Comments on: Why does the same job run slower on an exact clone? http://dboptimizer.com/2011/11/15/why-does-the-same-job-run-slower-on-an-exact-clone/ database performance, SQL tuning and data visualizatoin Sat, 03 Mar 2018 05:02:37 +0000 hourly 1 http://wordpress.org/?v=3.6 By: Kyle Hailey http://dboptimizer.com/2011/11/15/why-does-the-same-job-run-slower-on-an-exact-clone/#comment-12754 Kyle Hailey Fri, 18 Jan 2013 18:35:49 +0000 http://dboptimizer.com/?p=1787#comment-12754 @jsm: Carlos’ tools are awesome. Good timing on your post because our NoCOUG board is hosting Carlos for a 2 day sql tuning seminar with his tools!
https://nocoug.wildapricot.org/Events?eventId=605183
Feb 19-20, 2013
CarrAmerica Conference Center, Pleasanton, CA
8 Spaces left as of today

]]> By: jsm http://dboptimizer.com/2011/11/15/why-does-the-same-job-run-slower-on-an-exact-clone/#comment-12740 jsm Fri, 18 Jan 2013 16:06:06 +0000 http://dboptimizer.com/?p=1787#comment-12740 Take a look at Carlos Sierra’s SQLT tool downloadable from support. It has an “extract” method and a “compare” method. Compare will tell you EXACTLY what is different between the two extracts down to the nitty gritty of the optimizer, the plan, and the details of all the supporting statistics.

]]> By: Tom Robbins http://dboptimizer.com/2011/11/15/why-does-the-same-job-run-slower-on-an-exact-clone/#comment-11970 Tom Robbins Tue, 08 Jan 2013 02:29:04 +0000 http://dboptimizer.com/?p=1787#comment-11970 @Kyle Hailey
Thanks. for some reason, I had not looked at the dbwr parallel settings in the pfile/spfile….

]]> By: Kyle Hailey http://dboptimizer.com/2011/11/15/why-does-the-same-job-run-slower-on-an-exact-clone/#comment-10978 Kyle Hailey Wed, 19 Dec 2012 18:46:51 +0000 http://dboptimizer.com/?p=1787#comment-10978 @Tom Robbins: There is nothing inherent in a virtual database that would make parallel go to single threaded. Have you compared the init.ora/spfile parameters from the two databases? By default, Delphix will keep most of the init.ora parameters the same as the source but there are a few that can be removed or changed for various reasons. Also if you have created an init.ora template for your virtual databases then of course that could explain differences between the source database and virtual database.
The first things to look at
1. init.ora/spfile on source and virtual
2. table parallel settings on source and virtual
3. of course compare the insert code itself on source and virtual
Feel free to email me at kyle.hailey@delphix.com

]]> By: Tom Robbins http://dboptimizer.com/2011/11/15/why-does-the-same-job-run-slower-on-an-exact-clone/#comment-10967 Tom Robbins Wed, 19 Dec 2012 15:25:15 +0000 http://dboptimizer.com/?p=1787#comment-10967 Kyle,
We are using Delphix and it is great for creating a fresh test environment as a clone of production.

One thing we are struggling with, though, is a large INSERT statement of a 5GB table:
even though we have PARALLEL degree set to the default for the database, and we have enabled parallel DML, the INSERT still runs single-threaded, and therefore takes a LONG time (72 minutes).

Is there something in Delphix that causes this single-threadedness?

Thanks,
Tom

]]> By: Kyle Hailey http://dboptimizer.com/2011/11/15/why-does-the-same-job-run-slower-on-an-exact-clone/#comment-858 Kyle Hailey Wed, 16 Nov 2011 01:12:38 +0000 http://dboptimizer.com/?p=1787#comment-858 The database was cloned using Delphix

]]> By: Vishal Desai http://dboptimizer.com/2011/11/15/why-does-the-same-job-run-slower-on-an-exact-clone/#comment-857 Vishal Desai Tue, 15 Nov 2011 19:37:03 +0000 http://dboptimizer.com/?p=1787#comment-857 Hi Kyle,

What method was used to clone production database ?

Thanks,
Vishal

]]>