How important is high availability? The best person to ask is Tom. See, Tom had a single controller SAN about 10 years ago in his IT environment. The SAN housed 12 HDD’s in a fully populated model with a RAID 5 and hot spare as his primary methods of data protection. Tom thought that this would ensure his data was safe, and would also keep his environment up and running in the event of a hardware failure. What Tom failed to take into account was that his SAN only had one controller, one power supply, no redundant SAN and no Hot DR site. So when Tom’s only controller failed, he was surprised that his environment went down and his business came to a screeching halt. His sales people couldn’t enter orders, his accounting staff could not issue PO’s, and his frontline workers could not access customer information. When Tom contacted his SAN’s manufacturer, they said it was too late in the day to overnight him a new controller, meaning his new controller would not be there for two days. He knew his business would go on to lose thousands of dollars in revenue. In a word, Tom was screwed. So he did what any sane IT admin would do in this dire situation. He booked the first flight to Nepal and became a mountain climber. Sadly, he only purchased one winter coat, and after it was torn to shreds by an angry pangolin, he was left without anything to keep him warm. He once again failed himself by not having a redundant solution in place. Tom’s story does not end well, but yours can.
The moral of the story is that any IT administrator should make every effort to create a high availability environment.. This means maximizing sensible redundancy. Sensible redundancy means that the environment should have redundancy built in, but not in such a way that interferes with the environments efficiency or business practices. Sure, you could have 6 SANs with identical copies of data, spread out over each continent of Earth. But the likelihood of needing that many SANs at one time is virtually zero. In reality there are just a few boxes you want to tic, to make sure that the environment and its organization can remain functional in the event of a failure.
Have dual controllers in your SAN. In fact, have dual everything. Dual power supplies, dual sets of NIC ports, dual fans, etc.
Have a backup or DR SAN that can support the organization in the event that the original SAN is rendered unusable.
Consistently look for points of failure. Play out different scenarios in your head and understand how they would affect the infrastructure. I.e. if a controller fails, how long before the new one is shipped in, can I failover to the other controller seamlessly, should I have a spare controller on standby?
Having a high availability IT infrastructure does not have to be hard. Yes, it will be more expensive than doing everything on the cheap, but it will be far more costly when a failure occurs. All hardware will fail eventually, so account for it, plan for it and when the time comes, execute your plan. And for the love of all that is good, don’t end up like Tom. Visit StorTrends.com today!