10400455-002
©2008-14 Overland Storage, Inc.
86
SnapScale/RAINcloudOS 4.1 Administrator’s Guide
5 – Storage Options
Peer Sets and Recovery
Though data on peer sets is served indirectly by the unified cluster storage space, access to
files stored on a given peer set is dependent on the health of that peer set. When a drive in a
peer set fails, data is served from the remaining peer set member drives. If there is a spare
reserved for the cluster that does not exist on the same node as another active member of the
peer set and is not smaller than other members, the peer set can claim the drive and rebuild
the data (using the integrated RapidRebuild feature) onto that spare without administrator
intervention.
If a peer set is missing one drive but at least one other drive is available, the peer set continues
to be accessible but is in degraded mode. This table shows the different peer set statuses:
Peer Set Status
Failure Type
Data Availability
OK
The peer set drives are healthy and
connected.
Data is fully available for read and
write.
RapidRebuild
Spare made available to rebuild the peer
set using RapidRebuild.
Data is fully available for read and
write.
Degraded
One drive missing from the peer set
Data is fully available for read and
write.
Degraded –
Cannot repair;
no spares
The peer set cannot be repaired because
there are no spare drives.
Data is fully available for read and
write.
Degraded –
Cannot repair;
spares too small
The peer set cannot be repaired because
all eligible spares are too small.
Data is fully available for read and
write.
