Rebuilding Units - 3Ware 9650SE-2LPB-10 User Manual

Rebuilding Units

www.3ware.com
Notes:
If a unit that requires initialization has not previously been initialized, selecting Verify
Unit starts initialization. This is because fault-tolerant units cannot be verified until
after they are initialized.
If the unit is already in a state of rebuild, initialization, or verification, the unit cannot
be verified in 3BM. You must boot the system and let the task finish in the
background.
Rebuilding is the process of generating data on a new drive after it is put into
service to replace a failed drive in a fault tolerant unit.
If a hot spare is specified and a redundant unit degrades, it will be used to
automatically replace the failed drive in the redundant unit without
intervention on your part. The rebuild process will automatically be launched
as a background process at the next scheduled time. If scheduling is turned
off, the rebuild process will start almost immediately (within a couple of
minutes). If 3DM is running and E-mail notification is enabled, an event
notification will be sent to specified users when the unit degrades and again
when the rebuild process is complete.
If the Auto Rebuild policy is enabled (see "Setting the Auto Rebuild Policy"
on page 90), the firmware will attempt to rebuild a degraded unit with an
available drive or a failed drive.
If desired, you can manually replace the drive, rescan the controller, and start
the rebuild process. Manual rebuilds can be started from either 3BM or 3DM,
although the rebuild itself only happens when the operating system is running.
Rebuilds on multiple units can take place simultaneously.
If multiple drives are faulted in a RAID 10 configuration, the drives are
rebuilt simultaneously. In a 4-drive RAID 10 configuration, up to two drives
can be rebuilt. In a 6-drive configuration, up to three drives can be rebuilt. In
an 8-drive configuration, up to four drives can be rebuilt. In a 12-drive
configuration, up to six drives can be rebuilt.
:
Note
If both drives in a RAID 10 mirrored set are faulted, the data is not
recoverable. Up to half of the drives in a RAID 10 unit can become defective and
still have the user data retained, as long as the failed drives are only half of each
mirrored pair.
A RAID 5 unit can have one drive fail before becoming inoperable.
A RAID 50 unit can sustain multiple drive failures, as long there is only one
failed drive in each RAID 5 set.
Background Tasks
155