Re: While only running pg_dump, postgresql performs writes inside base/ directory. How can we stop this?

From: Spiros Ioannou <sivann(at)inaccess(dot)com>
To: Jeff Janes <jeff(dot)janes(at)gmail(dot)com>
Cc: "pgsql-general(at)postgresql(dot)org" <pgsql-general(at)postgresql(dot)org>
Subject: Re: While only running pg_dump, postgresql performs writes inside base/ directory. How can we stop this?
Date: 2013-12-13 15:19:00
Message-ID: CACKh8C8gm6sFXvL5BXRQnVM1pn44=C5pn31m0vUzToqcWt-Nsg@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

- autovacuum is not off in the pre-snapshot (production) database. It
auto-runs every 2 weeks to avoid transaction ID wraparound.
- I tried modifying pgdump to have a more relaxed transaction isolation
(READ UNCOMMITTED) instead of the default (serializable) but this didn't
help.
thanks for the info.

On 10 December 2013 21:13, Jeff Janes <jeff(dot)janes(at)gmail(dot)com> wrote:

> On Tue, Dec 10, 2013 at 12:43 AM, Spiros Ioannou <sivann(at)inaccess(dot)com>wrote:
>
>> Hi Jeff,
>> autovacuum is off on the DB running on the filesystem snapshot.
>>
>
>
> Which probably makes sense on the snapshot, but is it also off on the
> pre-snapshot database?
>
>
>> What "hint bits" do you suppose it is setting? It's running only one COPY
>> command for days. Do you have any suggestions to make it more "read only" ?
>>
>
>
> When a query sees a tuple that is still listed as part of an open
> transaction, it needs to figure out whether that transaction has now
> completed, and if so whether it committed or aborted. This can be quite a
> bit of work to do, so once complete it sets a hint bit locally to that
> tuple, so that the next visitor doesn't have to repeat the work. I don't
> believe that there is any way to turn this off, unless you want to run your
> own custom build of PostgreSQL.
>
> Cheers,
>
> Jeff
>

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Adrian Klaver 2013-12-13 15:30:27 Re: invisible dependencies on a table?
Previous Message rob stone 2013-12-13 12:46:38 Re: design for multiple time series