Building, implementing and executing a proper DR plan successfully is a challenging undertaking. It is a lot more complicated
than most experienced IT professionals and/or consultants think it is. This is because there are a LOT of moving parts to build a DR platform that’s going to fail over and allow the application to keep working. I’ve been bringing this point up in my sessions a lot recently. Our job in IT isn’t to build the cool, slick, sexy, solution. Our job is to make it so that the sales guy can sell widgets. Whatever widgets your company sells, the job of IT is to help the sales guy sell widgets. If the sales guy can’t sell widgets (and the shipping department can’t fulfill those orders and everything else that goes with selling widgets) then your company doesn’t get paid. If the company doesn’t get paid, then you don’t get paid. Then you have a pissed off spouse, and a pissed off mortgage company. And these aren’t good things to have. So let’s get back to talking about helping the sales team sell widgets.
HA Failover is pretty straightforward. There’s no data loss, everything is done using two-phase commit as it’s all inside the data center. So I don’t want to talk about that. I want to talk about when things really fall apart. The production site fails. Things are getting really interesting. Lets design for this.
Now let’s
just talk about what we need to think about.
- Active Directory
- IP Space
- Subnetting
- Connection String Issues
- Alerting
- Remote Access
- Connectivity
- End-user application access (web front end)
- Employee access (web / fat client)
That’s a lot of things that have to be thought about. You’ll notice that I haven’t even talked about the database stuff yet. Once we get into the database stuff starts getting more complex.
- Recovery Point Objective (RPO)
- Recovery Time Object (RTO)
- Am I using features that make availability groups not supportable (probably not on current versions)
- How many replicas do I need?
- Am I correctly licensed?
- Can I do this in the cloud?
- Should I do this in the cloud?
- How many values do I need to skip for sequences and IDENTITY columns?
This is clearly a complex topic. Because of this, we’ve put together a roundtable of experts on high availability and disaster recovery to have a roundtable discussion to talk about some of this complexities that there are that people stumble on. The webinar will be at 11 am Pacific Time / 2 pm Eastern Time on Tuesday, October 24th.
To register for the webinar so that we can remind you about it, click over to our registration page. Download the Outlook calendar entry and we’ll remind you when it’s time for the webcast. Can’t make it on the 24th? No problem. We’ll be recording the webcast and sharing making it available for viewing after for free.
When it comes to DR, you only get one chance. If you screw it up the company goes out of business so you really need to be taking your DR planning experience from the very best. And that’s who we have scheduled for our round table.
Denny
The post DR is more than just building an AG appeared first on SQL Server with Mr. Denny.
