Reallocated sector count continues to climb

arbar wrote:
At what point does one declare the drive failed?

If the disk fails really badly the NAS will let you know, but it could already be bad enough to get a replacement (depends whether it passes the manufacturer's tests or not).

arbar wrote:
How does one request a warranty replacement?

You can power down, remove the disk, put it in a PC and check the disk using vendor tools.

If you run the NAS while the disk is out of it and it comes up as OK and you wish to continue using it you will need to hot-add it (add it while the NAS is on). It will be wiped and re-added to the array.

Desktop x86 ReadyNAS have a "Test Disks" boot option as of 4.2.9. I don't know if the rackmount ReadyNAS have a "Test Disks" boot option.

I had a drive that eventually failed due to reallocated sector count. It was a Seagate drive, so I e-mailed Seagate my questions; here is a paraphrase of what they had to say:

• The drive has extra space to deal with bad sectors being found


• Each time a sector is marked bad you lose all data that was held in that sector


• There is a SMART threshold defined to let you know when this extra space has been used up (for me it seemed to be over 1000 sectors)


• Once the drive hits this SMART threshold SeaTools will return an error code and SeaGate will authorize an RMA (if the drive is in warranty)


• It is not uncommon for a drive to run into a few bad sectors


• These reallocated sectors can impact drive performance

This drive did not come with a ReadyNAS device though, so I would imagine your RMA situation is different.

The ReadyNAS device informed me once the SMART threshold had been triggered. It did bug me that I was potentially losing data each time a sector was reallocated, but that they did not consider it a problem until I had lost hundreds of sectors. I have since had another drive show reallocated sectors, but it stabilized at 37 and has not had another increase in the last six months.

If the reallocated sector count keeps growing I'd replace the drive, it'll fail eventually anyways so why not be proactive and replace it ahead of time.

Boog wrote:
I had a drive that eventually failed due to reallocated sector count. It was a Seagate drive, so I e-mailed Seagate my questions; here is a paraphrase of what they had to say: The drive has extra space to deal with bad sectors being found Each time a sector is marked bad you lose all data that was held in that sector

(Digging up an old thread which I found when searching for info, as one of my drives has started reallocating sectors)

So, the first thing I was wondering is, will the ReadyNAS use the parity data to recover the lost data when a sector is reallocated?

Boog wrote:
There is a SMART threshold defined to let you know when this extra space has been used up (for me it seemed to be over 1000 sectors) Once the drive hits this SMART threshold SeaTools will return an error code and SeaGate will authorize an RMA (if the drive is in warranty)

And the second thing is, is there any way to know whether or not a drive will fail the SeaTools test? Obviously it's not desirable to keep shutting the array down for 12 hours, plugging the drive into a PC and running the test. I could do that every day for a year and never have it flagged as failed, for all I know.

Boog wrote:
The ReadyNAS device informed me once the SMART threshold had been triggered.

So by this, do you mean it gave you a different error message to the normal one saying the reallocated sector count has increased from x to y? It would certainly address my previous question if the ReadyNAS knows at what point SeaTools will consider the drive failed and sends a different error message at that point.

steve_i wrote:
So, the first thing I was wondering is, will the ReadyNAS use the parity data to recover the lost data when a sector is reallocated?

no data is lost with reallocated sectors, this happens internally to the drive, the nas is not aware or care (other than the smart messages) that sectors are remapped.

And the second thing is, is there any way to know whether or not a drive will fail the SeaTools test? Obviously it's not desirable to keep shutting the array down for 12 hours, plugging the drive into a PC and running the test. I could do that every day for a year and never have it flagged as failed, for all I know.

Erm, other than the obviously fatal errors, not really... if there was then there wouldn't be much point to seatools, eh?

Honestly, ALL disks fail sooner or later. If a disk is giving errors, it will fail sooner.

If would even consider shutting down an array to test a single disk for 12 hours, then your better off replacing and testing it offline or using it as a spare or put into something not as critical until it does fail.

Boog wrote: The ReadyNAS device informed me once the SMART threshold had been triggered. So by this, do you mean it gave you a different error message to the normal one saying the reallocated sector count has increased from x to y? It would certainly address my previous question if the ReadyNAS knows at what point SeaTools will consider the drive failed and sends a different error message at that point.

I don't know if there is a different message when you run out of re-allocated sector room, however do you really want to keep your data on a disk for that long of a period to find out?

The drive is only 100 days old, so I feel if it's not worth carrying on using, I should be able to return it for replacement. But as I understand it, Seagate will only deem a drive faulty if SeaTools says it is. So it would be useful to know what threshold SeaTools uses for deciding a drive has too many reallocated sectors. From this thread, it sounds like it needs at least 100, possibly over 1000. So it's partly down to cost, and partly a point of principle. I want the drive to get to the point where I can return it to Seagate, otherwise if I just keep replacing them when they have reallocated sectors, but aren't deemed faulty by Seagate, they're being financially rewarded for making poor quality products.

A lot of this depends on the rate of accumulation of reallocated sectors. For example if you only have a dozen or so after 100 days and the count is not going up by a dozen or so every day, then I would say continue to use the drive. If the rate is already fairly high and increasing rapidly, I would say have a spare drive handy.

On the other hand, if you can return the drive to the vendor for replacement as a defective drive I would do so. When you RMA a drive to a manufacturer, they do not send you a brand new drive, they send you a refurbished drive of the same model if they have one. If they do not have one, then they may send you one of the same capacity which may not be on the HCL, or they may send you a drive of a higher capacity.

Over the past four years, I have had occasion to RMA two drives to Seagate. The first died about 45 days after setting up my NV+ almost four years ago, it just died. I RMA'd the drive, but also bought a brand new on to match the other drive in my array and get redundancy back. It took about 5 days for the drive to get to Texas (where RMA drives go), one day to check out and 3 days for the replacement to arrive. If was a refurbished 750GB drive that I have used ever since as a backup target for my NAS. This past summer, I lost a second Seagate, this one a 1TB drive out of my second NVX (backup target), again it just died, no warnings. The RMA experience was about the same, except this time the drive I received was the same model as the one I sent in, and even the same firmware version. I may use it in the backup unit, but not in my primary unit.

It is not recommended that you use refurbished drives in a RAID array. Although in a RAID composed of consumer drives rather than the much more expensive enterprise drives (as found in business units) they may not make that big a difference. After all a refurbished drive has been subjected to quality testing before being sent out, something no longer done on the original factory lines.

I recently replaced all four drives in my ReadyNas for newer larger drives. I bought 4 identical Seagate drives from Expansys. 2 out of the 4 drives started reporting reallocated sectors almost immediately. I got in touch with Expansys and explained the SMART errors reported by the ReadyNas. They agreed to exchange the drive so I RMA'ed the worse drive (highest reallocated sector count) back to them. Before returning the drive to Expansys I bought another identical drive (that's 5 in total) from them to hot swap out the drive that was going back. I'm glad I bought the "spare" drive as it took over a month for the replacement to arrive. I'm now just about to swap out the 2nd faulty drive with the replacement drive, having let SpinRite have a day checking it. Then I'll have to go through the whole rma process with Expansys again.

Whilst the process has been time consuming and a nuisance, Expansys were very good about it and the delay was Seagate's not theirs. I'm a bit miffed I had to buy another drive but at least I will eventually have a good spare. I'm also quite surprised that 2 out of 5 drives have been a problem. I've always rated Seagate but it seems their quality might not be as good as it once was.

My advice to anyone who starts getting SMART errors is to get in touch with the retailer where you brought the drive not the manufacturer.

I have no knowledge of Seagate's 'official' policy but I can tell you my experience with the incredibly unreliable 1.5TB Seagate 7200.11 series. I have RMAed about half a dozen of these over the last year or two. I pulled the drives whenever the reallocated sector count rose past 25 or 30. Seagate replaced every one (and paid for return shipping) without comment.

Come to think of it, I have two more sitting on the shelf I need to RMA...

Your mileage may (hopefully) vary.

For what it's worth, the T2B Seagate LP drives I've been replacing them with seem very good so far.

ephemeris wrote:
I pulled the drives whenever the reallocated sector count rose past 25 or 30. Seagate replaced every one (and paid for return shipping) without comment.

What did you do if SeaTools did not give you an "error code" - what did you put on the RMA?

I ask because I have 3 of them. #1 has 137 reallocated sectors, #2 just jumped from 14 to 4093 yesterday (!!!!), and #3 has been stable at 28. All are 3 days shy of their 3 year warranty running out. Yes, I am definitely getting a replacement for #2 (I am going to borrow a friend's new 2TB Seagate LP (on HCL) to use - it should only rebuild as a 1.5 - while RMA'ing my really bad drive). All have about 17,000 hours.

For what it is worth, I also have 2 WD 1TB "Greenpower" drives in the NAS with 19,000 and 22,000 hours on them. Both have zero bad sectors.