Press "Enter" to skip to content

Understanding Long Failover Times for Availability Groups

Sean Gallardy has answers to your Availability Group questions, as long as you ask the specific question in this post:

One of the most common issues I look at from day to day is some variation of the question. “Why did it take a long time for my AG/Database to failover?”. There are many different meanings for this innocuously simple looking statement, for example was it that the failover time was long or was it a long time bringing the database online, or was it that it took a long time because a failover wasn’t possible, and what *exactly* is a long time? Are we talking a long time means 10 seconds, 1 minute, 5 minutes, 30 minutes? To each different business and their needs, “long” dramatically fluctuates. I’d like to go through at a high level, some of the most common reasons that I troubleshoot and if they might apply to your environment. FYI, if you tell me 1 second is a long time then I’m going to point you toward different architectures with multiple layers of caches and front-end servers/services which isn’t going to be cheap, but that’s what you want so you’re _willing_ to pay for it, right? Yeah, I thought not.

Click through for several factors which may affect how long it takes for a failover to occur.