【issue description】
The customer uses two LUNs of 2T and 28T to create a VMware datastore. The VM is created on the VMware datastore. When the ESXi host is restarted during O&M, the datastore of the two LUNs cannot be mounted, and VM services are interrupted.
【root cause】
1. During the deployment, the customer plans to use the LUN combination of 2T and 28T to create a 30T datastore. However, during storage mapping, two 28T LUNs are mistakenly mapped to the ESXi host. After the customer discovers the LUNs, the LUN mapping of the wrong 28T LUN is removed, then the LUN of another 2T is mapped to the host instead.
2. Because the customer does not re-scan the disk on the ESXi after re-mapping, the mapped 28T LUN is the same as the host Lun Id of the newly mapped 2T LUN. When the customer creates the datastore, the ESXi writes the WWN of the mapped 28T LUN into the metadata of the 2T LUN.
3. When the host is restarted, the WWN recorded in the 2T LUN is inconsistent with the actual WWN when the ESXi is mounted to the datastore. As a result, the disk cannot be identified and the data storage fails to be mounted.
【solution】
Map the LUN whose WWN is inconsistent with the actual WWN to a recovery host, and manually restore the metadata area of the record WWN to be consistent with the actual WWN.
