Check out the latest free show: SSWUG TV – We’ve unveiled a new name for our video-based content, which will be one of many upcoming site revitalization efforts. In this episode, we have a two-part interview with SQL Azure MVP Ike Ellis and updates coming from the Microsoft Build conference.
[Watch the Show]
Featured Article(s)
The State of Database Administration 2011: Types, Trends and Technologies (Part 3)
Technology can be helpful. We need to adopt more intelligent automation of administrative tasks. This means getting more automated, more connected, more wireless, and more complex.
Featured White Paper(s)
How to Implement an Effective SharePoint Governance Plan
Moving past the "what" and "why" of governance, an even harder question to answer is how to implement SharePoint governance. … (read more)
September Survey – 7 Questions to WIN
We’re giving away not one, but two 2-year memberships to SSWUG.ORG – all you need to do is give us your thoughts about SQL Server and Bananas. No, really. It’s only 7 questions, and you can win…
Seriously, could use your feedback – please stop by the survey site and take our quick survey – I’ll be sure to share the results when the survey is completed in one week.
Auto-Pilot and Your Systems – Trust but Verify
Some excellent feedback on my note about watching over your automated systems.
Here are just a few of the responses – Terry wrote "With our Enterprise environment and minimum staff to cover it, we rely heavily on alerting. Any tasks or automation we employ must have some alerting built into it when things are not working. In today’s age of do twice as much with half as less, automation is the only way we can stay on top of the SQL environment we maintain."
Jeff wrote in with his take – "Auto-Pilot = Good Thing (w/conditions)
I’m in the middle of setting up such a system for a customer, for managing his various network administration tasks (mostly related to backup scenarios in this case). Windows Task Scheduler has proved woefully inadequate for the job.
An auto-pilot environment can be counted on reliably when each task logs its result (including a ‘last-run’ timestamp) to a central console. The key to success lies in a separate task that checks in regularly at the console and sends a summary report, and then a special notice if anything is amiss.
In other words, an auto-pilot to keep watch on all the other auto-pilots. When configured properly, this can be a very effective and reliable approach."
And lastly, Steve summed it up nicely – "Most of the systems that we have are running primarily on autopilot. The systems are set to email when problems occur but I never take that for granted. I daily review all systems on top of this to ensure that anything out of the ordinary is caught and addressed. The complexity and ever changing nature of the systems can cause unexpected results.
Assumptions are a dangerous habit to get into. Questioning those assumptions on a regular basis is the best way I’ve found to get ahead of any issues before they occur."
I think this is the key point, that you keep an eye on the systems, have a regular review cycle and that you make no assumptions about the workings (or non-workings) of long-running systems. Check ’em out and make sure.
It’s like SQL Server’s version of "Trust but Verify."