If an error occurs in the shared disk where the ZFS storage pool is placed or in the FC cable and the access to the ZFS storage pool cannot be recovered, the ZFS detector detects the error. The default timeout to detect the error is 300 seconds.
This timeout value can be changed by the RMS environment variable "HV_ZFS_CMD_TIMEOUT." However, if the value is too small, the ZFS detector may detect the error mistakenly caused by due to the temporary high load. In addition, if the value is too large, it may take time to detect the error. The value is set in seconds and can be set in the range of 30 - 600. If the value that is smaller than 30 or larger than 600 is set, 30 or 600 is assumed specified.
Note that this value is effective only for the ZFS detector and does not affect the behavior of other resource detectors.
An example is shown below.
Add the following definition to /opt/SMAW/SMAWRrms/bin/hvenv.local on all nodes constituting a cluster. If this file does not exist, create it.
export HV_ZFS_CMD_TIMEOUT=<Timeout value>After the setting, restart RMS. The setting change becomes effective after restarting RMS.