X-MimeOLE: Produced By Microsoft Exchange V6.5
Received: by onstor-exch02.onstor.net 
	id <01C76FD6.8B997D7B@onstor-exch02.onstor.net>; Mon, 26 Mar 2007 11:42:45 -0700
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Content-class: urn:content-classes:message
Subject: RE: TTE (Torture Test Environment) Plan
Date: Mon, 26 Mar 2007 11:40:34 -0700
Message-ID: <BB375AF679D4A34E9CA8DFA650E2B04E03036F15@onstor-exch02.onstor.net>
In-Reply-To: <BB375AF679D4A34E9CA8DFA650E2B04E019AF688@onstor-exch02.onstor.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: TTE (Torture Test Environment) Plan
Thread-Index: AcduZh1s6fLZq4lvSd6aoksjpBSDXgAqn/ngACtL6PAAA2bGnwACnxFw
From: "Raj Kumar" <raj.kumar@onstor.com>
To: "Jonathan Goldick" <jonathan.goldick@onstor.com>,
	"dl-Delorean" <dl-Delorean@onstor.com>

Replies..

-----Original Message-----
From: Jonathan Goldick=20
Sent: Monday, March 26, 2007 10:23 AM
To: Raj Kumar; dl-Delorean
Subject: Re: TTE (Torture Test Environment) Plan

The filesystem may not be the focus but you defined the configs pretty
closely.  If those config don't matter then why specify them? =20

** vsvr moves are effected by volume offline/online operations, and
inode cache timeouts. So I defined and documented these config
information to make easier to debug any issues.

Are there going to be no filesystem negative tests? If there are then it
likely does matter that the config is repeatable. =20

** Just wanted to have a FS rich configuration that is relevant to these
tests.

If this is strictly failovers due to crashes shouldn't we test the
largest supported cluster size and volume number?

** good point and we should. I will add this in the readmap.

Thanks.


-----Original Message-----
From: Raj Kumar
To: Jonathan Goldick; dl-Delorean
Sent: Mon Mar 26 08:48:53 2007
Subject: RE: TTE (Torture Test Environment) Plan

Replies..
_____________________________________________
From: Jonathan Goldick=20
Sent: Sunday, March 25, 2007 12:13 PM
To: Raj Kumar; dl-Delorean
Subject: RE: TTE (Torture Test Environment) Plan

Even after reading bullet describing the difference between this and
soak I'm really not clear on the value.  It may become clear when more
tests a re added but aside from inducing crashes I don't see what this
covers beyond soak.

** Soak doesn't cover any negative workflows. The purpose of this TTE is
to stress test our HA functionality with the repetitive vsvr
moves/crashes.=20


Are you intending to recreate/validate the objects in each file system
between runs so that you know that you start with exactly the same file
systems each time?  Unless you use an offline mirror to do this the
recreation time may take too long, or your file systems will have to be
artificially small.

** Sorry, didn't get you. The focus is not on the file system.


_____________________________________________
From: Raj Kumar=20
Sent: Saturday, March 24, 2007 3:45 PM
To: dl-Delorean
Subject: TTE (Torture Test Environment) Plan

Hi,

We are planning to implement a Torture Test Environment (TTE) as part of
our standard testing. The objective is to test our filer under heavy
stress and keep monitoring and recording our failure points (if there
should be any) and our success points, per submittals/releases basis. We
are planning to start the TTE for our Clustering/HA solutions.

Please have a look at the plan and let us know if you have any
suggestions/comments.

file://mightydog/QE/Projects/Delorean/TestPlans/TTE_Plan_Ver_0.1.doc

Thanks.

--kumar :-)
