cmvxserviced(1m)cmvxserviced(1m)NAMEcmvxserviced - monitor VxVM and CVM volumes for a high availability
package.
SYNOPSIScmvxserviced [-h|-v] [-O log_file] [-D log_level] [-t poll_interval]
volume_path...
DESCRIPTIONcmvxserviced monitors VxVM and CVM volumes. It runs as a service within
a package that depends on the monitored storage. When a monitored vol‐
ume fails or becomes inaccessible (see Scope of Monitoring), the ser‐
vice will exit, causing the package to fail on the current node. The
package's failover behavior depends on its configured settings.
cmvxserviced periodically probes each volume named by volume_path,
which must identify a block device file.
Scope of Monitoring
The VxVM Volume Monitor detects the following failures:
o Failure of the last link to a storage device or set of devices
critical to volume operation
o Failure of a storage device or set of devices critical to vol‐
ume operation
o An unexpected detachment or disablement of a volume
The VxVM Volume Monitor does not detect the following failures:
o Failure of a redundant link to a storage device or set of
devices if a functioning link remains
o Failure of a mirrored plex within a volume (assuming at least
one plex is functional)
o Corruption of data on a volume which VxVM or CVM regards as
enabled and active
Options
cmvxserviced supports the following options:
-h Displays the usage, as listed above, and
exits.
-v Displays the monitor version and exits.
NOTE Do not include the -h or -v parameters in
your service command; this will result in
immediate package failure at runtime.
-O log_file Specifies a file for logging (log messages are
printed to the console by default).
-D log_level Specifies the log level. The level of detail
logged is directly proportional to the numeri‐
cal value of the log level. That is, a log
level of 7 will provide the greatest amount of
log information. The default log level is 0.
-t poll_interval Specifies the interval between volume probes.
You can specify a polling interval of as lit‐
tle as 1 (one second), but bear in mind that a
short polling interval (less than 10 seconds)
may impair system performance if you are moni‐
toring a large number of volumes. HP recom‐
mends a polling interval of at least 10 sec‐
onds if 50 or more volumes are being monitored
by a single service command.
The default polling interval is 60 seconds.
EXAMPLES
/usr/sbin/cmvxserviced -O /pkg1/monlog.log -D 3
/dev/vx/dsk/cvm_dg0/lvol2
This command monitors a single volume, /dev/vx/dsk/cvm_dg0/lvol2, at
log level 3, with a polling interval of 60 seconds, and prints all log
messages to the file /pkg1/monlog.log.
There should be a one to one relationship between monitoring services
and log files. This provides a means to correlate log messages with the
originating monitor.
/usr/sbin/cmvxserviced /dev/vx/dsk/cvm_dg0/lvol1
dev/vx/dsk/cvm_dg0/lvol2
This command monitors two volumes at the default log level of 0, with a
polling interval of 60 seconds, and prints all log messages to the con‐
sole.
/usr/sbin/cmvxserviced -t 10 /dev/vx/dsk/cvm_dg2/lvol3
This command monitors a single volume at log level 0, with a polling
interval of 10 seconds, and prints all log messages to the console.
AUTHORcmvxserviced was developed by HP.
Requires Optional Serviceguard Software cmvxserviced(1m)