Amazon: Some data won't be recovered after cloud outage

Post mortem wait

Amazon says that about 0.07 per cent of the EBS storage volumes in the East Region of its infrastructure cloud are not "fully recoverable" following the extended outage that hit the service last Thursday.

The company has yet to fully explain the cause of the outage, but it still plans to publish a "post mortem" on the incident. "We are digging deeply into the root causes of this event," the company says in a post to its Amazon Web Services status dashboard.

In the early hours Pacific time on Thursday, Amazon said on its status page that it was investigating connectivity issues with its EC2 (Elastic Compute Cloud) service, which provides on-demand access to processing power via the web. The outage brought down many websites that run atop the service, including Quora, Sencha, Reddit, and FourSquare. According to one of the brief status messages from Amazon, the problem began with a "network event" that caused the service to re-mirror a large number of Elastic Block Store volumes in its East Region.

Amazon divides its "infrastructure cloud" into multiple geographic regions, and it guarantees 99.95 per cent availability within each region if you're using multiple "availability zones". Some regions – including the East Region, served up from Northern Virginia – are divided into these ostensibly separate zones, and Amazon has always said that these zones are "insulated" from each other's failures. But the East Region outage spread across multiple zones.

On Sunday, the company said that a "majority" of affected EBS volumes had been restored, but that it needed more time to restore data for some customers. But on Monday, it announced that some volumes would not be restored. "We have completed our remaining recovery efforts and though we've recovered nearly all of the stuck volumes, we've determined that a small number of volumes (0.07% of the volumes in our US-East Region) will not be fully recoverable," the company said.

It is in the process of contacting these customers.

For many – including Thorsten von Eicken, CTO of RightScale, an EC2 management service, and the employees of Scalr, an open source platform similar to RightScale – one of the chief problem is that Amazon has so far provided so little information about the outage. We await the post mortem with bated breath. Amazon has never said how its "availability zones" are designed. ®

Update: This story been updated to provide more detail on Amazon's uptime guarantee for EC2.