New server - disk disabled

[Husker_N7242C]

Hi everyone,

My old server had continuous issues with disks being disabled, yet being found healthy. I never could get to the bottom of it.

I rebuilt the server:

-New motherboard (brand new, old stock, SuperMicro X10-DRH-CT)

-New CPU(s) (used E5-2690V3 x2)

-New RAM (128GB DDR4 ECC)

-New HBAs (and SAS-Sata cables) (LSI 9207-8i x2)

-New USB Flash Drive (Sandisk Cruzer Blade 32GB)

-New Power Supply (850w Corsair)

 

I Ran it for two weeks on Windows (no HDDs) and run many hours of stress-testing to ensure the temps stay low nothing failed out of the box.

 

I transferred over the disks and flash and all seemed well for a week but I've just had disk2 disabled due to read errors. The drive isn't that old and the dashboard reports "Healthy"... but disabled.

 

I'd really appreciate it if anyone with more knowledge than me had a look at the diagnostics (attached) and could make some recommendations?

 

If you need any more info please let me know?

 

Dan

nas-diagnostics-20221202-1659.zip

Edited December 2, 2022 by Husker_N7242C

[JorgeB]

 

[Husker_N7242C]

Thanks @JorgeB as always for responding, however I've been through the process HERE to disable EPC and Low Current Spinup for all of these drives in the past. Is there anything else in the diagnostics that looks wrong?

I'll disable spin-down, add the drive back to the array and wait for it to happen again, that we I can be possitive it is/is not this known issue.

 

Just to double-check - if I stop the array, remove the disk, start the array and then stop the array again, re-add the disk and let the array rebuild, will I loose any data written to disk2 (currently emulated) since it was disabled, or does the drive get rebuilt from parity information instead?

Edited December 2, 2022 by Husker_N7242C

[JorgeB]

2 minutes ago, Husker_N7242C said:

Is there anything else in the diagnostics that looks wrong?

Nope, and the issue occurred right after spin up, suggesting that is the problem.

 

3 minutes ago, Husker_N7242C said:

I'll disable spin-down, add the drive back to the array and wait for it to happen again, that we I can be possitive it is/is not this known issue.

That's a good idea.

 

3 minutes ago, Husker_N7242C said:

if I stop the array, remove the disk, start the array and then stop the array again, re-add the disk and let the array rebuild, will I loose any data written to disk2 (currently emulated) since it was disabled, or does the drive get rebuilt from parity information instead?

It will rebuild the current emulated disk, including any data written to it after it got disabled, just make sure it's mounting and data looks correct before rebuilding on top.

[Husker_N7242C]

Thanks again @JorgeB, I've disabled spin-down now and will go through the process again for disabling EPC and low current spinup to be sure it is still applied. I was quite careful when I originally did this to be sure that all commands succeeded but you never know.

 

I'll mark your response as solution and get back in touch if re-applying the workarounds don't prevent this.

 

I want to avoid having the drives continuously spun-up as we are just coming into summer in Australia and the server's room isn't airconditioned so I try to keep it as dormant as possible during the middle of the day, no downloads, no transcoding, just a couple of VMs on the SSDs (with 12x 140mm Noctua fans) then let it all rip at night.