Article number
000003868
Affected Versions
All
Source Hypervisor
All

Promotion Does Not Progress While Recovery VMs Are Powered Off

Viewed 123 times

Summary

The promotion of the mirror disks is essentially the reading of all remaining data from the journals (still on the Recovery VRA machine) and writing said data to the mirror disk (now attached to the recovery VM). If, for whatever reason, a recovery VM is not powered on, promotion of all attached mirror disks will not be able to progress forward as the VM must be able to handle the write IOs to said disks. After a Failover/Move is committed, there may be Recovery VMs that fail to power on for one reason or another. One known issue is if the Recovery Host does not have ample CPU reservations for the recovery VMs that were just spun up. The error will look like the following:


7 out of 14 VMs started successfully
VM_8 failed to power on Failed setting VM vm-xxx to power state On Message: The host does not have sufficient CPU resources to satisfy the reservation.. Fault: Vim25Api.InsufficientHostCpuCapacityFault.
.
.
.
VM_14 failed to power on Failed setting VM vm-xxx to power state On Message: The host does not have sufficient CPU resources to satisfy the reservation.. Fault: Vim25Api.InsufficientHostCpuCapacityFault.



If an error similar to the above has occurred at the end of the Move/Failover Before Commit stage and/or the Recovery VMs are administratively powered off, promotion times will greatly increase.

Steps

1] Avoid powering off any Recovery VMs administratively until after the promotion is completed.

2] If there is an error such as the one mentioned above in the Symptoms section, resolve the environmental issue to avoid in future attempts and to be able to power on the necessary recovery VMs.