Re: Frequent failover

Ron <ronljohnsonjr@xxxxxxxxx> · Tue, 7 Nov 2023 07:09:17 -0600



    Have you scoured the Patroni and Postgresql logs?

    
    Have you increased the logging level?

    
    On 11/6/23 23:44, Rajesh Kumar wrote:

    
      Is there a way to narrow down the cause of
        failover? Yesterday itself failover happened 7 times.
      

        On Mon, 6 Nov, 2023, 1:34 PM
          kaido vaikla, <kaido.vaikla@xxxxxxxxx>
          wrote:

        
            Hi

              First of all, read a logs.

              $PGDATA/log

              #journalctl -u patroni 

              I had once failover flapping, because of master heavy
              load, caused by apliccation change. So, find a pattern.

              
              If physical replication, after failover you don't need run
              a vacuum manually if autovacuum is enabled.

              
              br

              Kaido

            
            On Mon, 6 Nov 2023 at
              04:53, Rajesh Kumar <rajeshkumar.dba09@xxxxxxxxx>
              wrote:

            
              Hi,
                

                When I issue command patronictl history,
                  I can see frequent failover. Like atleast weekly once.
                

                Every time this happened we have to do
                  vacuum analyse for all dba.
                

                So, I need assistance in how to avoid
                  frequent failovers and what is the solution for vacuum
                  part we r doing..
              
            
    -- 

      Born in Arizona, moved to Babylonia.