HP P2000 HP P2000 G3 MSA System CLI Reference Guide - Page 125

Cache write policy The cache policy setting for each volume determines when cached data is written to the disks. The ability to hold data in cache while it is being written to disk can increase storage device speed during sequential reads. • Write-back caching does not wait for data to be completely written to disk before signaling the host that the write is complete. This is the preferred setting for a fault-tolerant environment because it improves the performance of write operations and throughput. Write-back caching is enabled by default. • Write-through caching significantly impacts performance by waiting for data to be completely written to disk before signaling the host that the write is complete. Use this setting only when operating in an environment with low or no fault tolerance. You can configure the write policy to automatically change from write-back cache to write-through cache when certain environmental events occur, such as a fan failure. For details, see set auto-write-through-trigger. Cache optimization mode Before creating or modifying a volume, determine the appropriate cache optimization mode. The controller supports super-sequential optimization mode for sequential I/O and standard optimization mode for random I/O. The cache optimization mode setting for each volume optimizes the cache block size used by the controller: • For sequential optimization, the cache block size is 128 Kbyte. • For random optimization, the cache block size is 32 Kbyte. An appropriate cache block size improves performance when a particular application uses either large or small stripe sizes: • Video playback, multimedia post-production audio and video editing, and similar applications read and write large files in sequential order. • Transaction-based and database update applications read and write small files in random order. Since the cache block size works in conjunction with the default stripe size set by the cache optimization mode for each volume you create, these default stripe sizes are consistent with the cache block size setting. You can, however, specify a different stripe size for any volume at the time you create it. For more information, see create volume. Cache read-ahead size You can optimize a volume for sequential reads or streaming data by changing the amount of data read in advance after two back-to-back reads are made. Read ahead is triggered by two back-to-back accesses to consecutive logical block address (LBA) ranges. Read ahead can be forward (that is, increasing LBAs) or reverse (that is, decreasing LBAs). Increasing the read-ahead size can greatly improve performance for multiple sequential read streams. However, increasing read-ahead size will likely decrease random read performance. The default read-ahead size, which sets one chunk for the first access in a sequential read and one stripe for all subsequent accesses, works well for most users in most applications. The controllers treat non-RAID and RAID-1 vdisks internally as if they have a stripe size of 64 Kbyte, even though they are not striped. HP P2000 G3 MSA System CLI Reference Guide 125