I am certain we all have clients or customers that are, say, “high-maintenance”. For example, I have one individual who sends every email with that High Importance red exclamation point. This is most likely the same person who bullied their way to the front of the lunch line in grade school. “Pay attention to me, right now, and no one else. Drop everything you are doing, and focus only on my needs, which supercede anyone else.”
Honestly, it can be rather tiring. Esepcially since they are typically responsible for their own issues. Say, they cannot remember their password (oh, let me go look in the password table and find that for you), or they dropped a few indexes and now their queries are not performing well, etc. You get the idea.
Recently we had an issue that was “High Importance”. Seems that a particular box had much longer query times than others, and no one knew why. So, I looked into it a bit and found, as expected, much different query plans because this person does not understand how to compare apples to apples. I told them to get their schemas in order (i.e., decide which ones were the desired ones and deploy them to each box, so we could get a baseline), then get back to me.
Well, they did, only now things were even more critical because they had a production deployment coming up and the query performance was unacceptable. I agreed that it was unaccepatble, and set to work. And work I did, trying everything I could, measuring everything possible, trying to figure out what the problem could be. Turns out, the problem is OS related, not anything related to anything I can control. Something about Windows Server 2003 SP2 and the Scalable Networking Pack. For more gory details, go to:
http://www.microsoft.com/windowsserver2003/evaluation/news/bulletins/ws03net.mspx
and here:
http://msexchangeteam.com/archive/2007/07/18/446400.aspx
So, we spent hours trying to figure this out. Finally we spoke to one of the server guys who had been struggling with Exchange issues. When I described what I was seeing, he did not hesitate to suggest that we look into a few settings. So, he toggled the settings while I ran the query. We consistently got the query to run in 20 seconds, then 2 seconds, back to 20, then 2, until we were satisfied that we knew what was happening and more importantly what we should do about it.
If anyone else out there happens to be seeing some unexplained issues, have your server guys review the above two links and see what they can find. Hopefully I can save others from going Crazy.
But don’t expect to be thanked for your efforts, I am still waiting for someone to acknowledge our team for going out of our way on this one. We dropped everything in order to figure this one out, we were able to find the issue, and got nothing. Now, why exactly would I ever want to go through this again for them?
I wouldn’t. No one would.
A little positive energy goes a long way. If someone steps up to help you out, make certain you express your gratitude for their efforts. You’ll be surprised at what you get in return.
