Re: Effect of number of extents on Oracle I/O performance
From: Nuno Souto <dbvision_at_iinet.net.au>
Date: Sat, 14 Mar 2009 20:23:44 +1100
Message-ID: <49BB77A0.9090309_at_iinet.net.au>
Mark W. Farnham wrote,on my timestamp of 14/03/2009 10:58 AM:
Date: Sat, 14 Mar 2009 20:23:44 +1100
Message-ID: <49BB77A0.9090309_at_iinet.net.au>
Mark W. Farnham wrote,on my timestamp of 14/03/2009 10:58 AM:
> 2) When the extent size is very small (smaller than a multiblock read) then
> in cases where a multiblock read is interrupted by the extent boundary every
> time, then many small extents can be noticed. (If the extents are even a
> reasonable size, then the collision interrupting a multiblock read quickly
> becomes an infrequent event in a full table scan and the overhead disappears
> in the noise.
Seen this in some of the tablespaces in our DW. Unexplained slowdowns of FTS. Since then I've changed them to UNIFORM with size 50M or 100M, depending on the tables kept there. No more problems.
-- Cheers Nuno Souto in sunny Sydney, Australia dbvision_at_iinet.net.au -- http://www.freelists.org/webpage/oracle-lReceived on Sat Mar 14 2009 - 04:23:44 CDT