Google Drive Implementation Flaws

Google Drive Implementation Flaws

Designing software is hard.

And designing file synchronisation software is one of the more niggly of design jobs: tiny errors can lead to large scale data loss.

And Google is doing a fairly good job with its rollout of Drive File Stream. Not that we’ve not had some data loss, and some of it really pernicious, but yesterday’s illustrated a design oversight that really shouldn’t have happened.

We were observing a minor file sync issue with a few local machine files refusing to sync with the Google Drive cloud. Drive File Stream was doing a good job of recognising the error, throwing a well-crafted warning, and parking the troublesome files in a ‘lost and found’ folder in a smart location on the local hard drive.

We clicked the help menu in File Stream and followed the official (not community) troubleshooting steps one at a time, but none helped.

When we got to ‘disconnect and reconnect account’ is when disaster struck.

On disconnecting the account from the File Stream software, it permanently deleted the ‘lost and found’ folder. Including the sole copies of all our files which had not yet synced with our corporate file store.

I understand there is a potential privacy issue with retaining files that belong to an account that is no longer connected. But deleting files that are known to be un-synced is a very dim solution.

And Google has shown it knows how to solve similar problems in the same context. In particular, when a new account connects, the previous log file is ended, copied, stored safely, and a new one created.

So I could retrieve the log file for the disconnected account, but not the actual files it referred to.

9/10 Google. But really you need to be getting 10/10 for corporate file storage.