by
adam baum
Sessions 1 and 2 for me
Both of these sessions focused on DR. After going through the "what can we do now" spiel, VM broke out the good info. In this case, it's a new product called Site Recovery Manager. When thinking in terms of DR, VM has HA. Good for single host failures. What happens if you lose you entire data center? That's where this product comes in. Think of it as HA on steroids. SRM is a new product that is basically a VC plug-in. In fact, you'll get a new icon in VC when it's installed.
Besides being an HA type tool, SRM is also a planning tool. When you use this tool, prepare to do a lot of thinking. This is a good thing since you are really developing a semi-complete DR solution. Sort of DC in a box. What does is take to build a DC? You need to know all these things when using SRM.
SRM is going to validate your replication toplogy and that replication is working. It is going to force you to determine which of your guests are important (good for SLA mgmt), and it's going to force you to figure out how to turn things on. This last item is important. In a typical DC, your domain controllers, DNS, whatever may be physical. What happens if they are gone? If you had them in VM, you would need to make sure that they come up first. In the world of DR and police & fire, we use something called "run books". These tell us what to do, in what order, who needs to do it, and such in a disaster. SRM can export its run book so you can print it out. You can create multiple different plans (run books) based on the nature of the disaster. Once a one book is created, it's a simple matter of clicking the play button to put everything into action. Note: Be careful with this since it does talk to your storage system.
One interesting feature of SRM is that it has the ability to reconfigure the guests. It requires lots of thought because you will have to know the "to be" config before clicking the Play button. Items that can be configed: server name, IP info, VLAN, etc.
SRM looks like a good product and its due out in 1st half 2008. I Future versions will provide failback, support more DR technologies, application monitoring, an improved SDK, and better warnings on misconfigurations.
BTW, I got all of this from sessions 1. Sessions 2 was a boring regurgitation of session 1.
(Note: You must be logged in to post a comment.)