Re: [RFC] Should we fix postmaster to avoid slow shutdown?

>
> I also find others's ideas woth considering -- WAL-logging the stats files, type-specific stats files, etc. -- but I'm afraid those ideas would only be employed in a new major release, not in released versions. I'm asking for a remedy for a user (and potential users) who use older releases. And, I don't yet understand why patch would make the situation for existing users, and why stop writing files during immediate/abnormal shutdown requires other efforts.
>
I agree with this. I don't think we are going to change stats file
management in older versions. Hence after every crash recovery they
will be thrown away. In that case, why should stats collector in the
older versions wait for those files to be flushed to the disk? I
think, we should apply patch to shut down stats collector immediately
to the older versions and implement the stats file management solution
to the head. Why isn't that a feasible option?

--
Best Wishes,
Ashutosh Bapat
EnterpriseDB Corporation
The Postgres Database Company