|
This manual page is for Mac OS X Server version 10.6.3If you are running Mac OS X (client), this command is not available. If you are running a different version of Mac OS X Server, view the documentation locally:
Reading manual pagesManual pages are intended as a quick reference for people who already understand a technology.
|
watchdogtimerd(8) System Manager's Manual watchdogtimerd(8) NAME watchdogtimerd -- Mac OS X Server automatic reboot daemon (hardware watchdog tickler) SYNOPSIS watchdogtimerd [-d | -n | -R] [-i interval] [-c count] [-L facility] watchdogtimerd [-h | -v] DESCRIPTION watchdogtimerd is indirectly responsible for rebooting the server hardware if the machine hangs or pan-ics. panics. ics. When watchdogtimerd runs on supported hardware, it periodically resets the hardware watchdog timer. If the timer ever expires, the power management unit forces a hard reboot. Automatic reboot is disabled when watchdogtimerd quits cleanly, so it is imperative watchdogtimerd be terminated by sending a termination signal (SIGTERM), NOT a kill signal (SIGKILL)! watchdogtimerd is typically launched at boot time under the control of launchd(8) using the job label com.apple.watchdogtimerd. OPTIONS The following options are available: -c count --cycles=count Reboot the machine if the timer has not been reset for count intervals. This option is primarily used when debugging. Valid range of values is 3 through 480. Default is 10 with most configurations, resulting in a reboot within five minutes after failure. If the kernel's boot-args contained the "_panicd_ip" directive to enable panic dumps (as noted in the referenced Tech Note below), the default is raised to 30, resulting in a reboot after fifteen minutes to allow enough time for the kernel to send the dump information. -d --debug Print log strings to the terminal. -h --help Print usage summary and exit. -i interval --interval=interval Reset the watchdog timer every interval seconds. This option is primarily used when debugging. Valid range of values is 1 through 30. Default is 30. -L facility Enable logging to syslog(8) using the specified facility. The following values for facility are supported: daemon (the default), user, local[_-7]. (See syslog.conf(5) for their meanings.) This is equivalent to --syslog. -n --no-run Print log information to the terminal and quit after validating the hardware watchdog timer. -R --reboot Force a hard reboot by initializing the watchdog timer to (count * interval) sec-onds seconds onds and terminating. --syslog=facility Enable logging to syslog(8) using the specified facility. The following values for facility are supported: daemon (the default), auth, user, local[_-7]. (See syslog.conf(5) for their meanings.) This is equivalent to -L. -v --version Print build version and exit. FILES & FOLDERS /usr/sbin/watchdogtimerd - daemon /System/Library/LaunchDaemons/com.apple.watchdogtimerd.plist - launchd(8) configuration file. DIAGNOSTICS The watchdogtimerd utility exits 0 on success, and >0 if an error occurs. The following table lists possible exit codes, their equivalent constants in sysexits(3), and their corresponding failure condi-tion. condition. tion. code constant failure condition 64 EX_USAGE invalid arguments 69 EX_UNAVAILABLE automatic reboot timer unavailable 77 EX_NOPERM insufficient permissions Could not access automatic reboot timer. The executing machine does not support a hardware watchdog timer. Invalid interval. Valid intervals range from 1 to 30. Invalid number of cycles. Valid cycles range from 3 to 480. Unable to raise process priority. watchdogtimerd was unable to raise its priority. This may result in inadvertant reboots if other (typically real-time) processes consume all available cycles. watchdogtimerd must be executed as root. Because watchdogtimerd controls system-wide resources, it must be executed by the root user. SEE ALSO sysexits(3), syslog.conf(5), launchctl(8), launchd(8), syslogd(8) Apple's technical note TN2118: Kernel Core Dumps <http://developer.apple.com/technotes/tn2004/ tn2118.html> describes how to set up a crash dump server and how to send dumps from panicked machines. These instructions allow a site to record kernel panics for debugging or analysis while leveraging the higher service availability provided by watchdogtimerd. This approach is recommended for most sites instead of increasing the time before forcing a reboot. (As noted in the OPTIONS section, the minimum reboot time is extended if such dumping is configured.) KNOWN ISSUES None identified at this time. HISTORY watchdogtimerd replaces the hardware watchdog timer servicing of watchdog(8). The pressing question used to be, "Who watches the watchmen?" Now we know: launchd(8). Mac OS X Server 10.6 May 12, 2008 Mac OS X Server 10.6 |
The way to report a problem with this manual page depends on the type of problem: