I ran into an interesting vSphere Data Protection (VDP) error: Error Code 30931: Failed to initiate a backup or restore for a virtual machine. Probable cause is the datastore for the virtual machine is not accessible. Naturally, I search the VMware Knowledge Base (KB) and located this article: VMware vSphere Data Protection 5.1.x fails to start backup jobs (2037003). The KB article got me started, but I thought it would be worth a blog post to elaborate a bit and share the resolution I found.
As suggested in the KB article, I disconnected the virtual CDROM drive from the virtual machine that was causing the error (I had only a single virtual machine in the backup job that was failing). I tried the backup job again – this did not resolve the issue.
I followed that by opening an SSH session to the appliance – ssh firstname.lastname@example.org – where 192.168.10.51 is the IP address of my VDP appliance and running the command found in the KB article – echo “select * from proxy_datastore_mapping” | psql mcdb -p 5555 -U admin – which produced the output shown in the figure below. Running this command should produce approximately eight rows per datastore. This is because VDP has eight internal proxies. I had five datastores in my environment. As expected, I received 40 rows (five datastores x eight proxies).
I tried the backup job a few more times all resulting in failure. I decided to try “if all else fails, reboot (the VDP appliance)” and sure enough, that did the trick. I suspect the combination of removing the CDROM mapping from the virtual machine and rebooting the VDP appliance resolved the issue.
If you run into the same problem, please share your experience. Thank you.