Cisco UCS Infra. FW Update - One ESXi host loses connection the other doesn't

[chipskind]

I applied infrastructure and server FW updates to our Cisco UCS environment the other day.  I used Auto Install.

For the Infrastructure Firmware update, I expected to lose half my SAN fibre paths and half my network paths during each one of the Fabric Interconnect reboots.  I was pretty sure all hosts/blades (Windows Servers and ESXi Hosts) would ultimately stay up.

Well, all hosts stayed up except one ESXi host.

I have different Policies for Windows Servers and ESXi Hosts.  Both ESXi Hosts use the same policies.

I have taken a look at the ESXi hosts themselves via the KVM just to see if something would be wrong there.  Nothing.

I took a look in the vSphere vSwitch0 (Mangement/vMotion) of both hosts and still nothing inconsistent.

I made sure both ESXI servers were using the same Service Profiles/Policies.  They are.

Although the same, I took a look at the VIF Paths, and Network tabs but I found no differences.  I didn't expect to.

So if both use the same Service Profile Template and I have no Faults at all, where else should I look?  Am I missing something here?  Obviously I am.

Thanks

Chris

[Walter Dey]

Hi Chris

you are 100% correct ! nevertheless, something must be different in the ESXi setup, or something is wrong in the Northbound connectivity..

If the SP are derived from a template, and one is working, the other not, you can be sure, that it has nothing to do with UCS !

What is the Northbound connectivity ? vpc's ?

Walter.