Horton

Re: Performance degradation of REFRESH MATERIALIZED VIEW

Andres Freund <andres@anarazel.de>

View thread View on postgr.es
From: Andres Freund <andres@anarazel.de>
To: Tomas Vondra <tomas.vondra@enterprisedb.com>
Cc: Masahiko Sawada <sawada.mshk@gmail.com>, Michael Paquier <michael@paquier.xyz>, Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>, Kyotaro Horiguchi <horikyota.ntt@gmail.com>, PostgreSQL-development <pgsql-hackers@postgresql.org>, Pavan Deolasee <pavan.deolasee@gmail.com>, Anastasia Lubennikova <a.lubennikova@postgrespro.ru>, Jeff Janes <jeff.janes@gmail.com>, Paul Guo <guopa@vmware.com>
Date: 2021-05-18T18:44:30Z
Lists: pgsql-hackers 
Hi,

On 2021-05-18 20:34:08 +0200, Tomas Vondra wrote:
> Yeah, I see your point. I agree it's unfortunate there's no way to disable
> freezing during REFRESH MV. For most users that trade-off is probably fine,
> but for some cases (matviews refreshed often, or cases where it's fine to
> pay more but later) it may be an issue.
>
> From this POV, however, it may not be enough to optimize the current
> freezing code - it's always going to be a bit slower than before.

But the intrinsic overhead is *tiny*. Setting a few bits, with the other
costs amortized over a lot of pages. As far as I can tell the measurable
overhead is that the increased WAL logging - which is not necessary.

Greetings,

Andres Freund

Commits

  1. Fix pg_visibility regression failure with CLOBBER_CACHE_ALWAYS

  2. Revert most of 39b66a91bd

  3. Fix COPY FREEZE with CLOBBER_CACHE_ALWAYS