Re: Unexpectedly high disk space usage - Mailing list pgsql-general

From Scott Marlowe
Subject Re: Unexpectedly high disk space usage
Date
Msg-id CAOR=d=1XCz=_HM5hM=8H+o0Vy-C1r1D=f=BJbnm+TwHDE=xu2g@mail.gmail.com
Whole thread Raw
In response to Re: Unexpectedly high disk space usage  (Lists <lists@benjamindsmith.com>)
Responses Re: Unexpectedly high disk space usage RESOLVED (Manual reindex/vacuum)  (Lists <lists@benjamindsmith.com>)
List pgsql-general
On Thu, Nov 8, 2012 at 6:05 PM, Lists <lists@benjamindsmith.com> wrote:
> On 11/07/2012 12:42 PM, Tom Lane wrote:
>>>
>>> ... because it
>>> >occasionally causes transactions and queries to hang when an update
>>> >causes a vacuum mid-day, effectively taking us offline randomly.
>>
>> I suspect this claim is based on ancient and no longer very relevant
>> experience.
>
> Even so, if I felt the need to keep autovacuum off, what would I need to run
> regularly in order to keep things neat and tidy under the hood? Would a
> simple "vacuum" within each database suffice? Should I be logged in as the
> database owner or as an administrative user?

Just know that most of the time people think they need to turn off
autovacuum they usually need to tune it instead.  either more or less
agressive depending on why they think they need to turn it off.  If
it's consuming too much IO then reduce cost limit / increase cost
delay, if it's not aggressive enough, then reverse that and increase
cost limit and decrease cost delay.  If your IO subsystem can't keep
up, then turning off autovacuum or turning it down simply be delaying
the problem rather than solving it (i.e. throw more IO at it).


pgsql-general by date:

Previous
From: Alvaro Herrera
Date:
Subject: Re: Does PostgreSQL have complete functional test cases?
Next
From: Alvaro Herrera
Date:
Subject: Re: Does PostgreSQL have complete functional test cases?