Re: pg_class(relpersistence) of hash index - Mailing list pgsql-hackers

From Tom Lane
Subject Re: pg_class(relpersistence) of hash index
Date
Msg-id 23707.1416844080@sss.pgh.pa.us
Whole thread Raw
In response to pg_class(relpersistence) of hash index  (Antonin Houska <ah@cybertec.at>)
List pgsql-hackers
Antonin Houska <ah@cybertec.at> writes:
> While checking how BM_PERMANENT flag is set (in buffer header), I noticed that
> hash index has it set too. Shouldn't pg_class(relpersistence) be 'u' in this
> case?

See archives; we do not currently have a way to support unlogged indexes
on logged tables.  The whole hash-index mess would be better if we did,
but getting there is less than trivial.
        regards, tom lane



pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: no test programs in contrib
Next
From: Alex Shulgin
Date:
Subject: Re: Replication connection URI?