How to make a node to take over primary role in Identity manager.

Products

• VCF Operations/Automation (formerly VMware Aria Suite)

Issue/Introduction

Symptoms:

• Failover script failing to make specific node to take primary role.
• Sometimes a problematic node that is not keeping the delegateIP address may require that we designate a different node as the primary/master node instead of removing it from the cluster.

Environment

VMware Identity Manager 3.3.7

Resolution

Make sure to take non-memory quiesced snapshots of all the nodes in vCenter.

To configure a specific node as primary node in vIDM:

1. Stopping services:
    1.a Stop vIDM services on standby nodes followed by primary node.
    service horizion-workspace stop
    /etc/init.d/opensearch stop

    1.b Stop database services standby nodes followed by primary node.
    /etc/init.d/pgService stop
    service vpostgres stop

2. Hosts file:
    cat /etc/hosts on all nodes ensure that the hostname master is pointed to the correct IP in the hosts file:
    2.a. If the master hostname on any nodes needs to be corrected, follow the steps:
    vi /etc/hosts
        Arrow down to the line master
        press i to edit the line
        replace the old IP with the correct one
        press ESC to stop editing
        press :wq to save and exit the file

3. Recover conf file:
    3.a On the new primary database node:
    mv /db/data/recovery.conf /db/data/recovery.done

    3.b On the old primary database node:
    mv /db/data/recovery.done /db/data/recovery.conf

    3.c On the old primary database node, fix the file owner:
    chown postgres:users /db/data/recovery.conf

4. Delegate IP:
    4.a If the delegate IP is on old primary node:
    ifconfig eth0:0 down

    4.b On the new primary node:
    ifconfig eth0:0 inet delegateIP netmask <netmask>

5. Starting Database services:
    5.a Start the services on new primary node followed by standby nodes:
    service vpostgres start
    /etc/init.d/pgService start

8. Verification:
    su root -c "echo -e 'password'|/opt/vmware/vpostgres/current/bin/psql -h localhost -p 9999 -U pgpool postgres -c \"show pool_nodes\""
    su root -c "echo -e 'password'|/usr/local/bin/pcp_watchdog_info -p 9898 -h localhost -U pgpool"

9. If everything looks good start vIDM services on new Primary node followed by standby nodes.
    service horizion-workspace stop
    /etc/init.d/opensearch start