Over the years as server technology has improved the need for migrating to better performing servers, whether physical or virtual, has always been a challenge. Some of the challenges with migrating servers have been keeping the data current and minimizing downtime during the process. Once the migration or provisioning process has started users are not allowed to access the production server until the conversion is complete in order to create a duplicate image of that server. If a process is used where users can access the system, like with some P2V products then the production server remains online but the changes made after the conversion started are usually not included and a differential backup has to be applied once the server is migrated. This is the migration paradigm. How do you provision the most up to date changes without incurring significant downtime during the process?

The answer isn’t that difficult actually. There are several disaster recovery software products on the market that provide real-time replication and high availability of the operating system, application and associated data for disaster recovery purposes. In this scenario, if a server fails the IP and server name is failed over to a duplicate server and brought online with all the current data in minutes. This same process can be used to provision blade servers faster while minimizing downtime traditionally associated with previous migration processes. This subverts the migration paradigm by keeping the data consistent during the conversion process while at the same time minimizing the downtime usually required. If there are iSCSI devices currently be used in the environment then there are ways to create iSCSI bootable images of the servers and or workstations for even faster provisioning. If you don’t have an iSCSI device for bootable images then there are some products available that will help convert a server with DAS storage into an iSCSI SAN pretty cost effectively.

Certainly, many data centers are budgeting for converting many of their older x86 servers to blades and a virtual environment not only to reduce costs for power and cooling consumption but just overall ease of management. If you are one of the many IT managers that have this on your agenda for rolling out by year end or early next year there are options available to provision blades faster. Don’t think that your migration planning starts with coordinating a weekend when everyone can start late on a Friday and cross your fingers that the data transfers by Monday morning. Provision the server during the week with real-time replication or bootable disk images and when the data is current. Subvert the paradigm, flip the switch and get your weekends back for faster blade provisioning.