I’ve been hearing throughout my entire DBA career that 300 seconds is a good counter for PLE (Page Life Expectancy) to be above.
Paul Randal calls this “utter nonsense“. If anyone would have the right to say that, its Paul.
There was a time when this MIGHT have been a useful number…but the hardware has changed so much its no longer so. I “heard” that Paul actually made up that number just for documentation purposes…but that may just be a fun story. Ask Paul sometime.
Go here to find a good number: Page Life Expectancy isn’t what you think
While you are there, check out all of the posts that start with “SQL Server DBA Myth a Day” and see how many you thought were gospel truth 🙂
Thanks for reading,
Paul Randal says
No – definitely wasn’t me that made that number up.
Good to know…any idea where it came from?
Or the “Cost Threshold for Parallelism of 5 came from a desktop at Microsoft in 1997” story?