Why automatic disk copy from external to internal shelf
Applies to
- ONTAP 9
- Systems with internal storage:
- AFF-A200
- AFF-A220
- AFF-A320
- FAS25xx
- FAS26xx
- FAS27xx
- FAS28xx
Answer
- Some platforms have both internal disks and external disks.
- When an internal disk fails and if there is no spare disk available in the internal shelf, ONTAP will take an external (unpartitioned) disk for reconstruction.
- As soon as an internal disk is available as a spare (for instance after the replacement of the failed disk), ONTAP will copy the content of the external disk to the internal spare disk automatically.
- If all partitions of the external disk are spare the disk is automatically unpartitioned.
Example:
- Internal disk 0a.00.11 has failed and was replaced by an external spare disk 0b.01.1.
- After the failed 0a.00.11 gets physically replaced by a new disk, ONTAP will start copying the data from the external disk 0b.01.1 to the new internal disk 0a.00.11.
- As shown in the "
sysconfig -r
" output below:
Aggregate aggr0 (online, raid_dp) (block checksums)
Plex /aggr0/plex0 (online, normal, active, pool0)
RAID group /aggr0/plex0/rg0 (normal, block checksums)
RAID Disk Device HA SHELF BAY CHAN Pool Type RPM Used (MB/blks) Phys (MB/blks)
--------- ------ ------------- ---- ---- ---- ----- -------------- --------------
dparity 0a.00.1P2 0a 0 1 SA:A 0 SAS 10000 55176/113000448 55184/113016832
...
data 0b.01.1P2 0b 1 1 SA:B 0 SAS 10000 55176/113000448 55184/113016832(replacing,copy in progress)
-> copy 0a.00.11P2 0a 0 11 SA:A 0 SAS 15000 55176/113000448 55184/113016832 (copy 3% completed)
...
data 0a.00.19P2 0a 0 19 SA:A 0 SAS 10000 55176/113000448 55184/113016832
- This is the expected behavior, storage pool partitions would be copied from external storage to internal storage automatically when there are spares in the internal storage.
- To modify the default behavior and prevent the automatic copy from external storage to internal storage, the following command could be run:
::> options raid.lm.disk.copyback.enable
::> options raid.lm.disk.copyback.enable off
Additional Information
If the new aggregate is created using both external and internal shelves and there are some spare disks on an internal shelf, it will also be subject to copy back as described above.