Horton

Re: Handing off SLRU fsyncs to the checkpointer

Thomas Munro <thomas.munro@gmail.com>

View thread View on postgr.es
From: Thomas Munro <thomas.munro@gmail.com>
To: Jakub Wartak <Jakub.Wartak@tomtom.com>
Cc: "alvherre@2ndquadrant.com" <alvherre@2ndquadrant.com>, Robert Haas <robertmhaas@gmail.com>, pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2020-09-19T05:06:10Z
Lists: pgsql-hackers

Attachments

On Mon, Aug 31, 2020 at 8:50 PM Jakub Wartak <Jakub.Wartak@tomtom.com> wrote:
> - IO_URING - gives a lot of promise here I think, is it even planned to be shown for PgSQL14 cycle ? Or it's more like PgSQL15?

I can't answer that, but I've played around with the prototype quite a
bit, and thought quite a lot about how to port it to systems without
IO_URING, and I'm just as keen to see this happen as you are.

In the meantime, from the low-hanging-fruit department, here's a new
version of the SLRU-fsync-offload patch.  The only changes are a
tweaked commit message, and adoption of C99 designated initialisers
for the function table, so { [SYNC_HANDLER_CLOG] = ... } instead of
relying on humans to make the array order match the enum values.  If
there are no comments or objections, I'm planning to commit this quite
soon.

Commits

  1. Remove unused function prototypes.

  2. Defer flushing of SLRU files.

  3. Improve the vacuum error context phase information.

  4. Cache smgrnblocks() results in recovery.

  5. Refactor the fsync queue for wider use.

  6. Increase maximum number of clog buffers.

  7. Make the number of CLOG buffers adaptive, based on shared_buffers.

  8. Replace implementation of pg_log as a relation accessed through the