Intermittent error monitoring is specifiable only when the automatic failback function is enabled. To prevent an intermittent error from reducing I/O performance, we recommend that you monitor intermittent errors when automatic failback is enabled.
When intermittent error monitoring is enabled, you can specify intermittent error conditions. The default value for the intermittent error-monitoring interval is 30. The default value for the number of error occurrences is 3.
The system assumes that an intermittent error has occurred if the specified number of times that the error needs to occur is reached during the specified monitoring interval (minutes). A path that is assumed to have an intermittent error is excluded from performing an automatic failback. Intermittent error monitoring starts right when the path is recovered from the error by performing an automatic failback. Monitoring is performed on each, individual path.
When a value of 2 or more is specified for the number of times an error needs to occur, make sure that the condition shown in Setting up the automatic failback function is satisfied.
To determine whether a path is invalid for an automatic failback, you can use the results of the dlnkmgr command's view operation.
The following is an example of setting up intermittent error monitoring by using a command:
# /opt/DynamicLinkManager/bin/dlnkmgr set -iem on -intvl 20 -iemnum 2
To enable intermittent error monitoring, specify on. To disable intermittent error monitoring, specify off. When you specify on, you can use the -intvl and -iemnum parameters to specify intermittent error conditions (the conditions used by the system to determine whether an intermittent error is occurring). In the -intvl parameter, specify the monitoring interval for an intermittent error. In the -iemnum parameter, specify the number of times that the error is to occur. When these parameters are omitted, the default values of 30 and 3 are used, respectively.