Re: Non-pausing table scan on 9.6 replica? - Mailing list pgsql-general

From Andreas Kretschmer
Subject Re: Non-pausing table scan on 9.6 replica?
Date
Msg-id 1030367a-f503-3755-0edf-c2c12c498585@a-kretschmer.de
Whole thread Raw
In response to Re: Non-pausing table scan on 9.6 replica?  (Mark Fletcher <markf@corp.groups.io>)
List pgsql-general

Am 06.03.19 um 06:41 schrieb Mark Fletcher:
> Thank you for responding to my email.
>
> On Tue, Mar 5, 2019 at 9:20 PM Andreas Kretschmer 
> <andreas@a-kretschmer.de <mailto:andreas@a-kretschmer.de>> wrote:
>
>
>     have you set ```max_standby_streaming_delay``? The default is 30
>     seconds, which means that this will be the maximum time allowed for a
>     replication lag caused by a conflicting query.
>
>
> Yes, we've bumped that up a lot.

i tought so.


>     You can use ``hot_standby_feedback = on``, but the downside will
>     be more
>     bloat on the tables.
>
> I'm not sure I understand. I'm not worried about the query being 
> cancelled on the replica. max_standby_streaming_delay fixes that for 
> us. I'm worried about the streaming replication being paused while 
> this table scan is running. If the table scan takes several minutes, 
> then the replica becomes several minutes out of sync with the master. 
> I'd prefer that not to happen and I'm wondering if there's a way to do 
> that.
>

You can choose. max_standby_streaming_delay with delay in the streaming 
(hence the name) or hot_standby_feedback, with the downside of possible 
more bloat.


Regards, Andreas


-- 
2ndQuadrant - The PostgreSQL Support Company.
www.2ndQuadrant.com



pgsql-general by date:

Previous
From: Mark Fletcher
Date:
Subject: Re: Non-pausing table scan on 9.6 replica?
Next
From: Sameer Kumar
Date:
Subject: Re: Non-pausing table scan on 9.6 replica?