Re: During promotion, new master tries to archive same segment twice

From: "Phil Endecott" <spam_from_pgsql_lists(at)chezphil(dot)org>
To: "Adrian Klaver" <adrian(dot)klaver(at)aklaver(dot)com>, <pgsql-general(at)lists(dot)postgresql(dot)org>
Subject: Re: During promotion, new master tries to archive same segment twice
Date: 2018-08-16 15:48:27
Message-ID: 1534434507556@dmwebmail.dmwebmail.chezphil.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

Adrian Klaver wrote:
> On 08/16/2018 01:48 AM, Phil Endecott wrote:
>> Adrian Klaver wrote:
>>> On 08/15/2018 01:25 PM, Phil Endecott wrote:
>>>> Dear Experts,
>>>>
>
>>> The above is not clear to me. My best guess:
>>>
>
>> It's not part of the error for the archive command; it's just the
>> next thing in the log file.  Y is (1) trying write WAL segment
>> E8 to the Y-to-Z archive and failing because it is already there and
>> (2) trying to read WAL segment E9 from the X-to-Y archive and failing
>> because it doesn't exist (and never will).  (1) is the problem, (2) is
>> expected.
>
> Hmm, still not sure why they are both showing up in the same LOG: entry
> as the entry was for the failure of the archive_command.

It's not. It's just the next thing in the log file. It seems that
stderr output from scp appears raw in the log file.

> What is archive_mode set to on Y?

archive_mode = always

Regards, Phil.

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Data Ace 2018-08-16 16:10:02 Re: using graph model with PostgreSQL
Previous Message PO 2018-08-16 15:41:10 pg_upgrade (and recovery) pitfalls