Date: Thu, 24 Mar 2022 07:24:14 -0500
From: Justin Pryzby <pryzby@telsasoft.com>
To: Prajna Shetty <Prajna.Shetty@mindtree.com>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, pgsql-performance@postgresql.org,
	Beenu Sharma <Beenu.Sharma@mindtree.com>,
	Pranabesh Saha <Pranabesh.Saha@mindtree.com>,
	Surya Susai <Surya.Susai@mindtree.com>
Subject: Re: Performance issue post upgrade on Version 13 - Incorrect
 Estimation Cost choosing Hash Aggregate-Nested Left Loop Join
Message-ID: <20220324122414.GZ28503@telsasoft.com>
References: 
 <MAZPR01MB546945FEA4E21EE7C8E532FB92169@MAZPR01MB5469.INDPRD01.PROD.OUTLOOK.COM>
 <1872582.1647871148@sss.pgh.pa.us>
 <MAZPR01MB54690C89F014DCA4FB551B6592179@MAZPR01MB5469.INDPRD01.PROD.OUTLOOK.COM>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: 
 <MAZPR01MB54690C89F014DCA4FB551B6592179@MAZPR01MB5469.INDPRD01.PROD.OUTLOOK.COM>
User-Agent: Mutt/1.9.4 (2018-02-28)
Archived-At: 
 <https://www.postgresql.org/message-id/20220324122414.GZ28503%40telsasoft.com>
Precedence: bulk

On Tue, Mar 22, 2022 at 12:57:10PM +0000, Prajna Shetty wrote:
> 1.      We have performed Vacuum/Analyze/Reindex post Upgrade.
> 2.      Tweaked work_mem so it does not spill to Disk. We can Disk Usage But it is still using Hash Aggregate and came down from 5 minutes to 20 seconds. (Expected ~5 seconds). Attached plan after modifying work_mem
> 3.      Disabled Seqcan/ nestedloop
> 4.      Tweaked random_page_cost/seq_page_cost
> 5.      Set default_statistics_target=1000 and then run vacuum(analyze,verbose) on selected tables.
> 6.      We have also tested performance by increasing resources up to 4 vCPU and 32 GB RAM.

Would you provide your current settings ?
https://wiki.postgresql.org/wiki/Server_Configuration