Problem: During a failover, the recovery VM is not started when the protected VM is in a cluster (MSCS or Always On) with RDM or it has Thick Eager Zero VMDK disks
Viewed 1432 times
Multiwrite disks in clustering setups are required to be either set on an RDM disk or a Thick/Thin Eager Zero disk. However, ZVM can only instruct the vCenter to create a thick/thin lazy zero .vmdk file on the recovery site, due to limitations in the VMware environment. During a failover, after the recovery VM is created, vSphere will stop the recovery VM from booting as per the clustering disk provisioning requirements.
A similar error regarding one or more of the VM disks will appear:
"An error was received from the ESX host while powering on VM XXXX - testing recovery.
Failed to start the virtual machine.
Cannot open the disk '/vmfs/volumes/....../XXXX.vmdk' or one of the snapshot disks it depends on.
Thin/TBZ/Sparse disks cannot be opened in multiwriter mode"
The kept recovery disks should be changed to the Thick Eager Zero provisioning type.
The change of disk provisioning type can be done with an ESXi command-line tool: vmfkstools.
The needed command, which runs on the ESXi host is:
vmkfstools --eagerzero /vmfs/volumes/DatastoreName/VMName/VMName.vmdk
More details about the command line can be found in the following VMware KB:
After the change on the recovery disks, please then recreate the VPG and, in the process, use the kept recovery disks for the option of preseeded volume.