Re: minor bug - Mailing list pgsql-hackers

From Laurenz Albe
Subject Re: minor bug
Date
Msg-id d0af6fef161cf9a161f2b9377193b9099e73d17e.camel@cybertec.at
Whole thread Raw
In response to Re: minor bug  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: minor bug
List pgsql-hackers
On Tue, 2023-01-17 at 10:32 -0500, Tom Lane wrote:
> Laurenz Albe <laurenz.albe@cybertec.at> writes:
> > On Mon, 2023-01-16 at 19:59 +0100, Torsten Förtsch wrote:
> > > So, the timestamp displayed in the log message is certainly wrong.
>
> > If recovery stops at a WAL record that has no timestamp, you get this
> > bogus recovery stop time.  I think we should show the recovery stop time
> > only if time was the target, as in the attached patch.
>
> I don't think that is a tremendously useful definition: the user
> already knows what recoveryStopTime is, or can find it out from
> their settings easily enough.
>
> I seem to recall that the original idea was to report the timestamp
> of the commit/abort record we are stopping at.  Maybe my memory is
> faulty, but I think that'd be significantly more useful than the
> current behavior, *especially* when the replay stopping point is
> defined by something other than time.
>
> (Also, the wording of the log message suggests that that's what
> the reported time is supposed to be.  I wonder if somebody messed
> this up somewhere along the way.)

recoveryStopTime is set to recordXtime (the time of the xlog record)
a few lines above that patch, so this is useful information if it is
present.

I realized that my original patch might be a problem for translation;
here is an updated version that does not take any shortcuts.

Yours,
Laurenz Albe

Attachment

pgsql-hackers by date:

Previous
From: "wangw.fnst@fujitsu.com"
Date:
Subject: RE: Logical replication timeout problem
Next
From: Nikita Malakhov
Date:
Subject: Re: Inconsistency in vacuum behavior