On Wed, Nov 8, 2017 at 2:23 AM, Simon Riggs <si...@2ndquadrant.com> wrote: > On 31 October 2017 at 12:01, Michael Paquier <michael.paqu...@gmail.com> > wrote: >> While the mention about a manual checkpoint happening after a timed >> one will cause a full range of WAL segments to be recycled, it is not >> actually true that segments of the prior's prior checkpoint are not >> needed, because with your patch the segments of the prior checkpoint >> are getting recycled. So it seems to me that based on that the formula >> ought to use 1.0 instead of 2.0... > > I think the argument in the comment is right, in that > CheckPointDistanceEstimate is better if we use multiple checkpoint > cycles.
Yes, the theory behind is correct. No argument behind that. > But the implementation of that is bogus and multiplying by 2.0 > wouldn't make it better if CheckPointDistanceEstimate is wrong. Yes, this is wrong. My apologies if my words looked confusing. By reading your message I can see that our thoughts are on the same page. -- Michael -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers