Re: Improve handling of parameter differences in physical replication - Mailing list pgsql-hackers

From Peter Eisentraut
Subject Re: Improve handling of parameter differences in physical replication
Date
Msg-id 9f2b7b1e-7e41-4258-e86c-2611375ea9af@2ndquadrant.com
Whole thread Raw
In response to Re: Improve handling of parameter differences in physical replication  (Sergei Kornilov <sk@zsrv.org>)
Responses Re: Improve handling of parameter differences in physical replication  (Sergei Kornilov <sk@zsrv.org>)
List pgsql-hackers
On 2020-11-20 16:47, Sergei Kornilov wrote:
> Hmm... Good question. How about putting CheckForStandbyTrigger() in a wait loop, but reporting FATAL with an
appropriatemessage, such as "promotion is not possible because of insufficient parameter settings"?
 
> Also it suits me if we only document that we ignore promote here. I don't think this is an important case. And yes,
it'snot easy to allow promotion, since we have already updated control file.
 
> 
> Probably we need pause only after we reached consistency?

Here is an updated patch that implements both of these points.  I have 
used hot standby active instead of reached consistency.  I guess 
arguments could be made either way, but the original use case really 
cared about hot standby.

-- 
Peter Eisentraut
2ndQuadrant, an EDB company
https://www.2ndquadrant.com/

Attachment

pgsql-hackers by date:

Previous
From: "David G. Johnston"
Date:
Subject: Re: Add docs stub for recovery.conf
Next
From: Bruce Momjian
Date:
Subject: Re: Add docs stub for recovery.conf