Parity Check Errors - Auto Fix?

[David Bott]

Hello All, long time no type.

 

Ver: 4.7

 

Using Program: Monthly Parity Check

Description: This package installs a script that will schedule a monthly parity check on the 1st of the month at midnight.

Use NOCORRECT if you do not want parity to be automatically updated.

Use CORRECT if you do want it automatically updated. It is recommended that you NOT automatically correct parity, since it might be a data drive that is in error, and the parity drive might be correct.

Challenge is determining which is correct, and which is in error. unRAID normally assumes the data is correct and parity is wrong. Pressing the "Check" button on the web-interface will check AND update parity based on the data disks.

 

Each month my system runs full check using the above.  I had thought that when it finds a parity issue it auto fixes the issue and I would not see the error again upon the next run. (But as I read now it does not unless I tell it to and that can be a bad thing.)  So as you will see below, I get the same error each month (seeing it was not corrected.). 

 

Jun  1 00:00:01 Server kernel: mdcmd (301): check NOCORRECT (unRAID engine)

Jun  1 00:00:01 Server kernel:  (Routine)

Jun  1 00:00:01 Server kernel: md: recovery thread woken up ... (unRAID engine)

Jun  1 00:00:01 Server kernel: md: recovery thread checking parity... (unRAID engine)

Jun  1 00:00:01 Server kernel: md: using 1152k window, over a total of 1465138552 blocks. (unRAID engine)

Jun  1 00:00:11 Server kernel: md: parity incorrect: 6928 (Errors)

Jun  1 00:00:11 Server kernel: md: parity incorrect: 6936 (Errors)

Jun  1 00:00:11 Server kernel: md: parity incorrect: 15984 (Errors)

Jun  1 00:00:11 Server kernel: md: parity incorrect: 15992 (Errors)

Jun  1 07:16:48 Server kernel: md: sync done. time=26205sec rate=55910K/sec (unRAID engine)

Jun  1 07:16:48 Server kernel: md: recovery thread sync completion status: 0 (unRAID engine)

 

 

Jul  1 00:00:01 Server kernel: mdcmd (706): check NOCORRECT (unRAID engine)

Jul  1 00:00:01 Server kernel:  (Routine)

Jul  1 00:00:01 Server kernel: md: recovery thread woken up ... (unRAID engine)

Jul  1 00:00:01 Server kernel: md: recovery thread checking parity... (unRAID engine)

Jul  1 00:00:01 Server kernel: md: using 1152k window, over a total of 1465138552 blocks. (unRAID engine)

Jul  1 00:00:10 Server kernel: md: parity incorrect: 6928 (Errors)

Jul  1 00:00:10 Server kernel: md: parity incorrect: 6936 (Errors)

Jul  1 00:00:10 Server kernel: md: parity incorrect: 15984 (Errors)

Jul  1 00:00:10 Server kernel: md: parity incorrect: 15992 (Errors)

Jul  1 07:15:13 Server kernel: md: sync done. time=26111sec rate=56111K/sec (unRAID engine)

Jul  1 07:15:13 Server kernel: md: recovery thread sync completion status: 0 (unRAID engine)

 

Aug  1 00:00:01 Server kernel: mdcmd (1523): check NOCORRECT (unRAID engine)

Aug  1 00:00:01 Server kernel:  (Routine)

Aug  1 00:00:01 Server kernel: md: recovery thread woken up ... (unRAID engine)

Aug  1 00:00:01 Server kernel: md: recovery thread checking parity... (unRAID engine)

Aug  1 00:00:01 Server kernel: md: using 1152k window, over a total of 1465138552 blocks. (unRAID engine)

Aug  1 00:00:13 Server kernel: md: parity incorrect: 6928 (Errors)

Aug  1 00:00:13 Server kernel: md: parity incorrect: 6936 (Errors)

Aug  1 00:00:13 Server kernel: md: parity incorrect: 15984 (Errors)

Aug  1 00:00:13 Server kernel: md: parity incorrect: 15992 (Errors)

Aug  1 07:07:45 Server kernel: md: sync done. time=25663sec rate=57091K/sec (unRAID engine)

Aug  1 07:07:45 Server kernel: md: recovery thread sync completion status: 0 (unRAID engine)

 

Can someone please advise... is their a way to have it maybe then run the CHECK that will then fix the errors if errors are found?  Or how are users auto doing things like this.  Or are you not?

 

I just want to be sure my data is of cose safe.  (Currently running the manual CHECK function now to clean this up.)

 

Thank you for the advise.

[Joe L.]

The only way is to fix the errors, as you are now doing with the CORRECT option.

[David Bott]

Hi Joe...(Looking forward to seeing you Ilene.)

 

Thats what I thought...but same errors again...

 

Aug  1 10:49:47 Server kernel: mdcmd (1546): check CORRECT (unRAID engine)

Aug  1 10:49:47 Server kernel: md: recovery thread woken up ... (unRAID engine)

Aug  1 10:49:47 Server kernel: md: recovery thread checking parity... (unRAID engine)

Aug  1 10:49:47 Server kernel: md: using 1152k window, over a total of 1465138552 blocks. (unRAID engine)

Aug  1 10:49:48 Server kernel: md: parity incorrect: 6928 (Errors)

Aug  1 10:49:48 Server kernel: md: parity incorrect: 6936 (Errors)

Aug  1 10:49:48 Server kernel: md: parity incorrect: 15984 (Errors)

Aug  1 10:49:48 Server kernel: md: parity incorrect: 15992 (Errors)

Aug  1 17:55:02 Server kernel: md: sync done. time=25515sec rate=57422K/sec (unRAID engine)

Aug  1 17:55:02 Server kernel: md: recovery thread sync completion status: 0 (unRAID engine)

 

 

...Oh wait...will it still show the error count but they were actually corrected this time?

 

[JonathanM]

...Oh wait...will it still show the error count but they were actually corrected this time?

If you did it right, yes. Run a non-correcting check again if you want to make sure.

[Joe L.]

...Oh wait...will it still show the error count but they were actually corrected this time?

The syslog output was never changed to show any difference in correcting mode vs. non-correcting mode.

 

As already mentioned, another check (or any kind) should show no errors.  You can start one and cancel it within a minute or so if you see nothing in the syslog, since the errors showed themselves in the very first few seconds.

 

The consistent errors month to month show there is nothing to concern yourself about with your media files.  The very low addresses with the errors were in the housekeeping/journaling area of the file-system.

 

Joe L.

PS. Looking forward to seeing you and Brenda again too.