Description of problem:
Recently my customer opened a support case regarding an error message in their environment. I have also seen this same error in my lab and I, too, found it confusing at first since it doesn't actually tell you the real reason for the error.

# hosted-engine --vm-status
Unable to read vm.conf, please check ovirt-ha-agent logs

The customer reported that they were not able to start the hosted-engine because "the config file is missing" which, while accurate, is not the real reason. The real reason was that the Hosted Engine Storage Domain could not be mounted. After checking the storage connection and restarting the ovirt-ha-* services, the customer was able to launch the hosted-engine VM.


Version-Release number of selected component (if applicable):
4.0.6

How reproducible:
Very

Steps to Reproduce:
1. Create a hosted-engine environment hosted on NFS storage and cause an outage of some kind where the NFS mount doesn't get properly unmounted or can't be mounted at all.
2. Restart the ovirt-ha-agent and ovirt-ha-broker services.
3. No errors on ovirt-ha-* service restart. Only later in 'systemctl status ovirt-ha-agent ovirt-ha-broker' is the real error seen.
4. Run 'hosted-engine --vm-status'.

Actual results:
Error message is confusing and doesn't indicate a storage domain connection issue.

Expected results:
Error message should be more descriptive and provide users the proper commands to run (or perhaps a link to the attached kbase for more detailed resolution steps).

Additional info:
https://access.redhat.com/solutions/2973011