Leap Year Bug Causes Massive Microsoft Windows Azure Outage

The Microsoft Windows Azure compute cloud experienced a service disruption from 8:45 p.m. Eastern Feb. 28 to 5:57 a.m. Eastern Feb. 29, 2012, leaving many customers without access to their applications. This time Microsoft had Julius Caesar to blame: The outage apparently was due to a software bug that resulted in an “incorrect time calculation” for yesterday’s Leap Day.

Microsoft Corporate VP of Server and Cloud Bill Laing, the man in charge of Azure’s engineering organization, took to a blog entry to explain the situation and to beg forgiveness: “First let me apologize for any inconvenience this disruption has caused our customers,” Laing wrote.

Here’s Laing’s summary of the issue and how it was (mostly) resolved, from that blog entry:

At least as of time of writing, though, the Windows Azure Service Health Dashboard is reporting problems in distributed regions, though Laing promised fixes are coming to still-affected users ASAP. It’s worth noting that even at its peak, the disruption only impacted Microsoft Windows Azure compute, not storage.

An interesting side-note to this story: Vineet Jain, CEO of hybrid cloud storage vendor Egynyte, reached out to TalkinCloud with a statement indicating a belief that if Azure users had just kept some of their cloud infrastructure behind the firewall, they could have avoided this costly disruption in their business:

“By maintaining a behind the firewall presence and syncing that to the public cloud, companies are creating an insurance policy just for these situations,” Jain said.

That’s certainly food for thought. In the meanwhile, between this outage and a series of Microsoft Office 365 disruptions, Redmond has a long way to go before customers and partners can likely fully trust Microsoft’s cloud initiatives.