1SMARTD(8) SMART Monitoring Tools SMARTD(8)
2
6 smartd - SMART Disk Monitoring Daemon
7
10 smartd [options]
11
14 [This man page is generated for the Linux version of smartmontools. It
15 does not contain info specific to other platforms.]
16 smartd is a daemon that monitors the Self-Monitoring, Analysis and Re‐
18 porting Technology (SMART) system built into most ATA/SATA and SCSI/SAS
19 hard drives and solid-state drives. The purpose of SMART is to monitor
20 the reliability of the hard drive and predict drive failures, and to
21 carry out different types of drive self-tests. This version of smartd
22 is compatible with ACS-3, ACS-2, ATA8-ACS, ATA/ATAPI-7 and earlier
23 standards (see REFERENCES below).
24 smartd will attempt to enable SMART monitoring on ATA devices (equiva‐
26 lent to smartctl -s on) and polls these and SCSI devices every 30 min‐
27 utes (configurable), logging SMART errors and changes of SMART At‐
28 tributes via the SYSLOG interface. The default location for these SYS‐
29 LOG notifications and warnings is system-dependent (typically
30 /var/log/messages or /var/log/syslog). To change this default loca‐
31 tion, please see the '-l' command-line option described below.
32 In addition to logging to a file, smartd can also be configured to send
34 email warnings if problems are detected. Depending upon the type of
35 problem, you may want to run self-tests on the disk, back up the disk,
36 replace the disk, or use a manufacturer's utility to force reallocation
37 of bad or unreadable disk sectors. If disk problems are detected,
38 please see the smartctl manual page and the smartmontools web page/FAQ
39 for further guidance.
40 If you send a USR1 signal to smartd it will immediately check the sta‐
42 tus of the disks, and then return to polling the disks every 30 min‐
43 utes. See the '-i' option below for additional details.
44 smartd can be configured at start-up using the configuration file
46 /etc/smartmontools/smartd.conf (Windows: EXEDIR/smartd.conf). If the
47 configuration file is subsequently modified, smartd can be told to re-
48 read the configuration file by sending it a HUP signal, for example
49 with the command:
50 killall -HUP smartd.
51 On startup, if smartd finds a syntax error in the configuration file,
53 it will print an error message and then exit. However if smartd is al‐
54 ready running, then is told with a HUP signal to re-read the configura‐
55 tion file, and then find a syntax error in this file, it will print an
56 error message and then continue, ignoring the contents of the (faulty)
57 configuration file, as if the HUP signal had never been received.
58 When smartd is running in debug mode, the INT signal (normally gener‐
60 ated from a shell with CONTROL-C) is treated in the same way as a HUP
61 signal: it makes smartd reload its configuration file. To exit smartd
62 use CONTROL-\.
63 [Linux only] If smartd is started as a systemd(1) service and 'Type=No‐
65 tify' is specified in the service file, the service manager is notified
66 after successful startup. Other state changes are reported via systemd
67 notify STATUS messages. Notification of successful reloads (after HUP
68 signal) is not supported. To detect this process start-up type, smartd
69 checks whether the environment variable 'NOTIFY_SOCKET' is set. Note
70 that it is required to set the '-n' ('--nofork') option in the 'Exec‐
71 Start=/usr/sbin/smartd' command line if 'Type=Notify' is used.
72 On startup, in the absence of the configuration file /etc/smartmon‐
74 tools/smartd.conf, the smartd daemon first scans for all devices that
75 support SMART. The scanning is done as follows:
76 LINUX: Examine all entries "/dev/hd[a-t]" for IDE/ATA devices, and
78 "/dev/sd[a-z]", "/dev/sd[a-z][a-z]" for ATA/SATA or SCSI/SAS
79 devices. Disks behind RAID controllers are not included.
80 If directive '-d nvme' or no '-d' directive is specified, ex‐
82 amine all entries "/dev/nvme[0-99]" for NVMe devices.
83 smartd then monitors for all possible SMART errors (corresponding to
85 the '-a' Directive in the configuration file; see the smartd.conf(5)
86 man page).
87
90 -A PREFIX, --attributelog=PREFIX
91 Writes smartd attribute information (normalized and raw attri‐
92 bute values) to files 'PREFIX''MODEL-SERIAL.ata.csv' or 'PRE‐
93 FIX''VENDOR-MODEL-SERIAL.scsi.csv'. At each check cycle at‐
94 tributes are logged as a line of semicolon separated triplets of
95 the form "attribute-ID;attribute-norm-value;attribute-raw-
96 value;". For SCSI devices error counters and temperature
97 recorded in the form "counter-name;counter-value;". Each line
98 is led by a date string of the form "yyyy-mm-dd HH:MM:SS" (in
99 UTC).
100 MODEL and SERIAL are build from drive identify information, in‐
102 valid characters are replaced by underline.
103 If the PREFIX has the form '/path/dir/' (e.g.
105 '/var/lib/smartd/'), then files 'MODEL-SERIAL.ata.csv' are cre‐
106 ated in directory '/path/dir'. If the PREFIX has the form
107 '/path/name' (e.g. '/var/lib/misc/attrlog-'), then files 'nameM‐
108 ODEL-SERIAL.ata.csv' are created in directory '/path/'. The
109 path must be absolute, except if debug mode is enabled.
110 -B [+]FILE, --drivedb=[+]FILE
112 [ATA only] Read the drive database from FILE. The new database
113 replaces the built in database by default. If '+' is specified,
114 then the new entries prepend the built in entries. Please see
115 the smartctl(8) man page for further details.
116 -c FILE, --configfile=FILE
118 Read smartd configuration Directives from FILE, instead of from
119 the default location /etc/smartmontools/smartd.conf (Windows:
120 EXEDIR/smartd.conf). If FILE does not exist, then smartd will
121 print an error message and exit with nonzero status. Thus, '-c
122 /etc/smartmontools/smartd.conf' can be used to verify the exis‐
123 tence of the default configuration file.
124 By using '-' for FILE, the configuration is read from standard
126 input. This is useful for commands like:
127 echo /dev/sdb -m user@home -M test | smartd -c - -q onecheck
128 to perform quick and simple checks without a configuration file.
129 -C, --capabilities[=mail]
131 [Linux only] Use libcap-ng to drop unneeded Linux process capa‐
132 bilities(7). The following capabilities are kept in the effec‐
133 tive and permissive sets: CAP_SYS_ADMIN, CAP_SYS_RAWIO,
134 CAP_MKNOD. If the '-u, --warn_as_user' option (see below) is
135 used with a non-privileged user or group, the following capabil‐
136 ities are also kept: CAP_SETGID, CAP_SETUID. The capability
137 bounding set is cleared.
138 [NEW EXPERIMENTAL SMARTD FEATURE] Mail notification is no longer
140 suppressed if capabilities are dropped. It depends on the local
141 MTA whether mail could be send from a root process with all ca‐
142 pabilities dropped. It works with the postfix MTA.
143 If '--capabilities=mail' is specified, the following capabili‐
145 ties are added to the bounding set: CAP_SETGID, CAP_SETUID,
146 CAP_CHOWN, CAP_FOWNER, CAP_DAC_OVERRIDE. This allows one to
147 send mail with the exim MTA.
148 -d, --debug
150 Runs smartd in "debug" mode. In this mode, it displays status
151 information to STDOUT rather than logging it to SYSLOG and does
152 not fork(2) into the background and detach from the controlling
153 terminal. In this mode, smartd also prints more verbose infor‐
154 mation about what it is doing than when operating in "daemon"
155 mode. In this mode, the INT signal (normally generated from a
156 terminal with CONTROL-C) makes smartd reload its configuration
157 file. Please use CONTROL-\ to exit
158 -D, --showdirectives
160 Prints a list (to STDOUT) of all the possible Directives which
161 may appear in the configuration file /etc/smartmon‐
162 tools/smartd.conf, and then exits. These Directives are de‐
163 scribed in the smartd.conf(5) man page. They may appear in the
164 configuration file following the device name.
165 -h, --help, --usage
167 Prints usage message to STDOUT and exits.
168 -i N, --interval=N
170 Sets the interval between disk checks to N seconds, where N is a
171 decimal integer. The minimum allowed value is ten and the maxi‐
172 mum is the largest positive integer that can be represented on
173 your system (often 2^31-1). The default is 1800 seconds.
174 [NEW EXPERIMENTAL SMARTD FEATURE] The interval could be overrid‐
175 den with the '-c i=N' directive, see smartd.conf(5) man page.
176 Note that the superuser can make smartd check the status of the
178 disks at any time by sending it the SIGUSR1 signal, for example
179 with the command:
180 kill -SIGUSR1 <pid>
181 where <pid> is the process id number of smartd. One may also
182 use:
183 killall -USR1 smartd
184 for the same purpose.
185 -l FACILITY, --logfacility=FACILITY
187 Uses syslog facility FACILITY to log the messages from smartd.
188 Here FACILITY is one of local0, local1, ..., local7, or daemon
189 [default]. If this command-line option is not used, then by de‐
190 fault messages from smartd are logged to the facility daemon.
191 If you would like to have smartd messages logged somewhere other
193 than the default location, include (for example) '-l local3' in
194 its start up argument list. Tell the syslog daemon to log all
195 messages from facility local3 to (for example)
196 '/var/log/smartd.log'.
197 For more detailed information, please refer to the man pages for
199 the local syslog daemon, typically syslogd(8), syslog-ng(8) or
200 rsyslogd(8).
201 -n, --no-fork
203 Do not fork into background; this is useful when executed from
204 modern init methods like initng, minit, supervise or systemd.
205 -p NAME, --pidfile=NAME
207 Writes pidfile NAME containing the smartd Process ID number
208 (PID). To avoid symlink attacks make sure the directory to
209 which pidfile is written is only writable for root. Without
210 this option, or if the --debug option is given, no PID file is
211 written on startup. If smartd is killed with a maskable signal
212 then the pidfile is removed.
213 -q WHEN, --quit=WHEN
215 Specifies when, if ever, smartd should exit. The valid argu‐
216 ments are to this option are:
217 nodev - Exit if there are no devices to monitor, or if any er‐
219 rors are found at startup in the configuration file. This is
220 the default.
221 errors - Exit if there are no devices to monitor, or if any er‐
223 rors are found in the configuration file /etc/smartmon‐
224 tools/smartd.conf at startup or whenever it is reloaded.
225 nodevstartup - Exit if there are no devices to monitor at
227 startup. But continue to run if no devices are found whenever
228 the configuration file is reloaded.
229 never - Only exit if a fatal error occurs (no remaining system
231 memory, invalid command line arguments). In this mode, even if
232 there are no devices to monitor, or if the configuration file
233 /etc/smartmontools/smartd.conf has errors, smartd will continue
234 to run, waiting to load a configuration file listing valid de‐
235 vices.
236 nodev0 - [NEW EXPERIMENTAL SMARTD FEATURE] Same as 'nodev', ex‐
238 cept that the exit status is 0 if there are no devices to moni‐
239 tor.
240 nodev0startup - [NEW EXPERIMENTAL SMARTD FEATURE] Same as 'node‐
242 vstartup', except that the exit status is 0 if there are no de‐
243 vices to monitor.
244 errors,nodev0 - [NEW EXPERIMENTAL SMARTD FEATURE] Same as 'er‐
246 rors', except that the exit status is 0 if there are no devices
247 to monitor.
248 onecheck - Start smartd in debug mode, then register devices,
250 then check device's SMART status once, and then exit with zero
251 exit status if all of these steps worked correctly.
252 This last option is intended for 'distribution-writers' who want
254 to create automated scripts to determine whether or not to auto‐
255 matically start up smartd after installing smartmontools. After
256 starting smartd with this command-line option, the distribu‐
257 tion's install scripts should wait a reasonable length of time
258 (say ten seconds). If smartd has not exited with zero status by
259 that time, the script should send smartd a SIGTERM or SIGKILL
260 and assume that smartd will not operate correctly on the host.
261 Conversely, if smartd exits with zero status, then it is safe to
262 run smartd in normal daemon mode. If smartd is unable to moni‐
263 tor any devices or encounters other problems then it will return
264 with non-zero exit status.
265 showtests - Start smartd in debug mode, then register devices,
267 then write a list of future scheduled self tests to stdout, and
268 then exit with zero exit status if all of these steps worked
269 correctly. Device's SMART status is not checked.
270 This option is intended to test whether the '-s REGEX' direc‐
272 tives in smartd.conf will have the desired effect. The output
273 lists the next test schedules, limited to 5 tests per type and
274 device. This is followed by a summary of all tests of each de‐
275 vice within the next 90 days.
276 -r TYPE, --report=TYPE
278 Intended primarily to help smartmontools developers understand
279 the behavior of smartmontools on non-conforming or poorly-con‐
280 forming hardware. This option reports details of smartd trans‐
281 actions with the device. The option can be used multiple times.
282 When used just once, it shows a record of the ioctl() transac‐
283 tions with the device. When used more than once, the detail of
284 these ioctl() transactions are reported in greater detail. The
285 valid arguments to this option are:
286 ioctl - report all ioctl() transactions.
288 ataioctl - report only ioctl() transactions with ATA devices.
290 scsiioctl - report only ioctl() transactions with SCSI devices.
292 nvmeioctl - report only ioctl() transactions with NVMe devices.
294 Any argument may include a positive integer to specify the level
296 of detail that should be reported. The argument should be fol‐
297 lowed by a comma then the integer with no spaces. For example,
298 ataioctl,2 The default level is 1, so '-r ataioctl,1' and '-r
299 ataioctl' are equivalent.
300 -s PREFIX, --savestates=PREFIX
302 Reads/writes smartd state information from/to files 'PRE‐
303 FIX''MODEL-SERIAL.ata.state' or 'PREFIX''VENDOR-MODEL-SE‐
304 RIAL.scsi.state'. This preserves SMART attributes, drive min
305 and max temperatures (-W directive), info about last sent warn‐
306 ing email (-m directive), and the time of next check of the
307 self-test REGEXP (-s directive) across boot cycles.
308 MODEL and SERIAL are build from drive identify information, in‐
310 valid characters are replaced by underline.
311 If the PREFIX has the form '/path/dir/' (e.g.
313 '/var/lib/smartd/'), then files 'MODEL-SERIAL.ata.state' are
314 created in directory '/path/dir'. If the PREFIX has the form
315 '/path/name' (e.g. '/var/lib/misc/smartd-'), then files 'nameMO‐
316 DEL-SERIAL.ata.state' are created in directory '/path/'. The
317 path must be absolute, except if debug mode is enabled.
318 The state information files are read on smartd startup. The
320 files are always (re)written after reading the configuration
321 file, before rereading the configuration file (SIGHUP), before
322 smartd shutdown, and after a check forced by SIGUSR1. After a
323 normal check cycle, a file is only rewritten if an important
324 change (which usually results in a SYSLOG output) occurred.
325 -w PATH, --warnexec=PATH
327 Run the executable PATH instead of the default script when
328 smartd needs to send warning messages. PATH must point to an
329 executable binary file or script. The default script is
330 /etc/smartmontools/smartd_warning.sh.
331 -u USER[:GROUP], --warn-as-user=USER[:GROUP]
333 [NEW EXPERIMENTAL SMARTD FEATURE] Run the warning script as a
334 non-privileged user instead of root. The USER and optional
335 GROUP may be specified as numeric ids or names. If no GROUP is
336 specified, the default group of USER is used instead.
337 If a warning occurs, a child process is created with fork(2).
339 This process closes all inherited file descriptors, connects
340 stdio to /dev/null, changes the user and group ids, removes any
341 supplementary group ids and then calls the popen(3) function
342 from the standard library.
343 If '0:0' is specified, user and group are not changed, but the
345 remaining actions still apply.
346 If '-' is specified, popen(3) is called directly. This is the
348 default.
349 -V, --version, --license, --copyright
351 Prints version, copyright, license, home page and SVN revision
352 information for your copy of smartd to STDOUT and then exits.
353
356 smartd
357 Runs the daemon in forked mode. This is the normal way to run smartd.
358 Entries are logged to SYSLOG.
359 smartd -d -i 30
361 Run in foreground (debug) mode, checking the disk status every 30 sec‐
362 onds.
363 smartd -q onecheck
365 Registers devices, and checks the status of the devices exactly once.
366 The exit status (the shell $? variable) will be zero if all went well,
367 and nonzero if no devices were detected or some other problem was en‐
368 countered.
369
372 The syntax of the smartd.conf(5) file is discussed separately.
373
376 smartd will make log entries at loglevel LOG_INFO if the Normalized
377 SMART Attribute values have changed, as reported using the '-t', '-p',
378 or '-u' Directives. For example:
379 'Device: /dev/sda, SMART Attribute: 194 Temperature_Celsius changed
380 from 94 to 93'
381 Note that in this message, the value given is the 'Normalized' not the
382 'Raw' Attribute value (the disk temperature in this case is about 22
383 Celsius). The '-R' and '-r' Directives modify this behavior, so that
384 the information is printed with the Raw values as well, for example:
385 'Device: /dev/sda, SMART Attribute: 194 Temperature_Celsius changed
386 from 94 [Raw 22] to 93 [Raw 23]'
387 Here the Raw values are the actual disk temperatures in Celsius. The
388 way in which the Raw values are printed, and the names under which the
389 Attributes are reported, is governed by the various '-v Num,Descrip‐
390 tion' Directives described previously.
391 Please see the smartctl manual page for further explanation of the dif‐
393 ferences between Normalized and Raw Attribute values.
394 smartd will make log entries at loglevel LOG_CRIT if a SMART Attribute
396 has failed, for example:
397 'Device: /dev/sdc, Failed SMART Attribute: 5 Reallocated_Sector_Ct'
398 This loglevel is used for reporting enabled by the '-H', -f',
399 '-l selftest', and '-l error' Directives. Entries reporting failure of
400 SMART Prefailure Attributes should not be ignored: they mean that the
401 disk is failing. Use the smartctl utility to investigate.
402
405 When smartd makes log entries, these are time-stamped. The time stamps
406 are in the computer's local time zone, which is generally set using ei‐
407 ther the environment variable 'TZ' or using a time-zone file such as
408 /etc/localtime. You may wish to change the timezone while smartd is
409 running (for example, if you carry a laptop to a new time-zone and
410 don't reboot it). Due to a bug in the tzset(3) function of many unix
411 standard C libraries, the time-zone stamps of smartd might not change.
412 For some systems, smartd will work around this problem if the time-zone
413 is set using /etc/localtime. The work-around fails if the time-zone is
414 set using the 'TZ' variable (or a file that it points to).
415
418 The exit status (return value) of smartd can have the following values:
419 0: Daemon startup successful, or smartd was killed by a SIGTERM (or
421 in debug mode, a SIGQUIT).
422 1: Commandline did not parse.
424 2: There was a syntax error in the config file.
426 3: Forking the daemon failed.
428 4: Couldn't create PID file.
430 5: Config file does not exist (only returned in conjunction with
432 the '-c' option).
433 6: Config file exists, but cannot be read.
435 8: smartd ran out of memory during startup.
437 10: An inconsistency was found in smartd's internal data structures.
439 This should never happen. It must be due to either a coding or
440 compiler bug. Please report such failures to smartmontools de‐
441 velopers, see REPORTING BUGS below.
442 16: A device explicitly listed in /etc/smartmontools/smartd.conf
444 can't be monitored.
445 17: smartd didn't find any devices to monitor.
447 [NEW EXPERIMENTAL SMARTD FEATURE] This could be changed to 0
448 (success) with one of the '-q *nodev0*' options, see above.
449 254: When in daemon mode, smartd received a SIGINT or SIGQUIT. (Note
451 that in debug mode, SIGINT has the same effect as SIGHUP, and
452 makes smartd reload its configuration file. SIGQUIT has the
453 same effect as SIGTERM and causes smartd to exit with zero exit
454 status.
455 132 and above
457 smartd was killed by a signal that is not explicitly listed
458 above. The exit status is then 128 plus the signal number. For
459 example if smartd is killed by SIGKILL (signal 9) then the exit
460 status is 137.
461
464 /usr/sbin/smartd
465 full path of this executable.
466 /etc/smartmontools/smartd.conf
468 configuration file (see smartd.conf(5) man page).
469 /etc/smartmontools/smartd_warning.sh
471 script run on warnings (see '-w' option above and '-M exec' di‐
472 rective on smartd.conf(5) man page).
473 /etc/smartmontools/smartd_warning.d/
475 plugin directory for smartd warning script (see '-m' directive
476 on smartd.conf(5) man page).
477 /usr/share/smartmontools/drivedb.h
479 drive database (see '-B' option).
480 /etc/smartmontools/smart_drivedb.h
482 optional local drive database (see '-B' option).
483
486 Bruce Allen (project initiator),
487 Christian Franke (project manager, Windows port and all sort of
488 things),
489 Douglas Gilbert (SCSI subsystem),
490 Volker Kuhlmann (moderator of support and database mailing list),
491 Gabriele Pohl (wiki & development team support),
492 Alex Samorukov (FreeBSD port and more, new Trac wiki).
493 Many other individuals have made contributions and corrections, see AU‐
495 THORS, ChangeLog and repository files.
496 The first smartmontools code was derived from the smartsuite package,
498 written by Michael Cornwell and Andre Hedrick.
499
502 To submit a bug report, create a ticket in smartmontools wiki:
503 <https://www.smartmontools.org/>.
504 Alternatively send the info to the smartmontools support mailing list:
505 <https://listi.jpberlin.de/mailman/listinfo/smartmontools-support>.
506
509 smartd.conf(5), smartctl(8).
510 update-smart-drivedb(8).
511 systemd.exec(5).
512
515 Please see the following web site for more info: <https://www.smartmon‐
516 tools.org/>
517 An introductory article about smartmontools is Monitoring Hard Disks
519 with SMART, by Bruce Allen, Linux Journal, January 2004, pages 74–77.
520 See <https://www.linuxjournal.com/article/6983>.
521 If you would like to understand better how SMART works, and what it
523 does, a good place to start is with Sections 4.8 and 6.54 of the first
524 volume of the 'AT Attachment with Packet Interface-7' (ATA/ATAPI-7)
525 specification Revision 4b. This documents the SMART functionality
526 which the smartmontools utilities provide access to.
527 The functioning of SMART was originally defined by the SFF-8035i revi‐
529 sion 2 and the SFF-8055i revision 1.4 specifications. These are publi‐
530 cations of the Small Form Factors (SFF) Committee.
531 Links to these and other documents may be found on the Links page of
533 the smartmontools Wiki at <https://www.smartmontools.org/wiki/Links>.
534
537 smartmontools-7.3 2022-02-28 r5338
538 $Id: smartd.8.in 5333 2022-02-26 00:15:22Z dpgilbert $
539smartmontools-7.3 2022-02-28 SMARTD(8)