Administrator Guide

About reconstruction
If one or more disks fail in a disk group and spares of the appropriate size (same or larger) and type (same as the failed disks) are available,
the storage system automatically uses the spares to reconstruct the disk group. Disk group reconstruction does not require I/O to be
stopped, so volumes can continue to be used while reconstruction is in progress.
If no spares are available, reconstruction does not start automatically. The copyback process starts when the failed disk is replaced. If you
have configured the dynamic spares feature through the CLI, reconstruction will automatically start for disk groups. With dynamic spares
enabled, if a disk fails and you replace it with a compatible disk, the storage system rescans the bus, finds the new disk, automatically
designates it a spare, and starts reconstructing the disk group. See About spares.
For virtual storage, reconstruction of all disk groups uses a quick-rebuild feature. For more information on quick rebuild, see About quick
rebuild.
When a disk fails, its fault LED illuminates amber. When a spare is used as a reconstruction target, its activity LED blinks green. During
reconstruction, the fault LED and activity LEDs for all disks in the disk group blink. For descriptions of LED states, see the Deployment
Guide.
NOTE: Reconstruction can take hours or days to complete, depending on the disk group RAID level and size, disk speed,
utility priority, host I/O activity, and other processes running on the storage system.
At any time after disk failure, you can remove the failed disk and replace it with a new disk of the same type in the same slot.
The following steps describe the drive failure process that occurs when a drive fails in a disk group:
1. A drive fails.
2. An available compatible spare drive joins the disk group.
3. Reconstruction starts and the disk group status is VRSC/RCON.
4. The failed drive replaced by a new drive.
5. A copyback operation from the spare drive to the new drive begins. The status of the disk group is CPYBK.
6. When the copyback operation completes, the original spare drive exits the disk group and it becomes a spare drive again.
A drive may go missing from a slot because of accidental removal or bus/slot issues prevent it from being detected. The following steps
describe the drive failure process that occurs when a drive goes missing from a slot:
1. A drive goes missing from a slot.
2. An available compatible spare drive joins the disk group.
3. Reconstruction starts and the disk group status is VRSC/RCON.
4. The missing drive is placed back in its slot or the missing drive is detected and shows up. The status of the drive is LEFTOVER.
5. Metadata of the LEFTOVER drive is cleared and the drive joins the disk group.
NOTE:
If more than one drive in the disk group has a status of LEFTOVER, please contact technical support before
proceeding with any action.
6. A copyback operation from the spare drive to the drive that joined the disk group begins. The status of the disk group is CPYBK.
7. When the copyback operation completes, the original spare drive exits the disk group and it becomes a spare drive again.
About quick rebuild
Quick rebuild is a method for reconstructing a virtual disk group that is no longer fault-tolerant after a disk failure. This method takes
advantage of virtual storage knowledge of where user data is written to rebuild only the data stripes that contain user data.
Typically, storage is only partially allocated to volumes so the quick-rebuild process completes significantly faster than a standard RAID
rebuild. Data stripes that have not been allocated to user data are scrubbed in the background, using a lightweight process that allows
future data allocations to be more efficient.
After a quick rebuild, scrub starts on the disk group within a few minutes.
About performance statistics
You can view current or historical performance statistics for components of the storage system.
Current performance statistics for disks, disk groups, pools, tiers, host ports, controllers, and volumes are displayed in tabular format.
Current statistics show the current performance and are sampled immediately upon request.
Getting started
27