Cisco Hyperflex Upgrade Failing

[IC2317]

Hi All,

I am unable to do the cisco hyperflex (Contains 4 nodes) upgrade from  and keeps on getting below error. I have already checked firmware/software compatibility before upgrade and also ran pre-check script with no errors/warning in output.

On hyperflex connect, while upgrading I get one warning related to insufficient capacity in the cluster which i can skip. I don't have idea why this warning is coming even though my cluster can tolerate 1 host failure. However, I don't mind getting low performance during the upgrade as I always do this upgrade out of office hours. Good part is that Hyperflex upgrade allows me to skip this warning. When I skip this warning and go ahead with the upgrade I get above error which does not explains anything. 

One thing i noticed is that upgrade do copy and extract on all ESXi hosts except the one that is running leader storage controller VM. For example in the above image on ESX1, it gives message running cluster pre upgrade tasks while on ESX2,3,4 it successfully copy and extracts.

[Kirk J]

Without a more specific error, you'll need to open a TAC case.

Thanks,

Kirk..

[Steven Tardy]

Did you re-execute the same upgrade workflow?

Sometimes the workflow works on subsequent executions.

It doesn't need to copy/extract the files to ESXi01 because that is the current management IP. So no need to copy the file to "itself".

If a re-execution of the same workflow doesn't work. . . . Without more details, would likely need a Cisco TAC case like Kirk said.

[IC2317]

Thanks for the response. Yes, I have tried the workflow multiple times but it fails with same error. I have logged case with Cisco and they are suspecting that it is bug on version 4.0(2e), which is the one from which I am upgrading to 4.5(2b)

I shall update this thread once I have any further updates from TAC.

[CSOSTSupportSupport6291]

Any update on this.  We are experiencing something similar and are still waiting on Cisco to reply.

[Riaan van Niekerk]

This looks like a vSphere issue not an HX issue. Can you put one host manually in HX Maintenance Mode?

Do you perhaps have vSphere admission control enabled with too high a reservation % or a host as failover?

https://docs.vmware.com/en/VMware-vSphere/7.0/com.vmware.vsphere.avail.doc/GUID-53F6938C-96E5-4F67-9A6E-479F5A894571.html

”upgrading to 4.5(2b)” Any reason you are not upgrading to the latest 4.5, e.g. upgrading to 4.5(2e), or even 5.0(2e)?

Which server model, UCS firmware, vCenter version are you running? Which ESXi are you upgrading to (if you are).