Reallocated Sector Count - Best Practices

jlehtinen wrote:
I've got a set of 6 Seagate ST3500514NS 500GB - each drive seems to pick up an additional reallocated sector every week or two. Some of them are up to around ~30 reallocated sectors at this point. What is "normal" growth you have seen for reallocated sectors?

Answer is clear, 0 reallocated sectors, is "normal".
There isnt such thing as normal growth of reallocated sectors, its a faulty drive if it starts to reallocating.
Basically its ticking "timebomb", there isnt enough spare sectors to reallocate whole drive, on one day "boom" and youre data is just gone...

jlehtinen wrote:
What is the best practice for replacing drives due to reallocated sector count? At what number of reallocated sectors should I start worrying?

Actually you should start worrrying, after it starts rising, 1 is not a big deal, untill if it incriments, there is definately something wrong with your drive and you should back up your data and replace it.

jlehtinen wrote:
Interesting side note - in the same ReadyNAS, I have 6 WDC WD1002FBYS-02A6B0 1TB drives. These drives have yet to pick up a single reallocated sector, and have been running for the same time, and under the same load as the Seagate drives. Difference in quality maybe?

Different Manufacturer, slightly different materials and process...dropped on transport to store...etc...
(We once got whole box, 20 x 4TB Hitachi drives all DOA, so..)

NETGEAR recommends replacing hard drives at 50 reallocated sectors and 1 ATA error.

jlehtinen wrote:
I've got a set of 6 Seagate ST3500514NS 500GB - each drive seems to pick up an additional reallocated sector every week or two. Some of them are up to around ~30 reallocated sectors at this point.

While not a happy story, it's probably too early to panic (though not too early to start planning for replacement i.e. a drive or two that rapidly degenerates >50, or has other errors).

Of my 4 drives, one did very early-on accelerate rapidly to about 25 reallocated sectors and I bought an identical backup. Over the following several years it picked-up 3 or 4 more bad sectors, one again very recently. The other 3 are still at zero.

I suspect one day that the drive will deteriorate at a pace that will tell me to replace it... I hope so anyway. My spare drive is still in its static-free bag awaiting that day.

Thanks for the replies.

I was a little concerned about this, so I already have some replacement drives on hand. I've been checking daily on drive health to try to watch for a failure or any acceleration in the sector growth.

I guess there isn't much else I can do except wait until reallocated sectors reaches 50 on each of the drives, and I can RMA them one by one.

jlehtinen wrote:
Thanks for the replies. I was a little concerned about this, so I already have some replacement drives on hand. I've been checking daily on drive health to try to watch for a failure or any acceleration in the sector growth. I guess there isn't much else I can do except wait until reallocated sectors reaches 50 on each of the drives, and I can RMA them one by one.

Actually, you could replace those (one by one ofc) right now, then run Seatools on them and send to RMA right away.
Warranty already covers 1 Bad sector, you dont have to wait to get 50 of them

I contacted Seagate support about this. They told me Reallocated sectors are OK/normal, and the drive should still pass Seatools. As long as it passes Seatools any warranty claim I make will be rejected.

Since this NAS is in a production environment I can't easily pull the drive to run Seatools.

Any suggestions?

Seagate certainly would rather not replace drives...

You can return drives even if you can't run Seatools, and there are some manual failure codes you can enter. Seagate reserves the right to return your original drive if they think it has not really failed - though they have never done it to me.

Personally I wouldn't RMA a drive for 30 sectors, though I think Netgear's threshold of 50 sectors is reasonable. What is scary here is that the counts are increasing. Though perhaps the RMA is not really the question. The more important one is whether you can trust your production data on these drives.

Have you also looked at pending sectors? It makes sense to sum pending+reallocated - since pending counts are incremented on failed reads, reallocated sectors are failed rights.

The drives ARE running, they haven't failed yet. Even though it's a bad trend, it seems like Seagate might not honor an RMA until the drive has actually failed.

Have you ever returned a drive to Seagate just based off reallocated sector count? Did they honor the RMA?

I checked and pending sectors for all drives are 0 (that's good at least...)

Seagate might not honor an RMA, though in my experience they will.

In any event, keep in mind that once the reallocated counts begin to climb, they can grow extremely quickly. If both drives fail, you lose all the data in the array.

Just as a follow-up... I pulled and returned one drive (it was functioning, but had 30 reallocated sectors), and Seagate honored the RMA with no questions asked.