I started a 1-click ESXi HyperVisor upgrade of a 7-node cluster which progressed well and had already successfully upgraded 4 of the nodes, when one of the hosts wasn’t able to enter maintenance mode because one of the VMs was spending too long consolidating a snapshot. After many hours had passed, the remaining 1-click update tasks all appeared to have aborted.
I manually resolved the issue involving the guest, and then started the 1-click update again. It prompted for vCenter details, and then created new hypervisor update tasks to update the remainder of the ESXi hosts.
These also completed successfully.
All Hypervisor update Tasks appear to be completed, and all the ESXi hosts are updated.
All NCC checks also pass OK. The cluster is healthy, the Health Icon is green, and Data Resiliencey is OK.
Howerver, the CVMs all still believe there is a Hypervisor upgrade in progress:
nutanix@NTNX:~$ ncli host ls | egrep 'upgrade|99825'
Hypervisor Version : VMware ESXi 6.7.0 build-17499825
Under Maintenance Mode : false (Hypervisor upgrade)
Hypervisor Version : VMware ESXi 6.7.0 build-17499825
Under Maintenance Mode : false (Hypervisor upgrade)
Hypervisor Version : VMware ESXi 6.7.0 build-17499825
Under Maintenance Mode : false (Hypervisor upgrade)
Hypervisor Version : VMware ESXi 6.7.0 build-17499825
Under Maintenance Mode : false (Hypervisor upgrade)
Hypervisor Version : VMware ESXi 6.7.0 build-17499825
Under Maintenance Mode : false (Hypervisor upgrade)
Hypervisor Version : VMware ESXi 6.7.0 build-17499825
Under Maintenance Mode : false (Hypervisor upgrade)
Hypervisor Version : VMware ESXi 6.7.0 build-17499825
Under Maintenance Mode : false (Hypervisor upgrade)
nutanix@NTNX:~$
For example, none of them are willing to proceed with a cvm_shutdown because a they state a HyperVisor upgrade is apparently still in progress.
How can AOS be asked to understand that the HyperVisor upgrade has completed successfully?