Re: pgsql: Get rid of SET LOGGED indexes persistence kludge - Mailing list pgsql-committers

From Tom Lane
Subject Re: pgsql: Get rid of SET LOGGED indexes persistence kludge
Date
Msg-id 27238.1416073917@sss.pgh.pa.us
Whole thread Raw
In response to Re: pgsql: Get rid of SET LOGGED indexes persistence kludge  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-committers
Tom Lane <tgl@sss.pgh.pa.us> writes:
> Alvaro Herrera <alvherre@alvh.no-ip.org> writes:
>> Get rid of SET LOGGED indexes persistence kludge

> The buildfarm suggests that this patch fails under CLOBBER_CACHE_ALWAYS.

After a bit of study, the explanation is almost certainly that you can't
just do this in reindex_index:

+       /* Set the relpersistence of the new index */
+       iRel->rd_rel->relpersistence = persistence;

Probably the right fix is to add a persistence parameter to
RelationSetNewRelfilenode.  The existing patch seems pretty
klugy without that anyway.

            regards, tom lane


pgsql-committers by date:

Previous
From: Tom Lane
Date:
Subject: Re: pgsql: Get rid of SET LOGGED indexes persistence kludge
Next
From: Simon Riggs
Date:
Subject: pgsql: Emit msg re skipping ANALYZE for absent inh tree