[BBLISA] Whatever happened to Seagate?

[Rich Braun]

Is it time for a eulogy to Seagate?  My last job at DEC was in their storage
division, which got renamed to Avastor and then promptly sold to Quantum. Back
in the day, any time Seagate came out with a new product, our engineers
eagerly got hold of one and disassembled it, in search of the secrets behind
being #1 in the industry.

I still use Seagate as alternate-vendor for my home servers (alas, I no longer
maintain data center servers--AWS has killed off that whole business).  But
not for much longer, it seems.

On Saturday morning at 2:30am, my local Nagios gave the alert: disk dropped
out of the array.  It's one of two batches of the 3TB units that came out
around 2012-2013 (I can remember how long the industry sat on a 2TB
"standard", not coming out with higher capacities for what seemed like
forever, probably just two years or so but it's the longest capacity plateau I
can remember over the past 30 years).  Normally my ritual is to pull out the
failed drive, RMA it, and then over time swap others out a couple at a time as
my capacity needs require upgrades.

This time, array recovery was hobbled: I was able to resync about half the
partitions, but I discovered many more Offline_Unrecoverable type-198 errors:
I'd suffered a *triple-disk* failure in a 5-drive array! After dealing with
that crisis (somewhat), I looked at my other server and sure enough, found
errors on two of the Seagates there too.

Lesson learned: add smartctl error and temperature monitoring to my Nagios
arsenal.  Every time.

But is the other lesson to not buy Seagate?  Their consumer-grade warranty
stood at 5 years up until around 2012, when they led the whole industry
downhill into a 1-year standard warranty with an "enterprise" warranty of 3 or
5 years.  Backblaze is paying attention, and publishing reliability numbers
from its data centers Consumer Reports-style: take a look at
http://www.extremetech.com/extreme/175089-who-makes-the-most-reliable-hard-drives.
Sure enough, my 3TB Seagates are called out as worst-of-the-worst.

Western Digital bought out the Hitachi storage division, which worried me a
lot at the time, but so far hasn't compromised its top-tier quality.  So I'm
tossing out a bunch of Seagates in favor of Hitachi.  But regrettably, that
means my servers will have similar-vintage drives in years to come.  At least
I have Nagios watching them!

RIP, Seagate?  Do you share that sentiment?  (And, are you monitoring all your
drives' temperature and error counters?)

-rich