Sometimes we may observe that the gpbackup_manager replication failed with the error code 5056 returned from Data Domain, such as in the example below:
File copy failed to start on DDboost. Src file matt_lab/gpdb_6.4/20200514/20200514105028/gpbackup_1_20200514105028.gz Dest file matt_lab/gpdb_6.4_dr/20200514/20200514105028/gpbackup_1_20200514105028.gz. Err 5056 File copy failed to start on DDboost. Src file matt_lab/gpdb_6.4/20200514/20200514105028/gpbackup_20200514105028_toc.yaml Dest file matt_lab/gpdb_6.4_dr/20200514/20200514105028/gpbackup_20200514105028_toc.yaml. Err 5056 File copy failed to start on DDboost. Src file matt_lab/gpdb_6.4/20200514/20200514105028/gpbackup_20200514105028_metadata.sql Dest file matt_lab/gpdb_6.4_dr/20200514/20200514105028/gpbackup_20200514105028_metadata.sql. Err 5056
The error code 5056 from Data Domain means "Err: 5056-nfs mkdir failed (nfs: File already exists)", E.g, the backup set already existed on destination DD so source DD not able to create the folder. This usually due to the same replication has already been performed.
But in some cases, the previous replication may not be fully performed (for example there is a network outage caused replication half-done). To handle this situation, the workaround below can be used to clean up the backup set from destination Data Domain:
Remove the backup set directly from the destination Data Domain
Please refer to the admin guide of DD on how to delete files inside of DD file system
Remove the backup set on destination DD via gpbackup_manager
1. Create a YAML file with the info of target DD (please double confirm the yaml file is targeting to the destination DD), the example of yaml file can be found in [Using the DD Boost Storage Plugin with gpbackup, gprestore, and gpbackup_manager]
2. Use the command below to remove the backup on destination DD
$ gpbackup_manager delete-backup <backup_timestamp> --plugin-config <yaml_file_of_destination_DD>
3. Try the replication again.