So I've just had my third drive in a fortnight go bad.

Last week I took the precaution of adding a 2nd parity but had another go unmountable since then.  Now the drives failing have been old but I'm wondering if there is something going on here which I'm not noticing.  Posted diag from today and one from a v poory state last week.

 

Any advise / comments much appreciated..

[off to buy another couple of drives!]

server-diagnostics-20190223-0956.zip

server-diagnostics-20190216-1045.zip

Edited March 3, 20197 yr by fysmd
Mark solved - thanks!

Older diags disk7 dropped offline so no SMART.

 

On the newer ones SMART for the disabled disk17, looks OK, it might have been a cable issue, disk18 is failing though.

thanks, I noticed disk18 and have a replacement now - preclearing.

While starting a pre-clear, disk2 just went offline!!

 

It does seem that all these drives were on the same cable.  Bi of rearrangement and that cable's out of play now.

 

REALLY glad I added two parity now!

 

Other than drive18 with errors, would I be right thinking that they're likely OK?

 

Only took a quick look but other array drives looked mostly fine, there's a 2TB WD with a very high load cycle count, you could disabled parking with wdidle, but probably not much point now.

Sorry, that's not what I meant - I now have four drives which i removed from the array, likely cable related.  is there any reason to suspect them faulty?

 

One back at full strength I may run preclear a few times on them to test them but is there anything else?

 

Just had another drive go, disk2 now, diag attached

server-diagnostics-20190223-1439.zip

6 minutes ago, fysmd said:

is there any reason to suspect them faulty?

Post SMART reports for all 4.

2 minutes ago, fysmd said:

Just had another drive go, disk2

Also dropped offline, but previous SMART was fine, you can bring it online by rebooting and post new report to confirm, but it's likely still fine, same for the previous ones.

machine reboot rather than stop and restart array?

- damn, in the middle of preclears

 

Are you sure you have enough power for all those disks? Maybe you have an old power supply?

 

You have some hardware problem, and it's not the disks.

Just now, fysmd said:

machine reboot rather than stop and restart array?

Stopping array won't work, you can remove and re-insert the disk if the server supports hot plug.

2 minutes ago, trurl said:

Are you sure you have enough power for all those disks? Maybe you have an old power supply?

I would be suspicious of the Power Supply.   There have been several hardware problems lately which appear to have been fixed when the PS was replaced.  (PS are a much more complicated device now then they were back in the early 2000's and, perhaps, there has been some manufacturing cost cutting going on...) 

I would say if the 4 disks share a minSAS cable as the OP suggested start there, if not I agree the PSU is a good suspect.

the drives are split over two 550W PSUs but will tot up how many / which on each.

 

During rebuild two otherwise happy drives now report a >>LOT<< of read errors.

I'm replacing the PSU which supplies all so far affcted drives, is there anything lurking in the logs I dont see?

 

server-diagnostics-20190224-0822.zip

So, after more investigation I found the only component common to the misbehaving drives was one of these puppies:

https://www.scan.co.uk/products/silverstone-cp06-1-to-4-sata-power-adaptor-cable-with-capacitor

 

Removed and now over three hours into rebuild with no faults so far (everything's crossed).

Hi all,

Well, a week later, a full rebuild and much frustration later I think I can confirm that the SATA power cable with capacitor was my issue.  No issues reported since!  Thanks for the help and advise all!