The Rebuild (data reconstruction) function automatically restores redundancy of data if the level of redundancy is reduced because of a drive failure or a storage node failure.
Two types of Rebuild are available:
-
Normal Rebuild: Rebuilds all the data on the storage nodes and drives to be rebuilt. "Rebuild" refers to Normal Rebuild.
-
Fast Rebuild: Rebuilds only the necessary data among the storage nodes and drives to be rebuilt. Fast Rebuild requires less rebuild time than Normal Rebuild.
The following table lists the types of rebuild for each rebuild trigger.
|
Rebuild trigger event |
Rebuild type |
|
|---|---|---|
|
Drive failure 1 |
Normal Rebuild |
|
|
Updating the software 2 |
Fast Rebuild |
|
|
Performing storage node maintenance recovery |
Manual maintenance recovery after occurrence of a storage node failure |
Normal Rebuild |
|
Manual maintenance recovery after maintenance blockade |
Fast Rebuild |
|
|
Storage node auto recovery |
||
|
Replacing storage nodes |
Normal Rebuild |
|
|
1. In the event of failure of a drive when the storage node containing that drive does not have enough free space, the Rebuild will result in a failure due to insufficient drive capacity. 2. If software update is interrupted, Normal Rebuild may be initiated only for the storage nodes that were undergoing software update. |
||
-
If, during the Rebuild, another Rebuild is triggered by performing operations such as maintenance recovery of storage nodes, the Rebuild is temporarily suspended and then restarted automatically. If the event log indicating that the Rebuild started (KARS07000-I) is output after the event log indicating that the Rebuild is suspended (KARS07002-I), the Rebuild has been successfully restarted.
-
Even in cases where Fast Rebuild should apply, Normal Rebuild will be performed if the following conditions are met.
Also, Normal Rebuild is to be performed if one of the following conditions is detected after Fast Rebuild. In this case, event log KARS07102-I is output.
-
When the Rebuild is interrupted1
-
When drives of a storage node are replaced for maintenance recovery
-
When an error occurred in the internal Rebuild processing
-
When user data redundancy (dataRedundancy) is 0 (no redundancy) in HEPC 4D+2P2
-
If the storage cluster stops or power supply to the storage cluster is interrupted after a storage node is blocked for maintenance or after a storage node fails
1. When the Rebuild is interrupted due to storage node or drive blockage, power-off or shutdown of the storage cluster, suspension of the software update, or other cause. Normal Rebuild is to be performed if it starts after the interruption.
2. Normal Rebuild will continue until the redundancy level is restored to 2.
-
-
When failure in drives and failure in storage nodes in which the drives are installed occur at the same time, the following phenomena might occur.
-
Multiple Rebuild processings are performed.
-
Event log KARS07005-E (indicating that the free space for Rebuild is insufficient) is output.
If multiple Rebuild processings are performed, event log KARS07002-I (indicating cancellation of Rebuild) is output. After a while, the Rebuild is resumed. When event log KARS07001-I (indicating Rebuild completion) is output finally, the operation is successfully completed.
Even when event log KARS07005-E (indicating that the free space for Rebuild is insufficient) is output, the situation might be temporary. If event log KARS07001-I (indicating Rebuild completion) is output finally, the capacity insufficiency is resolved.
-