Re: pgsql: Teach autovacuum about multixact member wraparound.

On Fri, May 8, 2015 at 1:23 PM, Kevin Grittner <kgrittn(at)ymail(dot)com> wrote:
> Robert Haas <rhaas(at)postgresql(dot)org> wrote:
>
>> This patch leaves unsolved the problem of ensuring that emergency
>> autovacuums are triggered even when autovacuum=off. We'll need
>> to fix that via a separate patch.
>
> I think we also need something like your previously-posted
>
> multixact-truncate-race.patch as a follow-on. I'm not at all
> convinced that there is no possibility of truncating new data as
> things stand; and even if that is true now, it seems terribly
> fragile. (Only mentioning it because it wasn't noted in the commit
> message.)

I agree. I'm writing up a summary of open issues now. I didn't
mention it in the commit message because it's not related to the
autovacuum part of the problem.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company