Re: SCN Jumping issue

From: Oracle List <"Oracle>
Date: Fri, 27 May 2016 18:29:01 -0500
Message-Id: <9949929B-651A-4A5F-A67E-61B1EE7AA34F_at_yahoo.com>



Thanks Andrey. I'm just about realizing how serous this thing is.

Just tested and confirmed that a dblink query, no matter which direction, would sync the SCN on both the DBs to the same value (higher of the two DBs). In a huge network of DBs this can have a big ripple effect and that's what we're seeing today.

> On May 27, 2016, at 4:02 PM, Николаев Андрей Серапионович <Andrey.Nikolaev_at_rdtex.ru> wrote:
>
> Hello Deepak,
> Actually, there are a lot of Oracle bugs that could lead to SCN overflow.
> Usually, the toughest part of diagnostics is to find the Oracle instance
> that poison the entire database network with the abnormal SCNs.
>
> You may find useful the following script:
> https://andreynikolaev.wordpress.com/tools/scn_stats-sql/
>
> It calculates the SCN Headroom history, average daily SCN rates,
> and intrinsic average SCN generation rates for the database.
> The rogue database has the int_scn_rate >= 16K.
>
> I wrote this script some time ago following the ideas of famous Riyaj
> Shamsudeen post
> https://orainternals.wordpress.com/2012/01/19/scn-what-why-and-how/
>
> Good luck.
> Andrey Nikolaev
> Latch, mutex and beyond
> http://andreynikolaev.wordpress.com
>
> --
> http://www.freelists.org/webpage/oracle-l
>
>

--
http://www.freelists.org/webpage/oracle-l
Received on Sat May 28 2016 - 01:29:01 CEST

Original text of this message