Performing Failover of the Master Server

To perform failover of the master server

1. Stop CTERA Portal on all servers, by running the following command:

   ctera-portal-manage.sh stop

2. On all servers, update the master server IP address:
   1. Open portal.cfg, by running the following command:

      vim /etc/ctera/portal.cfg

   2. Locate the following line:

      CTERA_DB_ADDR=master IP address

   3. Replace master IP address (the IP address of the failed master server, which we will call server1) with the IP address of the warm standby that will become the new master server. (We will call this new master server server2)
3. If the failed master server (server1) had additional warm standbys:
   1. Open portal.cfg on each warm standby, and find the following line:

      CTERA_REPLICATION_FROM= master IP address

   2. Replace master IP address with the IP address of the new master server (server2).
4. On server2, run the following command:

   ctera-failover.sh become_master

5. Start server2 (the new master server), by running the following command:

   ctera-portal-manage.sh start

6. If a CTERA Cloud FS storage node was configured on the old master server, change the storage node's IP address using the CTERA Portal web interface:
   1. In the navigation pane of the Global Administration View, click Main > Storage Nodes.
   2. The Storage Nodes page appears.
   3. Select the CTERA Cloud FS storage node and click Edit. The Storage Node Wizard appears.
   4. Click Next.

      

   5. In the Host Address field, change the IP address.
   6. Complete the wizard.
7. If the old master server (server1) is repaired and you want to convert it to a warm standby of server2, run the following command on the old master server (server1):

   ctera-failover.sh become_replica

8. Restart all servers other than the master server, by running the following command:

   ctera-portal-manage.sh start