Data Domain: Understanding Disk Reconstruction
Summary: This article provides an understanding of the Data Domain Disk Reconstruction process.
Instructions
Provide an understanding of the Data Domain (DD) Disk Reconstruction process.
APPLIES TO
- All Data Domain systems
- All Data Domain Operation Systems (DDOS) Releases
Data Domain systems use RAID 6 to ensure robust protection of customer data, even if there are multiple disk failures. Monitoring disk reconstruction status is essential when a disk fails.
-
RAID 6 Protection: Supports multiple simultaneous disk failures without data loss (RAID 6 has double parity protection per disk group).
-
Reconstruction Rules:
- A maximum of two disk reconstructions can occur simultaneously (across two disk groups).
- Exception: If two disks fail within the same Disk Group (DG), only one reconstruction can proceed at a time. The second will begin only after the first completes.
-
Global Sparing Behavior:
- Data Domain uses global spares across all external enclosures.
- If all spares in one enclosure are consumed and not replaced, the system pulls a spare from another enclosure with the matching disk size.
- This may result in a Disk Group spanning multiple enclosures. While it is ideal to keep RAID and spare disks within the same enclosure, this configuration does not impact performance.
-
Reconfiguration Guidance:
- If global sparing has occurred, contact support to verify system health and assist with disk enclosure reconfiguration.
-
Controller Disk Behavior:
- Head Unit (Controller) disks do not reconstruct to external shelf spare disks.
-
Best Practice:
- Replace failed disks as soon as possible to maintain optimal protection and system stability.
Viewing Disk Reconstruction Status on Data Domain
To monitor disk reconstruction progress on a Data Domain system, use the following DD-CLI command:
# disk show state
OR
# storage show all
Command Output Includes:
-
- Current disk state
- Reconstruction activity
- Disk group undergoing reconstruction
- Percent completion of reconstruction
Special Case:
If a second disk failure occurs in the same disk group (for example, dg1 or dg2), the output indicates a reconstruction pending status. This means that the second reconstruction will not begin until the first has been completed.
sysadmin@hostname# disk show state
Enclosure Disk
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
--------- ---------------------------------------------
1 s . . .
2 . . R . . . . . . . . . . . F
3 F . . . . . . . . . . . . . R
4 . s . F . . . . . . . . . . .
--------- ---------------------------------------------
Legend State Count
------ ---------------------------- -----
. In Use Disks 44
s Spare Disks 1
R Spare (reconstructing) Disks 2
F Failed Disks 3
------ ---------------------------- -----
Total 49 disks
Disk Disks Progress Remaining
Group Reconstructing (%) (minutes)
----- -------------- -------- ---------
dg1 2.3 0 238
dg2 3.15 99 10
dg3 (Pending)
----- -------------- -------- ---------