Virtualization Bug Raises Downtime Concerns

We’ve blogged in the past about how virtualization adds to the complexity of production systems and – as a result – can cause problems for organizations with regard to downtime. Yesterday, the implications were made apparent when a VMWare software error prevented customers from powering up a virtual machine. Virtualization.info has a good summary of the issue along with updates since the problem was first reported.

The implications on downtime within the organization are obvious. Network Performance Daily points out that this is just another reason for end-users to blame the network for downtime.

“One of the things that we hear most often is that it’s hard to identify the problem being the application, server, or network. We can now officially add to this list: Application, server, network, or virtualization layer.”

The timing of this bug is also interesting as it comes when virtualization is gaining greater prominence in production environments for many companies. In fact, Virtual Strategy Magazine recently queried “Is Virtualization the Cloud of the Enterprise?” They had an interesting take on where virtualization is headed:

“As virtualization permeates more and more of the infrastructure, policy-based automation will continue to expand its presence within many of the major areas of management, including change management, configuration management, performance management and IT security management. Long term, the viability of effectively supporting service delivery in a highly dynamic environment depends upon incorporating centralized, policy-based management.”

But this is the future. In the meantime, we need to be worried about the data center of today. Virtualization may very well be how organizations handle their environments in the future, but for now they are likely blending virtual and physical infrastructures into a hybrid environment. The open question is if this incident will delay what some believe is an inevitable switch to fully virtualized environments. The CIO Weblog says that the VMWare problem is “considerably more disconcerting than bugs in other sorts of software.” He continues:

“Virtualization, after all, is very much about putting all of ones eggs in one basket, and if the basket has been woven by stoned hippies who have missed a few loops, then enterprises are going to start thinking twice about something which is an already disconcerting concept.”

We agree. The VMWare incident merely points out that – just as with physical environments – bugs and other software problems can have a wide ranging impact on availability. Virtualization by itself cannot replace things like architecting for availability, disaster recovery planning and effective testing.

Popularity: 5% [?]

Filed Under: Downtime, IT Operations, Virtualization