It is indeed a brand-new wait type that was added in a recent 2019 CU. I just added it to the waits library - see https://www.sqlskills.com/help/waits/sleep_bpool_steal/ for details....

Wow - that's messed up! All of us at SQLskills wish you a speedy recovery!

Anyone wants any of the non-sterilizable PLA ones, shoot me an email - paul@sqlskills.com. I ship in batches of 10s or 100s.

https://timradney.com/2020/04/13/3d-printing-for-covid19/?utm_source=ssc&utm_medium=pubemail

Hey Melissa - I can ship you some PLA ones no problem. Shoot me an email - paul@sqlskills.com. Thanks

We did a bunch of testing back in 1999 when the code was originally written, but I haven't done any large-scale empirical testing since then. Duration is a red herring...

Thanks for posting all this detail. It certainly looks like for this index, reorganize doesn't give you any gain, but I'm not seeing the proof of your generalization to never...

And btw, reorganize doesn't cause page splits. Random inserts (or row expansions) cause page splits. Just because reorganize isn't lowering the amount of free space, that doesn't mean it's causing...

Not taking it as a slam on me at all, nor is my arguing with you here a slam on you - I just don't like the generalization to never...

Correct. Reorg (and anything else) can't move rows to the right in the index structure (freeing up space), only to the left (taking up space). That's always been the design....

Jeff - can you explain this comment you made please: "NEVER USE REORGANIZE!!!! IT DOESN'T WORK THE WAY YOU THINK IT DOES AND IS A LEADING CAUSE OF PAGE SPLITS!!!"

Hi - unfortunately there's an error in each of your graphical representations of the index structure. For all non-leaf level pages on the left-hand edge of the b-tree, the first...

No problem. Look in your error log to see if there are any message 825s - that proves that it's EMC. See https://www.sqlskills.com/blogs/paul/a-little-known-sign-of-impending-doom-error-825/

You're getting transient bad page reads from your I/O subsystem - that explains the zero'd page header and the 'weird' behavior you're seeing. Time to get I/O subsystem diagnostics running...