I have been using unRAID since July of last year and never had an error occur during a parity check.

 

Last night unRAID ran it's monthly parity check and came up with 2 errors which it corrected. 

 

Since I've never had that occur before I'm wondering if it is normal, to occur on occasion or anything to worry about? 

 

The only thing that has happened in the past month is I replaced a bad drive that was rebuilt from parity.  I'm about to do the same thing today, only this time it's just to swap a 1 TB drive for a 2 TB. 

 

Here is the related lines from syslog.

 

Apr  1 00:00:01 TODD-Svr kernel: md: recovery thread woken up ... (unRAID engine)

Apr  1 00:00:01 TODD-Svr kernel: md: recovery thread checking parity... (unRAID engine)

Apr  1 00:00:01 TODD-Svr kernel: md: using 1152k window, over a total of 1953514552 blocks. (unRAID engine)

Apr  1 00:00:14 TODD-Svr kernel: md: parity incorrect: 15728 (Errors)

Apr  1 00:00:14 TODD-Svr kernel: md: parity incorrect: 15736 (Errors)

Apr  1 07:05:02 TODD-Svr kernel: md: sync done. time=25499sec rate=76611K/sec (unRAID engine)

Apr  1 07:05:02 TODD-Svr kernel: md: recovery thread sync completion status: 0 (unRAID engine)

 

Any insight is appreciated, thanks.

They are at low block addresses, so likely they are in the housekeeping area of the disk.

 

You will need to run a "correcting" parity calc to fix them, since the monthly check is usually in NOCORRECT mode.  (It reports them, and the syslog reads as if they were corrected, but they are not)

 

Perform that correction before changing any disks.

 

Joe L.

They are at low block addresses, so likely they are in the housekeeping area of the disk.

 

You will need to run a "correcting" parity calc to fix them, since the monthly check is usually in NOCORRECT mode.  (It reports them, and the syslog reads as if they were corrected, but they are not)

 

Perform that correction before changing any disks.

 

Joe L.

 

I take it the easiest way to do this is the "Check and Correct Parity" button in unMENU? 

 

What causes errors like that?  Could changing a disk this month cause it or does it just happen once in a while?  Any chance my data on that one disk has an error too?

 

 

They are at low block addresses, so likely they are in the housekeeping area of the disk.

 

You will need to run a "correcting" parity calc to fix them, since the monthly check is usually in NOCORRECT mode.  (It reports them, and the syslog reads as if they were corrected, but they are not)

 

Perform that correction before changing any disks.

 

Joe L.

 

I take it the easiest way to do this is the "Check and Correct Parity" button in unMENU? 

 

What causes errors like that?  Could changing a disk this month cause it or does it just happen once in a while?  Any chance my data on that one disk has an error too?

Or, the "Check" button on the stock unRAID interface. (as it really checks and corrects)

Meant to close this last weekend.  The parity update went fine.