Managing Serviceguard A.11.20, March 2013

About the Volume Monitor
Simply monitoring each physical disk in a Serviceguard cluster does not provide adequate
monitoring for volumes managed by Veritas Volume Manager from Symantec (VxVM) or logical
volumes managed by HP-UX Logical Volume Manager (LVM), because a physical volume failure
is not always a critical failure that triggers failover (for example, the failure of a single physical
disk within a mirrored volume is not considered critical). For this reason, it can be very difficult to
determine which physical disks must be monitored to ensure that a storage volume is functioning
properly. The HP Serviceguard Volume Monitor provides a means for simple and effective monitoring
of storage volumes.
IMPORTANT: Check the latest version of the release notes (at the address given in the preface
to this manual) for information about Serviceguard support for the volume monitor.
Using the Volume Monitor
NOTE: For LVM, using this monitor is an alternative to using Event Monitoring Service (EMS)
resource dependencies or System Fault Management/custom defined monitors in Serviceguard
using generic resources. EMS does not currently provide a monitor for VxVM.
• See “Using EMS to Monitor Volume Groups” (page 106).
• See “Using Generic Resources to Monitor Volume Groups” (page 106).
Configure the Volume Monitor as a service in a package that requires access to a VxVM or LVM
storage volume.
The package can be a failover package or multi-node package. For example, you can configure
the monitor as a service in a failover package to monitor a storage volume (or multiple storage
volumes) required by the package application. Because the root, boot, and swap volumes are
critical to the functioning of the node , the service should be configured with
service_fail_fast_enabled (page 248) set to yes.
When a monitored volume fails or becomes inaccessible, the monitor service will exit, causing the
package to fail on the current node. The package’s failure behavior depends on its configured
settings. For prompt recovery, HP recommends setting the value of service_restart (page 248)
for the monitoring service to none.
To ensure that a package requiring a storage volume does not attempt to start on or fail over to a
node where the storage volume is unavailable, the monitor service may be configured in a separate
package, and a package dependency may be used to ensure that the required package is running,
indicating the storage is available. Depending on the configuration, the monitor package could
be a multi-node or failover package, and would be required to be running by the storage
volume-dependent application package. Alternatively, if you are using EMS, please be aware that
EMS resource dependencies operate in this fashion – a package will not be run unless the specified
EMS resources conditions are met.
NOTE: When using the volume monitor to monitor LVM logical volumes, you need to make sure
that the logical volume timeout value is properly configured. This value should be configured to
be at least one second less than the poll-interval specified in the monitor service command.
I/O requests to logical volumes with no timeout set may block indefinitely. See “Setting Logical
Volume Timeouts” (page 184) for more information.
Command Syntax
The syntax for the monitoring command, cmvolmond, is as follows:
cmvolmond [-h, --help] [-v, --version]
[-O, --log-file <log_file>
[-D, --log-level <1-7>
134 Planning and Documenting an HA Cluster