Re: (some) Indexes ignored after long-running UPDATE and REINDEX at the same time (8.3.6)

Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> writes:

> Gregory Stark <stark(at)enterprisedb(dot)com> writes:
>> Marinos Yannikos <mjy(at)pobox(dot)com> writes:
>>> I had a strange problem this morning - I started a long-running UPDATE on a
>>> heavily indexed table with about 8m rows last night to test a trigger-based
>>> queue (PgQ):
>
>> I think you were bitten by a gotcha with newly created indexes and "heap-only"
>> updates.
>
> This would only be possible if it were 8.3 and the reindex script used
> REINDEX CONCURRENTLY. Neither fact is in evidence at this point.

Well it does say "8.3.6" in the subject.

We don't actually support REINDEX CONCURRENTLY... But if my other post is
right then it would still happen (unnecessarily) if the index was originally
built with REINDEX CONCURRENTLY and then reindexed.

--
Gregory Stark
EnterpriseDB http://www.enterprisedb.com
Ask me about EnterpriseDB's PostGIS support!