On Thu, Feb 11, 2016 at 11:05 AM, Alvaro Herrera
<alvherre@2ndquadrant.com> wrote:
> Thomas Munro wrote:
>
>> 4. You could look into whether all those multixacts with many member
>> are really expected. (Large numbers of concurrent FK checks or
>> explicit share locks on the same rows perhaps? A good proportion of
>> this happened on one day this week I see.)
>
> I think maybe this patch
> https://www.postgresql.org/message-id/20150620043650.GX133018%40postgresql.org
> should help with this. I expect to come back to this and get it pushed
> to 9.3 and 9.4 sometime ...
Ah, so the OP might get some improvement today by moving to 9.5, which
has the optimization already.
BTW in my message above I said vacuum_multixact_freeze_table_age when
I meant to say autovacuum_multixact_freeze_max_age (and its default is
400 million).
--
Thomas Munro
http://www.enterprisedb.com
