Path: news.easynews.com!newsfeed1.easynews.com!easynews.com!easynews!news.he.net!news-hog.berkeley.edu!ucberkeley!newsfeed.stanford.edu!postnews1.google.com!not-for-mail From: yong321@yahoo.com (Yong Huang) Newsgroups: comp.databases.oracle.server Subject: Re: log file sync vs. log file parallel write ? Date: 12 Dec 2002 11:54:13 -0800 Organization: http://groups.google.com/ Lines: 26 Message-ID: References: NNTP-Posting-Host: 192.94.8.250 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1039722853 22076 127.0.0.1 (12 Dec 2002 19:54:13 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 12 Dec 2002 19:54:13 GMT Xref: newsfeed1.easynews.com comp.databases.oracle.server:169548 X-Received-Date: Thu, 12 Dec 2002 12:53:47 MST (news.easynews.com) bchorng@yahoo.com (Bass Chorng) wrote in message news:... > I really can't tell the difference from the above statements. > But look at my v$session_event of a session, these 2 are way > different: > > EVENT TOTAL_WAITS TIME_WAITED > ----------------------------------------------------------------------- > log file parallel write 1576423 20189 > log file sync 1779889 337294 > > So is log file parallel write part of log file sync ? If so, > what's the difference called ? How can it be so big ( from > the above statements, the difference would be the time > writing to redo log buffer which should be very fast ) ? It's hard to compare these two events because they're experienced by different sessions: log file sync by user sessions, log file parallel write (as well as log file single write) by background sessions specifically LGWR. Total waits and time waited could be more or less when you compare them (not always one is higher than the other). Remember that when user sessions are not doing anything, LGWR could still be busy writing, because log writing is not always triggered by session commit. Also, don't forget the wait by LGWR writing logfile headers, which is sequential. Yong Huang