Customer Story: How high availability can prevent downtime

Around the office water cooler, while some people talk about the latest celebrity gossip, we talk about the “application failover guy.” Take that tidbit as you will…but here’s the story.

When we made the decision to build in high availability directly into the client (rather than rely solely on the server) we knew it would have big impact on the way applications handled various types of connection failures. But we had no idea it would be used the way “failover guy” did it!

Failover guy’s company had eight application servers but only one database. Every Friday, the company had to take the database down for maintenance. During the scheduled downtime, all of the application servers would crash. On Monday morning, the team had to restart everything. Needless to say, their process was less than stellar.

He decided to try something innovative with our drivers, tapping in to the high availability functionality in a really cool way. They wondered if our failover technology could benefit them. Since the logic for handling connection failures was completely client side (in the driver), they set the primary and alternate servers in the connect string to their single database. Then, using our connectionRetryCount and connectionRetryDelay options, they were able to set the driver to attempt a re-connect to the same database once every 5 minutes for an hour. We had not thought of using our failover to connect to the same server (never popped up as a use case for us) but nevertheless, we told them that it should work and that they should give it a shot. We got an excited call back saying that things were working great and that their servers just kind of paused during that downtime every Friday and the rebooting was a thing of the past; that’s just good technology! This is just one example of how our customers never cease to amaze us! Do you have a story similar to failover guy’s?