On Wed, Mar 28, 2018 at 12:57 PM, Michael Paquier <michael@paquier.xyz> wrote:
> On Tue, Mar 27, 2018 at 10:09:59PM -0400, Robert Haas wrote:
>>> I have to agree with Tom here. If you force pg_rewind to replay more
>>> WAL records from a checkpoint older than the checkpoint prior to where
>>> WAL has forked at promotion then you have a risk of losing data.
>>
>> Oh! I see now. Good point.
>
> Something that would address the issue would be to enforce a segment
> switch after each checkpoint, but that's a high price to pay on mostly
> idle systems with large WAL segments, which is not appealing either, and
> this even if the checkpoint skip logic has been fixed in v10 with the
> concept of "important" WAL records.
If the system is mostly idle would it really matter that much?
cheers
andrew
--
Andrew Dunstan https://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services