RE: [Todo item] Add entry creation timestamp column to pg_stat_replication

From: "myungkyu(dot)lim" <myungkyu(dot)lim(at)samsung(dot)com>
To: "'Michael Paquier'" <michael(at)paquier(dot)xyz>
Cc: "'Masahiko Sawada'" <sawada(dot)mshk(at)gmail(dot)com>, "'PostgreSQL-development'" <pgsql-hackers(at)postgresql(dot)org>, <woosung(dot)sohn(at)samsung(dot)com>, <don(dot)hong(at)samsung(dot)com>
Subject: RE: [Todo item] Add entry creation timestamp column to pg_stat_replication
Date: 2018-12-04 07:24:25
Message-ID: 03f901d48ba2$6398d430$2aca7c90$@samsung.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

>> I have been looking at this patch more in-depth, and you missed one
>> critical thing: hot standby feedback messages also include the
>> timestamp the client used when sending the message, so if we want to
>> track the latest time when a message has been sent we should track it
>> as much as the timestamp from status update messages.
>>
>> Fixing that and updating a couple of comments and variables, I am
>> finishing with the attached. Thoughts?

Thanks! I missed it..:(

>Another thing which is crossing my mind is if it would make sense to report
>the timestamp of the last HS feedback message and the timestamp of the last
>status update message into two separate columns. As the point of this
>field is to help with the debugging of mostly idle systems it seems to me
>that merging both is fine, but I'd like to hear extra opinions about that.

I think purpose of this field,
that react interval check and debugging on idle system.
So, merging both is better.
(Is 'Reply' and 'HSFeedback' worth measuring separately?)

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Michael Paquier 2018-12-04 07:36:04 Re: Use durable_unlink for .ready and .done files for WAL segment removal
Previous Message Fabien COELHO 2018-12-04 07:16:24 Re: doc - improve description of default privileges