Thanks for the reply.
For Solaris 8, I think from update 3 there is some workaround for this problem.
I was wondering because I have some customers who have a mount point per database, I am not sure for a busy database they would run into I/O problems?
The file level write locks still apply to many filesystems, even the ones which support direct IO.
However, some vendors have introducted concurrent direct IO option into their filesystems. Solaris has CIO since some update of Solaris 8. I've used CIO option on AIX 5.2 as well.
I haven't heard about CIO for Linux so I guess they don't have it.
Also, you have to manually enable concurrent IO in filesystem mount options in order to use it.
For Oracle 9.2 on AIX CIO stuff, read this article:
------------------------------ *From:* oracle-l-bounce@xxxxxxxxxxxxx [mailto: oracle-l-bounce@xxxxxxxxxxxxx] *On Behalf Of *LS Cheng *Sent:* Monday, September 11, 2006 16:38 *To:* _oracle_L_list *Subject:* UNIX file system read/write locks
I was wondering if UNIX/Linux filesystem read/write locks still applies in modern filesystems such as qfs/vxfs/OnlineJFS/ext2/ext3?