Veritas-bu

Re: [Veritas-bu] single initiator zoning

2009-02-27 16:45:16
Subject: Re: [Veritas-bu] single initiator zoning
From: Ed Wilts <ewilts AT ewilts DOT org>
To: "Geyer, Gregory" <Gregory.Geyer AT avnet DOT com>
Date: Fri, 27 Feb 2009 15:21:11 -0600
On Fri, Feb 27, 2009 at 1:48 PM, Geyer, Gregory <Gregory.Geyer AT avnet DOT com> wrote:
Hey Ed,
 
I'm wondering why you say categorically that it has nothing to do with performance?  That's not exactly what I get out of their doc, and so far not what I'm seeing in testing. 

My rationale is based on how zoning works.  I simply don't see how it could affect performance at all.  However, if you can get your SAN admin to re-zone and notice an improvement, I'd sure be interested in knowing about it.
 
Here's the quick and dirty test: 

And therein lies the problem - most of us can only do "quick and dirty" tests.  To prove or disprove the zoning/performance issue, it takes some very controlled tests that we usually don't have the time to set up.

 
A backup to a drive zoned to a few systems gets 30MB/Sec.  The exact same backup, same media server, to a drive zoned to all gets 7MB/Sec.  The only difference I can see is that the first is zoned to 3 systems and is not in the library.  The second is zoned to all HBAs (perhaps 60, 5 different OSes) and in the library.

Breaking that big zone out from 60 HBAs and all the tape drives to 60 individual zones should not, in theory, make any difference.  However, it is possible that having 60 HBAs see the same tape drive does make a difference.  That doesn't mean that zoning is the issue - it could simply be the number of HBAs seeing the same tape drive.

.../Ed

Ed Wilts, RHCE, BCFP, BCSD, SCSP, SCSE
ewilts AT ewilts DOT org
_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu