Postgres Pro
Downloads
Home   >   mailing lists

Setting autovacuum_vacuum_scale_factor to 0 a good idea ? - Mailing list pgsql-performance

From Sébastien Lorion
Subject Setting autovacuum_vacuum_scale_factor to 0 a good idea ?
Date
Msg-id CAGa5y0Nv==ObEnWr5BbR6mRWPTP21uLwduM2UYs3oxMdy+60sA@mail.gmail.com
Whole thread Raw
Responses Re: Setting autovacuum_vacuum_scale_factor to 0 a good idea ?  (Josh Berkus <josh@agliodbs.com>)
List pgsql-performance
Tree view
I am pondering about this... My thinking is that since *_scale_factor need to be set manually for largish tables (>1M), why not set autovacuum_vacuum_scale_factor and autovacuum_analyze_scale_factor, and increase the value of autovacuum_vacuum_threshold to, say, 10000, and autovacuum_analyze_threshold to 2500 ? What do you think ?

Also, with systems handling 8k-10k tps and dedicated to a single database, would there be any cons to decreasing autovacuum_naptime to say 15s, so that the system perf is less spiky ?

Sébastien

pgsql-performance by date:

Previous
From: "McKinzie, Alan (Alan)"
Date:
Subject: Are there known performance issues with defining all Foreign Keys as deferrable initially immediate
Next
From: Sébastien Lorion
Date:
Subject: wal_sync_method on FreeBSD 9.0 - ZFS