Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
From: Bob Lunney <bob_lunney@yahoo.com>
Mime-Version: 1.0 (1.0)
Subject: Re: AWS forcing PG upgrade from v9.6 a disaster
Date: Fri, 28 May 2021 18:09:50 -0400
Message-Id: <31E60671-57F1-4C45-AB20-DC3985FE3F8D@yahoo.com>
References: <4e238cc5-f63e-0435-63bf-2515500d5424@dunslane.net>
Cc: Jan Wieck <jan@wi3ck.info>,
 "Dean Gibson (DB Administrator)" <postgresql@mailpen.com>,
 Andrew Dunstan <andrew@dunslane.net>
In-Reply-To: <4e238cc5-f63e-0435-63bf-2515500d5424@dunslane.net>
To: pgsql-performance@lists.postgresql.org
Content-Length: 1963
Archived-At: 
 <https://www.postgresql.org/message-id/31E60671-57F1-4C45-AB20-DC3985FE3F8D%40yahoo.com>
Precedence: bulk

I recently did 20 upgrades from 9.6 to 12.4 and 12.5.   No issues and the up=
grade process uses pg_upgrade.  I don=E2=80=99t know if AWS modified it thou=
gh.=20

Bob

Sent from my PDP11

> On May 28, 2021, at 5:15 PM, Andrew Dunstan <andrew@dunslane.net> wrote:
>=20
> =EF=BB=BF
>> On 5/28/21 4:23 PM, Jan Wieck wrote:
>> On 5/28/21 2:48 PM, Dean Gibson (DB Administrator) wrote:
>>=20
>> What sticks out for me are these two scans, which balloon from 50-60
>> heap fetches to 1.5M each.
>>=20
>>>                       ->  Nested Loop  (cost=3D0.29..0.68 rows=3D1
>>> width=3D7) (actual time=3D0.003..0.004 rows=3D1 loops=3D1487153)
>>>                             Join Filter: ("_IsoCountry".iso_alpha2 =3D
>>> "_Territory".country_id)
>>>                             Rows Removed by Join Filter: 0
>>>                             ->  Index Only Scan using
>>> "_IsoCountry_iso_alpha2_key" on "_IsoCountry"  (cost=3D0.14..0.38
>>> rows=3D1 width=3D3) (actual time=3D0.001..0.002 rows=3D1 loops=3D1487153=
)
>>>                                   Index Cond: (iso_alpha2 =3D
>>> "_GovtRegion".country_id)
>>>                                   Heap Fetches: 1487153
>>>                             ->  Index Only Scan using
>>> "_Territory_pkey" on "_Territory"  (cost=3D0.14..0.29 rows=3D1 width=3D7=
)
>>> (actual time=3D0.001..0.001 rows=3D1 loops=3D1487153)
>>>                                   Index Cond: (territory_id =3D
>>> "_GovtRegion".territory_id)
>>>                                   Heap Fetches: 1550706
>>=20
>> How did you load the database? pg_dump -> psql/pg_restore?
>>=20
>> If so, did you perform a VACUUM FREEZE after the load?
>>=20
>>=20
>>=20
>=20
> Jan
>=20
>=20
> AIUI he did an RDS upgrade. Surely that's not doing a dump/restore? I
> assume you would know better than him or me what it actually does do :-)
>=20
>=20
> cheers
>=20
>=20
> andrew
>=20
>=20
> --
> Andrew Dunstan
> EDB: https://www.enterprisedb.com
>=20
>=20
>=20