Re: predefined role(s) for VACUUM and ANALYZE

On Tue, Sep 20, 2022 at 02:45:52PM +0900, Michael Paquier wrote:
> I have gone through the thread, and I'd agree with getting more
> granularity when it comes to assigning ACLs to relations rather than
> just an on/off switch for the objects of a given type would be nice.
> I've been looking at the whole use of AclMode and AclItem in the code,
> and I don't quite see why a larger size could have a noticeable
> impact. There are a few things that could handle a large number of
> AclItems, though, say for array operations like aclupdate(). These
> could be easily checked with some micro-benchmarking or some SQL
> queries that emulate a large number of items in aclitem[] arrays.

I performed a few quick tests with a couple thousand ACLs on my laptop, and
I'm consistently seeing a 4.3% regression.

> Any impact for the column sizes of the catalogs holding ACL
> information? Just asking while browsing the patch set.

Since each aclitem requires 16 bytes instead of 12, I assume so. However,
in my testing, I hit a "row is too big" error with the same number of
aclitems in a pg_class row before and after the change. I might be missing
something in my patch, or maybe I am misunderstanding how arrays of
aclitems are stored on disk.

> Some comments in utils/acl.h need a refresh as the number of lower and
> upper bits looked at from ai_privs changes.

Oops, I missed that one. I fixed it in the attached patch set.

--
Nathan Bossart
Amazon Web Services: https://aws.amazon.com