Hi,
have here 2*SG550XG as native stack. Attached are 3* vSphere 6.5 servers, via Etherchannel, redundunt cabeling across the stack members. This works fine.
After a FW upgrade the stack unit 2 ran into a boot loop. The good news was that the vSphere was up and running, so the redundancy worked.
I noticed today, that a couple of VM IPs were not pingable (but just combinations). This means, from the switch all IPs were pingable, from some host some IPs were not pingable to from other VMs/Hosts the IPs were pingable, so I think that the Etherchannel IP/MAC hash table was corrupt. I rebooted on host, this fixed some IPs, but some VMs were still not pingable.
After rebooting the stack and the hosts, all IPs are now pingable again.
How to analyze/avoid this kind of problems?
vSphere NIC teaming: Route based on IP hash
Load Balance Algorithm: IP/MAC Adress
Thanks
Henri