Postgres Pro
Downloads
Home   >   mailing lists

Re: Long-running and non-finishing VACUUM ANALYZE on large table - Mailing list pgsql-admin

From Tom Lane
Subject Re: Long-running and non-finishing VACUUM ANALYZE on large table
Date
Msg-id 4139.1443816949@sss.pgh.pa.us
Whole thread Raw
In response to Long-running and non-finishing VACUUM ANALYZE on large table  (Jan <pgsql.admin@j.mk-contact.de>)
Responses Re: Long-running and non-finishing VACUUM ANALYZE on large table  (Ben Primrose <bprimrose@tracelink.com>)
List pgsql-admin
Tree view 
Ben Primrose <bprimrose@tracelink.com> writes:
> Tom, how'd you determine the maintenance_work_mem?

From the numbers of tuples shown as being removed in each index pass.
The work-mem usage is 6 bytes per TID, and it'll stop and do an index
pass once we get close to full.

            regards, tom lane

pgsql-admin by date:

Previous
From: Tom Lane
Date:
Subject: Re: Long-running and non-finishing VACUUM ANALYZE on large table
Next
From: Ben Primrose
Date:
Subject: Re: Long-running and non-finishing VACUUM ANALYZE on large table