1PMSTAT(1) General Commands Manual PMSTAT(1)
2
6 pcp-vmstat, pmstat - high-level system performance overview
7
9 pcp [pcp options] vmstat [interval [samples]]
10 pmstat [-gLlPxz?] [-a archive] [-A align] [-h host] [-H file] [-n
12 pmnsfile] [-O offset] [-p port] [-s samples] [-S starttime] [-t inter‐
13 val] [-T endtime] [-Z timezone]
14 pcp-vmstat ...
16
18 pmstat provides a one line summary of system performance every interval
19 unit of time (the default is 5 seconds). pmstat is intended to monitor
20 system performance at the highest level, after which other tools may be
21 used to examine subsystems in which potential performance problems may
22 be observed in greater detail.
23 pcp-vmstat is a simple wrapper for use with the pcp(1) command, provid‐
25 ing a more familiar command line format for some users. It also
26 enables the extended reporting option by default, see the -x option
27 below.
28 Multiple hosts may be monitored by supplying more than one host with
30 multiple -h flags (for live monitoring) or by providing a name of the
31 hostlist file, where each line contain one host name, with -H, or mul‐
32 tiple -a flags (for retrospective monitoring from sets of archives).
33 By default, pmstat fetches metrics by connecting to the Performance
35 Metrics Collector Daemon (PMCD) on the local host. If the -L option is
36 specified, then pmcd(1) is bypassed, and metrics are fetched from PMDAs
37 on the local host using the standalone PM_CONTEXT_LOCAL variant of
38 pmNewContext(3). When the -h option is specified, pmstat connects to
39 the pmcd(1) on host and fetches metrics from there. As mentioned
40 above, multiple hosts may be monitored by supplying multiple -h flags.
41 Alternatively, if the -a option is used, the metrics are retrieved from
43 the Performance Co-Pilot archive log files identified by archive, which
44 is a comma-separated list of names, each of which may be the base name
45 of an archive or the name of a directory containing one or more ar‐
46 chives. Multiple sets of archives may be replayed by supplying multi‐
47 ple -a flags. When the -a flag is used, the -P flag may also be used
48 to pause the output after each interval.
49 Standalone mode can only connect to the local host, using a set of ar‐
51 chives implies a host name, and nominating a host precludes using an
52 archive, so the options -L, -a and -h are mutually exclusive.
53 pmstat may relinquish its own timing control, and operate under the
55 control of a pmtime(1) process that uses a GUI dialog to provide timing
56 control. In this case, either the -g option should be used to start
57 pmstat as the sole client of a new pmtime(1) instance, or -p should be
58 used to attach pmstat to an existing pmtime(1) instance via the IPC
59 channel identified by the port argument.
60 The -S, -T, -O and -A options may be used to define a time window to
62 restrict the samples retrieved, set an initial origin within the time
63 window, or specify a ``natural'' alignment of the sample times; refer
64 to PCPIntro(1) for a complete description of these options.
65
67 The available command line options are:
68 -a archive, --archive=archive
70 Performance metric values are retrieved from the set of Perfor‐
71 mance Co-Pilot (PCP) archive log files identified by the archive
72 argument, which is a comma-separated list of names, each of which
73 may be the base name of an archive or the name of a directory con‐
74 taining one or more archives.
75 -A align, --align=align
77 Force the initial sample to be aligned on the boundary of a natu‐
78 ral time unit align. Refer to PCPIntro(1) for a complete descrip‐
79 tion of the syntax for align.
80 -g, --guimode
82 Start pmstat as the client of a new pmtime(1) server process for
83 replay of archived performance data using the pmtime(1) graphical
84 user interface.
85 -h host, --host=host
87 Fetch performance metrics from pmcd(1) on host, rather than from
88 the default localhost.
89 -H path, --hostsfile=path
91 Specify the path to a file containing a set of hostnames where
92 pmcd(1) is running , rather than using the default localhost.
93 -K spec, --spec-local=spec
95 When fetching metrics from a local context (see -L), the -K option
96 may be used to control the DSO PMDAs that should be made accessi‐
97 ble. The spec argument conforms to the syntax described in
98 pmSpecLocalPMDA(3). More than one -K option may be used.
99 -l, --suffix
101 Prints the last 7 characters of a hostname in summaries involving
102 more than one host (when more than one -h option has been speci‐
103 fied on the command line).
104 -L, --local-PMDA
106 Use a local context to collect metrics from DSO PMDAs on the local
107 host without PMCD. See also -K.
108 -n pmnsfile, --namespace=pmnsfile
110 Load an alternative Performance Metrics Name Space (PMNS(5)) from
111 the file pmnsfile.
112 -O origin, --origin=origin
114 When reporting archived metrics, start reporting at origin within
115 the time window (see -S and -T). Refer to PCPIntro(1) for a com‐
116 plete description of the syntax for origin.
117 -p port, --guiport=port
119 Attach pmstat to an existing pmtime(1) time control process
120 instance via the IPC channel identified by the port argument.
121 This option is normally only used by other tools, e.g.
122 pmchart(1), when they launch pmstat with synchronized time con‐
123 trol.
124 -s samples, --samples=samples
126 The samples option defines the number of samples to be retrieved
127 and reported. If samples is 0 or -s is not specified, pmstat will
128 sample and report continuously - this is the default behavior.
129 -S starttime, --start=starttime
131 When reporting archived metrics, the report will be restricted to
132 those records logged at or after starttime. Refer to PCPIntro(1)
133 for a complete description of the syntax for starttime.
134 -t interval, --interval=interval
136 Set the reporting interval to something other than the default 1
137 second. The interval argument follows the syntax described in
138 PCPIntro(1), and in the simplest form may be an unsigned integer
139 (the implied units in this case are seconds).
140 -T endtime, --finish=endtime
142 When reporting archived metrics, the report will be restricted to
143 those records logged before or at endtime.
144 -x, --xcpu
146 The extended CPU metrics option, causes two additional CPU metrics
147 to be reported, namely wait for I/O ("wa") and virtualisation
148 steal time ("st").
149 -z, --hostzone
151 Change the reporting timezone to the local timezone at the host
152 that is the source of the performance metrics, as identified via
153 either the -h or -a options.
154 -Z timezone, --timezone=timezone
156 By default, pmtime reports the time of day according to the local
157 timezone on the system where pmstat is run. The -Z option changes
158 the timezone to timezone in the format of the environment variable
159 TZ as described in environ(7).
160
162 The output from pmstat is directed to standard output, and the columns
163 in the report are interpreted as follows:
164 loadavg The 1 minute load average.
166 memory The swpd column indicates average swap space used during the
168 interval, in Kbytes. The free column indicates average free
169 memory during the interval, in Kbytes. The buff column indi‐
170 cates average buffer memory in use during the interval, in
171 Kbytes. The cache column indicates average cached memory in
172 use during the interval, in Kbytes.
173 If the values become large, they are reported as Mbytes (m
175 suffix) or Gbytes (g suffix).
176 swap The metrics in this area of the kernel instrumentation are of
178 varying value. We try to report the average number of pages
179 that are paged in (pi) and out (po) per second during the
180 interval. If the corresponding page swapping metrics are
181 unavailable, we report the average rate per second of swap
182 operations in (si) and out (so) during the interval. It is
183 normal for the ``in'' values to be non-zero, but the system
184 is suffering memory stress if the ``out'' values are non-zero
185 over an extended period.
186 If the values become large, they are reported as thousands of
188 operations per second (K suffix) or millions of operations
189 per second (M suffix).
190 io The bi and bo columns indicate the average rate per second of
192 block input and block output operations (respectfully) during
193 the interval. Unless all file systems have a 1 Kbyte block
194 size, these rates do not directly indicate Kbytes trans‐
195 ferred.
196 If the values become large, they are reported as thousands of
198 operations per second (K suffix) or millions of operations
199 per second (M suffix).
200 system Interrupt rate (in) and context switch rate (cs). Rates are
202 expressed as average operations per second during the inter‐
203 val. Note that the interrupt rate is normally at least HZ
204 (the clock interrupt rate, usually 100) interrupts per sec‐
205 ond.
206 If the values become large, they are reported as thousands of
208 operations per second (K suffix) or millions of operations
209 per second (M suffix).
210 cpu Percentage of CPU time spent executing user and "nice user"
212 code (us), system and interrupt processing code (sy), idle
213 loop (id).
214 If any values for the associated performance metrics are unavailable,
216 the value appears as ``?'' in the output.
217
219 All are generated on standard error and are intended to be self-
220 explanatory.
221
223 $PCP_VAR_DIR/pmns/*
224 default PMNS specification files
225 $PCP_VAR_DIR/config/pmlogger/config.pmstat
227 pmlogger(1) configuration for creating an archive suitable for
228 replay with pmstat
229
231 Environment variables with the prefix PCP_ are used to parameterize the
232 file and directory names used by PCP. On each installation, the file
233 /etc/pcp.conf contains the local values for these variables. The
234 $PCP_CONF variable may be used to specify an alternative configuration
235 file, as described in pcp.conf(5).
236 For environment variables affecting PCP tools, see pmGetOptions(3).
238
240 PCPIntro(1), pmclient(1), pmtime(1), PMAPI(3), pmNewContext(3),
241 pcp.conf(5), pcp.env(5) and PMNS(5).
242Performance Co-Pilot PCP PMSTAT(1)