Windows Server 8: Part 5.2—Cluster Aware Patching
July 17, 2012 1 Comment
Slalom consultant and accomplished Microsoft systems developer Derek Martin sheds light on Windows Server 8 (WS8) through his insightful blog series focusing on his research within the technical preview documentation, personal experimentation with the product, and thoughts of how they can apply to the real world as soon as it is released to manufacturing (RTM).
One of my particular annoyances and pain points with Windows is patching. Oh how I hate patching. It happens at least one Tuesday of every month and while WS8 does improve greatly on reducing the attack surface, which reduces the need for restarts, by making the graphical user interface (GUI) optional, it still happens. Patching gets even more complicated when you are running lots of virtual machines (VMs) on a host. They have to get patched and then restarted, then the host has to get patched, the VMs paused (or moved if in a cluster) as the host restarts, then move the VMs back…agh.
Well, there are a variety of features that have been improved in this arena in WS8. At the base of all of these new clustering features is Cluster Aware Updating. Other technologies like Cluster Shared Volume improvements, SMB amazingness improvements, and the management tools help ease these burdens, but again, having your cluster service know what’s happening when you are patching is key. Here’s how it works:
- One of your cluster servers becomes an ‘orchestrator’—it scans and downloads patches to all of its brothers and sisters.
- It then moves VMs around, maintaining cluster quorum and, by the grace of the Hyper-V Resource Metering feature, keeps your SLAs in check by balancing and even powering down less important VMs if resources get pinched. Holy cow indeed!
- Once a cluster node is drained, it patches, restarts, and rejoins the cluster and the VMs are moved back onto the node and then particpates in helping out other nodes by taking on their VMs as directed from the big brother orchestration node.
Oh yeah, that’s hot. Again, when you pair this stuff with the high availability cluster shared volumes feature and the new transactional SMB goo, you’ve got yourself a self healing/patching private cloud.
Bear in mind, the example here with VMs is just one use case—this totally works with other clustering applications like SQL, Exchange, plain ‘ole highly available file shares, the new highly available DHCP stuff…the sky is the limit!