Thread: BUG #16851: Master DB was unreachable and both master,slave entered into Standby while performing the switchover
BUG #16851: Master DB was unreachable and both master,slave entered into Standby while performing the switchover
From
PG Bug reporting form
Date:
The following bug has been logged on the website: Bug reference: 16851 Logged by: Shaanthi Negathanan Email address: shaanthinegathanan.pm@gmail.com PostgreSQL version: 9.6.1 Operating system: Linux Description: We are facing an error while switchover the master to slave using efm promote dbname -switchover and both master and slave entered into standby mode. Did anyone faced this issue before or it is an bug? Below the error log Body: EFM node: 172.xxx.xx.xxx Cluster name: pg_xx_prod Database name: dbname VIP support: ENABLED An unexpected error has occurred on this node. Please check the agent log for more information. Error: Could not stop database: exit status: 1 error: Job for postgresql-9.6.service canceled. 1/24/21 11:49:32 PM com.enterprisedb.efm.nodes.EfmAgent stopMasterDatabase WARNING: Could not stop database. Will not be able to continue switchover Command: sudo /usr/efm-2.1/bin/efm_root_functions stopdbservice pg_apb_prod. Result: exit status: 1 error: Job for postgresql-9.6.service canceled. 1/24/21 11:49:32 PM com.enterprisedb.efm.nodes.EfmAgent startMonitoring SEVERE: Could not stop master database before promoting standby. Switchover cannot continue. 1/24/21 11:49:32 PM com.enterprisedb.efm.nodes.EfmAgent startMonitoring WARNING: Attempting to resume monitoring. 1/24/21 11:49:32 PM com.enterprisedb.efm.DBMonitor checkOnce WARNING: Could not connect to database at 172.xxx.xx.xxx: org.postgresql.util.PSQLException: FATAL: the database system is shutting down, state: 57P03 /24/21 11:49:32 PM com.enterprisedb.efm.nodes.EfmAgent startMonitoring WARNING: Can no longer connect to database. 1/24/21 11:49:32 PM com.enterprisedb.efm.utils.VipMonitor startMonitoring INFO: Starting vip monitor (to ensure this node isn't assigned the VIP) 1/24/21 11:49:32 PM com.enterprisedb.efm.nodes.EfmZombie <init> INFO: The auto resume period set to 0, so agent will stay IDLE until stopped or resumed manually. 1/24/21 11:49:32 PM com.enterprisedb.efm.nodes.EfmAgent setAgentDelegate INFO: setAgentDelegate: I was Master, will now be Idle. 1/24/21 11:49:32 PM com.enterprisedb.efm.nodes.EfmAgent setAgentDelegate INFO: Setting agent to type Idle 1/24/21 11:49:32 PM com.enterprisedb.efm.ClusterState removeAddress INFO: Removing master address dxxxxx-xxxxx 1/24/21 11:49:32 PM com.enterprisedb.efm.nodes.EfmAgent startMonitoring INFO: Setting state back to RUNNING instead of PROMOTE 1/24/21 11:51:33 PM com.enterprisedb.efm.nodes.EfmAgent doHandle INFO: responding to db status request with status: Idle%%172.19.146.217%%true%%false%%
Re: BUG #16851: Master DB was unreachable and both master,slave entered into Standby while performing the switchover
From
Magnus Hagander
Date:
On Wed, Feb 3, 2021 at 6:36 PM PG Bug reporting form <noreply@postgresql.org> wrote: > > The following bug has been logged on the website: > > Bug reference: 16851 > Logged by: Shaanthi Negathanan > Email address: shaanthinegathanan.pm@gmail.com > PostgreSQL version: 9.6.1 > Operating system: Linux > Description: > > We are facing an error while switchover the master to slave using > efm promote dbname -switchover and both master and slave entered into > standby mode. Did anyone faced this issue before or it is an bug? EFM is not a part of PostgreSQL -- it's a proprietary product from EnterpriseDB, and you should contact their support for help with it. -- Magnus Hagander Me: https://www.hagander.net/ Work: https://www.redpill-linpro.com/