Re: Refactoring the checkpointer's fsync request queue

On Thu, Feb 28, 2019 at 10:27 AM Shawn Debnath <sdn@amazon.com> wrote:
> We had a quick offline discussion to get on the same page and we agreed
> to move forward with Andres' approach above. Attached is patch v10.
> Here's the overview of the patch:

Thanks.  I will review, and try to rebase my undo patches on top of
this and see what problems I crash into.

> Ran make check-world and repeated the tests described in [1]. The
> numbers show a 12% drop in total time for single run of 1000 clients and
> ~62% drop in total time for 10 parallel runs with 200 clients:

Hmm, good but unexpected.  Will poke at this here.


--
Thomas Munro
https://enterprisedb.com