Exadata ACFS Cleanup: Safely Stop, Remove & Drop ACFS Volume (12c/19c Guide)

After using an ACFS filesystem on Exadata for Data Pump dumps, RMAN backups, or migration staging, it is critical to clean it up properly. Leaving the volume active wastes valuable space in the Disk Group.

This guide shows the exact, production-tested commands to safely stop, remove, and drop an ACFS volume so the space is immediately returned to the disk group.

Always follow this sequence. Skipping any step can leave orphaned resources or “device busy” errors.

Why Proper ACFS Cleanup on Exadata is Important?

• Releases every byte back to the Disk Group instantly
• Prevents space wastage
• Avoids Clusterware errors in future operations
• Keeps your Exadata storage clean and optimized
• Required after every temporary migration/staging use

Prerequisites

• Login as root for srvctl stop/remove
• Login as grid user for dropping the ASM volume
• Know the exact volume device name (e.g./dev/asm/acfs_2-323
• ASM instance must be running
• Root access on the Exadata compute node
• In Our Case We are using RECOC1 Disk Group

Step 1: Stop the ACFS Filesystem Resource

[root@ORACLEDBASECRETS01 ~]# /u01/app/19.0.0.0/grid/bin/srvctl stop filesystem -d /dev/asm/acfs_2-323

Step 2: Remove the Filesystem Resource from Clusterware

[root@ORACLEDBASECRETS01 ~]# /u01/app/19.0.0.0/grid/bin/srvctl remove filesystem -d /dev/asm/acfs_2-323

Step 3: Verify the Filesystem is No Longer Mounted

[root@ORACLEDBASECRETS01 ~]# df -h
Filesystem                           Size  Used Avail Use% Mounted on
devtmpfs                             1.2T     0  1.2T   0% /dev
tmpfs                                2.3T  4.3G  2.3T   1% /dev/shm
tmpfs                                1.2T   25M  1.2T   1% /run
tmpfs                                1.2T     0  1.2T   0% /sys/fs/cgroup
/dev/mapper/VGExaDb-LVDbSys1          15G  9.6G  5.4G  64% /
/dev/mapper/VGExaDb-LVDbVar1         2.0G  550M  1.4G  29% /var
/dev/mapper/VGExaDb-LVDbTmp          3.0G   54M  2.9G   2% /tmp
/dev/mapper/VGExaDb-LVDbOra1         200G   33G  168G  17% /u01
/dev/mapper/VGExaDb-LVDbHome         4.0G   62M  3.9G   2% /home
/dev/mapper/VGExaDb-LVDbVarLog        18G  2.8G   16G  16% /var/log
/dev/md24p1                          7.2G  183M  7.0G   3% /boot
/dev/mapper/VGExaDb-LVDbVarLogAudit  924M  140M  785M  16% /var/log/audit
/dev/md24p2                          254M  6.0M  249M   3% /boot/efi
tmpfs                                228G     0  228G   0% /run/user/1000
tmpfs                                228G     0  228G   0% /run/user/1001
tmpfs                                228G     0  228G   0% /run/user/0

Notice: /acfs_2 is no longer listed.

Step 4: Confirm the ASM Volume Device Still Exists

[root@ORACLEDBASECRETS01 ~]# ls -l /dev/asm/
total 0
brwxrwx--- 1 root asmadmin 251, 44033 Jul 18 20:14 acfs_2-323

Step 5: Drop the Volume from the Disk Group

[root@ORACLEDBASECRETS01 ~]# su - grid
[grid@ORACLEDBASECRETS01 ~]$ sqlplus / as sysasm

SQL*Plus: Release 19.0.0.0.0 - Production on Wed Jul 25 09:50:35 2025
Version 19.28.0.0.0

Copyright (c) 1982, 2025, Oracle.  All rights reserved.


Connected to:
Oracle Database 19c Enterprise Edition Release 19.0.0.0.0 - Production
Version 19.28.0.0.0

SQL> ALTER DISKGROUP RECOC1 DROP VOLUME acfs_2;

Diskgroup altered.

Success! The space is now immediately available again in the RECOC1 disk group.

Final Verification

SQL> SELECT name, total_mb/1024, free_mb/1024 
     FROM v$asm_diskgroup WHERE name = 'RECOC1';

NAME                           TOTAL_MB/1024 FREE_MB/1024 
------------------------------ ------------- ------------ 
RECOC1                            73798.3125   29861.7031 

SQL> SELECT name, volume_name FROM v$asm_volume WHERE volume_name = 'acfs_2';

no rows selected 

During a real 20 TB migration project on Exadata, we created an ACFS volume for Data Pump files and used the exact cleanup steps above. The entire cleanup took less than 2 minutes and returned 20 TB back to the RECOC1 disk group without any downtime or errors.

Key Takeaways

• Always use srvctl — never manually umount or delete device files
• Stop → Remove → Drop is the only safe order
• Run commands on any one node (Clusterware syncs automatically)
• Monitor RECOC1 disk group usage after cleanup
• Document volume names for future audits
• Test cleanup in non-production first

Conclusion

Safely cleaning up an ACFS volume on Exadata is quick and straightforward when you follow the correct sequence. This process ensures your RECOC1 disk group is freed up instantly and your Exadata environment remains optimized for production workloads.

Create ACFS on Oracle Exadata for Database Migration (Step-by-Step DBA Guide)

Mastering both creation and cleanup of ACFS gives you a powerful, zero-cost temporary storage solution for any large-scale database migration or data movement project.

Please feel free to ask if you have any questions. Thank you 🙂

Toufique Khan