Anyone run into an issue where they run a NOCORRECT parity check and it finds errrors, and run a CORRECT parity check after and those errors are no where to be found? I'm concerned if my parity is good or not at this point.

Is this scenario repeatable? Syslog(s) showing the events would be helpful.

it's shown once in this log although this isn't the first time it happened although i do not have any record of it unfortunately so you might have to just trust me lol

syslog-2013-08-20.zip

Was anything written to the array between the last 2 checks?

yes, plenty i would imagine.. so a true test would be to run 2 parity checks back to back with no data being written?

Yes. If the bad sectors were rewritten then the errors could be corrected.

previous syslog was up to the 20th's so i'll spare you that and just start from the 21st.

 

it happened again today. no data was written to the array yesterday or today.. parity check i ran yesterday morning found errors and paritty check i ran overnight found none. both NOCORRECT.

 

i do see something that i didn't notice before.. a brief power hit on the 21st.. could that cause it? i am not sure because the parity check after that was clean.

 

aug 21 @ 0530- 2 sec power hit

aug 23 @ 0901 - completed NOCORRECT w/ no errors

aug 24 @ 0245 - completed NOCORRECT w/ errors

aug24 @ 1523 - completed NOCORRECT w/ no errors

syslog-2013-08-24.txt

Yes. If the bad sectors were rewritten then the errors could be corrected.

 

No, that's not true.  Writing data to the array does NOT result in the computation of the correct parity -- it only UPDATES parity based solely on one bit from the disk being written to and the current value on the parity disk.    So if the parity value was wrong, it will still be wrong.    To compute the correct parity value would require spinning up and reading ALL disks.

 

As for what's happening here ... I'd suspect one of the following:

  =>  You're INTENDING to run no-correct checks, but are actually running correcting checks.    Are you running these from the main Web GUI or from an add-on?  [You may have discovered a glitch in an add-on]

 

  =>  You've got an intermittent memory error (which can cause many strange things to happen) ... and whether or not you get an error depends on whether or not that specific area of memory is used during the check.    Running MemTest for a LONG test (24 hours or so) MAY isolate this kind of issue.

 

  =>  One of your disks' cache memory is having occasional "hiccups" and transmitting a bad bit.  However, this should result in a disk error in the "Errors" column -- are you seeing any reported disk errors during these tests?

 

 

As for what's happening here ... I'd suspect one of the following:

  =>  You're INTENDING to run no-correct checks, but are actually running correcting checks.    Are you running these from the main Web GUI or from an add-on?  [You may have discovered a glitch in an add-on]

 

it's from simplefeatures' GUI (or is it just a theme?) in the syslog it says "check NOCORRECT" when it runs the parity check so i would think it is indeed running a NOCORRECT but i will disable SF and see if it makes a difference.

 

  =>  You've got an intermittent memory error (which can cause many strange things to happen) ... and whether or not you get an error depends on whether or not that specific area of memory is used during the check.    Running MemTest for a LONG test (24 hours or so) MAY isolate this kind of issue.

 

i did run a memtest for approx 16 hours with no errors. I'm not sure how many passes. I don't remember when exactly but not too long ago.. it was still within the month of august. another thing to add to my list of things to try.

 

  =>  One of your disks' cache memory is having occasional "hiccups" and transmitting a bad bit.  However, this should result in a disk error in the "Errors" column -- are you seeing any reported disk errors during these tests?

 

all drives are up and clean.. no errors.

 

some history:

 

feb 2013 - drive in disk2 went bad, was able to add it back to the array by unassigning, rebooting and assigning again

apr 26 2013 - disk2 failed again, replaced drive with a new one and rebuilt.. parity was good

aug 01 2013 - power outage.. massive amount of errors in back to back parity checks... ran memtest overnight, all clear.. was testing drives and put everything back together (no changes done at this point)and was getting clean parity checks all around (3 back to back checks)

aug 08 - disk2 failed again.. swapped SATA cable.. so far so good but after a while started getting intermittent parity errors.