Tom Lane wrote:
>
> maillist <maillist@tnss.com> writes:
> > there was a cron job running doing a vacum on the entire database
>
> I was about to say that then it's a known issue --- two ANALYZE
> operations running concurrently for the same table can generate
> a "simple_heap_insert" or "simple_heap_update" failure from trying
> to update the same pg_statistic row at the same time. This is fairly
> harmless (as long as one of 'em succeeds, who cares if the other
> doesn't?).
>
> However, looking back at your original message, you said the error
> mentioned "simple_heap_delete", didn't you? Hmm ... how old is your
> PG installation? ANALYZE hasn't used simple_heap_delete since 7.1...
>
> regards, tom lane
well it may be really old to you but in my world its middle of the road
I have pgsql
7.0.3 running on sco open server
7.0.2 running on redhat 6.2
7.1.3-4bp.2 running on rh 7.2 (where problem occured)
7.3.2-3 running on redhat 9.0 (where I'm finding all the wiz bang
improvments that are breaking my code)
whatever version runs on redhat as 2.1 replacing 7.0.3, allowing me to
retire 7.0.2
