>
>
>
> Live tuples = 19,766,480
> Analyze scale factor = 0.001
> Analyze thresh = 5000
> Thresh + live_tuples * factor = 24,766
>
> So an autovacuum analyze should trigger around 24K tuples modified, is
> this to little or too much?
This seems too much to me. Was there a specific problem
On Fri, 2019-11-01 at 18:58 +, Jason Ralph wrote:
> autovacuum_naptime = 1h
> autovacuum_vacuum_threshold = 1
> autovacuum_analyze_threshold = 5000
> autovacuum_vacuum_scale_factor = 0.002
> autovacuum_analyze_scale_factor = 0.001
These seem to be crazy settings.
Only once an hour you
I agree, this is excellent advice, I overlooked the fact that this is a sample
and the new rows may not even be included in this sample. I will adjust
accordingly.
-Original Message-
From: Jason Ralph
Sent: Friday, November 1, 2019 2:59 PM
To: pgsql-general@lists.postgresql.org
My thinking is opposite from what you have. I consider it important to very
aggressive on autovacuum because it only ever does the required amount of
work. If a tiny amount of work is needed, it does only that and is done.
Assuming it doesn't cause I/O concerns, do it as often as possible to
Hello list,
DB1=# select version();
-[ RECORD 1
]
version | PostgreSQL 11.5 on x86_64-pc-linux-gnu, compiled by gcc (GCC) 4.4.7
20120313 (Red Hat 4.4.7-23), 64-bit
I am sure this question has
