I have resisted creating a blog for two reasons: 1) there are never enough hours in the day, and 2) the areas that I know well enough to say something interesting about -- SQL Server performance and some other engine issues, mainly -- are esoteric, not exactly People magazine-type fare that would entertain the masses.
Recently, though, I stumbled across the SQL Query Optimizer team's blog (http://blogs.msdn.com/queryoptteam) and Ian Jose's blog (http://blogs.msdn.com/ianjo/). These are both good reads and really informative. These blogs are primarily about the SQL QO, and both manage to get very useful information out there that might otherwise remain locked up in Microsoft. I think the typical database developer or DBA would benefit from reading them, and that made me feel better about my second concern.
Which left my concern about not having enough time to do something new. But then I heard that Ken Henderson has somehow found the time to hold down a full-time development job, write best-selling books, and still maintain three different blogs. I was forced to abandon my theory that the gods have conspired to make our days too short, and concluded that maybe my first concern really boils down to crappy time management.
I guess we'll see how it goes.
Please see if you can provide information on the sp_blocker_pss90 if there is such thing? We got new waittype and the sp_blocker_pss80 is no longer working. I am sure a lot people out there waiting on this for SQL Server 2005.
Thanks a lot,
The problem for the sp_blocker_pss80 is we have the KB 271509 out there contains 2 create procedure statements, both create the same old sp_blocker_pss80 to select processes with waittype > 0, we will have all the system process returned with waittype 0x0060, 0x007f, 0x0080, 0x009d, 0x00a2, 0x00a9, 0x0080, 0x007e, 0x00ad, 0x00a9. They are not really waiting on any resource and we shouldn't have them listed. Is there a way we can eliminate these processes? These waittype are not showing in KB 822101 except 0x0081. But if you look the 0x81 in the kb 822101 it was wait for writelog, but look in the last waittype in the sysprocesses, it's CHECKPOINT_QUEUE for waittype 0x81. Looks like it's not the same in 2005 any more. Should both KB be updated for 2005? We need a way to eliminate all the waittype which is not really waiting on a resource so we can have no row returned where there is no wait or blocked process.
According to BOL the sysprocesses should no longer be used and we should use the dynamic view, would this work better for our blocking script?
where session_id > 49 AND wait_duration_ms > 0