-
Title
Entry with the same key already exists -
Description
Restore job fails with error "Entry with the same key already exists"
It is a 2 disk VM. Savepoint was created during COS-based backup. VM had a custom snapshot at that time.
Restore is executed in the COS mode, target is credentialed ESX host.
According to the log VM has been created, matching custom snapshot taken. -
Cause
Virtual disks have rather unusual names: server-000001.vmdk and server_1-000001.vmdk
Snapshots are correspondingly server-000001-000001.vmdk and server_1-000001-000001.vmdk
Although it is still acceptable from VMware standpoint, vRanger looks for the snapshot file names to only include one set of 00000x rather than two. -
Resolution
Both metadata files for this Savepoint should be modified.
1. Manifest.metadata:
<FileMapEntry SourceFile=....server-flat.vmdk for 1st disk
<FileMapEntry SourceFile=....server_1-flat.vmdk for 2nd disk
<FileMapEntry SourceFile=....server-000001-delta.vmdk for 1st disk delta
<FileMapEntry SourceFile=....server_1-000001-delta.vmdk for 2nd disk delta
2. VmConfig.metadata:
2.1 <SnapshotFiles>
........
<string>[datastore] server/server.vmdk</string>
<string>[datastore] server/server_1.vmdk</string>
</SnapshotFiles>
2.2 <device xsi:type="VirtualDisk"> should be changed correspondingly: <fileName> and <parent><fileName>Next step is to create a new Restore from Savepoint Manifest job.
-
Additional Information
To avoid problem in the future, disks attached to the original VM (if exists) must be renamed.