Re: Interpreting vacuum verbosity - Mailing list pgsql-general

From Tom Lane
Subject Re: Interpreting vacuum verbosity
Date
Msg-id 14232.1083950714@sss.pgh.pa.us
Whole thread Raw
In response to Re: Interpreting vacuum verbosity  ("Ed L." <pgsql@bluepolka.net>)
Responses Re: Interpreting vacuum verbosity
List pgsql-general
"Ed L." <pgsql@bluepolka.net> writes:
> No, our autovac logs the number of changes (upd+del for vac, upd+ins+del for
> analyze) on each round of checks, and we can see it was routinely
> performing when expected.  The number of updates/deletes just far exceeded
> the thresholds.  Vac threshold was 2000, and at times there might be
> 300,000 outstanding changes in the 10-30 minutes between vacuums.

Well, in that case you probably want a lot less than "10-30 minutes"
between vacuums, at least for this particular table.  I don't know how
one configures autovac for this, but I suppose it can be done ...

> max_fsm_relations = 1000 and max_fsm_pages = 10000.

Also you doubtless need max_fsm_pages a lot higher than that.  A
conservative setting would make it as big as your whole database,
eg for a 10Gb disk footprint use 10Gb/8K (something upwards of
a million) FSM page slots.

            regards, tom lane

pgsql-general by date:

Previous
From: Mark Harrison
Date:
Subject: any experience with multithreaded pg apps?
Next
From: Eduardo Pérez Ureta
Date:
Subject: Storing a file hash as primary key