Hard drive spin-up serious wearing factor on the long run, myth or fact?

[stuff_me_good]

Yeah, so if I choose to have snapRAID, flexRAID or unRAID in my server with 5-6 drives. All these raid modes have power management and they can put hard drives into sleep mode if not in use. This is actually the feature I'm hoping to use in the future.


But how much does these drives wear down when having few spin-ups and downs per day? I have been in the impression that the drives wear down more than being active 24/7. Is this true? Is there some data somewhere proving me one way or another?

 

[KingFatty]

There isn't very convenient data, but there is this google study:

http://research.google.com/pubs/pub32774.html

They don't seem to cite the "spin-up" as a factor.

Instead, things like scan error etc. are highly correlated with failure, see the conclusion. Access the full PDF on that site by clicking the button with the earth icon.

 

[code65536]

I expect that the drives used by Google's servers never spin down.

I've heard the same thing about drive spinups being hard on the spindle, but like you, I don't know if that's speculation or if it's actually based in fact. The somewhat screechy sound of a 3.5" drive spinning up always makes me wince a little, though, but that's just my aversion to the spinup sound.

 

[VirtualLarry]

I don't think that it matters nearly as much as it used to. Back in the day, the heads were stored on reserved inner cylinders, which had laser-textured media as a "landing zone". Back then, the heads actually landed on this zone when the drive was spun down, and took off from this zone once the drive started up and was up to speed.

But modern HDDs use an off-platter ramp to hold the heads, so now there's no wear / head-contact from starts and stops. So now, it's just a little extra wear + tear on the spindle motor, and not actual physical head slider wear.

 

[VirtualLarry]

* forum hiccup *

 

[glugglug]

The spin-down from various unraid utilities is not that aggressive to accumulate hundreds of thousands of cycles over the drive's lifetime.

The idiotic default spin down after 8 seconds of idle on WD Green drives IS.

Also, vibration from other drives is an issue in a server with many of them. So it makes sense to get WD RED or Seagate NAS drives for this.

 

[KingFatty]

When any of my drives fail, I disassemble them to get the magnets and take a look.

I've never had a failure because of the spindle motor failing. It's always been the head crashing directly into the middle of the disk, or the electronic board of the drive failing.

So it's just my personal anecdote, but over the years I just haven't been affected by the type of failure that would be caused by spin up/down. It's always something else.

 

[fuzzymath10]

My flexraid spins down after an hour of idling, including a daily update which requires that they all spin up at least once a day. Oldest hard drive is from 2008, and runs fine (incidentally, an infamous 1.5TB 7200.11)

My impression is that a good power supply goes a long way, but that might be anecdotal.

 

[Linflas]

My flexraid spins down after an hour of idling, including a daily update which requires that they all spin up at least once a day. Oldest hard drive is from 2008, and runs fine (incidentally, an infamous 1.5TB 7200.11)

My impression is that a good power supply goes a long way, but that might be anecdotal.

I'm still using one of those that actually had the problem a few years back. I recovered it by using a breadboard/serial port/thin piece of cardboard and going into the drive maintenance console to fix it and then updated the firmware. After that it has been fine.