Note
If the I/O is overloaded, make sure to set a monitoring disk.
If a file system is overloaded, it may be determined that the monitoring with hvdet_gmount described above has an error.
In a system where the file system load is expected to be high, use a monitoring disk to detect access errors (including a cable disconnection) of shared disk units. Other file systems are not monitored so that a glitch can be avoided. In addition, a shared disk can be monitored per operation so that the setting of 2. Disk area (LU) below is recommended.
To remove a disk from being monitored for the operation, set the MonitorOnly attribute of a resource to "1."
Shared disk device
Access errors (including a cable disconnection) can be detected per shared disk device.
For a single shared disk unit such as RAID as monitoring area, assign one disk area (LU) (for example /dev/sdf1). It needs 10 MB or more as a monitoring area.
Disk area (LU)
Access errors of a shared disk unit (such as LUN failure in addition to 1.) can be detected per operation.
For a single disk area, assign one slice for monitoring (for example, if you use /dev/sdd11 and /dev/sde for the operation, assign /dev/sdd2 and /dev/sde2 for monitoring). It needs 10 MB or more as a monitoring area.
