Spinning HDD Power on Hours

[T_Minus]

With recent post in Deals about used drives I was curious about power on hours of spinning drives and what you guys see\use at home and at work.

1. What's the max power on hours you've seen from a spinning hard drive? What was the RPM? Brand?
   
2. Do you "retire" drives after certain POH or only failures?

 

[pricklypunter]

I have a bunch of 3TB 7200 rpm HGST with 40K+ on them in various arrays in the UK, I don't lose any sleep over them. Unless I have an early failure, I don't even pay much attention to a disk until it hits 3yrs old. After that I will run a smart scan monthly and only replace it if bad pending counts begin rising.

 

[nephri]

It depends the purpose.
Recently i decommissioned NAS/MEDIA disks that was online for almost 10 years at home.
On a total of 4, only one having a large batch of bad sectors (that was customer hdd)

On more "crucial" storage pool, i use redudancy and let the failure to come.
For spinning disks, it's not so rare for me to reach 50K+ POH and 5-7 years.
But when i reach such POH, as soon as i can find a good deal (on ebay or other) i try to jump into for reducing the POH.

I saw the same article and let me a bit confusing.
My own experience is much more like this about "crash":
- When new and start their online live (the disk clash in the first year)
- When moving it (post employes playing footballs with thems)
- After many years (often after 7 years)

I think cycle counts may have an impact (it's only subjective, i haven't any statistics).
But my customer drive having bad sectors was the one set on a media device connected to the tv.
This device was powered on/off multiples times per day for 10 years (maybe 3 or 4 times per days) and i think it was a factor.

offtopic : For now, i'm more annoyed/confused with cold SSD (specially entreprise ones) that can loose data in a range of weeks or a year.
Now, you can't just unplug a SSD and expecting to keep data for a long term storage (but for a spinning disk you can)

 

[Markess]

The most I've ever seen was just under 7 years on a 2TB Hitachi (7K3000 SATA) from Ebay that came to me by mistake (was supposed to be a new 7K6000). To me, the hours weren't as impressive as when I fired it up, and the powercycle count was only 5. I got a refund and was told to keep the drive, but I didn't risk using it for anything.

 

[msg7086]

Having about 10 drives having 77k hours on them, Samsung HD501LJ desktop drives that were put in a core NAS for my company many years ago. (Please, don't ask why.) They were the same batch with mostly consecutive S/N. Running FreeNAS with UFS on HW RAID5 until I decommissioned it. (Again, please.)

Now they are sitting in my home as an extra cold backup.

 

[blunden]

I recently replaced an old WD20EARS (at 83 551 hours) that started showing a non-zero Pending Sector Count (10) at roughly 80k hours. I had no problems reading all the data from it when migrating to a new drive though so it probably had some more life left in it. I have an online backup of the important data on it so that's why I didn't replace it right away.

Late last year, I also replaced another old WD20EARS (at 79 815 hours) but it had started showing a non-zero Uncorrectable Sector Count too and it took multiple attempts to all the data off of it, but was ultimately successful in the end. That's when I started doing more complete backups.

I also have a WD20EARX at over 72k hours that runs just fine. The two I replaced were used though so the agressive head parking hadn't been disabled right away, meaning that one that had read issues had a Load/Unload Cycle Count of 140256 and the other one replaced recently a value of 37292 compared to 116 on the similar generation one that's still going strong.

As you can tell, I'm using consumer drives almost exclusively. The one exception is a Samsung PM863a 3.84TB SSD that I bought essentially unused.