SMF Timeout on Solaris

The Solaris SMF utility allots a set amount of time for each service to become fully enabled. For the acsls service, this time limit is calculated on the basis of the library configuration: the number of LSMs, the number of drives, and the number of CAPs. A large library configuration takes longer for ACSLS to recover than a smaller configuration, so a longer SMF timeout period is allotted to a larger configuration.

On rare occasions, a faulty LSM may take more time to come up than the SMF time limit allows. When the time-out period has expired, SMF will restart the operation. This action has the potential for the start sequence to go into an endless loop, preventing ACSLS from ever recovering during difficult startup conditions.

A special file, acsls_startup_policy, is intended for use in such situations. This file, located in the $ACS_HOME/data/external directory, when configured adds extra time for startup recovery, or to exempt any specific ACS from being recovered during the SMF startup sequence. Detailed configuration instructions are contained in the header remarks of acsls_startup_policy. By adjusting the start-up parameters in this file, you can avoid ACSLS startup problems introduced by an abnormal library startup condition.

For more information, see Diagnosing ACSLS Startup Problems.