(SOLVED) New build with SMART errors

[ccb87]

I'm building up my first unraid array and I have a bunch of old empty hard drives. 4 of them have SMART errors. Are all 4 dead or can I continue using them for 2-3 months until I replace them over the holidays? The array is still building now so everything is empty if I need to trash a hard drive. I've got a bunch of externals connected but it is not seeing the ones attached to an USB hub. Do they need to be directly connected for unraid to see them?

 

I'm attaching my diagnostics file for troubleshooting.

 

Thanks

husky-diagnostics-20201107-1615.zip

[trurl]

USB connections are too unreliable for use in the parity array or cache. 

 

On mobile now will look at Diagnostics later. 

[ccb87]

For the externals I was planning to use them as unassigned devices but the ones on the hub can't be seen by unraid. I know they work on the hub connected to this computer with another operating system. I do have one external in the array just to tinker with before I actually set up the server to use.

[Michael_P]

sdn and sdi don't look happy and should be replaced, sdl looks about the same vintage so is likely to follow suit. I wouldn't trust any of those as Seagates have always failed with data loss for me (saved by backups of course)

[ccb87]

I can see that sdn is pretty much dead but the other ones are still showing "PASSED" for the SMART checks. How can you tell if the hard drive is failing or not? 

[Decto]

4 hours ago, ccb87 said:

I can see that sdn is pretty much dead but the other ones are still showing "PASSED" for the SMART checks. How can you tell if the hard drive is failing or not? 

 

 

sdl is having issues with Hardware_ECC_Recovered errors, 006/100 is very low, also Raw_Read_Error_Rate and Seek_Error_Rate below 100 would indicate deterioration.

ID# ATTRIBUTE_NAME          FLAGS    VALUE WORST THRESH FAIL RAW_VALUE
1 Raw_Read_Error_Rate       POSR--   083   064   044    -    196122632
7 Seek_Error_Rate           POSR--   091   060   045    -    1254905430
195 Hardware_ECC_Recovered  -O-RC-   006   001   000    -    196122632

 

sdi doesn't look too bad yet, but given you have 2 similar drives, one of which has failed, and the second deteriorated I'd be keeping a close eye and checking I have backups.

 

Drives tend to either fail early due to manufacturing faults or run until they are mechanically worn out. 

At 3 years only and as NAS drives, it does seem a little odd though amazon.com has quite a few 1* reviews saying multiple drives failed around warranty ~3 years.  

[JorgeB]

2 hours ago, Decto said:

sdl is having issues with Hardware_ECC_Recovered errors, 006/100 is very low, also Raw_Read_Error_Rate and Seek_Error_Rate below 100 would indicate deterioration.

ID# ATTRIBUTE_NAME          FLAGS    VALUE WORST THRESH FAIL RAW_VALUE
1 Raw_Read_Error_Rate       POSR--   083   064   044    -    196122632
7 Seek_Error_Rate           POSR--   091   060   045    -    1254905430
195 Hardware_ECC_Recovered  -O-RC-   006   001   000    -    196122632

That's normal for Seagate drives, more info here.

[Michael_P]

7 hours ago, ccb87 said:

I can see that sdn is pretty much dead but the other ones are still showing "PASSED" for the SMART checks. How can you tell if the hard drive is failing or not? 

Dead giveaway is reallocated sectors and un-correctable sectors

 

   5 Reallocated_Sector_Ct   PO--CK   001   001   010    NOW  64880 (0 5)

197 Current_Pending_Sector  -O--C-   099   099   000    -    656
198 Offline_Uncorrectable   ----C-   099   099   000    -    656

 

Anything more than zero is cause for concern and for me is a candidate for immediate replacement - doubly so in the case of Seagate drives, which for me at least, have never failed gracefully allowing data to be retrieved (always lose a few files here and there, if not the whole drive's contents).

 

[ccb87]

Okay, I will dump the Seagates and get some new ones in there. Thanks for the help