Data Outage Hits Dropbox Users

Dropbox logoOn the evening of Friday 10th January 2014, Dropbox engineers accidentally deployed a software upgrade to their active data servers, bringing down Dropbox completely. While the service was partially restored within three hours, some users were still experiencing issues more than 24 hours after the initial outage. Dropbox has issued an apology and reassured users that their data is safe following the significant service outage over the weekend.

Following a high profile password hijacking incident in 2012, Dropbox was swift to try and quell anxieties that Friday’s outage had been from a data violation or distributed denial of service (DDoS) attack. Instead a software bug was to blame and a Dropbox spokesman explained that despite running two copy “slave” machines, the bug resulted in many master machines as well as their slaves going down concurrently, producing a loss of service.

In addition, the company has developed applications to speed up the recovery of substantial MySQL copies, a tool Dropbox said would be published into the open source community so others could profit.

Dropbox had learned from this weekend’s troubles and had taken steps to ensure the bug wouldn’t be replicated. These measures contain an added level of tests that require machines to locally check their state before running incoming commands, which gives machines the right to reject orders when they would create a “harmful” consequence.

As the marketplace for affordable cloud storage grows, the outage once again emphasised the value for cloud storage companies to keep data reachable constantly so that customer belief can be preserved. The outage also demonstrated the brittle nature of cloud storage.

Tags: , , , , ,