We recently purchased multiple M1000e enclosures with PowerEdge M620/M820 and two MXL switches installed in slots A1 and A2. Both MXL switches are connected to Cisco Nexus 5548 switches. We have followed configurations within below document:
Blades are running ESX5, all blade ports are setup as hybrid ports. During our testing we notice outage to some blades if one of the MXL switche's uplinks are down. Blades have connectivity to both A1 and A2 switches so in theory when A1 uplinks are down the blade would send traffic towards A2. But this is not happening, we have enabled beaconing feature but that also did not help. We are working with our account team but so far we have not found any solution. Just wondering if someone else has seen something similar.
The hosts (blade servers) may not see the links as down when a switch goes down, and depending on your vSwitch or dvSwitch configuration, it then wouldn't trigger a failover to one of the other links (to the remaining switch).
Member since 2003