Hacker News Reader

Comment by jfoster | Hacker News Reader

I had noticed one of the images in my Google Photos account seemingly not showing up. It was very noticeable because it was one of the "People" images in my Photos account. It's a bit concerning that a photo storage service seemingly can't maintain the integrity of their data. Makes me wonder about what other data issues there are, lurking in the depths of Google Drive...

View on HN

I personally found a corrupted file stored in online storage only, though it wasn't with Google, but with Microsoft's OneDrive.

I tried to open a Word document that I hadn't viewed for several years, but got an error that the file was damaged. It looks like this happens sometimes, after searching for the error online. Luckily I was able to recover an earlier version of the file (via version history), but it was alarming that it happened in the first place (also I'm not sure if important edits were made in the latest corrupted version).

It's unfortunate how many cloud services default to storing on the cloud only instead of also keeping a local copy, and don't even provide the option to opt out. Even if you choose the "keep offline" option for several services, for unknown reasons, this doesn't seem reliable (Maybe I'm 'using it wrong'? But in practice, I've found myself having to download files that I was sure I set to keep offline.)

Google is pretty proud of their record of not losing user data. As far as I know, they have never permanently lost even one byte of consumer data.

The same isn't true for GCP - there they lost some writes to customers persistent disks - but only ~50 megabytes worldwide, which is still rather good when you consider they store millions of terabytes.

It really depends what you consider lost. Thousands of users (at least) have been permanently and eternally cut off from their Google data after some AI system decided to axe someone's access to their own account. The loss is significantly over a byte. Or 50 MB.

It seems more likely that the consumer team isn't set up to detect data corruption events or process the resulting support tickets:

https://duckduckgo.com/?q=google+drive+data+corruption

> As far as I know, they have never permanently lost even one byte of consumer data.

It seems that every other day you learn about people being permanently locked out of of their Google account. Every byte of data stored in there is permanently lost to the person losing their account.

Perhaps I should add... "except by policy".

What if the policy is to lock accounts when data integrity errors are detected...

"We have always had better storage durability than Europa."

Note, I am not arguing that this is necessarily what happens. The thread full of contrary views on cloud storage just tickled this cynical take loose from wherever it was lodged.

They definitely had permanent data loss in App Engine around a decade ago, along with leaving any app active during the outage with copies of every database table to manually merge

> Google is pretty proud of their record of not losing user data.

As far as I know, they just haven't admitted to losing consumer data. Until they define that and put even a little bit of effort into checking whether it might hold true, they don't really have anything to be proud of.

There have been multiple mass-deletion events and Google has admitted to at least one of them - the infamous "only 0.02% of our users" incident, for example - but only because so many people were affected, the press picked up on it and google's PR firms couldn't keep it quiet.

> As far as I know, they have never permanently lost even one byte of consumer data.

Back in the aughts Google was infamous for disappearing mail, if not entire accounts just disappearing into thin air.