This thread leads me to a question. Does the Unix syncer come into play when direct IO is used? I believe it does with most FS caching, which means that even synchronous writes can't really be guaranteed, doesn't it? At least that's what I was lead to believe... Rich -----Original Message----- From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Hemant K Chitale Sent: Tuesday, September 26, 2006 9:28 AM To: juancarlosreyesp@xxxxxxxxx; Oracle-L Freelists Subject: Re: Important note about asynchronous commit I wouldn't know how to write a very reliable application that guarantees that it IS ALWAYS able to replay transactions in case of Instance failure. === oh yes ! There is one such : The Oracle RDBMS Engine itself. To date (except for one particular buggy 6.0.28 (or was it 6.0.29 ?) release), I have seen Oracle Instances CRASH AND RECOVER consistently. Play around with the guaranteed mechanism and you are playing dangerous with your data. [snip] -- //www.freelists.org/webpage/oracle-l