> I agree with all that. I don't have any data either, but I agree that
> AFAICT it seems to mostly be a problem for large (terabyte-scale)
> databases, or ones that are dreadfully short of I/O bandwidth. AWS,
> I'm looking at you.
Well, at this point, numerically I'd bet that more than 50% of our users
are on AWS, some other cloud, or some kind of iSCSI storage ... some
place where IO sucks. It's How Things Are Done Now.
Speaking for my own clientele, people run into issues, or think they
have issues, with autovacuum at databases as small as 100GB, as long as
they have sufficient write throughput. One really pathological case I
had to troubleshoot was a database which was only 200MB in size! (this
database contained counts of things, and was updated 10,000 times per
second).
Anyway, my goal with that wiki page -- which is on the wiki so that
others can add to it -- is to get all of the common chronic issues on
the table so that we don't inadvertently make one problem worse while
making another one better. Some of the solutions to FREEZE being
bandied around seemed likely to do that.
--
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com