Re: [PATCH]make pg_rewind to not copy useless WAL files

From: Michael Paquier <michael(dot)paquier(at)gmail(dot)com>
To: Alexander Korotkov <a(dot)korotkov(at)postgrespro(dot)ru>
Cc: chenhj <chjischj(at)163(dot)com>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: [PATCH]make pg_rewind to not copy useless WAL files
Date: 2017-09-17 00:33:33
Message-ID: CAB7nPqRtq4pZcs+R1i6endM_t7j7NLa4uf19Cern96DfEokChw@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Sun, Sep 17, 2017 at 3:19 AM, Alexander Korotkov
<a(dot)korotkov(at)postgrespro(dot)ru> wrote:
> Hi!
>
> On Sat, Sep 16, 2017 at 5:56 PM, chenhj <chjischj(at)163(dot)com> wrote:
>>
>> This patch optimizes the above mentioned issues, as follows:
>> 1. In the target data directory, do not delete the WAL files before the
>> divergence.
>> 2. When copying files from the source server, do not copy the WAL files
>> before the divergence and the WAL files after the current WAL insert
>> localtion.
>
>
> Looks like cool optimization for me. Please, add this patch to the next
> commitfest.

Agreed.

> Do you think this patch should modify pg_rewind tap tests too? It would be
> nice to make WAL files fetching more covered by tap tests. In particular,
> new tests may generate more WAL files and make sure that pg_rewind fetches
> only required files among them.

This looks mandatory to me. Using pg_switch_wal() and a minimum amount
of WAL generated you could just make the set of WAL segments skipped
minimal data.

I have not checked in details, but I think that the positions where
you are applying the filters are using the right approach.

! !(strncmp(path, "pg_wal", 6) == 0 && IsXLogFileName(path + 7) &&
Please use XLOGDIR here.
--
Michael

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Andreas Karlsson 2017-09-17 01:33:59 Re: postgres_fdw super user checks
Previous Message Douglas Doole 2017-09-16 23:29:54 Re: Add Roman numeral conversion to to_number