[Info-ingres] same select statement returns different results when repeatedly executed

[Dennis]

On 9/24/2014 4:19 AM, Tim wrote:
> On Tuesday, September 23, 2014 7:41:27 PM UTC+1, droesler wrote:
>> what should I be looking for to resolve this?
>
> Try Optimizedb.  We saw this in a situation where a lot of records
 > were added to a table outside of the range of the index when 
optimizdeb last ran.
> Rerunning optimizedb (or dropping statistics) allowed the query to run
 > successfully.  Turning off Parallel Query may also help?
> (We were attempting to delete all the matching rows, but several were left behind  > - this was bug 126831 - fixed in patch 14700, so may not be exactly 
the same?)
>

Re-indexing didn't seem to help but optimizedb did.  Sysmod failed after 
running optimizedb thinking there was still an exclusive lock being held 
somewhere, but bouncing ingres and running sysmod again worked.

I'm not sure how worried I should be about this.  Complete 
unload/reload, unload/destroydb/createdb/reload, or just watch and make 
sure there on no issues going forward.

Dennis