The answer is as varied as the shop, but I rely on empirical methods whenever possible. I hate guessing. I'm not going to cop-out with an "it depends", so I wrote this, just for you:What do you actually do ?
Not isolated, especially in larger IT shops with hundreds of instances. I worked for a shop with over 25 DBA’s, and the VP (a huge goof), saw that a reorg helped one database and mandated that all databases be reorged periodically, regardless of justification! It was right there is the DBA’s MBO’s . . . .Have you worked for various other IT managers that have also insisted you rebuild indexes or was this just some really bizarre isolated case ?
In theory, agreed. But when it's MY database, I like to be there.If you must, and you have the space, you should never need to be there for the rebuild.
It's more than superstition, it's ego too, IMHO. Some IT managers are over-confident jerks, with Ivy degrees and massive ego's (I call them "edicated idiots", all theory, no pragmatism). As they say: "You can always tell a Harvard man, but you can't tell him much":Prove why something doesn't work, and demonstrate it, and it should not matter who makes the reasons, unless of course they are blinded by superstition.
PERFECT!"...a tale told by an idiot, full of sound and fury, signifying nothing. "
Ah, but what about advice that is targeted to senior DBA's?especially if written in a generalistic manner without supporting caveats, evidence or practical examples.
If I did suggest that, I'll retract it today, as soon as I'm on the air again.When you write an article that suggests validating the structure of all indexes in a database, rebuild an index if the height is greater than "x",
That I stick by, but ONLY for indexes that experience range scans and index FFS, and NOT FOR BEGINNERS. Robin Schumacher has a reproduceable proof, if you want to try it yourself:use 32K block size for indexes, etc. etc.,
Oh, sorry, I'm in the middle of a reorg . . . . ; )Your link doesn't currently work so I can't comment.
Thanks. Ping me via e-mail, so I don't miss it, please . . . .any comprehensive review will likely have to wait until afterwards.