Repalaced a failing disk, getting persistent false "read error" emails for the new one

[chagercloud]

I had a disk with SMART errors (read errors) so I replaced it with a new disk. The new disk is in the array, rebuild went perfectly fine, and it's showing healthy and passes manual smart checks. The weekly parity check runs with zero errors. But the daily array status email, while showing the correct new serial number for that disk (disk 3) shows read errors for disk 3 and an overall status of FAIL. So I log into the web ui and everything is healthy and green, no read errors on any array member. I think the daily email is just broken.

 

Attached some screenshots showing the weekly and daily email and the current array status. Any ideas or suggestions?

[JorgeB]

Post a screenshot of main.

[chagercloud]

I see the 818 errors, I hadn't noticed that before. I will have to watch to see if that's increasing. Could that be left over (requiring clearing/acknowledgement) from before I replaced the disk?

 

[JorgeB]

IIRC the errors column is reset after any array stop/start, if that's correct those suggest that there were errors during the rebuild, post the diagnostics.

[chagercloud]

Attached!

michigan-diagnostics-20240418-0958.zip

[JorgeB]

There were read errors on disk3 during the last parity check, they are logged as a disk problem, but the data was successfully re-written, so the disk may be OK now, you can run an extended SMART test to confirm, or another parity check, but first, and to reset the errors, re-start the array, if that doesn't do it reboot.

[chagercloud]

Thanks a bunch, I'll give those things a try.

[chagercloud]

Just restarting the array resolved it, thanks. Actually, as soon as I stopped it, I got an "array turned good" email. I could have sworn that I've rebooted (possibly even had a power outage/rest) since the rebuild finished, but I guess not.