Wednesday, October 27, 2010

Fault Tolerence or Faulty Expectations?

Isn’t it amazing how often you purchase something that is supposed to be fault tolerant and then we act surprised when it breaks?  It may be an HBA setup, a server cluster or even a storage unit.  I know it sounds funny, but think about it.  Every manufacturer offers a “bullet proof” piece of hardware for a small fortune.  Maybe like 5% of the world can afford it -so they make a cheaper version that is almost as good...  The difference is that the lower priced unit usually comes with some sort of truncated Service Level Agreement or maybe an entirely different architecture that suits most of your needs at the price point that has to be hit.  All that being said, it never ceases to amaze me when a unit fails -the manufacturer notes that things would have been handled differently had an uplifted service agreement been purchased or if you had a different unit with more features.

My point here is that if you need to compromise on a solution, make sure that you have put the proper steps in place to make sure if something does occur, you are ready.  This starts with something as simple as a DR plan which notes who to call in an emergency and what priority should be given to an application when everything needs to come back up.  It may even dig deeper into your organization with restore procedures or failover steps.  The bottom line is that if you just buy a piece of hardware and forget to back it up, replicate it, or create a procedure to limit your exposure -then maybe you are putting too much faith in hardware.  Tier 1 anything is quite expensive.  You can make the difference of how successful a lower priced item succeeds in your environment.  Remember –Stuff breaks –it is how you respond to it that can make the difference…

What steps would you recommend for a disaster recovery plan?

1 comment: