Originally posted by: arappoport
one of my customers is facing some VM freeze when the workloads are increased and it looks to be pointing to the storage... I am wondering if IO multipathing could be responsible for it...
So the configuration is:
few ESXi vSphere 4.1.0 hosts in one HP c7000 blade enclosure
Emulex CNA FCoE 11100
HP Virtual Connect FlexFabric modules (a pair)
one FC uplink per VC FF module to one Brocade 300 switch (2 Fabrics)
HDS AMS2500, each controller connected with one link to each FC-SW
ESX hosts are BFS (Boot From SAN)
so we have 4 FC paths to any given datastore from any ESX host.
My question to the community is around the FC/FCoE multipathing management and your recommendations if it could lead to VMs freeze.
it's currently set to RR (round-robin) and I have asked the customer to try with one VC FF module disabled.
Many thanks for your answers. (Merci beaucoup!)