proposal: custom autovacuum entries

Started by Jordan Deitchabout 8 years ago3 messages
#1Jordan Deitch
jwdeitch@gmail.com

Hi hackers,

I would like to remove records from a time series table older than a
certain age. My understanding of current standard practice is to configure
an external scheduler like cron to perform the deletion.
My proposal is to allow clients to register functions on the tail end of
the autovacuum sequence.
Alternatively I would suggest some mechanism for scheduling within
PostgreSQL

Thanks,
Jordan Deitch

#2Tom Lane
tgl@sss.pgh.pa.us
In reply to: Jordan Deitch (#1)
Re: proposal: custom autovacuum entries

Jordan Deitch <jwdeitch@gmail.com> writes:

I would like to remove records from a time series table older than a
certain age. My understanding of current standard practice is to configure
an external scheduler like cron to perform the deletion.
My proposal is to allow clients to register functions on the tail end of
the autovacuum sequence.

There's already a mechanism for writing custom background workers.
I doubt that adding random user-defined stuff to what autovacuum has
to do is a better idea; that will mostly make autovac scheduling even
harder than it is now.

regards, tom lane

--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

#3Michael Paquier
michael.paquier@gmail.com
In reply to: Tom Lane (#2)
Re: proposal: custom autovacuum entries

On Thu, Oct 26, 2017 at 1:47 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:

Jordan Deitch <jwdeitch@gmail.com> writes:

I would like to remove records from a time series table older than a
certain age. My understanding of current standard practice is to configure
an external scheduler like cron to perform the deletion.
My proposal is to allow clients to register functions on the tail end of
the autovacuum sequence.

There's already a mechanism for writing custom background workers.
I doubt that adding random user-defined stuff to what autovacuum has
to do is a better idea; that will mostly make autovac scheduling even
harder than it is now.

I agree with Tom here, there is no point to have more facilities in
autovacuum so as it does extra work at its tail. There may be not even
any need to develop your own background worker. Have you looked at
pg_cron [1]https://github.com/citusdata/pg_cron -- Michael?

[1]: https://github.com/citusdata/pg_cron -- Michael
--
Michael

--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers