Adaptec 3200S User Guide - Page 98

parity data can be generated more efficiently. This results in better, write performance. However

Page 98 highlights

Storage Manager Cache Misses Read-Ahead Hits Write-Backs WriteThroughs The total number of sectors that were not accessed from the controller cache, thus were read directly from the disk. The number of the cache hits for data read requests satisfied by data held in the cache from previous disk read-ahead operations. The number of sectors written to disk that were held in the controller cache and written some time after the host Write command reported as completed. The number of sectors written directly to disk before the Write command ended. Environments with a large number of sequential reads should generate a high number of Read-Ahead Hits relative to Total Sectors. These hits reduce the number of seek operations and increase performance. The Read-Ahead Hits count can be increased by adding more cache memory to the controller. A high percentage of 4-KB or smaller I/O operations also indicates that increasing the controller cache would be beneficial. If cache hits are low, adding more cache RAM can also increase the hit count. Systems with a large number of disk writes also derive significant performance benefits from the controller cache. For RAID 0 and RAID 5 arrays, the default stripe size is set for optimal performance in most environments. However, some specific environments can benefit from smaller or larger stripes. For example, when a RAID 5 write operation accesses all drives, many stripe crossings occur. By using a larger stripe size, the RAID 5 parity data can be generated more efficiently. This results in better write performance. However, if the write crosses one or more stripes but does not involve all of the drives, performance is reduced by the larger stripe size. You should select the stripe size relative to both the I/O segment size and the number of drives in the array, so that most I/O operations either s Do not cross stripes and involve only a single drive, or s Cross many stripes and involve all the drives in the array. 3-53

  • 1
  • 2
  • 3
  • 4
  • 5
  • 6
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12
  • 13
  • 14
  • 15
  • 16
  • 17
  • 18
  • 19
  • 20
  • 21
  • 22
  • 23
  • 24
  • 25
  • 26
  • 27
  • 28
  • 29
  • 30
  • 31
  • 32
  • 33
  • 34
  • 35
  • 36
  • 37
  • 38
  • 39
  • 40
  • 41
  • 42
  • 43
  • 44
  • 45
  • 46
  • 47
  • 48
  • 49
  • 50
  • 51
  • 52
  • 53
  • 54
  • 55
  • 56
  • 57
  • 58
  • 59
  • 60
  • 61
  • 62
  • 63
  • 64
  • 65
  • 66
  • 67
  • 68
  • 69
  • 70
  • 71
  • 72
  • 73
  • 74
  • 75
  • 76
  • 77
  • 78
  • 79
  • 80
  • 81
  • 82
  • 83
  • 84
  • 85
  • 86
  • 87
  • 88
  • 89
  • 90
  • 91
  • 92
  • 93
  • 94
  • 95
  • 96
  • 97
  • 98
  • 99
  • 100
  • 101
  • 102
  • 103
  • 104
  • 105
  • 106
  • 107
  • 108
  • 109
  • 110
  • 111
  • 112
  • 113
  • 114
  • 115
  • 116
  • 117
  • 118
  • 119
  • 120
  • 121
  • 122
  • 123
  • 124
  • 125
  • 126
  • 127
  • 128
  • 129
  • 130
  • 131
  • 132
  • 133
  • 134
  • 135

3-53
Storage Manager
Environments with a large number of sequential reads should
generate a high number of Read-Ahead Hits relative to Total Sectors.
These hits reduce the number of seek operations and increase
performance. The Read-Ahead Hits count can be increased by
adding more cache memory to the controller. A high percentage of
4-KB or smaller I/O operations also indicates that increasing the
controller cache would be beneficial.
If cache hits are low, adding more cache RAM can also increase the
hit count. Systems with a large number of disk writes also derive
significant performance benefits from the controller cache.
For RAID 0 and RAID 5 arrays, the default stripe size is set for
optimal performance in most environments. However, some specific
environments can benefit from smaller or larger stripes. For
example, when a RAID 5 write operation accesses all drives, many
stripe crossings occur. By using a larger stripe size, the RAID 5
parity data can be generated more efficiently. This results in better
write performance. However, if the write crosses one or more stripes
but does not involve all of the drives, performance is reduced by the
larger stripe size.
You should select the stripe size relative to both the I/O segment
size and the number of drives in the array, so that most I/O
operations either
Do
not
cross stripes and involve only a single drive, or
Cross many stripes and involve all the drives in the array.
Cache Misses
The total number of sectors that were not accessed
from the controller cache, thus were read directly
from the disk.
Read-Ahead
Hits
The number of the cache hits for data read requests
satisfied by data held in the cache from previous
disk read-ahead operations.
Write-Backs
The number of sectors written to disk that were held
in the controller cache and written some time after
the host Write command reported as completed.
Write-
Throughs
The number of sectors written directly to disk before
the Write command ended.