Suppose you want to play around with a SQL Cluster, based on Microsoft Failover Clustering. In production environments this won’t do. There isn’t any space left neither in the test labs of your company so you need an alternative. However, at home you don’t have any more a real lab as well (when the first kid was about to arrive I emptied my own personal data center in order to restyle it into a baby room…).
All you’ve got is a pc with 8 GB RAM, some really fast disks and a quad core CPU. Nowadays a basic configuration. The system also runs Windows 7 since it’s the home pc as well. So installing Windows Server 2008 R2 for Hyper-V isn’t an option since some people, who’re also using the pc for their home stuff, won’t like that at all.
So now what? And YES, the SQL Cluster must be build. No matter what. Performance isn’t key here.
Hmm. Gladly enough I had a license for VMware Workstation 8.0. (I really love my own network :) ). So I downloaded it, installed it and now it was time for some good searching on the internet. Soon I bumped into these postings, all about creating a SQL Cluster based on VMware Workstation, 3 VMs (one of them DC and a virtualized iSCSI SAN based on Microsoft iSCSI Software Target 3.3):
And guess what? It all works very well. One advise though: when creating the virtual iSCSI SAN, there is an additional Step required in order to make it work:
- On the server hosting the iSCSI Target role one must open the Properties screen of the Target. The second tab enables one to add the iSCSI Initiators which are allowed to connect to the iSCSI Target. Without doing this the nodes won’t have access to the iSCSI Target. It took me some time to find this…
It’s funny to see that the software of two companies who are fighting each other for market share can work so good together and produce very good results :).