Backups fail after re-add of ESX(i) hosts to Virtual Center (87225)

Backups fail to pre-existing ESX(i) hosts which had successful backups performed on them before. The failures coincide with a re-add of the host(s) to the Virtual Center (i.e. remove/re-add).

Several causes can be at the root of this problem. Two of the most common are:

1) Some hardware component (SAN LUN reformat/re-add, new MAC address for NIC(s), etc) may have been the reason for the re-add of the host machine(s). Certain hardware changes make the Virtual Center "see" the ESX(i) host(s) as NEW hosts, even though they may have the same name(s) and IP address(es). Because of this, the Virtual Center database gives the hosts a new UUID (Unique Identifier), which does not match the earlier UUID before the changes occured. vRanger does not know what "happened" to the "old" UUID(s), and assumes they are standalone hosts still in existence, only not tied to the Virtual Center anymore (vRanger can have a mix of hosts within a Virtual Center and external standalone hosts in the same Inventory of the same Ranger database).

2) The ESX(i) host(s) were originally added to the VCenter by IP or by FQDN. When they were removed and re-added, they were added by the OPPOSITE method they were added by originally. Since the inventory database of vRanger does NOT delete db records for Inventory objects (it merely marks them "Is Deleted") instead, a NEW entry with a new UUID (since the IDs are unique, they cannot be re-used for this new entry, even though both entries represent the same host/resolve to the same hostname). Since everything Ranger does regarding API calls to the vCenter/host MOBs (Managed Object Browser) uses this UUID instead of a "friendly name" or IP address, this breaks the backup jobs. The vRanger machine is "looking" for a host UUID that is no longer present.

To fix:

1) Launch Tools--->Startup Wizard.

2) REMOVE the Virtual Center. Once this is done, you may see some "phantom" hosts appear in the inventory. If so, right-click them and select "Remove Credentials".

3) Once these are purged, re-add the Virtual Center, re-credential the hosts that show-up in the pane underneath the VC after adding it back. These are the "new" UUIDs.

4) This usually will involve reconstructing your backup jobs/starting over with a new Retention policy (since the retention policies are tied to the jobs, not the VMs themselves).

OR

To get around recreation of backup/replication jobs:

1) From the VMware client, turn-off VMware HA(if using) and VMWare DRS (if using)

2) Remove each host from the cluster (if clustered). If not clustered, remove each host from the VC database. This does not involve powering-off any running VMs. They are still up and running.

3) Re-add each host to the cluster and/or VCenter inventory by the ORIGINAL method (i.e. via IP or FQDN) they were first added.

4) Re-add the VC /hosts to Ranger by the SAME method.

4) Test an existing vRanger job when complete (do a refresh inventory first from the "My Inventory" screen. The job should now work, as the re-add by the correct name or IP should now use the correct UUID in the Ranger database.