IBM System Storage DS3500 Introduction And Implementation Manual page 76

7914DS3KPlanning_090710.fm
Figure 3-15 RAID 10. A Raid 1 array will only consist of one mirrored pair (#1 Mirrorset)
We recommend these best practises for using RAID 1/10 arrays:
Use a two drive RAID 1 array for the disks that contain your operating system. It is a good
choice, because the operating system can usually fit on one disk.
Use RAID 1 for transaction logs. Typically, the database server transaction log can fit on
one disk drive. In addition, the transaction log performs mostly sequential writes. Only
rollback operations cause reads from the transaction logs. Therefore, we can achieve a
high rate of performance by isolating the transaction log on its own RAID 1 array.
Use write caching on RAID 1/10 arrays. Because a RAID 1/10 write will not complete until
both writes have been completed on both disks in a mirrored pair, the performance of
writes can be improved through the use of a write cache. Always ensure that you use
battery-backed write cache.
The performance of RAID 10 is comparable to RAID 0 for sequential I/Os but RAID 10
provides data redundancy through disk mirroring.
Note: There are no guaranteed choices as to which type of RAID to use, because this is
very much dependent on the workload read and write activity. A good general guide might
be to consider using RAID 1 if random writes exceed about 25%, with a peak sustained I/O
rate that exceeds 50% of the storage subsystem's capacity
When comparing RAID Level 10 with RAID Level 5:
RAID 10 uses two write IOs to write a single block of data (one write IO to each drive in the
mirrored pair). RAID 5 requires two read IOs (read original data and parity) and then two
write IOs to write the same block of data. For this reason, random writes are significantly
faster on RAID 10 compared to RAID 5.
RAID 10 rebuilds take less time than RAID 5 rebuilds. If one drive fails, RAID 10 rebuilds it
by copying all the data on the mirrored drive to a replacement/hotspare drive. RAID 5
rebuilds a failed disk by merging the contents of the surviving disks in an array and writing
the result to a spare.
52
IBM System Storage DS3500: Introduction and Implementation Guide
Draft Document for Review March 28, 2011 12:24 pm