The recent earthquake in New Zealand and this morning's devastating earthquake and subsequent tsunami in Japan set off a lot of discussion on twitter about DR planning and testing.
I wrote a blog post on our SQL Server Magazine blog this morning asking people to think about their testing – see here.
I'm conducting two surveys here. One is asking about testing your DR strategy and the other is asking about whether the DR strategy considers human nature in the event of a natural disaster. Depending on the disaster, people won't necessarily be thinking about the company as their first priority.
Feel free to leave comments about your DR strategy and testing experiences and I'll include some of them when I editorialize the results in a week or two. And by all means send me private email if you want to vent in more depth! (But nothing sent in email will be included in the editorial.)
3 thoughts on “Surveys: DR plan testing and considering human nature”
I would be curious to see how this breaks along say a region (quake or hurricane prone, tornado alley vs less frequent areas), company size or even industry. I know some larger organizations I have worked with are very good at this, however most of the smaller shops give me that deer in the headlight look! Often times, many think it is important and just give lip service but when you come in and start asking for a plan and a test date (or even $$$), then they back down and start a real discussion. In this case, I often find they don’t even have a comprehensive backup strategy. We start there and work our way up the chain and start asking questions. That’s always an enlightening topic!
I’m retired, so perhaps I can be more blunt than most. In nearly 40 years working in data processing, I campaigned in most places for developing and testing a DR plan. In only one case did I succeed in getting management to agree and then fund development and testing of a DR plan.
In that case, we had been purchased by another company and the first thing the purchasing company did was to send in an audit team. I "casually" mentioned to the audit team that we didn’t have a DR plan, so that went into the audit report. That got us funding since local management wanted everything on the audit report fixed before the next audit.
The thing that is neglected most often is testing the plan. And testing is needed. People who would never put an untested program into production calmly have a DR plan that has never been tested. And, if you need the DR plan, it better work the first time. In the company referenced above, the first two times we tested our plan, we discovered we were missing critical items. It didn’t work until the third test.
It’s not quite true that there wasn’t a DR plan at the other companies, I had my own personal plan. It consisted of an updated resume and a letter that read
Roses are red
Violets are blue
You had a disaster
And I’m leaving you
(And, yes, I tested the resume several times and it seemed mostly to work OK <grin>)