Edge Private Cloud 4.19.01 - Unable to install PostGres Slave node for 5 node topology

priniren
Participant I

I have configured a 5 node Apigee Edge v4.19.01 and been able to onboard the organizations, create environments and test few sample API proxies successfully. ZK, CS, MS, OpenLDAP, Edge UI are up and running. The analytics components are installed and started only on the 4th node(Master) and we are able to view the analytics. However, the PostGres SQL and Slave in the 5th node are not getting installed. Status of the components on the 5th node(PostGres Slave node) is :

[root@machine-name user-name]# /opt/apigee/apigee-service/bin/apigee-all stat us
+ apigee-service apigee-postgresql status
apigee-service: apigee-postgresql: Not running (DEAD)
+ apigee-service apigee-qpidd status
apigee-service: apigee-qpidd: OK
+ apigee-service edge-qpid-server status
apigee-service: edge-qpid-server: OK
apigee-all: Error: status failed on [apigee-postgresql]

I have tried uninstalling and clearing the RPMs and repeated the installation. The issue persists. The following error is shown in the command prompt :

transaction log start point: 0/4000028 on timeline 1
pg_basebackup: starting background WAL receiver

0/32031 kB (0%), 0/1 tablespace (...-postgresql/pgdata//backup_label) Since this was fresh install, I attempted to set up the PostGres replication at install time using the link https://docs.apigee.com/private-cloud/v4.19.01/set-master-standby-replication-postgres#configuremast...

I am seeing some random errors in the "postgressql.log" on the Slave Node as shown below :

< 2019-02-04 11:48:32.120 UTC > ERROR: relation "analytics.group_id_seq" does not exist at character 16
..
< 2019-02-04 12:28:13.852 UTC > ERROR: base backup could not send data, aborting backup
< 2019-02-04 12:28:13.852 UTC > LOG: could not send data to client: Broken pipe
< 2019-02-04 12:28:13.852 UTC > FATAL: connection to client lost
< 2019-02-04 12:40:50.587 UTC > LOG: could not send data to client: Connection timed out
< 2019-02-04 12:40:50.588 UTC > ERROR: base backup could not send data, aborting backup

Referred to several community links which talk about "pg_hba.conf" file. It appears to have the IP address of the master and slave to allow replication.

Looking for clues to debug the above issue. Meanwhile, I have looked at https://docs.apigee.com/private-cloud/v4.19.01/set-master-standby-replication-postgres#configuremast... which talks about setting the replication after the installation. But I was attempting to configure it at install time. Please guide on how to configure the 5th node as "another PostGres SQL and server" and later set it up as a Slave.

The logs from Master and Slave are attached.

0 2 540
2 REPLIES 2

priniren
Participant I

@Paul Mibus@Carlos Frias Kindly assist

Not applicable

if it is a new setup can you just uninstall postgres both the components from 5th node and also clean all the data or catch its having. Then reinstall postgres components again and check whether both are running or not.