1PCP-ATOP(1) General Commands Manual PCP-ATOP(1)
2
3
4
6 pcp-atop - Advanced System and Process Monitor
7
9 Interactive Usage:
10
11 pcp [pcp options] atop [-aAcCdDfFgGmMnNopRsuvxy1] [-L linelen] [-Pla‐
12 bel[,label]...] [interval [samples]]
13
14 Writing and reading PCP archive folios:
15
16 pcp atop -w folio [-a] [-S] [interval [samples]]
17 pcp atop -r folio [-AcCdDfFgGmMnNopRsuvxy1] [-b hh:mm] [-e hh:mm] [-L
18 linelen] [-Plabel[,label]...] [interval [samples]]
19
21 The program pcp-atop is an interactive monitor to view various aspects
22 of load on a system. It shows the occupation of the most critical
23 hardware resources (from a performance point of view) on system level,
24 i.e. cpu, memory, disk and network.
25 It also shows which processes are responsible for the indicated load
26 with respect to cpu and memory load on process level. Disk load is
27 shown per process if "storage accounting" is active in the kernel.
28
29 Every interval (default: 10 seconds) information is shown about the
30 resource occupation on system level (cpu, memory, disks and network
31 layers), followed by a list of processes which have been active during
32 the last interval (note that all processes that were unchanged during
33 the last interval are not shown, unless the key 'a' has been pressed or
34 unless sorting on memory occupation is done). If the list of active
35 processes does not entirely fit on the screen, only the top of the list
36 is shown (sorted in order of activity).
37 The intervals are repeated till the number of samples (specified as
38 command argument) is reached, or till the key 'q' is pressed in inter‐
39 active mode.
40
41 When invoked via the pcp(1) command, the PCPIntro(1) options -h/--host,
42 -a/--archive, -O/--origin, -s/--samples, -t/--interval, -Z/--timezone
43 and several other pcp options become indirectly available. The long
44 option form of these is directly available. Additionally, the --hot‐
45 proc option can be used to request the per-process PCP metrics be used
46 instead of the default proc metrics from pmdaproc(1).
47
48 When pcp-atop is started, it checks whether the standard output channel
49 is connected to a screen, or to a file/pipe. In the first case it pro‐
50 duces screen control codes (via the ncurses library) and behaves inter‐
51 actively; in the second case it produces flat ASCII-output.
52
53 In interactive mode, the output of pcp-atop scales dynamically to the
54 current dimensions of the screen/window.
55 If the window is resized horizontally, columns will be added or removed
56 automatically. For this purpose, every column has a particular weight.
57 The columns with the highest weights that fit within the current width
58 will be shown.
59 If the window is resized vertically, lines of the process/thread list
60 will be added or removed automatically.
61
62 Furthermore in interactive mode the output of pcp-atop can be con‐
63 trolled by pressing particular keys. However it is also possible to
64 specify such key as flag on the command line. In that case pcp-atop
65 switches to the indicated mode on beforehand; this mode can be modified
66 again interactively. Specifying such key as flag is especially useful
67 when running pcp-atop with output to a pipe or file (non-interac‐
68 tively). These flags are the same as the keys that can be pressed in
69 interactive mode (see section INTERACTIVE COMMANDS).
70 Additional flags are available to support storage of pcp-atop data in
71 PCP archive format (see section PCP DATA STORAGE).
72
74 For the resource consumption on system level, pcp-atop uses colors to
75 indicate that a critical occupation percentage has been (almost)
76 reached. A critical occupation percentage means that is likely that
77 this load causes a noticeable negative performance influence for appli‐
78 cations using this resource. The critical percentage depends on the
79 type of resource: e.g. the performance influence of a disk with a busy
80 percentage of 80% might be more noticeable for applications/user than a
81 CPU with a busy percentage of 90%.
82 Currently pcp-atop uses the following default values to calculate a
83 weighted percentage per resource:
84
85 Processor
86 A busy percentage of 90% or higher is considered `critical'.
87
88 Disk
89 A busy percentage of 70% or higher is considered `critical'.
90
91 Network
92 A busy percentage of 90% or higher for the load of an interface is
93 considered `critical'.
94
95 Memory
96 An occupation percentage of 90% is considered `critical'. Notice
97 that this occupation percentage is the accumulated memory consump‐
98 tion of the kernel (including slab) and all processes; the memory
99 for the page cache (`cache' and `buff' in the MEM-line) and the
100 reclaimable part of the slab (`slrec`) is not implied!
101 If the number of pages swapped out (`swout' in the PAG-line) is
102 larger than 10 per second, the memory resource is considered
103 `critical'. A value of at least 1 per second is considered
104 `almost critical'.
105 If the committed virtual memory exceeds the limit (`vmcom' and
106 `vmlim' in the SWP-line), the SWP-line is colored due to overcom‐
107 mitting the system.
108
109 Swap
110 An occupation percentage of 80% is considered `critical' because
111 swap space might be completely exhausted in the near future; it is
112 not critical from a performance point-of-view.
113
114 These default values can be modified in the configuration file (see
115 separate man-page of pcp-atoprc(5)).
116
117 When a resource exceeds its critical occupation percentage, the con‐
118 cerning values in the screen line are colored red by default.
119 When a resource exceeded (default) 80% of its critical percentage (so
120 it is almost critical), the concerning values in the screen line are
121 colored cyan by default. This `almost critical percentage' (one value
122 for all resources) can be modified in the configuration file (see sepa‐
123 rate man-page of pcp-atoprc(5)).
124 The default colors red and cyan can be modified in the configuration
125 file as well (see separate man-page of pcp-atoprc(5)).
126
127 With the key 'x' (or flag -x), the use of colors can be suppressed.
128
130 GPU statistics can be gathered by pmdanvidia(1) which is a separate
131 data collection daemon process. It gathers cumulative utilization
132 counters of every Nvidia GPU in the system, as well as utilization
133 counters of every process that uses a GPU. When pcp-atop notices that
134 the daemon is active, it reads these GPU utilization counters with
135 every interval.
136
137 Find a description about the utilization counters in the section OUTPUT
138 DESCRIPTION.
139
141 When running pcp-atop interactively (no output redirection), keys can
142 be pressed to control the output. In general, lower case keys can be
143 used to show other information for the active processes and upper case
144 keys can be used to influence the sort order of the active
145 process/thread list.
146
147 g Show generic output (default).
148
149 Per process the following fields are shown in case of a window-
150 width of 80 positions: process-id, cpu consumption during the last
151 interval in system and user mode, the virtual and resident memory
152 growth of the process.
153
154 The subsequent columns depend on the used kernel:
155 When the kernel supports "storage accounting" (>= 2.6.20), the
156 data transfer for read/write on disk, the status and exit code are
157 shown for each process. When the kernel does not support "storage
158 accounting", the username, number of threads in the thread group,
159 the status and exit code are shown.
160 The last columns contain the state, the occupation percentage for
161 the chosen resource (default: cpu) and the process name.
162
163 When more than 80 positions are available, other information is
164 added.
165
166 m Show memory related output.
167
168 Per process the following fields are shown in case of a window-
169 width of 80 positions: process-id, minor and major memory faults,
170 size of virtual shared text, total virtual process size, total
171 resident process size, virtual and resident growth during last
172 interval, memory occupation percentage and process name.
173
174 When more than 80 positions are available, other information is
175 added.
176
177 For memory consumption, always all processes are shown (also the
178 processes that were not active during the interval).
179
180 d Show disk-related output.
181
182 When "storage accounting" is active in the kernel, the following
183 fields are shown: process-id, amount of data read from disk,
184 amount of data written to disk, amount of data that was written
185 but has been withdrawn again (WCANCL), disk occupation percentage
186 and process name.
187
188 s Show scheduling characteristics.
189
190 Per process the following fields are shown in case of a window-
191 width of 80 positions: process-id, number of threads in state
192 'running' (R), number of threads in state 'interruptible sleeping'
193 (S), number of threads in state 'uninterruptible sleeping' (D),
194 scheduling policy (normal timesharing, realtime round-robin, real‐
195 time fifo), nice value, priority, realtime priority, current pro‐
196 cessor, status, exit code, state, the occupation percentage for
197 the chosen resource and the process name.
198
199 When more than 80 positions are available, other information is
200 added.
201
202 v Show various process characteristics.
203
204 Per process the following fields are shown in case of a window-
205 width of 80 positions: process-id, user name and group, start date
206 and time, status (e.g. exit code if the process has finished),
207 state, the occupation percentage for the chosen resource and the
208 process name.
209
210 When more than 80 positions are available, other information is
211 added.
212
213 c Show the command line of the process.
214
215 Per process the following fields are shown: process-id, the occu‐
216 pation percentage for the chosen resource and the command line
217 including arguments.
218
219 e Show GPU utilization.
220
221 Per process at least the following fields are shown: process-id,
222 range of GPU numbers on which the process currently runs, GPU busy
223 percentage on all GPUs, memory busy percentage (i.e. read and
224 write accesses on memory) on all GPUs, memory occupation at the
225 moment of the sample, average memory occupation during the sample,
226 and GPU percentage.
227
228 When the pmdanvidia daemon does not run with root privileges, the
229 GPU busy percentage and the memory busy percentage are not avail‐
230 able on process level. In that case, the GPU percentage on
231 process level reflects the GPU memory occupation instead of the
232 GPU busy percentage (which is preferred).
233
234 o Show the user-defined line of the process.
235
236 In the configuration file the keyword ownprocline can be specified
237 with the description of a user-defined output-line.
238 Refer to the man-page of pcp-atoprc(5) for a detailed description.
239
240 y Show the individual threads within a process (toggle).
241
242 Single-threaded processes are still shown as one line.
243 For multi-threaded processes, one line represents the process
244 while additional lines show the activity per individual thread (in
245 a different color). Depending on the option 'a' (all or active
246 toggle), all threads are shown or only the threads that were
247 active during the last interval.
248 Whether this key is active or not can be seen in the header line.
249
250 u Show the process activity accumulated per user.
251
252 Per user the following fields are shown: number of processes
253 active or terminated during last interval (or in total if combined
254 with command `a'), accumulated cpu consumption during last inter‐
255 val in system and user mode, the current virtual and resident mem‐
256 ory space consumed by active processes (or all processes of the
257 user if combined with command `a').
258 When "storage accounting" is active in the kernel, the accumulated
259 read and write throughput on disk is shown. When the kernel mod‐
260 ule `netatop' has been installed, the number of received and sent
261 network packets are shown.
262 The last columns contain the accumulated occupation percentage for
263 the chosen resource (default: cpu) and the user name.
264
265 p Show the process activity accumulated per program (i.e. process
266 name).
267
268 Per program the following fields are shown: number of processes
269 active or terminated during last interval (or in total if combined
270 with command `a'), accumulated cpu consumption during last inter‐
271 val in system and user mode, the current virtual and resident mem‐
272 ory space consumed by active processes (or all processes of the
273 user if combined with command `a').
274 When "storage accounting" is active in the kernel, the accumulated
275 read and write throughput on disk is shown.
276 The last columns contain the accumulated occupation percentage for
277 the chosen resource (default: cpu) and the program name.
278
279 j Show the process activity accumulated per Docker container.
280
281 Per container the following fields are shown: number of processes
282 active or terminated during last interval (or in total if combined
283 with command `a'), accumulated cpu consumption during last inter‐
284 val in system and user mode, the current virtual and resident mem‐
285 ory space consumed by active processes (or all processes of the
286 user if combined with command `a').
287 When "storage accounting" is active in the kernel, the accumulated
288 read and write throughput on disk is shown.
289 The last columns contain the accumulated occupation percentage for
290 the chosen resource (default: cpu) and the Docker container id
291 (CID).
292
293 C Sort the current list in the order of cpu consumption (default).
294 The one-but-last column changes to ``CPU''.
295
296 E Sort the current list in the order of GPU utilization (preferred,
297 but only applicable when the pmdanvidia daemon runs under root
298 privileges) or the order of GPU memory occupation). The one-but-
299 last column changes to ``GPU''.
300
301 M Sort the current list in the order of resident memory consumption.
302 The one-but-last column changes to ``MEM''. In case of sorting on
303 memory, the full process list will be shown (not only the active
304 processes).
305
306 D Sort the current list in the order of disk accesses issued. The
307 one-but-last column changes to ``DSK''.
308
309 N Sort the current list in the order of network bandwidth (received
310 and transmitted). The one-but-last column changes to ``NET''.
311
312 A Sort the current list automatically in the order of the most busy
313 system resource during this interval. The one-but-last column
314 shows either ``ACPU'', ``AMEM'', ``ADSK'' or ``ANET'' (the preced‐
315 ing 'A' indicates automatic sorting-order). The most busy
316 resource is determined by comparing the weighted busy-percentages
317 of the system resources, as described earlier in the section COL‐
318 ORS.
319 This option remains valid until another sorting-order is explic‐
320 itly selected again.
321 A sorting-order for disk is only possible when "storage account‐
322 ing" is active.
323
324 Miscellaneous interactive commands:
325
326 ? Request for help information (also the key 'h' can be pressed).
327
328 V Request for version information (version number and date).
329
330 R Gather and calculate the proportional set size of processes (tog‐
331 gle). Gathering of all values that are needed to calculate the
332 PSIZE of a process is a relatively time-consuming task, so this
333 key should only be active when analyzing the resident memory con‐
334 sumption of processes.
335
336 x Suppress colors to highlight critical resources (toggle).
337 Whether this key is active or not can be seen in the header line.
338
339 z The pause key can be used to freeze the current situation in order
340 to investigate the output on the screen. While pcp-atop is
341 paused, the keys described above can be pressed to show other
342 information about the current list of processes. Whenever the
343 pause key is pressed again, pcp-atop will continue with the next
344 sample.
345
346 i Modify the interval timer (default: 10 seconds). If an interval
347 timer of 0 is entered, the interval timer is switched off. In
348 that case a new sample can only be triggered manually by pressing
349 the key 't'.
350
351 t Trigger a new sample manually. This key can be pressed if the
352 current sample should be finished before the timer has exceeded,
353 or if no timer is set at all (interval timer defined as 0). In
354 the latter case pcp-atop can be used as a stopwatch to measure the
355 load being caused by a particular application transaction, without
356 knowing on beforehand how many seconds this transaction will last.
357
358 When viewing the contents of an archive folio, this key can be
359 used to show the next sample from the folio.
360
361 T When viewing the contents of an archive folio, this key can be
362 used to show the previous sample from the folio.
363
364 b When viewing the contents of an archive folio, this key can be
365 used to move to a certain timestamp within the file (either for‐
366 ward or backward).
367
368 r Reset all counters to zero to see the system and process activity
369 since boot again.
370
371 When viewing the contents of an archive, this key can be used to
372 rewind to the beginning of the file again.
373
374 U Specify a search string for specific user names as a regular
375 expression. From now on, only (active) processes will be shown
376 from a user which matches the regular expression. The system sta‐
377 tistics are still system wide. If the Enter-key is pressed with‐
378 out specifying a name, (active) processes of all users will be
379 shown again.
380 Whether this key is active or not can be seen in the header line.
381
382 I Specify a list with one or more PIDs to be selected. From now on,
383 only processes will be shown with a PID which matches one of the
384 given list. The system statistics are still system wide. If the
385 Enter-key is pressed without specifying a PID, all (active) pro‐
386 cesses will be shown again.
387 Whether this key is active or not can be seen in the header line.
388
389 P Specify a search string for specific process names as a regular
390 expression. From now on, only processes will be shown with a name
391 which matches the regular expression. The system statistics are
392 still system wide. If the Enter-key is pressed without specifying
393 a name, all (active) processes will be shown again.
394 Whether this key is active or not can be seen in the header line.
395
396 / Specify a specific command line search string as a regular expres‐
397 sion. From now on, only processes will be shown with a command
398 line which matches the regular expression. The system statistics
399 are still system wide. If the Enter-key is pressed without speci‐
400 fying a string, all (active) processes will be shown again.
401 Whether this key is active or not can be seen in the header line.
402
403 J Specify a Docker container id of 12 (hexadecimal) characters.
404 From now on, only processes will be shown that run in that spe‐
405 cific Docker container (CID). The system statistics are still
406 system wide. If the Enter-key is pressed without specifying a
407 container id, all (active) processes will be shown again.
408 Whether this key is active or not can be seen in the header line.
409
410 S Specify search strings for specific logical volume names, specific
411 disk names and specific network interface names. All search
412 strings are interpreted as a regular expressions. From now on,
413 only those system resources are shown that match the concerning
414 regular expression. If the Enter-key is pressed without specify‐
415 ing a search string, all (active) system resources of that type
416 will be shown again.
417 Whether this key is active or not can be seen in the header line.
418
419 a The `all/active' key can be used to toggle between only show‐
420 ing/accumulating the processes that were active during the last
421 interval (default) or showing/accumulating all processes.
422 Whether this key is active or not can be seen in the header line.
423
424 G By default, pcp-atop shows/accumulates the processes that are
425 alive and the processes that are exited during the last interval.
426 With this key (toggle), showing/accumulating the processes that
427 are exited can be suppressed.
428 Whether this key is active or not can be seen in the header line.
429
430 f Show a fixed (maximum) number of header lines for system resources
431 (toggle). By default only the lines are shown about system
432 resources (CPUs, paging, logical volumes, disks, network inter‐
433 faces) that really have been active during the last interval.
434 With this key you can force pcp-atop to show lines of inactive
435 resources as well.
436 Whether this key is active or not can be seen in the header line.
437
438 F Suppress sorting of system resources (toggle). By default system
439 resources (CPUs, logical volumes, disks, network interfaces) are
440 sorted on utilization.
441 Whether this key is active or not can be seen in the header line.
442
443 1 Show relevant counters as an average per second (in the format
444 `..../s') instead of as a total during the interval (toggle).
445 Whether this key is active or not can be seen in the header line.
446
447 l Limit the number of system level lines for the counters per-cpu,
448 the active disks and the network interfaces. By default lines are
449 shown of all CPUs, disks and network interfaces which have been
450 active during the last interval. Limiting these lines can be use‐
451 ful on systems with huge number CPUs, disks or interfaces in order
452 to be able to run pcp-atop on a screen/window with e.g. only 24
453 lines.
454 For all mentioned resources the maximum number of lines can be
455 specified interactively. When using the flag -l the maximum number
456 of per-cpu lines is set to 0, the maximum number of disk lines to
457 5 and the maximum number of interface lines to 3. These values
458 can be modified again in interactive mode.
459
460 k Send a signal to an active process (a.k.a. kill a process).
461
462 q Quit the program.
463
464 PgDn Show the next page of the process/thread list.
465 With the arrow-down key the list can be scrolled downwards with
466 single lines.
467
468 ^F Show the next page of the process/thread list (forward).
469 With the arrow-down key the list can be scrolled downwards with
470 single lines.
471
472 PgUp Show the previous page of the process/thread list.
473 With the arrow-up key the list can be scrolled upwards with single
474 lines.
475
476 ^B Show the previous page of the process/thread list (backward).
477 With the arrow-up key the list can be scrolled upwards with single
478 lines.
479
480 ^L Redraw the screen.
481
483 In order to store system and process level statistics for long-term
484 analysis (e.g. to check the system load and the active processes run‐
485 ning yesterday between 3:00 and 4:00 PM), pcp-atop can store the system
486 and process level statistics in the PCP archive format, as an archive
487 folio (see mkaf(1)).
488 All information about processes and threads is stored in the archive.
489 The interval (default: 10 seconds) and number of samples (default:
490 infinite) can be passed as last arguments. Instead of the number of
491 samples, the flag -S can be used to indicate that pcp-atop should fin‐
492 ish anyhow before midnight.
493
494 A PCP archive can be read and visualized again with the -r option. The
495 argument is a comma-separated list of names, each of which may be the
496 base name of an archive or the name of a directory containing one or
497 more archives. If no argument is specified, the file
498 $PCP_LOG_DIR/pmlogger/HOST/YYYYMMDD is opened for input (where YYYYMMDD
499 are digits representing the current date, and HOST is the hostname of
500 the machine being logged). If a filename is specified in the format
501 YYYYMMDD (representing any valid date), the file $PCP_LOG_DIR/pmlog‐
502 ger/HOST/YYYYMMDD is opened. If a filename with the symbolic name y is
503 specified, yesterday's daily logfile is opened (this can be repeated so
504 'yyyy' indicates the logfile of four days ago).
505 The samples from the file can be viewed interactively by using the key
506 't' to show the next sample, the key 'T' to show the previous sample,
507 the key 'b' to branch to a particular time or the key 'r' to rewind to
508 the begin of the file.
509 When output is redirected to a file or pipe, pcp-atop prints all sam‐
510 ples in plain ASCII. The default line length is 80 characters in that
511 case; with the flag -L followed by an alternate line length, more (or
512 less) columns will be shown.
513 With the flag -b (begin time) and/or -e (end time) followed by a time
514 argument of the form HH:MM, a certain time period within the archive
515 can be selected.
516
518 The first sample shows the system level activity since boot (the
519 elapsed time in the header shows the time since boot). Note that par‐
520 ticular counters could have reached their maximum value (several times)
521 and started by zero again, so do not rely on these figures.
522
523 For every sample pcp-atop first shows the lines related to system level
524 activity. If a particular system resource has not been used during the
525 interval, the entire line related to this resource is suppressed. So
526 the number of system level lines may vary for each sample.
527 After that a list is shown of processes which have been active during
528 the last interval. This list is by default sorted on cpu consumption,
529 but this order can be changed by the keys which are previously
530 described.
531
532 If values have to be shown by pcp-atop which do not fit in the column
533 width, another format is used. If e.g. a cpu-consumption of 233216 mil‐
534 liseconds should be shown in a column width of 4 positions, it is shown
535 as `233s' (in seconds). For large memory figures, another unit is cho‐
536 sen if the value does not fit (Mb instead of Kb, Gb instead of Mb, Tb
537 instead of Gb, ...). For other values, a kind of exponent notation is
538 used (value 123456789 shown in a column of 5 positions gives 123e6).
539
541 The system level information consists of the following output lines:
542
543 PRC Process and thread level totals.
544 This line contains the total cpu time consumed in system mode
545 (`sys') and in user mode (`user'), the total number of processes
546 present at this moment (`#proc'), the total number of threads
547 present at this moment in state `running' (`#trun'), `sleeping
548 interruptible' (`#tslpi') and `sleeping uninterruptible'
549 (`#tslpu'), the number of zombie processes (`#zombie'), the number
550 of clone system calls (`clones'), and the number of processes that
551 ended during the interval (`#exit') when process accounting is
552 used. Instead of `#exit` the last column may indicate that process
553 accounting could not be activated (`no procacct`).
554 If the screen-width does not allow all of these counters, only a
555 relevant subset is shown.
556
557 CPU CPU utilization.
558 At least one line is shown for the total occupation of all CPUs
559 together.
560 In case of a multi-processor system, an additional line is shown
561 for every individual processor (with `cpu' in lower case), sorted
562 on activity. Inactive CPUs will not be shown by default. The
563 lines showing the per-cpu occupation contain the cpu number in the
564 field combined with the wait percentage.
565
566 Every line contains the percentage of cpu time spent in kernel
567 mode by all active processes (`sys'), the percentage of cpu time
568 consumed in user mode (`user') for all active processes (including
569 processes running with a nice value larger than zero), the per‐
570 centage of cpu time spent for interrupt handling (`irq') including
571 softirq, the percentage of unused cpu time while no processes were
572 waiting for disk I/O (`idle'), and the percentage of unused cpu
573 time while at least one process was waiting for disk I/O (`wait').
574 In case of per-cpu occupation, the cpu number and the wait per‐
575 centage (`w') for that cpu. The number of lines showing the per-
576 cpu occupation can be limited.
577
578 For virtual machines, the steal-percentage (`steal') shows the
579 percentage of cpu time stolen by other virtual machines running on
580 the same hardware.
581 For physical machines hosting one or more virtual machines, the
582 guest-percentage (`guest') shows the percentage of cpu time used
583 by the virtual machines. Notice that this percentage overlaps the
584 user-percentage!
585
586 When PMC performance monitoring counters are supported by the CPU
587 and the kernel (and pmdaperfevent(1) runs with root privileges),
588 the number of instructions per CPU cycle (`ipc') is shown. The
589 first sample always shows the value 'initial', because the coun‐
590 ters are just activated at the moment that pcp-atop is started.
591 When the CPU busy percentage is high and the IPC is less than 1.0,
592 it is likely that the CPU is frequently waiting for memory access
593 during instruction execution (larger CPU caches or faster memory
594 might be helpful to improve performance). When the CPU busy per‐
595 centage is high and the IPC is greater than 1.0, it is likely that
596 the CPU is instruction-bound (more/faster cores might be helpful
597 to improve performance).
598 Furthermore, per CPU the effective number of cycles (`cycl') is
599 shown. This value can reach the current CPU frequency if such CPU
600 is 100% busy. When an idle CPU is halted, the number of effective
601 cycles can be (considerably) lower than the current frequency.
602 Notice that the average instructions per cycle and number of
603 cycles is shown in the CPU line for all CPUs.
604 See also: http://www.brendangregg.com/blog/2017-05-09/cpu-utiliza‐
605 tion-is-wrong.html
606
607 In case of frequency scaling, all previously mentioned CPU per‐
608 centages are relative to the used scaling of the CPU during the
609 interval. If a CPU has been active for e.g. 50% in user mode dur‐
610 ing the interval while the frequency scaling of that CPU was 40%,
611 only 20% of the full capacity of the CPU has been used in user
612 mode.
613
614 If the screen-width does not allow all of these counters, only a
615 relevant subset is shown.
616
617 CPL CPU load information.
618 This line contains the load average figures reflecting the number
619 of threads that are available to run on a CPU (i.e. part of the
620 runqueue) or that are waiting for disk I/O. These figures are
621 averaged over 1 (`avg1'), 5 (`avg5') and 15 (`avg15') minutes.
622 Furthermore the number of context switches (`csw'), the number of
623 serviced interrupts (`intr') and the number of available CPUs are
624 shown.
625
626 If the screen-width does not allow all of these counters, only a
627 relevant subset is shown.
628
629 GPU GPU utilization (Nvidia).
630 Read the section GPU STATISTICS GATHERING in this document to find
631 the details about the activation of the pmdanvidia daemon.
632
633 In the first column of every line, the bus-id (last nine charac‐
634 ters) and the GPU number are shown. The subsequent columns show
635 the percentage of time that one or more kernels were executing on
636 the GPU (`gpubusy'), the percentage of time that global (device)
637 memory was being read or written (`membusy'), the occupation per‐
638 centage of memory (`memocc'), the total memory (`total'), the mem‐
639 ory being in use at the moment of the sample (`used'), the average
640 memory being in use during the sample time (`usavg'), the number
641 of processes being active on the GPU at the moment of the sample
642 (`#proc'), and the type of GPU.
643
644 If the screen-width does not allow all of these counters, only a
645 relevant subset is shown.
646 The number of lines showing the GPUs can be limited.
647
648 MEM Memory occupation.
649 This line contains the total amount of physical memory (`tot'),
650 the amount of memory which is currently free (`free'), the amount
651 of memory in use as page cache including the total resident shared
652 memory (`cache'), the amount of memory within the page cache that
653 has to be flushed to disk (`dirty'), the amount of memory used for
654 filesystem meta data (`buff'), the amount of memory being used for
655 kernel mallocs (`slab'), the amount of slab memory that is
656 reclaimable (`slrec'), the resident size of shared memory includ‐
657 ing tmpfs (`shmem`), the resident size of shared memory (`shrss`)
658 the amount of shared memory that is currently swapped (`shswp`),
659 the amount of memory that is currently claimed by vmware's balloon
660 driver (`vmbal`), the amount of memory that is claimed for huge
661 pages (`hptot`), and the amount of huge page memory that is really
662 in use (`hpuse`).
663
664 If the screen-width does not allow all of these counters, only a
665 relevant subset is shown.
666
667 SWP Swap occupation and overcommit info.
668 This line contains the total amount of swap space on disk (`tot')
669 and the amount of free swap space (`free').
670 Furthermore the committed virtual memory space (`vmcom') and the
671 maximum limit of the committed space (`vmlim', which is by default
672 swap size plus 50% of memory size) is shown. The committed space
673 is the reserved virtual space for all allocations of private mem‐
674 ory space for processes. The kernel only verifies whether the
675 committed space exceeds the limit if strict overcommit handling is
676 configured (vm.overcommit_memory is 2).
677
678 PAG Paging frequency.
679 This line contains the number of scanned pages (`scan') due to the
680 fact that free memory drops below a particular threshold and the
681 number times that the kernel tries to reclaim pages due to an
682 urgent need (`stall').
683 Also the number of memory pages the system read from swap space
684 (`swin') and the number of memory pages the system wrote to swap
685 space (`swout') are shown.
686
687 PSI Pressure Stall Information.
688 This line contains three percentages per category: average pres‐
689 sure percentage over the last 10, 60 and 300 seconds (separated by
690 slashes).
691 The categories are: CPU for 'some' (`cs'), memory for 'some'
692 (`ms'), memory for 'full' (`mf'), I/O for 'some' (`is'), and I/O
693 for 'full' (`if').
694
695 LVM/MDD/DSK
696 Logical volume/multiple device/disk utilization.
697 Per active unit one line is produced, sorted on unit activity.
698 Such line shows the name (e.g. VolGroup00-lvtmp for a logical vol‐
699 ume or sda for a hard disk), the busy percentage i.e. the portion
700 of time that the unit was busy handling requests (`busy'), the
701 number of read requests issued (`read'), the number of write
702 requests issued (`write'), the number of KiBytes per read
703 (`KiB/r'), the number of KiBytes per write (`KiB/w'), the number
704 of MiBytes per second throughput for reads (`MBr/s'), the number
705 of MiBytes per second throughput for writes (`MBw/s'), the average
706 queue depth (`avq') and the average number of milliseconds needed
707 by a request (`avio') for seek, latency and data transfer.
708 If the screen-width does not allow all of these counters, only a
709 relevant subset is shown.
710
711 The number of lines showing the units can be limited per class
712 (LVM, MDD or DSK) with the 'l' key or statically (see separate
713 man-page of pcp-atoprc(5)). By specifying the value 0 for a par‐
714 ticular class, no lines will be shown any more for that class.
715
716 NFM Network Filesystem (NFS) mount at the client side.
717 For each NFS-mounted filesystem, a line is shown that contains the
718 mounted server directory, the name of the server (`srv'), the
719 total number of bytes physically read from the server (`read') and
720 the total number of bytes physically written to the server
721 (`write'). Data transfer is subdivided in the number of bytes
722 read via normal read() system calls (`nread'), the number of bytes
723 written via normal read() system calls (`nwrit'), the number of
724 bytes read via direct I/O (`dread'), the number of bytes written
725 via direct I/O (`dwrit'), the number of bytes read via memory
726 mapped I/O pages (`mread'), and the number of bytes written via
727 memory mapped I/O pages (`mwrit').
728
729 NFC Network Filesystem (NFS) client side counters.
730 This line contains the number of RPC calls issues by local pro‐
731 cesses (`rpc'), the number of read RPC calls (`read`) and write
732 RPC calls (`rpwrite') issued to the NFS server, the number of RPC
733 calls being retransmitted (`retxmit') and the number of authoriza‐
734 tion refreshes (`autref').
735
736 NFS Network Filesystem (NFS) server side counters.
737 This line contains the number of RPC calls received from NFS
738 clients (`rpc'), the number of read RPC calls received (`cread`),
739 the number of write RPC calls received (`cwrit'), the number of
740 Megabytes/second returned to read requests by clients (`MBcr/s`),
741 the number of Megabytes/second passed in write requests by clients
742 (`MBcw/s`), the number of network requests handled via TCP
743 (`nettcp'), the number of network requests handled via UDP
744 (`netudp'), the number of reply cache hits (`rchits'), the number
745 of reply cache misses (`rcmiss') and the number of uncached
746 requests (`rcnoca'). Furthermore some error counters indicating
747 the number of requests with a bad format (`badfmt') or a bad
748 authorization (`badaut'), and a counter indicating the number of
749 bad clients (`badcln').
750
751 NET Network utilization (TCP/IP).
752 One line is shown for activity of the transport layer (TCP and
753 UDP), one line for the IP layer and one line per active interface.
754 For the transport layer, counters are shown concerning the number
755 of received TCP segments including those received in error
756 (`tcpi'), the number of transmitted TCP segments excluding those
757 containing only retransmitted octets (`tcpo'), the number of UDP
758 datagrams received (`udpi'), the number of UDP datagrams transmit‐
759 ted (`udpo'), the number of active TCP opens (`tcpao'), the number
760 of passive TCP opens (`tcppo'), the number of TCP output retrans‐
761 missions (`tcprs'), the number of TCP input errors (`tcpie'), the
762 number of TCP output resets (`tcpor'), the number of UDP no ports
763 (`udpnp'), and the number of UDP input errors (`udpie').
764 If the screen-width does not allow all of these counters, only a
765 relevant subset is shown.
766 These counters are related to IPv4 and IPv6 combined.
767
768 For the IP layer, counters are shown concerning the number of IP
769 datagrams received from interfaces, including those received in
770 error (`ipi'), the number of IP datagrams that local higher-layer
771 protocols offered for transmission (`ipo'), the number of received
772 IP datagrams which were forwarded to other interfaces (`ipfrw'),
773 the number of IP datagrams which were delivered to local higher-
774 layer protocols (`deliv'), the number of received ICMP datagrams
775 (`icmpi'), and the number of transmitted ICMP datagrams (`icmpo').
776 If the screen-width does not allow all of these counters, only a
777 relevant subset is shown.
778 These counters are related to IPv4 and IPv6 combined.
779
780 For every active network interface one line is shown, sorted on
781 the interface activity. Such line shows the name of the interface
782 and its busy percentage in the first column. The busy percentage
783 for half duplex is determined by comparing the interface speed
784 with the number of bits transmitted and received per second; for
785 full duplex the interface speed is compared with the highest of
786 either the transmitted or the received bits. When the interface
787 speed can not be determined (e.g. for the loopback interface),
788 `---' is shown instead of the percentage.
789 Furthermore the number of received packets (`pcki'), the number of
790 transmitted packets (`pcko'), the line speed of the interface
791 (`sp'), the effective amount of bits received per second (`si'),
792 the effective amount of bits transmitted per second (`so'), the
793 number of collisions (`coll'), the number of received multicast
794 packets (`mlti'), the number of errors while receiving a packet
795 (`erri'), the number of errors while transmitting a packet
796 (`erro'), the number of received packets dropped (`drpi'), and the
797 number of transmitted packets dropped (`drpo').
798 If the screen-width does not allow all of these counters, only a
799 relevant subset is shown.
800 The number of lines showing the network interfaces can be limited.
801
802 IFB Infiniband utilization.
803 For every active Infiniband port one line is shown, sorted on
804 activity. Such line shows the name of the port and its busy per‐
805 centage in the first column. The busy percentage is determined by
806 taking the highest of either the transmitted or the received bits
807 during the interval, multiplying that value by the number of lanes
808 and comparing it against the maximum port speed.
809 Furthermore the number of received packets divided by the number
810 of lanes (`pcki'), the number of transmitted packets divided by
811 the number of lanes (`pcko'), the maximum line speed (`sp'), the
812 effective amount of bits received per second (`si'), the effective
813 amount of bits transmitted per second (`so'), and the number of
814 lanes (`lanes').
815 If the screen-width does not allow all of these counters, only a
816 relevant subset is shown.
817 The number of lines showing the Infiniband ports can be limited.
818
820 Following the system level information, the processes are shown from
821 which the resource utilization has changed during the last interval.
822 These processes might have used cpu time or issued disk or network
823 requests. However a process is also shown if part of it has been paged
824 out due to lack of memory (while the process itself was in sleep
825 state).
826
827 Per process the following fields may be shown (in alphabetical order),
828 depending on the current output mode as described in the section INTER‐
829 ACTIVE COMMANDS and depending on the current width of your window:
830
831 AVGRSZ The average size of one read-action on disk.
832
833 AVGWSZ The average size of one write-action on disk.
834
835 CID Container ID (Docker) of 12 hexadecimal digits, referring to
836 the container in which the process/thread is running. If a
837 process has been started and finished during the last inter‐
838 val, a `?' is shown because the container ID is not part of
839 the standard process accounting record.
840
841 CMD The name of the process. This name can be surrounded by
842 "less/greater than" signs (`<name>') which means that the
843 process has finished during the last interval.
844 Behind the abbreviation `CMD' in the header line, the current
845 page number and the total number of pages of the
846 process/thread list are shown.
847
848 COMMAND-LINE
849 The full command line of the process (including arguments). If
850 the length of the command line exceeds the length of the
851 screen line, the arrow keys -> and <- can be used for horizon‐
852 tal scroll.
853 Behind the verb `COMMAND-LINE' in the header line, the current
854 page number and the total number of pages of the
855 process/thread list are shown.
856
857 CPU The occupation percentage of this process related to the
858 available capacity for this resource on system level.
859
860 CPUNR The identification of the CPU the (main) thread is running on
861 or has recently been running on.
862
863 CTID Container ID (OpenVZ). If a process has been started and fin‐
864 ished during the last interval, a `?' is shown because the
865 container ID is not part of the standard process accounting
866 record.
867
868 DSK The occupation percentage of this process related to the total
869 load that is produced by all processes (i.e. total disk
870 accesses by all processes during the last interval).
871 This information is shown when per process "storage account‐
872 ing" is active in the kernel.
873
874 EGID Effective group-id under which this process executes.
875
876 ENDATE Date that the process has been finished. If the process is
877 still running, this field shows `active'.
878
879 ENTIME Time that the process has been finished. If the process is
880 still running, this field shows `active'.
881
882 ENVID Virtual environment identified (OpenVZ only).
883
884 EUID Effective user-id under which this process executes.
885
886 EXC The exit code of a terminated process (second position of col‐
887 umn `ST' is E) or the fatal signal number (second position of
888 column `ST' is S or C).
889
890 FSGID Filesystem group-id under which this process executes.
891
892 FSUID Filesystem user-id under which this process executes.
893
894 GPU When the pmdanvidia daemon does not run with root privileges,
895 the GPU percentage reflects the GPU memory occupation percent‐
896 age (memory of all GPUs is 100%).
897 When the pmdanvidia daemon runs with root privileges, the GPU
898 percentage reflects the GPU busy percentage.
899
900 GPUBUSY Busy percentage on all GPUs (one GPU is 100%).
901 When the pmdanvidia daemon does not run with root privileges,
902 this value is not available.
903
904 GPUNUMS Comma-separated list of GPUs used by the process during the
905 interval. When the comma-separated list exceeds the width of
906 the column, a hexadecimal value is shown.
907
908 MAJFLT The number of page faults issued by this process that have
909 been solved by creating/loading the requested memory page.
910
911 MEM The occupation percentage of this process related to the
912 available capacity for this resource on system level.
913
914 MEMAVG Average memory occupation during the interval on all used
915 GPUs.
916
917 MEMBUSY Busy percentage of memory on all GPUs (one GPU is 100%), i.e.
918 the time needed for read and write accesses on memory.
919 When the pmdanvidia daemon does not run with root privileges,
920 this value is not available.
921
922 MEMNOW Memory occupation at the moment of the sample on all used
923 GPUs.
924
925 MINFLT The number of page faults issued by this process that have
926 been solved by reclaiming the requested memory page from the
927 free list of pages.
928
929 NICE The more or less static priority that can be given to a
930 process on a scale from -20 (high priority) to +19 (low prior‐
931 ity).
932
933 NPROCS The number of active and terminated processes accumulated for
934 this user or program.
935
936 PID Process-id.
937
938 POLI The policies 'norm' (normal, which is SCHED_OTHER), 'btch'
939 (batch) and 'idle' refer to timesharing processes. The poli‐
940 cies 'fifo' (SCHED_FIFO) and 'rr' (round robin, which is
941 SCHED_RR) refer to realtime processes.
942
943 PPID Parent process-id.
944
945 PRI The process' priority ranges from 0 (highest priority) to 139
946 (lowest priority). Priority 0 to 99 are used for realtime
947 processes (fixed priority independent of their behavior) and
948 priority 100 to 139 for timesharing processes (variable prior‐
949 ity depending on their recent CPU consumption and the nice
950 value).
951
952 PSIZE The proportional memory size of this process (or user).
953 Every process shares resident memory with other processes.
954 E.g. when a particular program is started several times, the
955 code pages (text) are only loaded once in memory and shared by
956 all incarnations. Also the code of shared libraries is shared
957 by all processes using that shared library, as well as shared
958 memory and memory-mapped files. For the PSIZE calculation of
959 a process, the resident memory of a process that is shared
960 with other processes is divided by the number of sharers.
961 This means, that every process is accounted for a proportional
962 part of that memory. Accumulating the PSIZE values of all
963 processes in the system gives a reliable impression of the
964 total resident memory consumed by all processes.
965 Since gathering of all values that are needed to calculate the
966 PSIZE is a relatively time-consuming task, the 'R' key (or
967 '-R' flag) should be active. Gathering these values also
968 requires superuser privileges (otherwise '?K' is shown in the
969 output).
970
971 RDDSK When the kernel maintains standard io statistics (>= 2.6.20):
972 The read data transfer issued physically on disk (so reading
973 from the disk cache is not accounted for).
974 Unfortunately, the kernel aggregates the data tranfer of a
975 process to the data transfer of its parent process when termi‐
976 nating, so you might see transfers for (parent) processes like
977 cron, bash or init, that are not really issued by them.
978
979 RGID The real group-id under which the process executes.
980
981 RGROW The amount of resident memory that the process has grown dur‐
982 ing the last interval. A resident growth can be caused by
983 touching memory pages which were not physically created/loaded
984 before (load-on-demand). Note that a resident growth can also
985 be negative e.g. when part of the process is paged out due to
986 lack of memory or when the process frees dynamically allocated
987 memory. For a process which started during the last interval,
988 the resident growth reflects the total resident size of the
989 process at that moment.
990
991 RSIZE The total resident memory usage consumed by this process (or
992 user). Notice that the RSIZE of a process includes all resi‐
993 dent memory used by that process, even if certain memory parts
994 are shared with other processes (see also the explanation of
995 PSIZE).
996
997 RTPR Realtime priority according the POSIX standard. Value can be
998 0 for a timesharing process (policy 'norm', 'btch' or 'idle')
999 or ranges from 1 (lowest) till 99 (highest) for a realtime
1000 process (policy 'rr' or 'fifo').
1001
1002 RUID The real user-id under which the process executes.
1003
1004 S The current state of the (main) thread: `R' for running (cur‐
1005 rently processing or in the runqueue), `S' for sleeping inter‐
1006 ruptible (wait for an event to occur), `D' for sleeping non-
1007 interruptible, `Z' for zombie (waiting to be synchronized with
1008 its parent process), `T' for stopped (suspended or traced),
1009 `W' for swapping, and `E' (exit) for processes which have fin‐
1010 ished during the last interval.
1011
1012 SGID The saved group-id of the process.
1013
1014 ST The status of a process.
1015 The first position indicates if the process has been started
1016 during the last interval (the value N means 'new process').
1017
1018 The second position indicates if the process has been finished
1019 during the last interval.
1020 The value E means 'exit' on the process' own initiative; the
1021 exit code is displayed in the column `EXC'.
1022 The value S means that the process has been terminated unvol‐
1023 untarily by a signal; the signal number is displayed in the in
1024 the column `EXC'.
1025 The value C means that the process has been terminated unvol‐
1026 untarily by a signal, producing a core dump in its current
1027 directory; the signal number is displayed in the column `EXC'.
1028
1029 STDATE The start date of the process.
1030
1031 STTIME The start time of the process.
1032
1033 SUID The saved user-id of the process.
1034
1035 SWAPSZ The swap space consumed by this process (or user).
1036
1037 SYSCPU CPU time consumption of this process in system mode (kernel
1038 mode), usually due to system call handling.
1039
1040 THR Total number of threads within this process. All related
1041 threads are contained in a thread group, represented by pcp-
1042 atop as one line or as a separate line when the 'y' key (or -y
1043 flag) is active.
1044
1045 TID Thread-id. All threads within a process run with the same PID
1046 but with a different TID. This value is shown for individual
1047 threads in multi-threaded processes (when using the key 'y').
1048
1049 TRUN Number of threads within this process that are in the state
1050 'running' (R).
1051
1052 TSLPI Number of threads within this process that are in the state
1053 'interruptible sleeping' (S).
1054
1055 TSLPU Number of threads within this process that are in the state
1056 'uninterruptible sleeping' (D).
1057
1058 USRCPU CPU time consumption of this process in user mode, due to pro‐
1059 cessing the own program text.
1060
1061 VDATA The virtual memory size of the private data used by this
1062 process (including heap and shared library data).
1063
1064 VGROW The amount of virtual memory that the process has grown during
1065 the last interval. A virtual growth can be caused by e.g.
1066 issueing a malloc() or attaching a shared memory segment.
1067 Note that a virtual growth can also be negative by e.g. issue‐
1068 ing a free() or detaching a shared memory segment. For a
1069 process which started during the last interval, the virtual
1070 growth reflects the total virtual size of the process at that
1071 moment.
1072
1073 VPID Virtual process-id (within an OpenVZ container). If a process
1074 has been started and finished during the last interval, a `?'
1075 is shown because the virtual process-id is not part of the
1076 standard process accounting record.
1077
1078 VSIZE The total virtual memory usage consumed by this process (or
1079 user).
1080
1081 VSLIBS The virtual memory size of the (shared) text of all shared
1082 libraries used by this process.
1083
1084 VSTACK The virtual memory size of the (private) stack used by this
1085 process
1086
1087 VSTEXT The virtual memory size of the (shared) text of the executable
1088 program.
1089
1090 WRDSK When the kernel maintains standard io statistics (>= 2.6.20):
1091 The write data transfer issued physically on disk (so writing
1092 to the disk cache is not accounted for). This counter is
1093 maintained for the application process that writes its data to
1094 the cache (assuming that this data is physically transferred
1095 to disk later on). Notice that disk I/O needed for swapping
1096 is not taken into account.
1097 Unfortunately, the kernel aggregates the data tranfer of a
1098 process to the data transfer of its parent process when termi‐
1099 nating, so you might see transfers for (parent) processes like
1100 cron, bash or init, that are not really issued by them.
1101
1102 WCANCL When the kernel maintains standard io statistics (>= 2.6.20):
1103 The write data transfer previously accounted for this process
1104 or another process that has been cancelled. Suppose that a
1105 process writes new data to a file and that data is removed
1106 again before the cache buffers have been flushed to disk.
1107 Then the original process shows the written data as WRDSK,
1108 while the process that removes/truncates the file shows the
1109 unflushed removed data as WCANCL.
1110
1112 With the flag -P followed by a list of one or more labels (comma-sepa‐
1113 rated), parseable output is produced for each sample. The labels that
1114 can be specified for system-level statistics correspond to the labels
1115 (first verb of each line) that can be found in the interactive output:
1116 "CPU", "cpu", "CPL", "GPU", "MEM", "SWP", "PAG", "PSI", "LVM", "MDD",
1117 "DSK", "NFM", "NFC", "NFS", "NET" and "IFB".
1118 For process-level statistics special labels are introduced: "PRG" (gen‐
1119 eral), "PRC" (cpu), "PRE" (GPU), "PRM" (memory), "PRD" (disk, only if
1120 "storage accounting" is active).
1121 With the label "ALL", all system and process level statistics are
1122 shown.
1123
1124 For every interval all requested lines are shown whereafter pcp-atop
1125 shows a line just containing the label "SEP" as a separator before the
1126 lines for the next sample are generated.
1127 When a sample contains the values since boot, pcp-atop shows a line
1128 just containing the label "RESET" before the lines for this sample are
1129 generated.
1130
1131 The first part of each output-line consists of the following six
1132 fields: label (the name of the label), host (the name of this machine),
1133 epoch (the time of this interval as number of seconds since 1-1-1970),
1134 date (date of this interval in format YYYY/MM/DD), time (time of this
1135 interval in format HH:MM:SS), and interval (number of seconds elapsed
1136 for this interval).
1137
1138 The subsequent fields of each output-line depend on the label:
1139
1140 CPU Subsequent fields: total number of clock-ticks per second for
1141 this machine, number of processors, consumption for all CPUs
1142 in system mode (clock-ticks), consumption for all CPUs in user
1143 mode (clock-ticks), consumption for all CPUs in user mode for
1144 niced processes (clock-ticks), consumption for all CPUs in
1145 idle mode (clock-ticks), consumption for all CPUs in wait mode
1146 (clock-ticks), consumption for all CPUs in irq mode (clock-
1147 ticks), consumption for all CPUs in softirq mode (clock-
1148 ticks), consumption for all CPUs in steal mode (clock-ticks),
1149 consumption for all CPUs in guest mode (clock-ticks) overlap‐
1150 ping user mode, frequency of all CPUs and frequency percentage
1151 of all CPUs.
1152
1153 cpu Subsequent fields: total number of clock-ticks per second for
1154 this machine, processor-number, consumption for this CPU in
1155 system mode (clock-ticks), consumption for this CPU in user
1156 mode (clock-ticks), consumption for this CPU in user mode for
1157 niced processes (clock-ticks), consumption for this CPU in
1158 idle mode (clock-ticks), consumption for this CPU in wait mode
1159 (clock-ticks), consumption for this CPU in irq mode (clock-
1160 ticks), consumption for this CPU in softirq mode (clock-
1161 ticks), consumption for this CPU in steal mode (clock-ticks),
1162 consumption for this CPU in guest mode (clock-ticks) overlap‐
1163 ping user mode, frequency of all CPUs, frequency percentage of
1164 all CPUs, instructions executed by all CPUs and cycles for all
1165 CPUs.
1166
1167 CPL Subsequent fields: number of processors, load average for last
1168 minute, load average for last five minutes, load average for
1169 last fifteen minutes, number of context-switches, and number
1170 of device interrupts.
1171
1172 GPU Subsequent fields: GPU number, bus-id string, type of GPU
1173 string, GPU busy percentage during last second (-1 if not
1174 available), memory busy percentage during last second (-1 if
1175 not available), total memory size (KiB), used memory (KiB) at
1176 this moment, number of samples taken during interval, cumula‐
1177 tive GPU busy percentage during the interval (to be divided by
1178 the number of samples for the average busy percentage, -1 if
1179 not available), cumulative memory busy percentage during the
1180 interval (to be divided by the number of samples for the aver‐
1181 age busy percentage, -1 if not available), and cumulative mem‐
1182 ory occupation during the interval (to be divided by the num‐
1183 ber of samples for the average occupation).
1184
1185 MEM Subsequent fields: page size for this machine (in bytes), size
1186 of physical memory (pages), size of free memory (pages), size
1187 of page cache (pages), size of buffer cache (pages), size of
1188 slab (pages), dirty pages in cache (pages), reclaimable part
1189 of slab (pages), total size of vmware's balloon pages (pages),
1190 total size of shared memory (pages), size of resident shared
1191 memory (pages), size of swapped shared memory (pages), huge
1192 page size (in bytes), total size of huge pages (huge pages),
1193 and size of free huge pages (huge pages).
1194
1195 SWP Subsequent fields: page size for this machine (in bytes), size
1196 of swap (pages), size of free swap (pages), 0 (future use),
1197 size of committed space (pages), and limit for committed space
1198 (pages).
1199
1200 PAG Subsequent fields: page size for this machine (in bytes), num‐
1201 ber of page scans, number of allocstalls, 0 (future use), num‐
1202 ber of swapins, and number of swapouts.
1203
1204 PSI Subsequent fields: PSI statistics present on this system (n or
1205 y), CPU some avg10, CPU some avg60, CPU some avg300, CPU some
1206 accumulated microseconds during interval, memory some avg10,
1207 memory some avg60, memory some avg300, memory some accumulated
1208 microseconds during interval, memory full avg10, memory full
1209 avg60, memory full avg300, memory full accumulated microsec‐
1210 onds during interval, I/O some avg10, I/O some avg60, I/O some
1211 avg300, I/O some accumulated microseconds during interval, I/O
1212 full avg10, I/O full avg60, I/O full avg300, and I/O full
1213 accumulated microseconds during interval.
1214
1215 LVM/MDD/DSK
1216 For every logical volume/multiple device/hard disk one line is
1217 shown.
1218 Subsequent fields: name, number of milliseconds spent for I/O,
1219 number of reads issued, number of sectors transferred for
1220 reads, number of writes issued, and number of sectors trans‐
1221 ferred for write.
1222
1223 NFM Subsequent fields: mounted NFS filesystem, total number of
1224 bytes read, total number of bytes written, number of bytes
1225 read by normal system calls, number of bytes written by normal
1226 system calls, number of bytes read by direct I/O, number of
1227 bytes written by direct I/O, number of pages read by memory-
1228 mapped I/O, and number of pages written by memory-mapped I/O.
1229
1230 NFC Subsequent fields: number of transmitted RPCs, number of
1231 transmitted read RPCs, number of transmitted write RPCs, num‐
1232 ber of RPC retransmissions, and number of authorization
1233 refreshes.
1234
1235 NFS Subsequent fields: number of handled RPCs, number of received
1236 read RPCs, number of received write RPCs, number of bytes read
1237 by clients, number of bytes written by clients, number of RPCs
1238 with bad format, number of RPCs with bad authorization, number
1239 of RPCs from bad client, total number of handled network
1240 requests, number of handled network requests via TCP, number
1241 of handled network requests via UDP, number of handled TCP
1242 connections, number of hits on reply cache, number of misses
1243 on reply cache, and number of uncached requests.
1244
1245 NET First one line is produced for the upper layers of the TCP/IP
1246 stack.
1247 Subsequent fields: the verb "upper", number of packets
1248 received by TCP, number of packets transmitted by TCP, number
1249 of packets received by UDP, number of packets transmitted by
1250 UDP, number of packets received by IP, number of packets
1251 transmitted by IP, number of packets delivered to higher lay‐
1252 ers by IP, and number of packets forwarded by IP.
1253
1254 Next one line is shown for every interface.
1255 Subsequent fields: name of the interface, number of packets
1256 received by the interface, number of bytes received by the
1257 interface, number of packets transmitted by the interface,
1258 number of bytes transmitted by the interface, interface speed,
1259 and duplex mode (0=half, 1=full).
1260
1261 IFB Subsequent fields: name of the InfiniBand interface, port num‐
1262 ber, number of lanes, maximum rate (Mbps), number of bytes
1263 received, number of bytes transmitted, number of packets
1264 received, and number of packets transmitted.
1265
1266 PRG For every process one line is shown.
1267 Subsequent fields: PID (unique ID of task), name (between
1268 brackets), state, real uid, real gid, TGID (group number of
1269 related tasks/threads), total number of threads, exit code (in
1270 case of fatal signal: signal number + 256), start time
1271 (epoch), full command line (between brackets), PPID, number of
1272 threads in state 'running' (R), number of threads in state
1273 'interruptible sleeping' (S), number of threads in state
1274 'uninterruptible sleeping' (D), effective uid, effective gid,
1275 saved uid, saved gid, filesystem uid, filesystem gid, elapsed
1276 time (hertz), is_process (y/n), OpenVZ virtual pid (VPID),
1277 OpenVZ container id (CTID) and Docker container id (CID).
1278
1279 PRC For every process one line is shown.
1280 Subsequent fields: PID, name (between brackets), state, total
1281 number of clock-ticks per second for this machine, CPU-con‐
1282 sumption in user mode (clockticks), CPU-consumption in system
1283 mode (clockticks), nice value, priority, realtime priority,
1284 scheduling policy, current CPU, sleep average, TGID (group
1285 number of related tasks/threads) and is_process (y/n).
1286
1287 PRE For every process one line is shown.
1288 Subsequent fields: PID, name (between brackets), process
1289 state, GPU state (A for active, E for exited, N for no GPU
1290 user), number of GPUs used by this process, bitlist reflecting
1291 used GPUs, GPU busy percentage during interval, memory busy
1292 percentage during interval, memory occupation (KiB) at this
1293 moment cumulative memory occupation (KiB) during interval, and
1294 number of samples taken during interval.
1295
1296 PRM For every process one line is shown.
1297 Subsequent fields: PID, name (between brackets), state, page
1298 size for this machine (in bytes), virtual memory size
1299 (Kbytes), resident memory size (Kbytes), shared text memory
1300 size (Kbytes), virtual memory growth (Kbytes), resident memory
1301 growth (Kbytes), number of minor page faults, number of major
1302 page faults, virtual library exec size (Kbytes), virtual data
1303 size (Kbytes), virtual stack size (Kbytes), swap space used
1304 (Kbytes), TGID (group number of related tasks/threads),
1305 is_process (y/n) and proportional set size (Kbytes) if in 'R'
1306 option is specified.
1307
1308 PRD For every process one line is shown.
1309 Subsequent fields: PID, name (between brackets), state, obso‐
1310 leted kernel patch installed ('n'), standard io statistics
1311 used ('y' or 'n'), number of reads on disk, cumulative number
1312 of sectors read, number of writes on disk, cumulative number
1313 of sectors written, cancelled number of written sectors, TGID
1314 (group number of related tasks/threads) and is_process (y/n).
1315 If the standard I/O statistics (>= 2.6.20) are not used, the
1316 disk I/O counters per process are not relevant. The counters
1317 'number of reads on disk' and 'number of writes on disk' are
1318 obsoleted anyhow.
1319
1320 PRN For every process one line is shown.
1321 Subsequent fields: PID, name (between brackets), state, kernel
1322 module 'netatop' loaded ('y' or 'n'), number of TCP-packets
1323 transmitted, cumulative size of TCP-packets transmitted, num‐
1324 ber of TCP-packets received, cumulative size of TCP-packets
1325 received, number of UDP-packets transmitted, cumulative size
1326 of UDP-packets transmitted, number of UDP-packets received,
1327 cumulative size of UDP-packets transmitted, number of raw
1328 packets transmitted (obsolete, always 0), number of raw pack‐
1329 ets received (obsolete, always 0), TGID (group number of
1330 related tasks/threads) and is_process (y/n).
1331
1333 By sending the SIGUSR1 signal to pcp-atop a new sample will be forced,
1334 even if the current timer interval has not exceeded yet. The behavior
1335 is similar to pressing the `t` key in an interactive session.
1336
1337 By sending the SIGUSR2 signal to pcp-atop a final sample will be forced
1338 after which pcp-atop will terminate.
1339
1341 To monitor the current system load interactively with an interval of 5
1342 seconds:
1343
1344 pcp atop 5
1345
1346 To monitor the system load and write it to a file (in plain ASCII) with
1347 an interval of one minute during half an hour with active processes
1348 sorted on memory consumption:
1349
1350 pcp atop -M 60 30 > /log/pcp-atop.mem
1351
1352 Store information about the system and process activity in a PCP ar‐
1353 chive folio with an interval of ten minutes during an hour:
1354
1355 pcp atop -w /tmp/pcp-atop 600 6
1356
1357 View the contents of this file interactively:
1358
1359 pcp atop -r /tmp/pcp-atop
1360
1361 View the processor and disk utilization of this file in parseable for‐
1362 mat:
1363
1364 pcp atop -PCPU,DSK -r /tmp/pcp-atop.folio
1365
1366 View the contents of today's standard logfile interactively:
1367
1368 pcp atop -r
1369
1370 View the contents of the standard logfile of the day before yesterday
1371 interactively:
1372
1373 pcp atop -r yy
1374
1375 View the contents of the standard logfile of 2014, June 7 from 02:00 PM
1376 onwards interactively:
1377
1378 pcp atop -r 20140607 -b 14:00
1379
1381 pcp-atop is based on the source code of the atop(1) command from
1382 https://atoptool.nl, maintained by Gerlof Langeveld
1383 (gerlof.langeveld@atoptool.nl), and aims to be command line and output
1384 compatible with it as much as possible. Some features of that atop
1385 command are not available in pcp-atop.
1386
1387 Some features of pcp-atop (such as reporting on the Apache HTTP daemon,
1388 Infiniband, NFS client mounts, hardware event counts and GPU statis‐
1389 tics) are only activated if the corresonding PCP metrics are available.
1390 Refer to the documentation for pmdaapache(1), pmdainfiniband(1), pmdan‐
1391 fsclient(1), pmdanvidia(1) and pmdaperfevent(1) for further details on
1392 activating these metrics.
1393
1395 /etc/atoprc
1396 Configuration file containing system-wide default values. See
1397 related man-page.
1398
1399 ~/.atoprc
1400 Configuration file containing personal default values. See
1401 related man-page.
1402
1404 Environment variables with the prefix PCP_ are used to parameterize the
1405 file and directory names used by PCP. On each installation, the file
1406 /etc/pcp.conf contains the local values for these variables. The
1407 $PCP_CONF variable may be used to specify an alternative configuration
1408 file, as described in pcp.conf(5).
1409
1410 For environment variables affecting PCP tools, see pmGetOptions(3).
1411
1413 PCPIntro(1), pcp(1), pcp-atopsar(1), pmdaapache(1), pmdainfiniband(1),
1414 pmdanfsclient(1), pmdanvidia(1), pmdaproc(1), mkaf(1), pmlogger(1),
1415 pmlogger_daily(1) and pcp-atoprc(5).
1416
1417
1418
1419Performance Co-Pilot PCP PCP-ATOP(1)