Best Practice for Sync Between Primary and Failover

Although you can configure a cron job to run configsync in the background, in your transactional code elsewhere, you still won't have any feedback that the configsync process is successful or not (unless you don't care if the new configuration actually can load on the peer machine). As you may have known, configsync is an involved process in which configuration from the source box must be packaged up and transferred to the peer box, and the new configuration must be loaded successfully on the peer. If you have a large configuration set with 1000s of VSes, pools and servers..., this may take some time to finish. It's the reason configsync is not seen as something that is done on a regular basis. Even when new servers are added/removed from a LB pool, typically configsync should only be done when everything is stabilized. The CLI, GUI, and iControl all use the same mechanism to configsync, so your cron job (which will likely use a CLI command) will also take approx the same amount of time to finish.

 

 

In the next release of BIGIP, we'll have an auto-configsync feature, in which if enabled, configsync will be automatically done when it detects that something has changed.

 

 

Thanks,

 

Loc