I've been buying Seagate 4TB EXOS SAS 'wiped' drives on ebay in box lots
of ten for $150 to $200. That's $15-$20 for a 4TB drive. The one box had
2 DOA drives, the rest are performing great. I have a dozen of these
running in a Dell R720XD Rack server (RAID controller reflashed; can't
think of the name of the site whose tutorial I followed.) Boot disk is
an intel 2.5" SATA 200GB SSD, also used, from ebay. That ~10 yr old
machine isn't as energy efficient as new hardware, but I'm not working
it hard and did not notice any definite change in the monthly
fluctuations in my power bill. If I run benchmarks on its (two sockets)
CPUs, its many fans rev and you'd think I have a drone flying in the
basement.
The first two months it was in operation I had several occasions with
two drives at a time falling out of the zpool, usually upon reboot. Was
able to 'dd' wipe the start and end of the dropped drives and 'zpool
replace' them back into the pool. Oddly, I had the same problem on a
different project where I installed four new 20TB WD Red Pro on a new
AsRock Rack mobo running Debian 12, headless. After two months and 3-4
HDD dropouts (always in pairs but not the same HDDs each time), things
stabilized and have been running solid since. Throughout all that, the
zpool always retained some redundancy. The nfs share never glitched,
except in the new setup they had a failing client spamming the nfs
server and collapsing it with kernel panic. That stopped when I removed
their funky client machine.
Downside is these are SAS drives so non-server hardware needs PCIe card
and SAS cables like:
https://www.amazon.com/OIKWAN-Mini-SAS-SAS-Cable-Connector-SATA-Power/dp/B0CNPQTCS7
https://www.ebay.com/itm/194910024856
My first two SAS drives I hand wired the power adapters (SATA -> SAS)
because I didn't have the right adapters on hand. Switched the machine
on and got a loud buzz and lots of smoke somewhere out of the HDDs.
Thankfully my pc was non the worse for it, but I wrote "SMOKED" with a
black marker on those two HDDs and never tried powering them again.
Just now looked at smartctl numbers for the 12 HDDs, the newest one says
30K and the oldest 70K, spinning hours. I've spun them some 6K hours
since I bought them. Mfr. date reported from 2013 to 2018. One of them
has always reported 3 re-allocated sectors, so what, this is RAIDz3, and
at my own risk. To date, all projects I sold to customers contained all
brand new drives.
On 2/22/25 06:29, gene heskett wrote:
On 2/21/25 11:42, Stefan Monnier wrote:
That was 2+ years ago, and 2T's were brand new.
With a lot of emphasis on the "+" I guess, since I bought my first 2½"
2TB HDD in 2012.
Stefan
I was shopping in the 3.5" drives at newegg IIRC, 2T was the biggest,
and my 3rd woof died in 2020, and those were bought after she passed,
sometime in 2021 or 2022. For $129/copy IIRC. They both just
disappeared off the end of a sata-iii cable within 36 hours of each
other with only 2 weeks spin time on them.. NOS for sure, marked as
made in 2014, so they were at least 7 damned years on somebody's shelf
when I bought the pair of them. But it took a 16mm projector lens to
read all that in the drive label. There was a time when seagate made
good hard drives. One of my cnc'd machines has a 250G in it, shut off
only for new installs, still running wheezy. No reallocated sectors,
the last time I looked, at probably 90K+ spinning hours its fine. I'd
query it, but smartctl seems to have been removed, so what now serves
that purpose of interrogating a drive on wheezy?? Thanks Stefan.
.
Cheers, Gene Heskett, CET.
--
Thank You!
Titus Newswanger
Curtiss WI
