Re: delayed block cleanout (?) for active tx

  • From: "Alberto Dell'Era" <alberto.dellera@xxxxxxxxx>
  • To: "Shamsudeen, Riyaj" <RS2273@xxxxxxx>
  • Date: Fri, 22 Jun 2007 14:17:33 +0200

In short, every time this select statement is accessing the block, block is
cleaned out and updated with current SCN, even though there is nothing to be
cleaned out. Essentially, confirming what you stated in the first place
"null" cleanout. I don't know why though.

Thanks Riyaj,

your great analysis convinced me completely.

About the "why", all we know is that updating the csc scn  avoids
another block cleanout (and visit to the rollback segments) if the
block is requested at the same scn. The only scenario where
it could make a big difference is e.g. for the root block of the inner
table index on a nested loop; maybe it checks the current block
(tries to cleanout it) before searching for the CR copy everytime.

But I'm only guessing here.

Thanks again for your explanation,
Alberto

--
Alberto Dell'Era
"the more you know, the faster you go"
--
//www.freelists.org/webpage/oracle-l


Other related posts: