Well, I'm annoyed

vbimport

#1

After this, it turns out that they have (almost) mastered the art of predicting HDD failures.

The 3TB (backup/non-essential) HDD in my little work server has died, literally 6 weeks after the 2 year manufacturer warranty ran out …

I’ve struck WD off my list of HDD suppliers for the next 5 years, which is basically when I expect that HDDs will be completely replaced by SSDs. (WD will still be on my $hitlist permanently though).

Smart notification that (38+ bad sectors had been relocated) by the monitoring software was issued 10 minutes after I’d rebooted the server after the HDD failed and the server crapped itself when the HDD stopped responding to requests.

Thanks WD, smart, microsoft :slight_smile:
It’s made my day.


#2
  1. It would be nice if failures were more predictable, wouldn’t it? Like if maybe the OS could learn the warning signs of a rapidly degrading drive without having to rely in SMART data alone. (And maybe the drive could have been better about updating its SMART status).

  2. I wonder at what stage of the drive production process HDD manufacturers realized that their devices were growing less reliable… Early on when they were actually designing them, or late in the process pretty much as they were being boxed for retail? And why can’t they get reliability back up? Can’t imagine having drives fail prematurely is the best way to get more business for your brand.


#3

This is why i set up a weekly scheduled task on my NAS that does an extensive SMART analysis on all disks and mails me the result.

It will not prevent me from sudden crashed, but at least i get a little automated notification about the global performance.