Re: pg_basebackup failing

From: Michael Paquier <michael(dot)paquier(at)gmail(dot)com>
To: Alan Nilsson <anilsson(at)apple(dot)com>
Cc: PostgreSQL mailing lists <pgsql-general(at)postgresql(dot)org>
Subject: Re: pg_basebackup failing
Date: 2014-01-15 03:00:36
Message-ID: CAB7nPqQ6H7ZgL6dDsTE8WvUXtcpkuEknOr+3sTx1J7GuU=GiEA@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

On Wed, Jan 15, 2014 at 6:53 AM, Alan Nilsson <anilsson(at)apple(dot)com> wrote:
> Could someone give me some insight to the following error message:
>
> [mqsql06:/Volumes/SQL_Set] _postgres% pg_basebackup --user=replicate
> --host=mqsql03 -xP -Fp --pgdata=pgsql
> 19439890/65873894 kB (29%), 1/1 tablespace
> pg_basebackup: could not get transaction log end position from server:
> ERROR: archive member "base/17040/25491" too large for tar format
>
> I find the error message confusing since I have the format explicitly set to
> plain.
pg_basebackup fetches a full backup from server using tar all the time
for performance purposes to reduce the amount of data sent through the
stream. pg_basebackup contains some code to decode this tarball and
generate a plain format from it.
The error you are seeing is triggered because this relation file
exceeds MAX_TAR_MEMBER_FILELEN or 8GB for a single tar member, which
is as well the norm for tar.

> Is there a solution/workaround?
Not use a pg_basebackup but an external backup solution or a custom
script. Break this relation file into several pieces by doing some
partitions on it.
> Can I break up the referenced file somehow?
Partitions on the parent table.
Regards,
--
Michael

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message ambilalmca 2014-01-15 05:27:27 Re: How to know server status variable in postgresql?
Previous Message Michael Paquier 2014-01-15 02:48:41 Re: non-zero xmax yet visible