If the UNIX agent is unable to connect to a collector, it spools the session data to local storage. When a collector becomes available, it then sends the spooled data to that collector.
By default, the minimum amount of allocated disk space that must be available to the offline database before spooling stops and warnings are posted to the agent error log is 10%. You can change this percentage by assigning a different value to
spool.diskspace.min in the
/etc/centrifyda/centrifyda.conf file. For example, to change the minimum to 15%, set the following value:
If the threshold is reached and a collector is still not available, the agent stops spooling data, and further audit data is lost. If this happens frequently or unexpectedly, you may want to increase the disk space allocation.