Hi Everyone,
I've been doing disaster recovery tests on my Hyper-V cluster and have found that it takes around 30 seconds for my ISCSI traffic to fail over. Is this normal or should it be instant.
Background info:
SAN: Dell 3220i, dual/redundant controllers, 8 1 GbE ports total.
Cluster: 3 nodes running Windows 2012 R2
Server NIC setup: 4, 2 port cards. 4 ports for ISCSI traffic, 2 ports for live migration, 2 ports for network connectivity.
Switches: 2 switches for ISCSI traffic to provide redundancy. All NIC ports/cards are split between the switches
Simple Test:
Remove a node from the cluster. Could leave it in the cluster but wanted to eliminate as many variables as i can.
Remove the 2 ISCSI cables from switch 1 that run to the server simulating a switch failure. Leave the other 2 ISCSI cables that are in the second switch connected.
During this time a ping will constantly run to the server. However, if you try to RDP into the machine it will log in and sit on a black screen for abaout 30 seconds until it sorts out the ISCSI traffic. if i have a virtual machine running on the node while performing the test, the VM will be unresponsive for the 30 seconds as well. After this period of time everything will work fine.
Any insight would be great!