On Wed, Feb 27, 2019 at 12:39 PM Stephen Eilert <cont...@stepheneilert.com> wrote:
> Are you running Vacuum on the slave node? It has to run on the master. > > Thanks, > > – Stephen > Thanks Stephen for your advice Iam running the VACUUM command from 2 of the linux instances using the below command. We were running some scripts from these linux boxes and surprised to see 1 works the other does not so wondering why does it happen. * Command:* *VACUUM VERBOSE myTableTest;* 1) Running from Linux Instance1 Command Iam running -> psql -h clusterName -U myUserName -d myPostgresDB There is no .pgpass setup on this linux instance so I had to manually enter the password here. VACUUM on the table Iam running does not work and throws an error as ERROR: cannot execute VACUUM during recovery 2) Running from Linux Instance2 Command Iam running -> psql -d myPostgresDB -h clusterName -U myUserName This option had a .pgpass file at the root in this linux instance and I did not pass any password here, the *VACUUM on the table Iam running works.* *pgpass is setup here and contains -> clusterName:5432:myPostgresDB:myUserName:myDBPassword* Thanks - Kran. > On Feb 27, 2019, 6:43 AM -0800, github kran <githubk...@gmail.com>, wrote: > > Hello Team, > > We are using a PostgreSQL 9.6 and seeing the below error while trying to > run a VACUUM on one of our live tables running in Production. We wanted to > clean up some DEAD tuples on the table. > > > *Command*: VACUUM (ANALYZE,VERBOSE) table_name. > > ERROR: cannot execute VACUUM during recovery > > Thanks > Kran > >
