[SOLVED] How to fix "uncommitted xmin from before xid cutoff needs to be frozen, automatic vacuum of table "db.pg_catalog.pg

How to fix "uncommitted xmin from before xid cutoff needs to be frozen, automatic vacuum of table "db.pg_catalog.pg_largeobject"

PostgreSQL Database error log generate this error all day and still continue error to next day

[23523] ERROR:  uncommitted xmin 53354897 from before xid cutoff 210760077 needs to be frozen
[23523] CONTEXT:  automatic vacuum of table "xxxx.pg_catalog.pg_largeobject"
[23523] ERROR:  uncommitted xmin 53354897 from before xid cutoff 210760077 needs to be frozen
[23523] CONTEXT:  automatic vacuum of table "xxxx.pg_catalog.pg_largeobject_metadata"

The error are involve system catalogs (pg_catalog.pg_largeobject, pg_catalog.pg_largeobject_metadata).

I need help about how to fix it or what will be affected if I disable autovacuum on these 2 tables.

Note:

DB : PostgreSQL version 11.6
OS : Red Hat Enterprise Linux Server release 7.8

Solution

You are experiencing data corruption, and if you don't take action, you are headed for disaster: if autovacuum keeps failing (as it will), you will eventually get close enough to transaction ID wraparound that your database will stop accepting transactions.

Create a new database cluster, dump the corrupted cluster with pg_dumpall, restore it into the new cluster and remove the old one.

You are running an old minor release (current is 11.10), so you are missing about a year of bug fixes. The cause could be a software bug or (more often) a hardware problem.