HELP! Not sure how to proceed. Scrub shows degraded and faulted drive(s)

[sfatula]

You could, but I would do before it was created (or at least loaded). They take a good part of a day depending on drive size. The idea is to not put faulty stuff into production. Even with brand new drives. My thought is why create more work by having to recover stuff, etc. when it can be easily prevented by one day (or portion thereof) of tests. Otherwise, you can be left guessing or spending days of tests to isolate some exact cause.

 

[Redcoat]

What tests/scrub should I perform on these disks? Should these tests be done prior to building them into a pool?

In the resource section here find "Uncle Fester's Guide" that is maintained by forum member @danb35. It has a section on HDD Validation which describes a SMART test and Badblocks test combination that many of us here have used to test and exercise new drives before putting them in to service. One aspect covered in detail is the use of TMUX to run badblocks tests on multiple drives simultaneously and untended - helpful when you have several large drives to qualify.

I can't imagine building a new pool of large drives without completing this work first and ensuring that all drives are "good to go out of the box".

As there seem to have been few reports here of early drive failures of HDD's so tested AFAIK, I believe that infant mortality issues have likely been avoided.

 

[Daisuke]

Should I still do the isolation test on the new drive that I put in, assuming I test with a single drive pool?

I would definitely test one disk only, properly formatted. You want to make sure your server hardware is functional, then you know for sure you deal with bad disks.

Edit: I did not know about the Uncle Fester's Guide resource, thank you @Redcoat.