Backgrounder: 3 nodes Windows 2008 R2 enterprise Edition failover cluster with CSV + hyper-v. Networks contains iscsi, host only, VM only, heartbeat + live migration.
Trouble: I have some windows and Linux based (CentOS) VMs. When I live migrate a Linux VM, it sometimes takes anywhere between 5 seconds to 60 seconds before it becomes accessible via network again. To be precise, I always continuously ping the VM from another VM while doing a live migration of a VM. That is how I noticed how long it took to become accessible again. However, after live migration, I can immediately use hyper-v manager or failover manager to access that VM. What have caused the delay? In theory, there should be no interruption. Could I misconfigure something? The cluster is otherwise healthy and has been working fine.
Also I do have NIC teaming configured at the host level for VM only traffic. This network is used by VMs only, and no cluster communication is allowed to use it. no sure if this would make a difference.
Please enlighten me!
Cheers,
Bo
want to learn C#