For a couple of weeks now I have been watching the daily report for one of my servers (@joeschmuck report) for increases in a value 1 MultiZoneError. The last two days, though, report 0 errors on that drive.
I don’t believe I have ever read of such an error “fixing itself”. The change in report occurred from a long test. The first report of the error resulted from a long test.
Anyone have any thoughts about this?
I do believe it is possible, however typically you would see it when a drive is failing and the failures just keep on coming.
Mutli-Zone Error “Rate” is an average over a period of operations. I suspect each manufacturer has their own calculations but here is one for Seek Error Rate as an example:
The normalized value is only computed when the number of bits in the “BitsRead” count is greater than 10^10. The count is cleared at power-on reset and wraps to 10^10 when it exceeds 10^12.
So yes, while it is rare, it can self-heal.
Thank you, Joe. I’ll keep on monitoring it and report back.
That is why I will never tell a person that the drive is messed up solely from a “rate” value. However I have seen that be the only indication of pending doom, so if it were to keep climbing fairly fast (a few a day), that would be a concern.
Not too many people would have realized they might have a problem. You did what people should do, be observant and wait to see a trend.
Take care!