HP 3PAR StoreServ 7400 2-node HP 3PAR System Reporter 3.1.0 Software User& - Page 208

Service Time Histogram Buckets, The statpd Table

Page 208 highlights

Table 78 Common Performance Metrics for a Sample Row (continued) Metric Write KBytes/sec Total KBytes/sec Read service time (millisec) Write service time (millisec) Total service time (millisec) Read IO Size (KBytes) Write IO Size (KBytes) Total IO Size (KBytes) Queue length SQL Expression CASE WHEN d_now > 0 THEN d_wbytes * 1000.0/d_now ELSE 0 END CASE WHEN d_now > 0 THEN (d_rbytes + d_wbytes)* 1000.0/d_now ELSE 0 END CASE WHEN d_rcount > 0 THEN d_rticks / (d_rcount * 1000.0) ELSE 0 END CASE WHEN d_wcount > 0 THEN d_wticks / (d_wcount * 1000.0) ELSE 0 END CASE WHEN (d_rcount + d_wcount) > 0 THEN (d_rticks + d_wticks) / ((d_rcount + d_wcount) * 1000.0) ELSE 0 END CASE WHEN d_rcount > 0 THEN d_rbytes / (d_rcount * 1000.0) ELSE 0 END CASE WHEN d_wcount > 0 THEN d_wbytes / (d_wcount * 1000.0) ELSE 0 END CASE WHEN (d_rcount + d_wcount) > 0 THEN (d_rbytes + d_wbytes) / ((d_rcount + d_wcount) * 1000.0) ELSE 0 END CASE WHEN qlen > 0 THEN qlen ELSE 0 END The check for qlen > 0 is because qlen < 0 indicates that qlen is invalid for that sample. Busy% AVG(CASE WHEN ((d_now > 0) AND (d_busy < busy)) THEN (d_busy * 100.0/d_now) ELSE 0 END) Service Time Histogram Buckets The service time range corresponding to each of the rhist and whist buckets depends on the version of the HP 3PAR OS (os_rev column in the system table, see "The system Table" (page 193)). For os_rev prior to 2.2.2, the nth bucket (rhistn and whistn) correspond to service times from the previous bucket to (2n*4.096/mhz) millisec where mhz is the mhz column in the system_3 table (see "The system Table" (page 193)). This means that service time histograms from systems with different CPU frequencies could not be correctly aggregated. Beginning with the 2.2.2 release, the buckets were made independent of the CPU frequency, and the nth bucket (rhistn and whistn) corresponds to service times from the previous bucket to (2n-7) millisec. IO Size Histogram Buckets There are 16 IO Size buckets correspond IO sizes 512 bytes to 16 MiBytes. The nth bucket holds the count for IO Sizes from the previous bucket's max size to 512 * 2n bytes. The statpd Table The statpd__4 table holds physical disk (PD) performance data samples. Table 79 statpd__4 Table Column tsecs sys_uid pdid port_n Type Comment INTEGER NOT NULL Sample time in secs. This is the Unix time and is the number of seconds since 0:00:00 Jan 1, 1970 GMT. INTEGER NOT NULL A unique identifier for the system instance refers to the uniq_id for the system instance in the system_3 table. INTEGER NOT NULL The physical disk ID. INTEGER NOT NULL The node that the disk's primary port is connected to (the "n" part of n:s:p). 208 Database Schema

  • 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
  • 136
  • 137
  • 138
  • 139
  • 140
  • 141
  • 142
  • 143
  • 144
  • 145
  • 146
  • 147
  • 148
  • 149
  • 150
  • 151
  • 152
  • 153
  • 154
  • 155
  • 156
  • 157
  • 158
  • 159
  • 160
  • 161
  • 162
  • 163
  • 164
  • 165
  • 166
  • 167
  • 168
  • 169
  • 170
  • 171
  • 172
  • 173
  • 174
  • 175
  • 176
  • 177
  • 178
  • 179
  • 180
  • 181
  • 182
  • 183
  • 184
  • 185
  • 186
  • 187
  • 188
  • 189
  • 190
  • 191
  • 192
  • 193
  • 194
  • 195
  • 196
  • 197
  • 198
  • 199
  • 200
  • 201
  • 202
  • 203
  • 204
  • 205
  • 206
  • 207
  • 208
  • 209
  • 210
  • 211
  • 212
  • 213
  • 214
  • 215
  • 216
  • 217
  • 218
  • 219
  • 220
  • 221
  • 222
  • 223
  • 224
  • 225
  • 226
  • 227
  • 228
  • 229
  • 230
  • 231

Table 78 Common Performance Metrics for a Sample Row
(continued)
SQL Expression
Metric
CASE WHEN d_now > 0 THEN d_wbytes * 1000.0/d_now ELSE 0 END
Write KBytes/sec
CASE WHEN d_now > 0 THEN (d_rbytes + d_wbytes)* 1000.0/d_now ELSE
0 END
Total KBytes/sec
CASE WHEN d_rcount > 0 THEN d_rticks / (d_rcount * 1000.0) ELSE 0
END
Read service time
(millisec)
CASE WHEN d_wcount > 0 THEN d_wticks / (d_wcount * 1000.0) ELSE 0
END
Write service time
(millisec)
CASE WHEN (d_rcount + d_wcount) > 0 THEN (d_rticks + d_wticks) /
((d_rcount + d_wcount) * 1000.0) ELSE 0 END
Total service time
(millisec)
CASE WHEN d_rcount > 0 THEN d_rbytes / (d_rcount * 1000.0) ELSE 0
END
Read IO Size (KBytes)
CASE WHEN d_wcount > 0 THEN d_wbytes / (d_wcount * 1000.0) ELSE 0
END
Write IO Size (KBytes)
CASE WHEN (d_rcount + d_wcount) > 0 THEN (d_rbytes + d_wbytes) /
((d_rcount + d_wcount) * 1000.0) ELSE 0 END
Total IO Size (KBytes)
CASE WHEN qlen > 0 THEN qlen ELSE 0 END
Queue length
The check for qlen > 0 is because qlen < 0 indicates that qlen is invalid for that sample.
AVG(CASE WHEN ((d_now > 0) AND (d_busy < busy)) THEN (d_busy *
100.0/d_now) ELSE 0 END)
Busy%
Service Time Histogram Buckets
The service time range corresponding to each of the rhist and whist buckets depends on the version
of the HP 3PAR OS (
os_rev
column in the system table, see
“The system Table” (page 193)
). For
os_rev
prior to 2.2.2, the nth bucket (
rhistn
and
whistn
) correspond to service times from the
previous bucket to (2n*4.096/mhz) millisec where
mhz
is the mhz column in the
system_3
table
(see
“The system Table” (page 193)
). This means that service time histograms from systems with
different CPU frequencies could not be correctly aggregated. Beginning with the 2.2.2 release,
the buckets were made independent of the CPU frequency, and the nth bucket (
rhistn
and
whistn
)
corresponds to service times from the previous bucket to (2n-7) millisec.
IO Size Histogram Buckets
There are 16 IO Size buckets correspond IO sizes 512 bytes to 16 MiBytes. The nth bucket holds
the count for IO Sizes from the previous bucket’s max size to 512 * 2n bytes.
The statpd Table
The
statpd_<res>_4
table holds physical disk (PD) performance data samples.
Table 79 statpd_<res>_4 Table
Comment
Type
Column
Sample time in secs. This is the Unix time and is the number of seconds
since 0:00:00 Jan 1, 1970 GMT.
INTEGER NOT NULL
tsecs
A unique identifier for the system instance refers to the uniq_id for the
system instance in the system_3 table.
INTEGER NOT NULL
sys_uid
The physical disk ID.
INTEGER NOT NULL
pdid
The node that the disk’s primary port is connected to (the “n” part of
n:s:p).
INTEGER NOT NULL
port_n
208
Database Schema