1PCP-IOSTAT(1) General Commands Manual PCP-IOSTAT(1)
2
6 pmiostat, pcp-iostat - performance metrics i/o statistics tool
7
9 pcp iostat [-A align --align=TIME] [-a archive --archive=FILE] [-G
10 method --aggregate=method] [-h host --host=HOST] [-O offset --ori‐
11 gin=TIME] [-S starttime --start=TIME] [-s samples --samples=N] [-T end‐
12 time --finish=TIME] [-t interval --interval=DELTA] [-P precision --pre‐
13 cision=N] [-R pattern --regex=pattern] [-u --no-interpolate] [-Z time‐
14 zone --timezone=TZ] [-z --hostzone] [-? --help] [-x [dm][,t][,h][,noi‐
15 dle]]
16
18 pcp-iostat reports I/O statistics for scsi devices (by default) or
19 device-mapper devices (if the -x dm option is specified). By default
20 pcp-iostat reports live data for the local host but can also report for
21 a remote host (-h) or from a previously captured PCP archive (-a).
22 The -S, -T, -O and -A options may be used to define a time window to
24 restrict the samples retrieved, set an initial origin within the time
25 window, or specify a ``natural'' alignment of the sample times; refer
26 to PCPIntro(1) for a complete description of these options.
27 The other options which control the source, timing and layout of the
29 information reported by pcp-iostat are as follows:
30 -a Performance metric values are retrieved from the Performance Co-
32 Pilot (PCP) archive log files identified by the argument archive,
33 which is a comma-separated list of names, each of which may be the
34 base name of an archive or the name of a directory containing one
35 or more archives. See also -u.
36 -G Specifies that statistics for device names matching the regular
38 expression specified with the -R regex option should be aggregated
39 according to method. Note this is aggregation based on matching
40 device names (not temporal aggregation). When -G is used, the
41 device name column is reported as method(regex), e.g. if -G sum
42 -R 'sd(a|b)$' is specified, the device column will be
43 sum(sd(a|b)$) and summed statistics for sda and sdb will be
44 reported in the remaining columns. If -G is specified but -R is
45 not specified, then the default regex is .*, i.e. matching all
46 device names. If method is sum then the statistics are summed.
47 This includes the %util column, which may therefore exceed 100% if
48 more than one device name matches. If method is avg then the sta‐
49 tistics are summed and then averaged by dividing by the number of
50 matching device names. If method is min or max, the minimum or
51 maximum statistics for matching devices are reported, respect‐
52 fully.
53 -h Current performance metric values are retrieved from the nominated
55 host machine.
56 -s The argument samples defines the number of samples to be retrieved
58 and reported. If samples is 0 or -s is not specified, pcp-iostat
59 will sample and report continuously (in real time mode) or until
60 the end of the set of PCP archives (in archive mode).
61 -t The default update interval may be set to something other than the
63 default 1 second. The interval argument follows the syntax
64 described in PCPIntro(1), and in the simplest form may be an
65 unsigned integer (the implied units in this case are seconds).
66 The -t option is particularly useful when replaying large sets of
67 archives (-a option) that span several hours or even days. In
68 this case specifying a large interval (e.g. 1h for 1 hour) will
69 reduce the volume of data reported and the i/o statistics will be
70 averaged (interpolated) over the reporting interval (unless the -u
71 option is specified, see below).
72 -R This restricts the report to device names matching regex. The
74 regex pattern is searched as a perl style regular expression, and
75 will match any portion of a device name. e.g. '^sd[a-zA-Z]+' will
76 match all device names starting with 'sd' followed by one or more
77 alphabetic characters. e.g. '^sd(a|b)$' will only match 'sda' and
78 'sdb'. e.g. 'sda$' will match 'sda' but not 'sdab'. See also the
79 -G option for aggregation options.
80 -P This indicates the number of decimals to print. The default preci‐
82 sion N may be set to something other than the default 2 decimals.
83 Note that the avgrq-sz and avgqu-sz fields are always reported
84 with N+1 decimals of precision. These fields typically have val‐
85 ues less than 1.0.
86 -u When replaying a set of archives, by default values are reported
88 according to the selected sample interval (-t option), not accord‐
89 ing to the actual record intervals in the set of archives. To
90 this effect PCP interpolates the values to be reported based on
91 the records in the set of archives, and is particularly useful
92 when the -t option is used to replay a set of archives with a
93 longer sampling interval than the underlying interval the set of
94 archives was originally recorded with. With the -u option, unin‐
95 terpolated reporting is enabled - every value is reported accord‐
96 ing to the native recording interval in the set of archives. When
97 the -u option is specified, the -t option makes no sense and is
98 incompatible because the replay interval is always the same as the
99 recording interval in the set of archive. In addition, -u only
100 makes sense when replaying a set of archives, see -a above, and so
101 if -u is specified then -a must also be specified.
102 -Z By default, pcp-iostat reports the time of day according to the
104 local timezone on the system where pcp-iostat is run. The -Z
105 option changes the timezone to timezone in the format of the envi‐
106 ronment variable TZ as described in environ(7).
107 -z Change the reporting timezone to the local timezone at the host
109 that is the source of the performance metrics, as identified via
110 either the -h or -a options. When replaying a PCP archive that
111 was captured in a foreign timezone, the -z option would almost
112 always be used (the default reporting timezone is the local time‐
113 zone, which may not be the same as the timezone of the PCP ar‐
114 chive).
115 -x Specifies a comma separated list of one or more extended reporting
117 options as follows:
118 dm - report statistics for device-mapper logical devices instead
119 of scsi devices,
120 t - prefix every line in the report with a timestamp in ctime(3)
121 format,
122 h - omit the heading, which is otherwise reported every 24 sam‐
123 ples,
124 noidle - Do not display statistics for idle devices.
125
127 The columns in the pcp-iostat report have the following interpretation
128 :
129 Timestamp
131 When the -x t option is specified, this column is the timestamp
132 in ctime(3) format.
133 Device Specifies the scsi device name, or if -x dm is specified, the
135 device-mapper logical device name. When -G is specified, this
136 is replaced by the aggregation method and regular expression -
137 see the -G and -R options above.
138 rrqm/s The number of read requests expressed as a rate per-second that
140 were merged during the reporting interval by the I/O scheduler.
141 wrqm/s The number of write requests expressed as a rate per-second that
143 were merged during the reporting interval by the I/O scheduler.
144 r/s The number of read requests completed by the device (after
146 merges), expressed as a rate per second during the reporting
147 interval.
148 w/s The number of write requests completed by the device (after
150 merges), expressed as a rate per second during the reporting
151 interval.
152 rkB/s The average volume of data read from the device expressed as
154 KBytes/second during the reporting interval.
155 wkB/s The average volume of data written to the device expressed as
157 KBytes/second during the reporting interval.
158 avgrq-sz
160 The average I/O request size for both reads and writes to the
161 device expressed as Kbytes during the reporting interval.
162 avgqu-sz
164 The average queue length of read and write requests to the
165 device during the reporting interval.
166 await The average time in milliseconds that read and write requests
168 were queued (and serviced) to the device during the reporting
169 interval.
170 r_await
172 The average time in milliseconds that read requests were queued
173 (and serviced) to the device during the reporting interval.
174 w_await
176 The average time in milliseconds that write requests were queued
177 (and serviced) to the device during the reporting interval.
178 %util The percentage of time during the reporting interval that the
180 device was busy processing requests. A value of 100% indicates
181 device saturation.
182
184 $PCP_VAR_DIR/pmns/*
185 default PMNS specification files
186
188 Environment variables with the prefix PCP_ are used to parameterize the
189 file and directory names used by PCP. On each installation, the file
190 /etc/pcp.conf contains the local values for these variables. The
191 $PCP_CONF variable may be used to specify an alternative configuration
192 file, as described in pcp.conf(5).
193
195 pcp(1), PCPIntro(1), iostat2pcp(1), pmcd(1), pmchart(1), pmlogger(1),
196 pcp.conf(5) and pcp.env(5).
197
199 All are generated on standard error and are intended to be self-
200 explanatory.
201Performance Co-Pilot PCP PCP-IOSTAT(1)