rddmon(8) rddmon(8)
NAME
rddmon - daemon for monitoring status changes in PXRC and PXRE RAID
systems
SYNOPSIS
rddmon -s [-a cycletime]
rddmon -H [-a cycletime]
rddmon -t
DESCRIPTION
The rddmon RAID monitoring daemon is generally initiated at system
startup with the S72rddmon start script. It cyclically calculates
(default = 300 sec.) various statuses of the different PXRC and PXRE
systems listed in /dev/autoconf.
Status changes are output as console messages and logged in a log
file. Follow-up activities can then be initiated based on the log file
entries using an action script. The overall objective is to provide
information early on for the system support specialist to enable the
problem to be rectified.
The rddmon daemon monitors the following statuses:
- LUN status (optimal/reconstructing/degraded/dead) of the PXRC and
PXRE LUNs listed in /dev/autoconf.
- The components registered by the EMU (Environmental Monitoring
Unit), i.e. power supplies, fans, temperature sensor and UPS (Unin-
terruptible Power Supply).
- The age of the cache battery is also calculated with PXRE systems;
an appropriate message is output shortly before and after the
expected life span has been reached.
OPTIONS
-s Starts the rddmon daemon.
-H Sends a SIGHUP signal to the active rddmon to initiate an immedi-
ate status check, for example following a configuration change.
If the rddmon daemon is not active, it is started with -H; in
this case, the -a option is evaluated.
-a cycletime
Polling cycle of the rddmon daemon; the default is 300 seconds.
Only works with the -s and -H options, if a daemon is not yet
active.
Page 1 Reliant UNIX 5.44 Printed 11/98
rddmon(8) rddmon(8)
-t The active daemon is terminated; this has the same effect as the
kill -9 rddmonPID command, but is more comfortable to use, for
example if you want to vary the cycle time for the purpose of
testing.
SIGNALS
The following signals can be used to influence the rddmon daemon dur-
ing the life cycle:
SIGHUP
Initiates an immediate status check, see the -H option. If the
rddmon daemon is not yet active, it is started; in this case, the
-a option is also evaluated for setting the cycle time.
SIGTERM
Terminates the active daemon, see the -t option. This is only
useful if you want to vary the cycle time for testing purposes.
OUTPUT
A complete and up-to-date list of log file entries generated by
rddmon, including a description of errors and recommended actions, can
be created with the
getemm -c rdd
command [see getemm(1M)].
Individual error descriptions can be queried with the
getemm rdd:number
command.
You should use SYSADM to view the log file.
EXIT STATUS
0 Normal termination of parent process which starts the daemon and
termination of the daemon following rddmon -t.
1 Abnormal termination:
- The daemon could not be started
- Nothing to monitor in /dev/autoconf
- With rddmon -t: No daemon active
2 Parameter error, user is not root.
Page 2 Reliant UNIX 5.44 Printed 11/98
rddmon(8) rddmon(8)
FILES
This device node contains the RAID systems to be monitored:
/dev/autoconf
Action description file for LOG3:
/opt/log3/filters/events.rdd
Message catalogs for the getemm command:
/usr/lib/locale/ENUS.ASCII/LCMESSAGES/SIemM/rdd.a
/usr/lib/locale/ENUS.ASCII/LCMESSAGES/SIemM/rdd.d
/usr/lib/locale/ENUS.ASCII/LCMESSAGES/SIemM/rdd.n
/usr/lib/locale/ENUS.ASCII/LCMESSAGES/SIemM/rdd.r
/usr/lib/locale/ENUS.ASCII/LCMESSAGES/SIemM/rdd.t
SEE ALSO
getemm(1M).
Page 3 Reliant UNIX 5.44 Printed 11/98