DVS SpycerBox Ultra, Hardware Manual

Page: 42 / 72

5-2
SpycerBox Ultra Hardware Guide
5.1 Hard Disk Maintenance
This section deals with the possible event of a hard disk failure. First,
some introductory information about RAID are provided. After that fol-
low descriptions how to act when a hard disk fails.
5.1.1 Introduction to RAID
In a system where huge amounts of data are processed, large storage
capacities combined with high data throughputs are mandatory. To
provide both at the same time it is common practice in the area of video
and digital film to configure several hard disks together into a stripe set
or RAID. The IT world has defined several ’levels’ of RAID, most of
them providing some kind of data protection.
The RAID feature makes the DVS system tolerant of disk failures. Even
with a broken disk operations can still be continued and, once the failed
disk has been replaced, the missing data can be recovered easily. The
data protection is provided by RAID controllers installed inside the sys-
tem. Each controller independently administers the data protection for
the set of hard disks that is connected to it.
The system disk array for the system/metadata hard disks comprises
four disks. They are connected to a RAID controller that applies a
RAID 1 to the data, meaning that they are protected by a mirroring of
the data between two disks. If one of the hard disks fails, the missing
data can be restored with the mirrored information stored on the other
disk.
The main storage of the DVS system comprises up to 24 hard disks.
These are connected to three RAID controllers which connect a set of
eight hard disks each and apply a RAID 5 to the data. With this, the
data is striped across these hard disks during write procedures. At the
same time the information necessary to rebuild a failed hard disk (parity
information) is generated and written across the disks as well. With the
parity information written, one hard disk per disk set (RAID pack) can
fail and your data will still be recoverable due to the information stored
on the other disks.
You can identify the hard disks that are connected to a RAID controller
easily at the DVS system because every three rows represents one disk
set/RAID pack:
If the second disk fails in the meantime, the data will be unre-
coverable.
If a second disk within the same disk set fails in the meantime,
the data will be unrecoverable.