Vmotion Failure after copy.

[carl robinson]

I have 2 ESXi Host on a pair of 10G fex's via port-channel without LACP.

(2) vmotion Nics, (2) Data Nics.

We can see the vmotion copy the images successfully over the vmotion Nics.

I was able to see the RARP broadcast, (VSphere , "Switch Notify" is enabled) using the ethanalyzer inband through the upstream 5548, but I never see a unicast response from the new host?
When I clear the mac from the table, the Vmotion completes immediately.
Otherwise, we have to wait until the Mac timer expires. (300sec)

This makes me think the receiving host is not seeing the broadcast or the switch is not seeing the unicast response.
Any suggestions?

Thank you in Advance...

[Kirk J]

Might also want to fire up pktcap-uw tool on ESXi (https://kb.vmware.com/s/article/2051814 ) and confirm what the ESXi host is/is not seeing.  You have to run two sessions at same time to capture both rx/tx at same time

See note at link:

• pktcap defaults to inbound traffic only.
• In vSphere 6.5 and earlier, specify the direction of traffic using --dir 0 for inbound and --dir 1 for outbound. You can’t specify traffic going both ways at the same time. However, in vSphere 6.7 and later, you can specify the direction of traffic using --dir 0 for inbound, --dir 1 for outbound, or --dir 2 for both.
• Two (or more) separate traces can be run in parallel but need to be merged later in wireshark.

Kirk...

[carl robinson]

Thank you, Kirk.

I seem to be loosing layer 2 information in the captures.

We tested again with a single nic and it worked flawlessly.

So the standard vswitch is seeing an issue with the standard port-channel.

We are not running lacp.