Subject: Re: max_wal_size
To: pgsql-docs@lists.postgresql.org
References: <7020bbeb-cde5-b203-9803-2facd964688b@postgrespro.ru>
 <5d9bf1bd-d322-ba24-ce89-fa6029eb6615@a-kretschmer.de>
From: Egor Rogov <e.rogov@postgrespro.ru>
Message-ID: <d472283a-91a8-70dd-d877-cfb3ef03abde@postgrespro.ru>
Date: Wed, 27 May 2020 13:28:14 +0300
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101
 Thunderbird/60.9.0
MIME-Version: 1.0
In-Reply-To: <5d9bf1bd-d322-ba24-ce89-fa6029eb6615@a-kretschmer.de>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Content-Language: en-US
Precedence: bulk

On 27.05.2020 13:16, Andreas Kretschmer wrote:
>
> Am 27.05.20 um 12:11 schrieb Pavel Luzanov:
>> Hello,
>>
>> About parameter max_wal_size the documentation 
>> <https://www.postgresql.org/docs/current/runtime-config-wal.html#GUC-MAX-WAL-SIZE> 
>> says: "Maximum size to let the WAL grow to between automatic WAL 
>> checkpoints."
>> Is it correct?
>>
>> The size between automatic WAL checkpoints or the size of all WAL 
>> files in the pg_wal directory?
>
> size between checkpoints, size of all wal's can be higher, see 
> wal_keep_segments.
>

It is clear that total size can be higher due to wal_keep_segments, 
replications slots, archive_command etc.

But does max_wal_size relate to the (a) two subsequent checkpoints (as 
you may read the wording in the docs) or (b) to the past completed 
checkpoint AND the current yet-uncompleted checkpoint (which together 
may be up to TWO checkpoint intervals)?

If I read CalculateCheckpointSegments() in xlog.c correctly, the answer 
is (b), and hence the documentation is misleading.

Regards,
Egor.