power down one VSA node, and one ESXi host (with the PSOD problem) failed to switch paths and crashed completely.