Random parity errors in the middle of 5 parity checks.

[YourNightmar3]

A couple days ago i replaced my original 4TB parity drive for an 8TB parity drive, and i added the 4tb that was original the parity to the array, and i also added another 8TB drive to the array. I did it like this:

1. Shutdown server. (No hotswap)
2. Add two 8TB drives to the server.
3. Power on server
4. Reset config, then assign all array drives like normal and replace the parity with one of the 8TB drives, leaving the original parity (4tb) and the second 8tb drive unassigned. The array is unchanged, only the parity drive has been changed.
5. Rebuild parity
6. Add the 4tb and 8tb drives to the array with preclearing.
7. Perform a parity check.

Everything went normal. No parity errors or anything during this process. Now here's my parity check log for the past 5 days:

 

Nr        Date                    Duration                Speed        Status    Errors
1        2020-11-03, 04:47:29    18 hr, 54 min, 49 sec    117.5 MB/s    OK        0
2        2020-11-01, 18:54:50    18 hr, 54 min, 49 sec    117.5 MB/s    OK        58
3        2020-10-28, 17:52:33    8 hr, 6 min, 1 sec        274.4 MB/s    OK        0
4        2020-10-28, 06:01:04    13 hr, 59 min, 54 sec    158.8 MB/s    OK        0
5        2020-10-27, 15:58:14    18 hr, 31 min, 33 sec    120.0 MB/s    OK        0

(if the above is not readable due to formatting, i have included a screenshot in the attachments)

 

I added a "nr" column to be able to name it here. So here's what's going on:

• Looking at the delivery date of the two new drives, Nr 5 is the original parity rebuild that i described in step 5 above.
• Nr 4 is a parity check described in step 7 above.
• I am honestly not sure what nr 3 is. The speed is suddenly much higher and the duration much lower. I do not know what caused this and i do not know what this parity check is and why it was performed.
• Nr 2 is the monthly first day of the month parity check, suddenly with 58 errors. I changed absolutely nothing between the first 3 parity checks and the monthly one. The server didn't power off, and i did not change any settings or anything. Auto correct errors was ON so it corrected them (which i later read is not always a good idea).
• I performed another parity check to see if more errors would occur immediately after, which has 0 errors again. This is nr 1 in the list above.

 

What is the explanation for this? Why is there suddenly 58 errors in the scheduled check but is nothing wrong with my own manual checks? Do i need to be worried? How do i make sure really nothing is wrong? I have included the diagnostics in the attachments of this post as well.

tower-diagnostics-20201103-1007.zip

Edited November 3, 2020 by YourNightmar3

[YourNightmar3]

Am i allowed to bump? I'm not really sure why all other posts except this one are being answered. Did i do something wrong?

[JorgeB]

These issues aren't easy to diagnose, I would start by running memtest for a few hours, ideally at least 24, if no errors run a couple of more checks.