How To Reclaim ASM Disks

The project I have been working on is migrating GoldenGate trails from DBFS and ACFS. Now that the migration has completed and the DBFS database has been dropped, the next step is to reclaim ASM disks. DBFS is running on 2 nodes RAC.

1. Initial Environment Assessment

Before proceeding with disk reclamation, it is essential to verify the current state of the ASM instance and ensure the target diskgroups are identified correctly.

Checking ASM Version and Current Diskgroup State

From node1 - Check asmcmd version:

$ asmcmd -V asmcmd version 12.1.0.2.0 

List ASM diskgroup:

$ asmcmd lsdg State    Type    Rebal  Sector  Block  AU       Total_MB  Free_MB  Req_mir_free_MB  Usable_file_MB  Offline_disks  Voting_files  Name MOUNTED  EXTERN  N         512   4096  4194304    307184    51044                0           51044              0             N  ACFS_DATA/ MOUNTED  EXTERN  N         512   4096  4194304   9289488   793500                0          793500              0             N  DATA/ MOUNTED  EXTERN  N         512   4096  4194304    153584    83580                0           83580              0             N  DBFS_DATA/ MOUNTED  EXTERN  N         512   4096  1048576     15344     9832                0            9832              0             Y  GRID/ 

2. Pre-Reclamation Verification

Reclaiming disks requires ensuring that no processes are currently accessing the files within the diskgroup and identifying which physical disks are associated with the target.

Verifying Open Files and Identifying Candidate Disks

Verify no open files from DBFS_DATA:

$ asmcmd lsof -G DBFS_DATA DB_Name  Instance_Name  Path 

Check current candidate disks:

$ asmcmd lsdsk --candidate Path /dev/mapper/dbfs_data02p1 

List specific disks currently assigned to DBFS_DATA:

$ asmcmd lsdsk -G DBFS_DATA Path /dev/mapper/dbfs_data01p1 

3. Step-by-Step Reclamation on Node 1

The first phase of removal involves unmounting the diskgroup from the first node in the RAC cluster.

Unmounting the Target Diskgroup

Unmount disk for DBFS_DATA:

$ asmcmd umount -f DBFS_DATA $ asmcmd lsdg State    Type    Rebal  Sector  Block  AU       Total_MB  Free_MB  Req_mir_free_MB  Usable_file_MB  Offline_disks  Voting_files  Name MOUNTED  EXTERN  N         512   4096  4194304    307184    51044                0           51044              0             N  ACFS_DATA/ MOUNTED  EXTERN  N         512   4096  4194304   9289488   793500                0          793500              0             N  DATA/ MOUNTED  EXTERN  N         512   4096  1048576     15344     9832                0            9832              0             Y  GRID/ 

4. Final Cleanup and Removal on Node 2

Once unmounted on the first node, we move to the second node to verify the state and permanently drop the diskgroup from the ASM metadata.

Final Verification and Dropping the Diskgroup

From node2 - List ASM diskgroup:

$ asmcmd lsdg State    Type    Rebal  Sector  Block  AU       Total_MB  Free_MB  Req_mir_free_MB  Usable_file_MB  Offline_disks  Voting_files  Name MOUNTED  EXTERN  N         512   4096  4194304    307184    51044                0           51044              0             N  ACFS_DATA/ MOUNTED  EXTERN  N         512   4096  4194304   9289488   793500                0          793500              0             N  DATA/ MOUNTED  EXTERN  N         512   4096  4194304    153584    83580                0           83580              0             N  DBFS_DATA/ MOUNTED  EXTERN  N         512   4096  1048576     15344     9832                0            9832              0             Y  GRID/ 

Verify no open files from DBFS_DATA:

$ asmcmd lsof -G DBFS_DATA DB_Name  Instance_Name  Path 

Drop DBFS_DATA and verify the updated diskgroup list:

$ asmcmd dropdg -r DBFS_DATA $ asmcmd lsdg State    Type    Rebal  Sector  Block  AU       Total_MB  Free_MB  Req_mir_free_MB  Usable_file_MB  Offline_disks  Voting_files  Name MOUNTED  EXTERN  N         512   4096  4194304    307184    51044                0           51044              0             N  ACFS_DATA/ MOUNTED  EXTERN  N         512   4096  4194304   9289488   793500                0          793500              0             N  DATA/ MOUNTED  EXTERN  N         512   4096  1048576     15344     9832                0            9832              0             Y  GRID/ 

5. Conclusion

Check candidate disk: /dev/mapper/dbfs_data01p1 is now successfully listed as a candidate

$ asmcmd lsdsk --candidate Path /dev/mapper/dbfs_data01p1 /dev/mapper/dbfs_data02p1 

In summary, don't forget to reclaim disks after dropping databases so they can be repurposed for other storage needs within your infrastructure.

Oracle Database Consulting Services

Ready to optimize your Oracle Database for the future?