1SAR(1) Linux User's Manual SAR(1)
2
6 sar - Collect, report, or save system activity information.
7
9 sar [ -A ] [ -B ] [ -b ] [ -C ] [ -D ] [ -d ] [ -F [ MOUNT ] ] [ -H ] [
10 -h ] [ -p ] [ -q ] [ -r [ ALL ] ] [ -S ] [ -t ] [ -u [ ALL ] ] [ -V ] [
11 -v ] [ -W ] [ -w ] [ -y ] [ -z ] [ --dec={ 0 | 1 | 2 } ] [ --dev=
12 dev_list ] [ --fs= fs_list ] [ --help ] [ --human ] [ --iface=
13 iface_list ] [ --sadc ] [ -I { int_list | SUM | ALL } ] [ -P { cpu_list
14 | ALL } ] [ -m { keyword [,...] | ALL } ] [ -n { keyword [,...] | ALL }
15 ] [ -j { SID | ID | LABEL | PATH | UUID | ... } ] [ -f [ filename ] |
16 -o [ filename ] | -[0-9]+ ] [ -i interval ] [ -s [ hh:mm[:ss] ] ] [ -e
17 [ hh:mm[:ss] ] ] [ interval [ count ] ]
18
20 The sar command writes to standard output the contents of selected
21 cumulative activity counters in the operating system. The accounting
22 system, based on the values in the count and interval parameters,
23 writes information the specified number of times spaced at the speci‐
24 fied intervals in seconds. If the interval parameter is set to zero,
25 the sar command displays the average statistics for the time since the
26 system was started. If the interval parameter is specified without the
27 count parameter, then reports are generated continuously. The col‐
28 lected data can also be saved in the file specified by the -o filename
29 flag, in addition to being displayed onto the screen. If filename is
30 omitted, sar uses the standard system activity daily data file (see
31 below). By default all the data available from the kernel are saved in
32 the data file.
33 The sar command extracts and writes to standard output records previ‐
35 ously saved in a file. This file can be either the one specified by the
36 -f flag or, by default, the standard system activity daily data file.
37 It is also possible to enter -1, -2 etc. as an argument to sar to dis‐
38 play data of that days ago. For example, -1 will point at the standard
39 system activity file of yesterday.
40 Standard system activity daily data files are named saDD or saYYYYMMDD,
42 where YYYY stands for the current year, MM for the current month and DD
43 for the current day. They are the default files used by sar only when
44 no filename has been explicitly specified. When used to write data to
45 files (with its option -o), sar will use saYYYYMMDD if option -D has
46 also been specified, else it will use saDD. When used to display the
47 records previously saved in a file, sar will look for the most recent
48 of saDD and saYYYYMMDD, and use it.
49 Standard system activity daily data files are located in the
51 /var/log/sa directory by default. Yet it is possible to specify an
52 alternate location for them: If a directory (instead of a plain file)
53 is used with options -f or -o then it will be considered as the direc‐
54 tory containing the data files.
55 Without the -P flag, the sar command reports system-wide (global among
57 all processors) statistics, which are calculated as averages for values
58 expressed as percentages, and as sums otherwise. If the -P flag is
59 given, the sar command reports activity which relates to the specified
60 processor or processors. If -P ALL is given, the sar command reports
61 statistics for each individual processor and global statistics among
62 all processors. Offline processors are not displayed.
63 You can select information about specific system activities using
65 flags. Not specifying any flags selects only CPU activity. Specifying
66 the -A flag selects all possible activities.
67 The default version of the sar command (CPU utilization report) might
69 be one of the first facilities the user runs to begin system activity
70 investigation, because it monitors major system resources. If CPU uti‐
71 lization is near 100 percent (user + nice + system), the workload sam‐
72 pled is CPU-bound.
73 If multiple samples and multiple reports are desired, it is convenient
75 to specify an output file for the sar command. Run the sar command as
76 a background process. The syntax for this is:
77 sar -o datafile interval count >/dev/null 2>&1 &
79 All data are captured in binary form and saved to a file (datafile).
81 The data can then be selectively displayed with the sar command using
82 the -f option. Set the interval and count parameters to select count
83 records at interval second intervals. If the count parameter is not
84 set, all the records saved in the file will be selected. Collection of
85 data in this manner is useful to characterize system usage over a
86 period of time and determine peak usage hours.
87 Note: The sar command only reports on local activities.
89
92 -A This is equivalent to specifying -bBdFHqSvwWy -I SUM -m ALL -n
93 ALL -r ALL -u ALL. This option also implies specifying -I ALL
94 -P ALL unless these options are explicitly set on the command
95 line.
96 -B Report paging statistics. The following values are displayed:
98 pgpgin/s
100 Total number of kilobytes the system paged in from disk
101 per second.
102 pgpgout/s
104 Total number of kilobytes the system paged out to disk
105 per second.
106 fault/s
108 Number of page faults (major + minor) made by the system
109 per second. This is not a count of page faults that gen‐
110 erate I/O, because some page faults can be resolved with‐
111 out I/O.
112 majflt/s
114 Number of major faults the system has made per second,
115 those which have required loading a memory page from
116 disk.
117 pgfree/s
119 Number of pages placed on the free list by the system per
120 second.
121 pgscank/s
123 Number of pages scanned by the kswapd daemon per second.
124 pgscand/s
126 Number of pages scanned directly per second.
127 pgsteal/s
129 Number of pages the system has reclaimed from cache
130 (pagecache and swapcache) per second to satisfy its mem‐
131 ory demands.
132 %vmeff
134 Calculated as pgsteal / pgscan, this is a metric of the
135 efficiency of page reclaim. If it is near 100% then
136 almost every page coming off the tail of the inactive
137 list is being reaped. If it gets too low (e.g. less than
138 30%) then the virtual memory is having some difficulty.
139 This field is displayed as zero if no pages have been
140 scanned during the interval of time.
141 -b Report I/O and transfer rate statistics. The following values
143 are displayed:
144 tps
146 Total number of transfers per second that were issued to
147 physical devices. A transfer is an I/O request to a
148 physical device. Multiple logical requests can be com‐
149 bined into a single I/O request to the device. A trans‐
150 fer is of indeterminate size.
151 rtps
153 Total number of read requests per second issued to physi‐
154 cal devices.
155 wtps
157 Total number of write requests per second issued to phys‐
158 ical devices.
159 dtps
161 Total number of discard requests per second issued to
162 physical devices.
163 bread/s
165 Total amount of data read from the devices in blocks per
166 second. Blocks are equivalent to sectors and therefore
167 have a size of 512 bytes.
168 bwrtn/s
170 Total amount of data written to devices in blocks per
171 second.
172 bdscd/s
174 Total amount of data discarded for devices in blocks per
175 second.
176 -C When reading data from a file, tell sar to display comments that
178 have been inserted by sadc.
179 -D Use saYYYYMMDD instead of saDD as the standard system activity
181 daily data file name. This option works only when used in con‐
182 junction with option -o to save data to file.
183 -d Report activity for each block device. When data are displayed,
185 the device specification devM-n is generally used (DEV column).
186 M is the major number of the device and n its minor number.
187 Device names may also be pretty-printed if option -p is used or
188 persistent device names can be printed if option -j is used (see
189 below). Statistics for all devices are displayed unless a
190 restricted list is specified using option --dev= (see corre‐
191 sponding option entry). Note that disk activity depends on sadc
192 options -S DISK and -S XDISK to be collected. The following val‐
193 ues are displayed:
194 tps
196 Total number of transfers per second that were issued to
197 physical devices. A transfer is an I/O request to a
198 physical device. Multiple logical requests can be com‐
199 bined into a single I/O request to the device. A trans‐
200 fer is of indeterminate size.
201 rkB/s
203 Number of kilobytes read from the device per second.
204 wkB/s
206 Number of kilobytes written to the device per second.
207 dkB/s
209 Number of kilobytes discarded for the device per second.
210 areq-sz
212 The average size (in kilobytes) of the I/O requests that
213 were issued to the device.
214 Note: In previous versions, this field was known as
215 avgrq-sz and was expressed in sectors.
216 aqu-sz
218 The average queue length of the requests that were issued
219 to the device.
220 Note: In previous versions, this field was known as
221 avgqu-sz.
222 await
224 The average time (in milliseconds) for I/O requests
225 issued to the device to be served. This includes the time
226 spent by the requests in queue and the time spent servic‐
227 ing them.
228 %util
230 Percentage of elapsed time during which I/O requests were
231 issued to the device (bandwidth utilization for the
232 device). Device saturation occurs when this value is
233 close to 100% for devices serving requests serially. But
234 for devices serving requests in parallel, such as RAID
235 arrays and modern SSDs, this number does not reflect
236 their performance limits.
237 --dec={ 0 | 1 | 2 }
239 Specify the number of decimal places to use (0 to 2, default
240 value is 2).
241 --dev=dev_list
243 Specify the block devices for which statistics are to be dis‐
244 played by sar. dev_list is a list of comma-separated device
245 names.
246 -e [ hh:mm[:ss] ]
248 Set the ending time of the report. The default ending time is
249 18:00:00. Hours must be given in 24-hour format. This option
250 can be used when data are read from or written to a file
251 (options -f or -o).
252 -F [ MOUNT ]
254 Display statistics for currently mounted filesystems. Pseudo-
255 filesystems are ignored. At the end of the report, sar will dis‐
256 play a summary of all those filesystems. Use of the MOUNT
257 parameter keyword indicates that mountpoint will be reported
258 instead of filesystem device. Statistics for all filesystems are
259 displayed unless a restricted list is specified using option
260 --fs= (see corresponding option entry). Note that filesystems
261 statistics depend on sadc option -S XDISK to be collected.
262 The following values are displayed:
264 MBfsfree
266 Total amount of free space in megabytes (including space
267 available only to privileged user).
268 MBfsused
270 Total amount of space used in megabytes.
271 %fsused
273 Percentage of filesystem space used, as seen by a privi‐
274 leged user.
275 %ufsused
277 Percentage of filesystem space used, as seen by an
278 unprivileged user.
279 Ifree
281 Total number of free file nodes in filesystem.
282 Iused
284 Total number of file nodes used in filesystem.
285 %Iused
287 Percentage of file nodes used in filesystem.
288 -f [ filename ]
290 Extract records from filename (created by the -o filename flag).
291 The default value of the filename parameter is the current stan‐
292 dard system activity daily data file. If filename is a direc‐
293 tory instead of a plain file then it is considered as the direc‐
294 tory where the standard system activity daily data files are
295 located. The -f option is exclusive of the -o option.
296 --fs=fs_list
298 Specify the filesystems for which statistics are to be displayed
299 by sar. fs_list is a list of comma-separated filesystem names
300 or mountpoints.
301 -H Report hugepages utilization statistics. The following values
303 are displayed:
304 kbhugfree
306 Amount of hugepages memory in kilobytes that is not yet
307 allocated.
308 kbhugused
310 Amount of hugepages memory in kilobytes that has been
311 allocated.
312 %hugused
314 Percentage of total hugepages memory that has been allo‐
315 cated.
316 kbhugrsvd
318 Amount of reserved hugepages memory in kilobytes.
319 kbhugsurp
321 Amount of surplus hugepages memory in kilobytes.
322 -h Make the output of sar easier to read by a human. Options
324 --human and -p (pretty-print) are enabled implicitly with this
325 option. This option may be especially useful when displaying
326 e.g., network interfaces or block devices statistics.
327 --help Display a short help message then exit.
329 --human
331 Print sizes in human readable format (e.g. 1.0k, 1.2M, etc.)
332 The units displayed with this option supersede any other default
333 units (e.g. kilobytes, sectors...) associated with the metrics.
334 -I { int_list | SUM | ALL }
336 Report statistics for interrupts. int_list is a list of comma-
337 separated values or range of values (e.g., 0-16,35,400-). The
338 SUM keyword indicates that the total number of interrupts
339 received per second is to be displayed. The ALL keyword indi‐
340 cates that statistics from all interrupts, including potential
341 APIC interrupt sources, are to be reported. Note that interrupt
342 statistics depend on sadc option "-S INT" to be collected.
343 -i interval
345 Select data records at seconds as close as possible to the num‐
346 ber specified by the interval parameter.
347 --iface=iface_list
349 Specify the network interfaces for which statistics are to be
350 displayed by sar. iface_list is a list of comma-separated
351 interface names.
352 -j { SID | ID | LABEL | PATH | UUID | ... }
354 Display persistent device names. Use this option in conjunction
355 with option -d. Options ID, LABEL, etc. specify the type of the
356 persistent name. These options are not limited, only prerequi‐
357 site is that directory with required persistent names is present
358 in /dev/disk. If persistent name is not found for the device,
359 the device name is pretty-printed (see option -p below). Option
360 SID tries to get a stable identifier to use as the device name.
361 A stable identifier won't change across reboots for the same
362 physical device. If it exists, this identifier is normally the
363 WWN (World Wide Name) of the device, as read from the
364 /dev/disk/by-id directory.
365 -m { keyword [,...] | ALL }
367 Report power management statistics. Note that these statistics
368 depend on sadc's option "-S POWER" to be collected.
369 Possible keywords are CPU, FAN, FREQ, IN, TEMP and USB.
371 With the CPU keyword, statistics about CPU are reported. The
373 following value is displayed:
374 MHz
376 Instantaneous CPU clock frequency in MHz.
377 With the FAN keyword, statistics about fans speed are reported.
379 The following values are displayed:
380 rpm
382 Fan speed expressed in revolutions per minute.
383 drpm
385 This field is calculated as the difference between cur‐
386 rent fan speed (rpm) and its low limit (fan_min).
387 DEVICE
389 Sensor device name.
390 With the FREQ keyword, statistics about CPU clock frequency are
392 reported. The following value is displayed:
393 wghMHz
395 Weighted average CPU clock frequency in MHz. Note that
396 the cpufreq-stats driver must be compiled in the kernel
397 for this option to work.
398 With the IN keyword, statistics about voltage inputs are
400 reported. The following values are displayed:
401 inV
403 Voltage input expressed in Volts.
404 %in
406 Relative input value. A value of 100% means that voltage
407 input has reached its high limit (in_max) whereas a value
408 of 0% means that it has reached its low limit (in_min).
409 DEVICE
411 Sensor device name.
412 With the TEMP keyword, statistics about devices temperature are
414 reported. The following values are displayed:
415 degC
417 Device temperature expressed in degrees Celsius.
418 %temp
420 Relative device temperature. A value of 100% means that
421 temperature has reached its high limit (temp_max).
422 DEVICE
424 Sensor device name.
425 With the USB keyword, the sar command takes a snapshot of all
427 the USB devices currently plugged into the system. At the end of
428 the report, sar will display a summary of all those USB devices.
429 The following values are displayed:
430 BUS
432 Root hub number of the USB device.
433 idvendor
435 Vendor ID number (assigned by USB organization).
436 idprod
438 Product ID number (assigned by Manufacturer).
439 maxpower
441 Maximum power consumption of the device (expressed in
442 mA).
443 manufact
445 Manufacturer name.
446 product
448 Product name.
449 The ALL keyword is equivalent to specifying all the keywords
451 above and therefore all the power management statistics are
452 reported.
453 -n { keyword [,...] | ALL }
455 Report network statistics.
456 Possible keywords are DEV, EDEV, FC, ICMP, EICMP, ICMP6, EICMP6,
458 IP, EIP, IP6, EIP6, NFS, NFSD, SOCK, SOCK6, SOFT, TCP, ETCP, UDP
459 and UDP6.
460 With the DEV keyword, statistics from the network devices are
462 reported. Statistics for all network interfaces are displayed
463 unless a restricted list is specified using option --iface= (see
464 corresponding option entry). The following values are dis‐
465 played:
466 IFACE
468 Name of the network interface for which statistics are
469 reported.
470 rxpck/s
472 Total number of packets received per second.
473 txpck/s
475 Total number of packets transmitted per second.
476 rxkB/s
478 Total number of kilobytes received per second.
479 txkB/s
481 Total number of kilobytes transmitted per second.
482 rxcmp/s
484 Number of compressed packets received per second (for
485 cslip etc.).
486 txcmp/s
488 Number of compressed packets transmitted per second.
489 rxmcst/s
491 Number of multicast packets received per second.
492 %ifutil
494 Utilization percentage of the network interface. For
495 half-duplex interfaces, utilization is calculated using
496 the sum of rxkB/s and txkB/s as a percentage of the
497 interface speed. For full-duplex, this is the greater of
498 rxkB/S or txkB/s.
499 With the EDEV keyword, statistics on failures (errors) from the
501 network devices are reported. Statistics for all network inter‐
502 faces are displayed unless a restricted list is specified using
503 option --iface= (see corresponding option entry). The following
504 values are displayed:
505 IFACE
507 Name of the network interface for which statistics are
508 reported.
509 rxerr/s
511 Total number of bad packets received per second.
512 txerr/s
514 Total number of errors that happened per second while
515 transmitting packets.
516 coll/s
518 Number of collisions that happened per second while
519 transmitting packets.
520 rxdrop/s
522 Number of received packets dropped per second because of
523 a lack of space in linux buffers.
524 txdrop/s
526 Number of transmitted packets dropped per second because
527 of a lack of space in linux buffers.
528 txcarr/s
530 Number of carrier-errors that happened per second while
531 transmitting packets.
532 rxfram/s
534 Number of frame alignment errors that happened per second
535 on received packets.
536 rxfifo/s
538 Number of FIFO overrun errors that happened per second on
539 received packets.
540 txfifo/s
542 Number of FIFO overrun errors that happened per second on
543 transmitted packets.
544 With the FC keyword, statistics about fibre channel traffic are
546 reported. Note that fibre channel statistics depend on sadc's
547 option "-S DISK" to be collected. The following values are dis‐
548 played:
549 FCHOST
551 Name of the fibre channel host bus adapter (HBA) inter‐
552 face for which statistics are reported.
553 fch_rxf/s
555 The total number of frames received per second.
556 fch_txf/s
558 The total number of frames transmitted per second.
559 fch_rxw/s
561 The total number of transmission words received per sec‐
562 ond.
563 fch_txw/s
565 The total number of transmission words transmitted per
566 second.
567 With the ICMP keyword, statistics about ICMPv4 network traffic
569 are reported. Note that ICMPv4 statistics depend on sadc's
570 option "-S SNMP" to be collected. The following values are dis‐
571 played (formal SNMP names between square brackets):
572 imsg/s
574 The total number of ICMP messages which the entity
575 received per second [icmpInMsgs]. Note that this counter
576 includes all those counted by ierr/s.
577 omsg/s
579 The total number of ICMP messages which this entity
580 attempted to send per second [icmpOutMsgs]. Note that
581 this counter includes all those counted by oerr/s.
582 iech/s
584 The number of ICMP Echo (request) messages received per
585 second [icmpInEchos].
586 iechr/s
588 The number of ICMP Echo Reply messages received per sec‐
589 ond [icmpInEchoReps].
590 oech/s
592 The number of ICMP Echo (request) messages sent per sec‐
593 ond [icmpOutEchos].
594 oechr/s
596 The number of ICMP Echo Reply messages sent per second
597 [icmpOutEchoReps].
598 itm/s
600 The number of ICMP Timestamp (request) messages received
601 per second [icmpInTimestamps].
602 itmr/s
604 The number of ICMP Timestamp Reply messages received per
605 second [icmpInTimestampReps].
606 otm/s
608 The number of ICMP Timestamp (request) messages sent per
609 second [icmpOutTimestamps].
610 otmr/s
612 The number of ICMP Timestamp Reply messages sent per sec‐
613 ond [icmpOutTimestampReps].
614 iadrmk/s
616 The number of ICMP Address Mask Request messages received
617 per second [icmpInAddrMasks].
618 iadrmkr/s
620 The number of ICMP Address Mask Reply messages received
621 per second [icmpInAddrMaskReps].
622 oadrmk/s
624 The number of ICMP Address Mask Request messages sent per
625 second [icmpOutAddrMasks].
626 oadrmkr/s
628 The number of ICMP Address Mask Reply messages sent per
629 second [icmpOutAddrMaskReps].
630 With the EICMP keyword, statistics about ICMPv4 error messages
632 are reported. Note that ICMPv4 statistics depend on sadc's
633 option "-S SNMP" to be collected. The following values are dis‐
634 played (formal SNMP names between square brackets):
635 ierr/s
637 The number of ICMP messages per second which the entity
638 received but determined as having ICMP-specific errors
639 (bad ICMP checksums, bad length, etc.) [icmpInErrors].
640 oerr/s
642 The number of ICMP messages per second which this entity
643 did not send due to problems discovered within ICMP such
644 as a lack of buffers [icmpOutErrors].
645 idstunr/s
647 The number of ICMP Destination Unreachable messages
648 received per second [icmpInDestUnreachs].
649 odstunr/s
651 The number of ICMP Destination Unreachable messages sent
652 per second [icmpOutDestUnreachs].
653 itmex/s
655 The number of ICMP Time Exceeded messages received per
656 second [icmpInTimeExcds].
657 otmex/s
659 The number of ICMP Time Exceeded messages sent per second
660 [icmpOutTimeExcds].
661 iparmpb/s
663 The number of ICMP Parameter Problem messages received
664 per second [icmpInParmProbs].
665 oparmpb/s
667 The number of ICMP Parameter Problem messages sent per
668 second [icmpOutParmProbs].
669 isrcq/s
671 The number of ICMP Source Quench messages received per
672 second [icmpInSrcQuenchs].
673 osrcq/s
675 The number of ICMP Source Quench messages sent per second
676 [icmpOutSrcQuenchs].
677 iredir/s
679 The number of ICMP Redirect messages received per second
680 [icmpInRedirects].
681 oredir/s
683 The number of ICMP Redirect messages sent per second
684 [icmpOutRedirects].
685 With the ICMP6 keyword, statistics about ICMPv6 network traffic
687 are reported. Note that ICMPv6 statistics depend on sadc's
688 option "-S IPV6" to be collected. The following values are dis‐
689 played (formal SNMP names between square brackets):
690 imsg6/s
692 The total number of ICMP messages received by the inter‐
693 face per second which includes all those counted by
694 ierr6/s [ipv6IfIcmpInMsgs].
695 omsg6/s
697 The total number of ICMP messages which this interface
698 attempted to send per second [ipv6IfIcmpOutMsgs].
699 iech6/s
701 The number of ICMP Echo (request) messages received by
702 the interface per second [ipv6IfIcmpInEchos].
703 iechr6/s
705 The number of ICMP Echo Reply messages received by the
706 interface per second [ipv6IfIcmpInEchoReplies].
707 oechr6/s
709 The number of ICMP Echo Reply messages sent by the inter‐
710 face per second [ipv6IfIcmpOutEchoReplies].
711 igmbq6/s
713 The number of ICMPv6 Group Membership Query messages
714 received by the interface per second [ipv6IfIcmpInGroup‐
715 MembQueries].
716 igmbr6/s
718 The number of ICMPv6 Group Membership Response messages
719 received by the interface per second [ipv6IfIcmpInGroup‐
720 MembResponses].
721 ogmbr6/s
723 The number of ICMPv6 Group Membership Response messages
724 sent per second [ipv6IfIcmpOutGroupMembResponses].
725 igmbrd6/s
727 The number of ICMPv6 Group Membership Reduction messages
728 received by the interface per second [ipv6IfIcmpInGroup‐
729 MembReductions].
730 ogmbrd6/s
732 The number of ICMPv6 Group Membership Reduction messages
733 sent per second [ipv6IfIcmpOutGroupMembReductions].
734 irtsol6/s
736 The number of ICMP Router Solicit messages received by
737 the interface per second [ipv6IfIcmpInRouterSolicits].
738 ortsol6/s
740 The number of ICMP Router Solicitation messages sent by
741 the interface per second [ipv6IfIcmpOutRouterSolicits].
742 irtad6/s
744 The number of ICMP Router Advertisement messages received
745 by the interface per second [ipv6IfIcmpInRouterAdvertise‐
746 ments].
747 inbsol6/s
749 The number of ICMP Neighbor Solicit messages received by
750 the interface per second [ipv6IfIcmpInNeighborSolicits].
751 onbsol6/s
753 The number of ICMP Neighbor Solicitation messages sent by
754 the interface per second [ipv6IfIcmpOutNeighborSolicits].
755 inbad6/s
757 The number of ICMP Neighbor Advertisement messages
758 received by the interface per second [ipv6IfIcmpInNeighb‐
759 orAdvertisements].
760 onbad6/s
762 The number of ICMP Neighbor Advertisement messages sent
763 by the interface per second [ipv6IfIcmpOutNeighborAdver‐
764 tisements].
765 With the EICMP6 keyword, statistics about ICMPv6 error messages
767 are reported. Note that ICMPv6 statistics depend on sadc's
768 option "-S IPV6" to be collected. The following values are dis‐
769 played (formal SNMP names between square brackets):
770 ierr6/s
772 The number of ICMP messages per second which the inter‐
773 face received but determined as having ICMP-specific
774 errors (bad ICMP checksums, bad length, etc.)
775 [ipv6IfIcmpInErrors]
776 idtunr6/s
778 The number of ICMP Destination Unreachable messages
779 received by the interface per second [ipv6IfIcmpInDestUn‐
780 reachs].
781 odtunr6/s
783 The number of ICMP Destination Unreachable messages sent
784 by the interface per second [ipv6IfIcmpOutDestUnreachs].
785 itmex6/s
787 The number of ICMP Time Exceeded messages received by the
788 interface per second [ipv6IfIcmpInTimeExcds].
789 otmex6/s
791 The number of ICMP Time Exceeded messages sent by the
792 interface per second [ipv6IfIcmpOutTimeExcds].
793 iprmpb6/s
795 The number of ICMP Parameter Problem messages received by
796 the interface per second [ipv6IfIcmpInParmProblems].
797 oprmpb6/s
799 The number of ICMP Parameter Problem messages sent by the
800 interface per second [ipv6IfIcmpOutParmProblems].
801 iredir6/s
803 The number of Redirect messages received by the interface
804 per second [ipv6IfIcmpInRedirects].
805 oredir6/s
807 The number of Redirect messages sent by the interface by
808 second [ipv6IfIcmpOutRedirects].
809 ipck2b6/s
811 The number of ICMP Packet Too Big messages received by
812 the interface per second [ipv6IfIcmpInPktTooBigs].
813 opck2b6/s
815 The number of ICMP Packet Too Big messages sent by the
816 interface per second [ipv6IfIcmpOutPktTooBigs].
817 With the IP keyword, statistics about IPv4 network traffic are
819 reported. Note that IPv4 statistics depend on sadc's option "-S
820 SNMP" to be collected. The following values are displayed (for‐
821 mal SNMP names between square brackets):
822 irec/s
824 The total number of input datagrams received from inter‐
825 faces per second, including those received in error
826 [ipInReceives].
827 fwddgm/s
829 The number of input datagrams per second, for which this
830 entity was not their final IP destination, as a result of
831 which an attempt was made to find a route to forward them
832 to that final destination [ipForwDatagrams].
833 idel/s
835 The total number of input datagrams successfully deliv‐
836 ered per second to IP user-protocols (including ICMP)
837 [ipInDelivers].
838 orq/s
840 The total number of IP datagrams which local IP user-pro‐
841 tocols (including ICMP) supplied per second to IP in
842 requests for transmission [ipOutRequests]. Note that
843 this counter does not include any datagrams counted in
844 fwddgm/s.
845 asmrq/s
847 The number of IP fragments received per second which
848 needed to be reassembled at this entity [ipReasmReqds].
849 asmok/s
851 The number of IP datagrams successfully re-assembled per
852 second [ipReasmOKs].
853 fragok/s
855 The number of IP datagrams that have been successfully
856 fragmented at this entity per second [ipFragOKs].
857 fragcrt/s
859 The number of IP datagram fragments that have been gener‐
860 ated per second as a result of fragmentation at this
861 entity [ipFragCreates].
862 With the EIP keyword, statistics about IPv4 network errors are
864 reported. Note that IPv4 statistics depend on sadc's option "-S
865 SNMP" to be collected. The following values are displayed (for‐
866 mal SNMP names between square brackets):
867 ihdrerr/s
869 The number of input datagrams discarded per second due to
870 errors in their IP headers, including bad checksums, ver‐
871 sion number mismatch, other format errors, time-to-live
872 exceeded, errors discovered in processing their IP
873 options, etc. [ipInHdrErrors]
874 iadrerr/s
876 The number of input datagrams discarded per second
877 because the IP address in their IP header's destination
878 field was not a valid address to be received at this
879 entity. This count includes invalid addresses (e.g.,
880 0.0.0.0) and addresses of unsupported Classes (e.g.,
881 Class E). For entities which are not IP routers and
882 therefore do not forward datagrams, this counter includes
883 datagrams discarded because the destination address was
884 not a local address [ipInAddrErrors].
885 iukwnpr/s
887 The number of locally-addressed datagrams received suc‐
888 cessfully but discarded per second because of an unknown
889 or unsupported protocol [ipInUnknownProtos].
890 idisc/s
892 The number of input IP datagrams per second for which no
893 problems were encountered to prevent their continued pro‐
894 cessing, but which were discarded (e.g., for lack of buf‐
895 fer space) [ipInDiscards]. Note that this counter does
896 not include any datagrams discarded while awaiting re-
897 assembly.
898 odisc/s
900 The number of output IP datagrams per second for which no
901 problem was encountered to prevent their transmission to
902 their destination, but which were discarded (e.g., for
903 lack of buffer space) [ipOutDiscards]. Note that this
904 counter would include datagrams counted in fwddgm/s if
905 any such packets met this (discretionary) discard crite‐
906 rion.
907 onort/s
909 The number of IP datagrams discarded per second because
910 no route could be found to transmit them to their desti‐
911 nation [ipOutNoRoutes]. Note that this counter includes
912 any packets counted in fwddgm/s which meet this 'no-
913 route' criterion. Note that this includes any datagrams
914 which a host cannot route because all of its default
915 routers are down.
916 asmf/s
918 The number of failures detected per second by the IP re-
919 assembly algorithm (for whatever reason: timed out,
920 errors, etc) [ipReasmFails]. Note that this is not nec‐
921 essarily a count of discarded IP fragments since some
922 algorithms can lose track of the number of fragments by
923 combining them as they are received.
924 fragf/s
926 The number of IP datagrams that have been discarded per
927 second because they needed to be fragmented at this
928 entity but could not be, e.g., because their Don't Frag‐
929 ment flag was set [ipFragFails].
930 With the IP6 keyword, statistics about IPv6 network traffic are
932 reported. Note that IPv6 statistics depend on sadc's option "-S
933 IPV6" to be collected. The following values are displayed (for‐
934 mal SNMP names between square brackets):
935 irec6/s
937 The total number of input datagrams received from inter‐
938 faces per second, including those received in error
939 [ipv6IfStatsInReceives].
940 fwddgm6/s
942 The number of output datagrams per second which this
943 entity received and forwarded to their final destinations
944 [ipv6IfStatsOutForwDatagrams].
945 idel6/s
947 The total number of datagrams successfully delivered per
948 second to IPv6 user-protocols (including ICMP) [ipv6IfS‐
949 tatsInDelivers].
950 orq6/s
952 The total number of IPv6 datagrams which local IPv6 user-
953 protocols (including ICMP) supplied per second to IPv6 in
954 requests for transmission [ipv6IfStatsOutRequests]. Note
955 that this counter does not include any datagrams counted
956 in fwddgm6/s.
957 asmrq6/s
959 The number of IPv6 fragments received per second which
960 needed to be reassembled at this interface [ipv6IfStat‐
961 sReasmReqds].
962 asmok6/s
964 The number of IPv6 datagrams successfully reassembled per
965 second [ipv6IfStatsReasmOKs].
966 imcpck6/s
968 The number of multicast packets received per second by
969 the interface [ipv6IfStatsInMcastPkts].
970 omcpck6/s
972 The number of multicast packets transmitted per second by
973 the interface [ipv6IfStatsOutMcastPkts].
974 fragok6/s
976 The number of IPv6 datagrams that have been successfully
977 fragmented at this output interface per second [ipv6IfS‐
978 tatsOutFragOKs].
979 fragcr6/s
981 The number of output datagram fragments that have been
982 generated per second as a result of fragmentation at this
983 output interface [ipv6IfStatsOutFragCreates].
984 With the EIP6 keyword, statistics about IPv6 network errors are
986 reported. Note that IPv6 statistics depend on sadc's option "-S
987 IPV6" to be collected. The following values are displayed (for‐
988 mal SNMP names between square brackets):
989 ihdrer6/s
991 The number of input datagrams discarded per second due to
992 errors in their IPv6 headers, including version number
993 mismatch, other format errors, hop count exceeded, errors
994 discovered in processing their IPv6 options, etc.
995 [ipv6IfStatsInHdrErrors]
996 iadrer6/s
998 The number of input datagrams discarded per second
999 because the IPv6 address in their IPv6 header's destina‐
1000 tion field was not a valid address to be received at this
1001 entity. This count includes invalid addresses (e.g., ::0)
1002 and unsupported addresses (e.g., addresses with unallo‐
1003 cated prefixes). For entities which are not IPv6 routers
1004 and therefore do not forward datagrams, this counter
1005 includes datagrams discarded because the destination
1006 address was not a local address [ipv6IfStatsInAddr‐
1007 Errors].
1008 iukwnp6/s
1010 The number of locally-addressed datagrams received suc‐
1011 cessfully but discarded per second because of an unknown
1012 or unsupported protocol [ipv6IfStatsInUnknownProtos].
1013 i2big6/s
1015 The number of input datagrams that could not be forwarded
1016 per second because their size exceeded the link MTU of
1017 outgoing interface [ipv6IfStatsInTooBigErrors].
1018 idisc6/s
1020 The number of input IPv6 datagrams per second for which
1021 no problems were encountered to prevent their continued
1022 processing, but which were discarded (e.g., for lack of
1023 buffer space) [ipv6IfStatsInDiscards]. Note that this
1024 counter does not include any datagrams discarded while
1025 awaiting re-assembly.
1026 odisc6/s
1028 The number of output IPv6 datagrams per second for which
1029 no problem was encountered to prevent their transmission
1030 to their destination, but which were discarded (e.g., for
1031 lack of buffer space) [ipv6IfStatsOutDiscards]. Note that
1032 this counter would include datagrams counted in fwddgm6/s
1033 if any such packets met this (discretionary) discard cri‐
1034 terion.
1035 inort6/s
1037 The number of input datagrams discarded per second
1038 because no route could be found to transmit them to their
1039 destination [ipv6IfStatsInNoRoutes].
1040 onort6/s
1042 The number of locally generated IP datagrams discarded
1043 per second because no route could be found to transmit
1044 them to their destination [unknown formal SNMP name].
1045 asmf6/s
1047 The number of failures detected per second by the IPv6
1048 re-assembly algorithm (for whatever reason: timed out,
1049 errors, etc.) [ipv6IfStatsReasmFails]. Note that this is
1050 not necessarily a count of discarded IPv6 fragments since
1051 some algorithms can lose track of the number of fragments
1052 by combining them as they are received.
1053 fragf6/s
1055 The number of IPv6 datagrams that have been discarded per
1056 second because they needed to be fragmented at this out‐
1057 put interface but could not be [ipv6IfStatsOutFragFails].
1058 itrpck6/s
1060 The number of input datagrams discarded per second
1061 because datagram frame didn't carry enough data [ipv6IfS‐
1062 tatsInTruncatedPkts].
1063 With the NFS keyword, statistics about NFS client activity are
1065 reported. The following values are displayed:
1066 call/s
1068 Number of RPC requests made per second.
1069 retrans/s
1071 Number of RPC requests per second, those which needed to
1072 be retransmitted (for example because of a server time‐
1073 out).
1074 read/s
1076 Number of 'read' RPC calls made per second.
1077 write/s
1079 Number of 'write' RPC calls made per second.
1080 access/s
1082 Number of 'access' RPC calls made per second.
1083 getatt/s
1085 Number of 'getattr' RPC calls made per second.
1086 With the NFSD keyword, statistics about NFS server activity are
1088 reported. The following values are displayed:
1089 scall/s
1091 Number of RPC requests received per second.
1092 badcall/s
1094 Number of bad RPC requests received per second, those
1095 whose processing generated an error.
1096 packet/s
1098 Number of network packets received per second.
1099 udp/s
1101 Number of UDP packets received per second.
1102 tcp/s
1104 Number of TCP packets received per second.
1105 hit/s
1107 Number of reply cache hits per second.
1108 miss/s
1110 Number of reply cache misses per second.
1111 sread/s
1113 Number of 'read' RPC calls received per second.
1114 swrite/s
1116 Number of 'write' RPC calls received per second.
1117 saccess/s
1119 Number of 'access' RPC calls received per second.
1120 sgetatt/s
1122 Number of 'getattr' RPC calls received per second.
1123 With the SOCK keyword, statistics on sockets in use are reported
1125 (IPv4). The following values are displayed:
1126 totsck
1128 Total number of sockets used by the system.
1129 tcpsck
1131 Number of TCP sockets currently in use.
1132 udpsck
1134 Number of UDP sockets currently in use.
1135 rawsck
1137 Number of RAW sockets currently in use.
1138 ip-frag
1140 Number of IP fragments currently in queue.
1141 tcp-tw
1143 Number of TCP sockets in TIME_WAIT state.
1144 With the SOCK6 keyword, statistics on sockets in use are
1146 reported (IPv6). Note that IPv6 statistics depend on sadc's
1147 option "-S IPV6" to be collected. The following values are dis‐
1148 played:
1149 tcp6sck
1151 Number of TCPv6 sockets currently in use.
1152 udp6sck
1154 Number of UDPv6 sockets currently in use.
1155 raw6sck
1157 Number of RAWv6 sockets currently in use.
1158 ip6-frag
1160 Number of IPv6 fragments currently in use.
1161 With the SOFT keyword, statistics about software-based network
1163 processing are reported. The following values are displayed:
1164 total/s
1166 The total number of network frames processed per second.
1167 dropd/s
1169 The total number of network frames dropped per second
1170 because there was no room on the processing queue.
1171 squeezd/s
1173 The number of times the softirq handler function termi‐
1174 nated per second because its budget was consumed or the
1175 time limit was reached, but more work could have been
1176 done.
1177 rx_rps/s
1179 The number of times the CPU has been woken up per second
1180 to process packets via an inter-processor interrupt.
1181 flw_lim/s
1183 The number of times the flow limit has been reached per
1184 second. Flow limiting is an optional RPS feature that
1185 can be used to limit the number of packets queued to the
1186 backlog for each flow to a certain amount. This can help
1187 ensure that smaller flows are processed even though much
1188 larger flows are pushing packets in.
1189 With the TCP keyword, statistics about TCPv4 network traffic are
1191 reported. Note that TCPv4 statistics depend on sadc's option
1192 "-S SNMP" to be collected. The following values are displayed
1193 (formal SNMP names between square brackets):
1194 active/s
1196 The number of times TCP connections have made a direct
1197 transition to the SYN-SENT state from the CLOSED state
1198 per second [tcpActiveOpens].
1199 passive/s
1201 The number of times TCP connections have made a direct
1202 transition to the SYN-RCVD state from the LISTEN state
1203 per second [tcpPassiveOpens].
1204 iseg/s
1206 The total number of segments received per second, includ‐
1207 ing those received in error [tcpInSegs]. This count
1208 includes segments received on currently established con‐
1209 nections.
1210 oseg/s
1212 The total number of segments sent per second, including
1213 those on current connections but excluding those contain‐
1214 ing only retransmitted octets [tcpOutSegs].
1215 With the ETCP keyword, statistics about TCPv4 network errors are
1217 reported. Note that TCPv4 statistics depend on sadc's option
1218 "-S SNMP" to be collected. The following values are displayed
1219 (formal SNMP names between square brackets):
1220 atmptf/s
1222 The number of times per second TCP connections have made
1223 a direct transition to the CLOSED state from either the
1224 SYN-SENT state or the SYN-RCVD state, plus the number of
1225 times per second TCP connections have made a direct tran‐
1226 sition to the LISTEN state from the SYN-RCVD state
1227 [tcpAttemptFails].
1228 estres/s
1230 The number of times per second TCP connections have made
1231 a direct transition to the CLOSED state from either the
1232 ESTABLISHED state or the CLOSE-WAIT state [tcpEstabRe‐
1233 sets].
1234 retrans/s
1236 The total number of segments retransmitted per second -
1237 that is, the number of TCP segments transmitted contain‐
1238 ing one or more previously transmitted octets [tcpRe‐
1239 transSegs].
1240 isegerr/s
1242 The total number of segments received in error (e.g., bad
1243 TCP checksums) per second [tcpInErrs].
1244 orsts/s
1246 The number of TCP segments sent per second containing the
1247 RST flag [tcpOutRsts].
1248 With the UDP keyword, statistics about UDPv4 network traffic are
1250 reported. Note that UDPv4 statistics depend on sadc's option
1251 "-S SNMP" to be collected. The following values are displayed
1252 (formal SNMP names between square brackets):
1253 idgm/s
1255 The total number of UDP datagrams delivered per second to
1256 UDP users [udpInDatagrams].
1257 odgm/s
1259 The total number of UDP datagrams sent per second from
1260 this entity [udpOutDatagrams].
1261 noport/s
1263 The total number of received UDP datagrams per second for
1264 which there was no application at the destination port
1265 [udpNoPorts].
1266 idgmerr/s
1268 The number of received UDP datagrams per second that
1269 could not be delivered for reasons other than the lack of
1270 an application at the destination port [udpInErrors].
1271 With the UDP6 keyword, statistics about UDPv6 network traffic
1273 are reported. Note that UDPv6 statistics depend on sadc's
1274 option "-S IPV6" to be collected. The following values are dis‐
1275 played (formal SNMP names between square brackets):
1276 idgm6/s
1278 The total number of UDP datagrams delivered per second to
1279 UDP users [udpInDatagrams].
1280 odgm6/s
1282 The total number of UDP datagrams sent per second from
1283 this entity [udpOutDatagrams].
1284 noport6/s
1286 The total number of received UDP datagrams per second for
1287 which there was no application at the destination port
1288 [udpNoPorts].
1289 idgmer6/s
1291 The number of received UDP datagrams per second that
1292 could not be delivered for reasons other than the lack of
1293 an application at the destination port [udpInErrors].
1294 The ALL keyword is equivalent to specifying all the keywords
1296 above and therefore all the network activities are reported.
1297 -o [ filename ]
1299 Save the readings in the file in binary form. Each reading is in
1300 a separate record. The default value of the filename parameter
1301 is the current standard system activity daily data file. If
1302 filename is a directory instead of a plain file then it is con‐
1303 sidered as the directory where the standard system activity
1304 daily data files are located. The -o option is exclusive of the
1305 -f option. All the data available from the kernel are saved in
1306 the file (in fact, sar calls its data collector sadc with the
1307 option "-S ALL". See sadc(8) manual page).
1308 -P { cpu_list | ALL }
1310 Report per-processor statistics for the specified processor or
1311 processors. cpu_list is a list of comma-separated values or
1312 range of values (e.g., 0,2,4-7,12-). Note that processor 0 is
1313 the first processor, and processor all is the global average
1314 among all processors. Specifying the ALL keyword reports sta‐
1315 tistics for each individual processor, and globally for all pro‐
1316 cessors. Offline processors are not displayed.
1317 -p Pretty-print device names. Use this option in conjunction with
1319 option -d. By default names are printed as devM-n where M and n
1320 are the major and minor numbers for the device. Use of this
1321 option displays the names of the devices as they (should) appear
1322 in /dev. Name mappings are controlled by /etc/sysconfig/sys‐
1323 stat.ioconf.
1324 -q Report queue length and load averages. The following values are
1326 displayed:
1327 runq-sz
1329 Run queue length (number of tasks waiting for run time).
1330 plist-sz
1332 Number of tasks in the task list.
1333 ldavg-1
1335 System load average for the last minute. The load aver‐
1336 age is calculated as the average number of runnable or
1337 running tasks (R state), and the number of tasks in unin‐
1338 terruptible sleep (D state) over the specified interval.
1339 ldavg-5
1341 System load average for the past 5 minutes.
1342 ldavg-15
1344 System load average for the past 15 minutes.
1345 blocked
1347 Number of tasks currently blocked, waiting for I/O to
1348 complete.
1349 -r [ ALL ]
1351 Report memory utilization statistics. The ALL keyword indicates
1352 that all the memory fields should be displayed. The following
1353 values may be displayed:
1354 kbmemfree
1356 Amount of free memory available in kilobytes.
1357 kbavail
1359 Estimate of how much memory in kilobytes is available for
1360 starting new applications, without swapping. The esti‐
1361 mate takes into account that the system needs some page
1362 cache to function well, and that not all reclaimable slab
1363 will be reclaimable, due to items being in use. The
1364 impact of those factors will vary from system to system.
1365 kbmemused
1367 Amount of used memory in kilobytes (calculated as total
1368 installed memory - kbmemfree - kbbuffers - kbcached -
1369 kbslab).
1370 %memused
1372 Percentage of used memory.
1373 kbbuffers
1375 Amount of memory used as buffers by the kernel in kilo‐
1376 bytes.
1377 kbcached
1379 Amount of memory used to cache data by the kernel in
1380 kilobytes.
1381 kbcommit
1383 Amount of memory in kilobytes needed for current work‐
1384 load. This is an estimate of how much RAM/swap is needed
1385 to guarantee that there never is out of memory.
1386 %commit
1388 Percentage of memory needed for current workload in rela‐
1389 tion to the total amount of memory (RAM+swap). This num‐
1390 ber may be greater than 100% because the kernel usually
1391 overcommits memory.
1392 kbactive
1394 Amount of active memory in kilobytes (memory that has
1395 been used more recently and usually not reclaimed unless
1396 absolutely necessary).
1397 kbinact
1399 Amount of inactive memory in kilobytes (memory which has
1400 been less recently used. It is more eligible to be
1401 reclaimed for other purposes).
1402 kbdirty
1404 Amount of memory in kilobytes waiting to get written back
1405 to the disk.
1406 kbanonpg
1408 Amount of non-file backed pages in kilobytes mapped into
1409 userspace page tables.
1410 kbslab
1412 Amount of memory in kilobytes used by the kernel to cache
1413 data structures for its own use.
1414 kbkstack
1416 Amount of memory in kilobytes used for kernel stack
1417 space.
1418 kbpgtbl
1420 Amount of memory in kilobytes dedicated to the lowest
1421 level of page tables.
1422 kbvmused
1424 Amount of memory in kilobytes of used virtual address
1425 space.
1426 -S Report swap space utilization statistics. The following values
1428 are displayed:
1429 kbswpfree
1431 Amount of free swap space in kilobytes.
1432 kbswpused
1434 Amount of used swap space in kilobytes.
1435 %swpused
1437 Percentage of used swap space.
1438 kbswpcad
1440 Amount of cached swap memory in kilobytes. This is mem‐
1441 ory that once was swapped out, is swapped back in but
1442 still also is in the swap area (if memory is needed it
1443 doesn't need to be swapped out again because it is
1444 already in the swap area. This saves I/O).
1445 %swpcad
1447 Percentage of cached swap memory in relation to the
1448 amount of used swap space.
1449 -s [ hh:mm[:ss] ]
1451 Set the starting time of the data, causing the sar command to
1452 extract records time-tagged at, or following, the time speci‐
1453 fied. The default starting time is 08:00:00. Hours must be
1454 given in 24-hour format. This option can be used only when data
1455 are read from a file (option -f).
1456 --sadc Indicate which data collector is called by sar. If the data
1458 collector is sought in PATH then enter "which sadc" to know
1459 where it is located.
1460 -t When reading data from a daily data file, indicate that sar
1462 should display the timestamps in the original local time of the
1463 data file creator. Without this option, the sar command displays
1464 the timestamps in the user's locale time.
1465 -u [ ALL ]
1467 Report CPU utilization. The ALL keyword indicates that all the
1468 CPU fields should be displayed. The report may show the follow‐
1469 ing fields:
1470 %user
1472 Percentage of CPU utilization that occurred while execut‐
1473 ing at the user level (application). Note that this field
1474 includes time spent running virtual processors.
1475 %usr
1477 Percentage of CPU utilization that occurred while execut‐
1478 ing at the user level (application). Note that this field
1479 does NOT include time spent running virtual processors.
1480 %nice
1482 Percentage of CPU utilization that occurred while execut‐
1483 ing at the user level with nice priority.
1484 %system
1486 Percentage of CPU utilization that occurred while execut‐
1487 ing at the system level (kernel). Note that this field
1488 includes time spent servicing hardware and software
1489 interrupts.
1490 %sys
1492 Percentage of CPU utilization that occurred while execut‐
1493 ing at the system level (kernel). Note that this field
1494 does NOT include time spent servicing hardware or soft‐
1495 ware interrupts.
1496 %iowait
1498 Percentage of time that the CPU or CPUs were idle during
1499 which the system had an outstanding disk I/O request.
1500 %steal
1502 Percentage of time spent in involuntary wait by the vir‐
1503 tual CPU or CPUs while the hypervisor was servicing
1504 another virtual processor.
1505 %irq
1507 Percentage of time spent by the CPU or CPUs to service
1508 hardware interrupts.
1509 %soft
1511 Percentage of time spent by the CPU or CPUs to service
1512 software interrupts.
1513 %guest
1515 Percentage of time spent by the CPU or CPUs to run a vir‐
1516 tual processor.
1517 %gnice
1519 Percentage of time spent by the CPU or CPUs to run a
1520 niced guest.
1521 %idle
1523 Percentage of time that the CPU or CPUs were idle and the
1524 system did not have an outstanding disk I/O request.
1525 -V Print version number then exit.
1527 -v Report status of inode, file and other kernel tables. The fol‐
1529 lowing values are displayed:
1530 dentunusd
1532 Number of unused cache entries in the directory cache.
1533 file-nr
1535 Number of file handles used by the system.
1536 inode-nr
1538 Number of inode handlers used by the system.
1539 pty-nr
1541 Number of pseudo-terminals used by the system.
1542 -W Report swapping statistics. The following values are displayed:
1544 pswpin/s
1546 Total number of swap pages the system brought in per sec‐
1547 ond.
1548 pswpout/s
1550 Total number of swap pages the system brought out per
1551 second.
1552 -w Report task creation and system switching activity.
1554 proc/s
1556 Total number of tasks created per second.
1557 cswch/s
1559 Total number of context switches per second.
1560 -y Report TTY devices activity. The following values are displayed:
1562 rcvin/s
1564 Number of receive interrupts per second for current
1565 serial line. Serial line number is given in the TTY col‐
1566 umn.
1567 xmtin/s
1569 Number of transmit interrupts per second for current
1570 serial line.
1571 framerr/s
1573 Number of frame errors per second for current serial
1574 line.
1575 prtyerr/s
1577 Number of parity errors per second for current serial
1578 line.
1579 brk/s
1581 Number of breaks per second for current serial line.
1582 ovrun/s
1584 Number of overrun errors per second for current serial
1585 line.
1586 -z Tell sar to omit output for any devices for which there was no
1588 activity during the sample period.
1589
1592 The sar command takes into account the following environment variables:
1593 S_COLORS
1596 When this variable is set, display statistics in color on the
1597 terminal. Possible values for this variable are never, always
1598 or auto (the latter is the default).
1599 Please note that the color (being red, yellow, or some other
1601 color) used to display a value is not indicative of any kind of
1602 issue simply because of the color. It only indicates different
1603 ranges of values.
1604 S_COLORS_SGR
1607 Specify the colors and other attributes used to display statis‐
1608 tics on the terminal. Its value is a colon-separated list of
1609 capabilities that defaults to
1610 C=33;22:H=31;1:I=32;22:M=35;1:N=34;1:R=31;22:Z=34;22. Supported
1611 capabilities are:
1612 C= SGR (Select Graphic Rendition) substring for comments
1615 inserted in the binary daily data files.
1616 H= SGR substring for percentage values greater than or equal
1619 to 75%.
1620 I= SGR substring for item names or values (eg. network
1623 interfaces, CPU number...)
1624 M= SGR substring for percentage values in the range from 50%
1627 to 75%.
1628 N= SGR substring for non-zero statistics values.
1631 R= SGR substring for restart messages.
1634 Z= SGR substring for zero values.
1637 S_TIME_DEF_TIME
1640 If this variable exists and its value is UTC then sar will save
1641 its data in UTC time (data will still be displayed in local
1642 time). sar will also use UTC time instead of local time to
1643 determine the current daily data file located in the /var/log/sa
1644 directory. This variable may be useful for servers with users
1645 located across several timezones.
1646 S_TIME_FORMAT
1649 If this variable exists and its value is ISO then the current
1650 locale will be ignored when printing the date in the report
1651 header. The sar command will use the ISO 8601 format (YYYY-MM-
1652 DD) instead. The timestamp will also be compliant with ISO 8601
1653 format.
1654
1656 sar -u 2 5
1657 Report CPU utilization for each 2 seconds. 5 lines are dis‐
1658 played.
1659 sar -I 14 -o int14.file 2 10
1661 Report statistics on IRQ 14 for each 2 seconds. 10 lines are
1662 displayed. Data are stored in a file called int14.file.
1663 sar -r -n DEV -f /var/log/sa/sa16
1665 Display memory and network statistics saved in daily data file
1666 'sa16'.
1667 sar -A
1669 Display all the statistics saved in current daily data file.
1670
1672 /proc filesystem must be mounted for the sar command to work.
1673 All the statistics are not necessarily available, depending on the ker‐
1675 nel version used. sar assumes that you are using at least a 2.6 ker‐
1676 nel.
1677
1679 /var/log/sa/saDD
1680 /var/log/sa/saYYYYMMDD
1681 The standard system activity daily data files and their default
1682 location. YYYY stands for the current year, MM for the current
1683 month and DD for the current day.
1684 /proc and /sys contain various files with system statistics.
1686
1688 Sebastien Godard (sysstat <at> orange.fr)
1689
1691 sadc(8), sa1(8), sa2(8), sadf(1), sysstat(5), pidstat(1), mpstat(1),
1692 iostat(1), vmstat(8)
1693 https://github.com/sysstat/sysstat
1695 http://pagesperso-orange.fr/sebastien.godard/
1697Linux SEPTEMBER 2019 SAR(1)