Running Jetstress on SCC nodes


I am building an Exchange 2007 SP1 cluster, A/A/P, using SCC.

The CMS's are Exch1 and Exch2. The CNO's are ServerA, ServerB, ServerC - All Windows 2008 servers.

I need to run Jetstress to ensure the disk subsystem/SAN can cope with the requirements, however am not sure if I need to run the Jetstress on ServerA and ServerB, or on each physical node?

Secondly, I understand that we'll be using Node Majority for our Quorum, but I still can't get how this works in my head. Does anyone have a simple explanation/link?

Many thanks in advance.
Who is Participating?

[Product update] Infrastructure Analysis Tool is now available with Business Accounts.Learn More

I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

Ok so JetStress First.  Personally when I have done this in the past I run it on the Virtuals EXCH1 and EXCH2 in your case.  However, then I make sure and fail over to the passive and test it there as well.

The backend storage is the same no matter when so testing that singularly is ok, however if you don't test on the passive node when it has to take over, that is not the time to find you have a bad storage controller.  I figure it is always better to test everything just to be sure.

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
kam_ukAuthor Commented:

Thanks for answering...

>"Personally when I have done this in the past I run it on the Virtuals EXCH1 and EXCH2 in your case"
But I haven't created the virtuals yet since I haven't installed Exchange on the servers yet?
Glad to help.  I just don't like leaving a server untested just in case.

MNS is a little harder.  I will see if I can find you a good doc that breaks it down.  Seems like most either get how ti works or it is cloudy to say the least so you are certainly not alone!!
Big Business Goals? Which KPIs Will Help You

The most successful MSPs rely on metrics – known as key performance indicators (KPIs) – for making informed decisions that help their businesses thrive, rather than just survive. This eBook provides an overview of the most important KPIs used by top MSPs.

Jamie McKillopIT ManagerCommented:

Since you have a A/A/P, I would run jetstress against the two active nodes simultaneously. That will give an accurate representation of the system load you will experience. Since all nodes should be identical, there is no need to move the virtual servers to other nodes and run it again.

Node Majority works by having each node "vote" on whether to initiate a failover. At least 50 percent of the nodes must be availbale to make a quorum. This prevents a single node that is isolated from taking control of the cluster. Let's say node 3 is the passive node and its network card fails. It no longer sees the active node as up but it won't take control of the cluster because it can't communiate with either of the other nodes to form a quorum. Now let's say node 1 fails. Node 3 will see node 1 went down. It will ask node 2 if it also sees node one as down. They both agree node 1 is down and both vote to move the virtual server to node 3. 2 out of 3 is a majority vote and the virtual server is moved.

JJ brings a good point except I still like to test the passive node simply because all nodes identical does not mean identical hardware.  I have had servers fail JetStress due to the servers Storage Controller.
Jamie McKillopIT ManagerCommented:
When you build a cluster all your nodes should be completely identical. That is the assumption I'm going off. Using mixed hardware is not recommended, but if you are, you should certainly run jetstress against each unique configuration.

Agreed. But even identical hardware means a controller in each machine.  Bascially if you have QLogic Fiber Channel cards for instance you are going to have 3 of them (possibly 6 for redundancy and such of course)  But if you only test 2 of the 3 you are not testing that third card, even though they are all the same Type of card.  Doe that make more sense?

To JJ's point though they should all be the same for sure!!  You don't want 3 different type of controllers.

Jamie McKillopIT ManagerCommented:
So, you are talking about identical brand/model of controller in each node? If so, you are referring to testing for manufacturer's defect. I see what you are doing but jetstress is a tool intended to lest the performance limits of a systems. It isn't realy intended to diagnose hardware faults. The hardware manufacturer should have tools that are better suited for system burn in and testing for hardware defects that you should run prior to jetstress.

Indeed I agree.  However to test the performance limits of the system still relys on that controller in the box.  I would agree manufacturers should have tools.

I just don't personally like leaving a machine worth of variables not tested.  Server BUS, Controllers, cables...   I have seen boxes test at X and then you failover and test the other it is only performing at 1/2 X...   I figure it is very easy to test why risk it and just test everything.
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today

From novice to tech pro — start learning today.