Re: 'redo synch time overhead ...' statistics
From: Timur Akhmadeev <timur.akhmadeev_at_gmail.com>
Date: Thu, 3 Dec 2015 10:57:03 +0300
Message-ID: <CACGsLC+18XwSSctcnL0AmCQJRwDi+qmvC18O-SkbP758CYn1Lw_at_mail.gmail.com>
On Thu, Dec 3, 2015 at 6:28 AM, Chitale, Hemant K <Hemant-K.Chitale_at_sc.com> wrote:
Date: Thu, 3 Dec 2015 10:57:03 +0300
Message-ID: <CACGsLC+18XwSSctcnL0AmCQJRwDi+qmvC18O-SkbP758CYn1Lw_at_mail.gmail.com>
On Thu, Dec 3, 2015 at 6:28 AM, Chitale, Hemant K <Hemant-K.Chitale_at_sc.com> wrote:
> 12614085.8
Thanks for the note. As I understand Oracle tries to measure difference
between 'log file sync' and associated 'log file parallel write' wait and
if it is small, it means that most of the time was spent in actually
writing stuff to disk. High values suggest time lost in inter process
communication - on CPU/runqueue.
I'm not sure this new statistics will make a difference on an instance
level, but on a session level - most certainly.
-- Regards Timur Akhmadeev -- http://www.freelists.org/webpage/oracle-lReceived on Thu Dec 03 2015 - 08:57:03 CET