Many customers have expressed an interest in maintaining a hot spare Application server in case of a catastrophic hardware failure in their Network Configuration Manager production Application server. Because of the distributed architecture of Network Configuration Manager software, and hot backup capabilities of the backup utilities, a hot spare server can be maintained in a ready standby state, able to take over the production application role in a matter of minutes.
To set up a hot spare scenario, two identical Application servers are required to be installed withunique IP addresses on the same subnet. Both servers should have identical versions of the Network Configuration Manager Application (or combination) Server software installed. However, only the primary, or production server, is configured to communicate with any distributed Device servers.
Once the primary and hot spare Application servers are configured and running, the root crontab can be edited on the primary server to copy the nightly backup bundle to the [Product directory] /data-image directory on the hot spare server using rcp or scp commands. (Note that there are a variety of ways you can copy the backup file from one server to the other, originating the command from either the primary or the hot spare server. This is only one example).
Once the backup bundle is on the hot spare server, it can be restored using the [Product directory] /tools/restore command (see the “Backup utilities” on page 224) making the network environment on the hot spare identical to the production server. This process can be repeated nightly through the use of cron, so that the two servers stay in sync.
The refresh rate of the sync process is up to your requirements. You may choose to sync the network data every four hours, or more frequently. There is a performance penalty on the production server during a backup, and large environments should take care when determining the frequency of backups.
Once the sync process is in place, if the primary server ever goes down, it is a simple process of executing a ‘takeover’ script on the hot spare sever that will change its IP address to that of the primary, and restart network services. Once the hot spare server has successfully obtained the IP address of the primary, it will communicate with all Device servers and user clients.