I have a clustered Hyper-V environment with four nodes and three PowerVault MD3xxx storage arrays for shared storage.
The servers are using standard Ethernet connections for the iSCSI fabric (not HBAs). MPIO is configured via the Dell storage wizard.
The problem that I seem to be having is that upon rebooting a node, the Ethernet connections seem to be taken down before the iSCSI initiator is ready for it – so the connectivity to any of the LUNs that it happens to own is lost unexpectedly.
Is there a way of preventing this so that a shutdown will be more graceful? In future I will be making sure that a node that is going to be rebooted has any storage it owns shifted to another node first, but I was wondering if this should really happen automatically.
Best Answer
To answer the question - yes this should happen automatically. I have an estate with a number of Hyper-V servers, MD3xxxi arrays, etc. and have not seen an issue with rebooting nodes.
However, what is the error or issue you are seeing on a shutdown that makes you say "lost unexpectedly" ?
Also, how are you presenting the shared storage? Is it a CSV or are the LUNs shared between VMs that are in a windows cluster? (I note you have tagged mscs).