I recently posted some comments about some guidance offered by Microsoft when talking about the CXPACKET wait type. Nothing I said was wrong, no. But, evidently there are a few whip smart and agile people who work for Microsoft. Specifically, Rick Byham, who works for the team tasked with taking care of the SQL Server Books Online. Why am I calling Mr. Byham smart and agile. Evidently he saw the blog post and has completely updated the description for CXPACKET at MSDN:
Occurs with parallel query plans when trying to synchronize the query processor exchange iterator. If waiting is excessive and cannot be reduced by tuning the query (such as adding indexes), consider adjusting the cost threshold for parallelism or lowering the degree of parallelism.
WHOOP! That’s awesome work. Thanks Mr. Byham.
Of course, one of my best and most popular recent blog posts is now completely incomprehensible if people follow the link to MSDN, but I can live with that. Well done.
And the caution in this case? You never know who is going to read this stuff, so try to be nice when offering criticisms. I could be a little more respectful with my criticisms, especially since I’ve put my full share of mistakes and weak advice out in books, articles and on this blog. My apologies if my flippant approach ruffled too many feathers. I got two lessons out of one blog post.
[…] UPDATE: While this was good information, things sometimes do move quickly on the internet. Microsoft has updated their entry on CXPACKET waits. Read all about it here. […]
I’ve read your post and the original MSDN entry. Personally I think Rick is paying you a compliment. The original info from MSDN was accurate but it assumed many things. You tried to remove the assumption part by asking people to wait and figure out exactly what they are looking at before jumping. Rick changed the MSDN wording to reflect that philosophy, even if it was in (a lot) less words. You are both right but I appreciate you giving us the back-story. Everything makes more sense when you understand the context.
Thanks.
I’m still surprised that people are reading this stuff.
Grant, I agree with the previous poster. Thanks to you for pointing the original confusion out, thanks to Rick Byham for jumping on it and updating the description, and then kudos to you again for admitting that your are entirely human…
Dontcha hate/love when that happens?
While I think nice is…nice…I think what’s most important is that the criticisms and warnings be *constructive*.
I don’t like reading posts that point out errors, even crazy bad ones, in a way that is mocking, bullying and “I’m so smartish”, but offer no solution or recommendation to making it better.
Nice is still good. Constructive is even better.
Karen,
Touche!
You’re right (of course). Nice wasn’t really the right word. Respectful, constructive, these are much better words for the situation. I do feel a landed a little hard in the original post, especially since it was actually read by someone as opposed to another comment just tossed out into the ether.
I did get a thanks from Rick Byham because I offered an alternative, not just complaints.
Hmmm… I should point him at my SQL Howlers article on Simple-Talk.
Thanks Will. I’m quite impressed with the fact that Microsoft can actually act in such a nimble fashion. So frequently it feels like this stuff languishes through multiple versions.