So admittedly I've gotten a bit complacent with my drive monitoring.  I've had the server for years and only ever one red ball and no data loss... so, pretty lucky.  I have a 10 drive array + parity + warm spare.  Something made me check drive status today, and I was greeted with this.  Uh oh.

 

The full smart reports (in the same order as the image), are attached.

 

I'm trying to figure what drives need replaced immediately, if any, and what can wait.  Parity, Disk 4, and Disk 10 all have red warnings, but pass SMART and my parity check on the first of the month reported no errors.  Frustratingly, Parity and Disk 10 are the newest, but at least they're under warranty.  Also, strangely, the warm spare that supposedly sits spun down unused has a high load count even though it doesn't contain even a byte of data.  I guess something is spinning it up?

 

As a note in case it matters, Disk 4 and Warm Spare were purchased at the same time from a single retailer.  The rest should all be from different batches.

 

Replace Disk 4, then Parity, then 10?  Disk 4 then wait and be ready with some more spares?  Any suggestions appreciated.  I'm a little lost on this stuff.  I've read what everything means, but I'm not sure how to translate that into when I need to take action.

unRaid_Drive_Status_-_062014.txt

There's nothing inherently "bad" about reallocated sectors -- modern disks are DESIGNED for that.  The reason they have spare sectors is for precisely that reason ... so any failed sectors can be reallocated to a spare.  "Pending" simply means these sectors will be either reallocated or removed from the pending count the next time they're written to (depending on the results of the write).

 

My experience is that you can ignore high fly writes on Seagate drives -- they all seem to report these, but they're not a problem (I've seen some very high counts).

 

If UnRAID isn't showing any errors on the disks, I'd just be sure you have available spares, but just leave them in the system.  HOWEVER, I'd note the current pending and reallocated sector counts, and see if these are growing => if so, then I'd replace the disk.

 

One final note:  You may want to do an advance replacement RMA for the two drives that are under warranty ... or at least note the warranty expiration date and see if the SMART data changes in the next few months => if so, be sure to RMA them before the warranty expires.

 

 

So I've been keeping an eye on the drives over the past few days according to your advice, and the sectors pending allocation on one of the drives has grown by two since I posted.  I decided to grab a set of 3 4TB drives, so with those and the spare, I'll swap out that failing drive and buy myself some time to get Seagate to replace the two less-critical drive failures at their own pace.  The warranty is up in about 2 months, so I have to get going on it.

 

Thanks for all your help.

While I agree that reallocated sectors are not a death sentence to drives or data, my experience in supporting many users is that once they start to show up, they keep growing and growing with every parity check. I describe the effect as like an unraveling or growing pothole. If the count holds steady at a reasonably low number (say less than 100) even after repeated parity checks, I'd monitor it but not rush to replace it. But if it starts growing by one or two every check I'd look to RMA or replace it.

 

That multizone error I've seen a number of times. Don't fully understand but have seen it crop up  drives that start exhibiting flakey behavior. I'd monitor it but not be too concerned.