Re: Handing off SLRU fsyncs to the checkpointer

On Wed, Aug 12, 2020 at 6:06 PM Thomas Munro <thomas(dot)munro(at)gmail(dot)com> wrote:
> [patch]

Bitrot, rebased, no changes.

> Yeah, the combined effect of these two patches is better than I
> expected. To be clear though, I was only measuring the time between
> the "redo starts at ..." and "redo done at ..." messages, since I've
> been staring at the main recovery code, but there are also some more
> fsyncs before (SyncDataDirectory()) and after (RemoveOldXlogFiles())
> that are unaffected. I think it's probably possible to do something
> about those too, but that's another topic.

... and of course the end-of-recovery checkpoint; in my tests this
wasn't materially changed since there isn't actually very much CLOG,
it's just that we avoided syncing it block at a time and getting
rescheduled. FWIW I put a very simple test here:
https://github.com/macdice/redo-bench, YMMV.