From: | Rui DeSousa <rui(at)crazybean(dot)net> |
---|---|
To: | Kurt Roeckx <kurt(at)roeckx(dot)be> |
Cc: | Laurenz Albe <laurenz(dot)albe(at)cybertec(dot)at>, pgsql-admin(at)lists(dot)postgresql(dot)org |
Subject: | Re: Append only tables |
Date: | 2020-03-24 20:59:39 |
Message-ID: | 1B98792C-304F-4248-8B3F-7435F32B1551@crazybean.net |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-admin |
> On Mar 24, 2020, at 2:41 PM, Kurt Roeckx <kurt(at)roeckx(dot)be> wrote:
>
> I think I'll set autovacuum_freeze_max_age a lot lower than the default
> 200 M.
Hmm… not sure what you really want to accomplish?
Administrating a high volume database and went the other way; with it set to 800M; however, auto vacuum set aggressively. I don’t want to see wrap around/freeze vacuums occurring as they are more aggressive than auto vacuum, a good indicator that vacuuming is not occurring enough, and will block where normal auto vacuums will yield. Really shouldn’t run into blocking issue unless issuing DDL statements; which the application shouldn’t do anyway. It does happen though; despite not allowing DDLs an applicatoin developer thought it would be wise to embed statistics updating on a table. The result is that the application would be denied but it still required a lock on the table to do so; the result is the freeze operation blocked the application.
From | Date | Subject | |
---|---|---|---|
Next Message | Kurt Roeckx | 2020-03-24 22:02:02 | Re: Append only tables |
Previous Message | Andres Freund | 2020-03-24 19:42:59 | Re: PG12 autovac issues |