Hacker News Reader

Yev from Backblaze here -> no, any drives that come DOA wouldn't have made it past burn-in and into our fleet, so they aren't represented.

View on HN

I really wish you tracked this data, because my colleagues and I in various parts of this industry have been noticing some trends...

With failure rates of in-service drives mostly being similar, I start to become much more interested in data about a brand/product's quality control process.

Yev -> that's really interesting. Do you want to publish your data and maybe we can compare notes ;-)

I just have anecdotes from about 4-5 people buying maybe in the 150TB range of disks per year. Mostly only Seagate or HGST.

Not nearly as clean (and representative) a datsset as y'all would have.

Is the takeaway there that you need to be doing burn-in runs too?

You can't burn in drives that can't even be used in the first place. They literally did not function straight out of the packaging.

Sounds like it could have been some kind of incompatibility or firmware/software bug, either on the machines you were using or the drive itself....

Possibly 3.3V issue? (I don't know Exos affects this)

https://www.instructables.com/How-to-Fix-the-33V-Pin-Issue-i...

That only affects WD drives, these are Seagates.

Negative.

Reply on news.ycombinator.com