Re: DBWR_IO_SLAVES and Asynchronous I/O

From: Yong Huang <yong321_at_yahoo.com>
Date: 28 May 2001 22:05:34 -0700
Message-ID: <b3cb12d6.0105282105.c03ba4a@posting.google.com>

Your description is very clear! One comment I want to make is that because of file write lock (inode lock), it's a good idea to use several files for one tablespace instead of one file, so that in case of multiple writers writing to different blocks that are so far apart as to reside in two files, there's no file lock contention.

Yong Huang
yong321_at_yahoo.com

"Spencer" <spencerp_at_swbell.net> wrote in message news:<S%jQ6.734$My3.270852_at_nnrp3.sbc.net>...
> note that multiple processes means more overhead (more
> context switches), and with filesystems (on hp-ux) only one
> process can obtain a write lock to a file at a time, which
> means that multiple processes writing to the same file at
> the same time are going to block each other.
Received on Tue May 29 2001 - 00:05:34 CDT