SiteStats aggregator job running multiple times on same events

In testing SiteStats for use at our institution, I disabled real-time event aggregation and ran the StatsAggregateJob on all the events in the database. During these tests, often the job would run on the same data for multiple executions in a row.

I traced it to the fact that the the id inserted into SST_JOB_RUN table was smaller than existing entries.  That causes the incorrect behavior because the job orders the results by ID to determine the sakai_event id to start from on the next run.