Re: Speedup of relation deletes during recovery

On Wed, Apr 18, 2018 at 10:44 AM, Michael Paquier <michael(at)paquier(dot)xyz> wrote:
> On Wed, Apr 18, 2018 at 12:46:58AM +0900, Fujii Masao wrote:
>> Yes, I think. And, I found that smgrdounlinkfork() is also dead code.
>> Per the discussion [1], this unused function was left intentionally.
>> But it's still dead code since 2012, so I'd like to remove it. Patch attached.
>
> Indeed, it's close to six years and the status is the same. So let's
> drop it. I have been surrounding the area to see if any modules
> actually use those, particularly on github, but I could not find
> callers.
>
> The patch looks logically fine to me. In your first message, you
> mentioned that the replay time increased a lot. Do you have numbers to
> share with some large settings of shared_buffers?

No. But after my colleague truncated more than one hundred tables on
the server with shared_buffers = 300GB, the recovery could not finish
even after 10 minutes since the startup of the recovery. So I had to
shutdown the server immediately, set shared_buffers to very small
temporarily and start the server to cause the recovery to finish soon.

> It would be better to wait for v12 branch to open before pushing
> anything, as the focus is on stabililizing things on v11.

Yes, so I added this to next CommitFest.

Regards,

--
Fujii Masao