Random replication jobs suddenly start failing immediately after leaving the "queued" state and entering the "running" state with the error "An internal error occurred during execution, please contact Quest support if the error persists. Error Message: Unable to successfully load and create host system." Recreating/re-seeding a pre-existing job also results in the error. Recreating jobs"from scratch" does not fix the issue.
VMWare may have a lock on the system files of the VM (notably the .vmx configuration file), due to locking operations that occur during automated VMTools installations, or during failover/load-balancing operations in VMWare DRS/HA, etc. When the configuration files are locked, the snapshots needed to be taken for vRanger are not possible, as the .vmx file needs to be modified by the snapshot procedure at the beginning of the replication (and cannot be modified due to the lock on the file).
1) If possible, vMotion (a simple vMotion is all that is needed -- i.e. just change to a different host) should break the lock.
2) In Datastore Browser, download and edit the .vmx file for the VM. Look at the "IDE 0:0" mapping. It may show a mapping to the VMTools install .iso image.
3) If the link to the .iso image is present, the simplest way to now remove it is to (via the "Edit Settings" page in the VMWare client) CREATE an iso mapping to the virtual CDROM of the machine. Commit the change, then re-enter the "Edit Settings" page and remove it. This will "fix" the .vmx file.
4) If there is no mapping showing, the lock may be due other VMware functions, like DRS or HA migration. Check the VMware Knowledge Base for any information. In the meantime, vMotioning the Guest OS in question usually will fix other problems not associated with an incorrect mapped CDROM .iso file.
There may be other causes of this .vmx locking issue.