Error: Timeout for content=(number) in production only

Question

Running Portal 3.0.9.8.2. Apps Server Sun 450R; Database Server Sun 450R. Environments are same for both Development and Production. Similar performance navigating the site(s), load is not an issue, however Development returns Oracle Text(Intermedia) searches within 2-3 seconds while Production continually hits the 'Portal Content Area' provider timeout of 15 seconds and displays 'Error: Timeout for content=(15???)'. We use ctx_schedule to run nightly 'SYNC' and 'OPTIMIZE' jobs on all 6 indexes. This error only appears when searching, never when navigating and only in production. Any advice on where to start looking for problems?
Regards, Joshua

Answer 1

...update...
The 15 second timeouts can be alleviated by increasing the 'Portal Content Area' provider timeout value, however, that is just a symptom. Our problem is that the searches should not be taking 15 seconds to return.
We thought we were scheduling our 'SYNC' and 'OPTIMIZE' jobs to run nightly by issuing the following command for each of our indexes...
exec ctx_schedule.startup ('PORTAL30.WWSBR_THING_CTX_INDX', 'SYNC', 1440, TO_DATE('4/17/2002 01:00:00', 'MM/DD/YYYY HH24:MI:SS'));
exec ctx_schedule.startup ('PORTAL30.WWSBR_THING_CTX_INDX', 'OPTIMIZE FULL', 1440, TO_DATE('4/17/2002 02:00:00', 'MM/DD/YYYY HH24:MI:SS'));
However, after checking the interval column in the dba_jobs table we found that each job was scheduled to run every minute (we hope this explains our sudden performance problem). Select what,interval from dba_jobs returns...
ctxsys.ctx_ddl.sync_index('PORTAL30.WWSBR_THING_CTX_INDX');|sysdate + (1 / 24 / 60)
ctxsys.ctx_ddl.optimize_index('PORTAL30.WWSBR_THING_CTX_INDX', 'FULL');|sysdate + (1 / 24 / 60)
Why, when we pass in 1440 as the wait_time parameter, do the jobs get scheduled to run every minute?