RE: Slow catchup of 2PC (twophase) transactions on replica in LR - Mailing list pgsql-hackers

From Hayato Kuroda (Fujitsu)
Subject RE: Slow catchup of 2PC (twophase) transactions on replica in LR
Date
Msg-id OSBPR01MB2552F7859F0EF334B9E8BC03F5E32@OSBPR01MB2552.jpnprd01.prod.outlook.com
Whole thread Raw
In response to Re: Slow catchup of 2PC (twophase) transactions on replica in LR  (Peter Smith <smithpb2250@gmail.com>)
List pgsql-hackers
Dear Peter,

Thanks for reviewing! New patch is available in [1].

> I'm having second thoughts about how these patches mention the option
> values "on|off". These are used in the ALTER SUBSCRIPTION document
> page for 'two_phase' and 'failover' parameters, and then those
> "on|off" get propagated to the code comments, error messages, and
> tests...
> 
> Now I see that on the CREATE SUBSCRIPTION page [1], every boolean
> parameter (even including 'two_phase' and 'failover') is described in
> terms of "true|false" (not "on|off").

Hmm. But I could sentences like "The default value is off,...". Also, in alter_subscription.sgml,
"on|off" notation has already been used. Not sure, but I felt there are no rules around here.

> In hindsight, it is probably better to refer only to true|false
> everywhere for these boolean parameters, instead of sometimes using
> different values like on|off.
> 
> What do you think?

It's OK for me to make message/code comments consistent. Not sure the documentation,
but followed only my part.

[1]:
https://www.postgresql.org/message-id/OSBPR01MB2552F66463EFCFD654E87C09F5E32%40OSBPR01MB2552.jpnprd01.prod.outlook.com

Best Regards,
Hayato Kuroda
FUJITSU LIMITED
https://www.fujitsu.com/ 


pgsql-hackers by date:

Previous
From: Aleksander Alekseev
Date:
Subject: Re: Postgres and --config-file option
Next
From: Jelte Fennema-Nio
Date:
Subject: Re: First draft of PG 17 release notes