At 02:58 PM 9/10/02 , Duncan Greenwood wrote:

According to the docs for the stat system call, a write should update both
mtime and ctime. I can't see how Oracle could avoid this this happening,
even intentionally.

A similar discussion has been going on over in the PostgreSQL mailing list,
with regard to their write ahead logs (WAL).  There used to be an option to
open the WAL with either a synchronous or an asynchronous write, with the
expected trade off between performance and security of data.

Several operating systems now support modes that are somewhere in between,
that will allow synchronous writes of the data, but delayed (or even
suppressed) writes of the metadata.  So the timestamps in the inode may not
reflect reality, and you might not care since the WALs are frequently the
performance bottleneck.

Check, for example, the O_DSYNC flag to open(2) under Solaris.

        -crl
--
Chad R. Larson (CRL22)    chad AT eldocomp DOT com
  Eldorado Computing, Inc.   602-604-3100
     5353 North 16th Street, Suite 400
       Phoenix, Arizona   85016-3228

--
Note: To sign off this list, send a "signoff" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=