Excessive Hi Fly Writes

vbimport

#1

I have 4 Seagate ST4000VN000 drives that have excessive Hi Fly Writes.

I burned them in for 90 hours and have the following results,

drive 1 202 hits
drive 2 213 hits
drive 3 285 hits
drive 4 471 hits

all other smart parameters are ok.

Is this ok, or does it indicate a problem?

I also tested one drive by copying 500 MB of date to it, and the counter
increased by 3.

This is Seagates new NAS drive, could it have a couple of bugs that need
to be worked out?


#2

Where (in what, powered by what, etc) and how (physical orientation, etc) are the drives mounted?

Also, what are the other SMART attributes?

Also also, when you did the 500 MB copy, was it all sequential data that required no real seeking, or was it multiple smaller files?


#3

Smart report is attached.

The drive was mounted horizontal in a case with 15 other drives on a Corsair TX750 watt power supply.

The copy was large files 30 @ 40 GB files, the drive was on a piece of foam
rubber to see if vibration was a problem.

Smart Rpt.txt (4.67 KB)


#4

Long post of suggestions/ideas:

The drive has reached a max of 49 degrees Celsius over its lifetime, possibly in your case. This isn’t amazingly hot, but could be cooler. Allegedly, heat can affect the fly height, and it’s possible that continuous writing to areas close to one another allowed for increased heat outside of what the drives actually like.

Outside of that, only the high fly writes look even remotely curious. All other error indicators seem to say that the drive can perform normally.

I would try heavy testing with one drive activated/powered up at a time (or maybe 2 at a time), with plenty of breathing room. If you still get an increase in HFW but everything else looks good, I would keep the drives but keep an eye on other parameters from time to time. I would also make note of any logging errors that the host operating system might issue regarding a true I/O problem.

And of course if anything changes and you get no additional high-fly writes, something in your setup needs to be tweaked.

You might also try to run one or two of the drives in a vertical orientation, or possibly even upside-down, to see if the drive acts any differently. Plus see if many small files results in a significant increase in the HFW count. (Maybe this would force many seeks, allowing the read/write heads to be in proper position each time?)

Either way, the drives should ideally be fine. If the high fly writes–self-corrected by the drive–concern you, then maybe seek Seagate drives from a previous generation until the kinks of this generation are worked out. Or seek drives from Western Digital or Hitachi/HGST (I don’t know how much technology is shared between the two). Maybe a comparable Samsung would perform without whatever causes the increasing count, though again, I’m not sure how much technology is shared between Seagate and Samsung at this point.


#5

My NAS drives (WD Red 4 TB) don’t even report the High Flying Writes parameter, so it’s difficult to say if you should worry about this or not.

The NAS drives are supposedly built to work in environments, where there are multiple drives running in the same cabinet, so it’s difficult to imagine that they cannot handle being horizontal, vertical, up-side down or close together.


#6

[QUOTE=DrageMester;2715243]The NAS drives are supposedly built to work in environments, where there are multiple drives running in the same cabinet, so it’s difficult to imagine that they cannot handle being horizontal, vertical, up-side down or close together.[/QUOTE]

You speak the truth. :iagree:


#7

[QUOTE=DrageMester;2715243]My NAS drives (WD Red 4 TB) don’t even report the High Flying Writes parameter, so it’s difficult to say if you should worry about this or not.

The NAS drives are supposedly built to work in environments, where there are multiple drives running in the same cabinet, so it’s difficult to imagine that they cannot handle being horizontal, vertical, up-side down or close together.[/QUOTE]

Perhaps I am off base on this. but I wonder why Seagate allows this parameter to be read if they were not concerned about it?


#8

Thanks for all your help guys.