Hi All,

Every time I run a parity check I get a different number of errors each time (never 0). I dug through some earlier posts of people having similar issues and have ran a correcting parity check followed immediately by a non-correcting parity check and it gave me a different number of errors.

Please see attached for Diagnostics, screenshot of my parity check history is also attached.

sdchonitower-diagnostics-20240825-1714.zip

Start by running memtest.

Okay, I ran it for 24 hours and 5 passes and it showed 0 errors.

I pressed ESC to close instead of c, so it did not save the HTML output. My bad on that one....

 

EDIT:
Apparently I was using an ancient version of memtest, version 4.37. I'm going to run it again using a newer version just to be safe.

Edited August 27, 20241 yr by sdchoni

MemTest86-Report-20240827-141537.html
Pass with 0 errors.
 

Unfortunately Memtest is only definitive if it finds errors, but if there are no obvious RAM issues, try this:

 

reboot to clear the log

run a correcting check

without rebooting run a non correcting check

 

Then post the diags.

Alright, so according to my archived notifications (which I believe is correct) I have run a correcting parity check, non-correcting, correcting, and non-correcting (each with a different amount of errors each time).



According to History, I have run 3 non-correcting parity checks?



I have attached my diagnostics. The Server has been on since I ran MemTest on Aug 27th

sdchonitower-diagnostics-20240903-1341.zip

You can ignore the notifications, there's an old bug, history should always be correct.

 

Diags only show one check, possibly due to mover log spam, but since the last three found a different number of errors, there's a problem, RAM would still be my main suspect, since you have 4 sticks, I would remove two, run a correcting check, then a non correcting one, if the latter finds new errors repeat with the other RAM pair.

Alright, so I stuck in 2 sticks, ran a parity check, got 3 errors. Ran a correcting parity check got 0 errors!

I then put in the other 2 sticks of ram, ran a parity check, and got 0 errors.

 

Maybe it's my DIMM slots or maybe I just needed to re-seat my ram?

 

I guess I'll put all 4 sticks back in, try running another parity check and see what happens?

I tried downloading the diagnostics, but it crashes my browser after it uses up all of my ram.

It is not that unusual to find running with more RAM sticks can put extra total load on the memory controller so you get errors even though each stick checks out fine.

I put all 4 sticks back in and I get errors again (more specifically: I ran a correcting parity check and got  1  error

, ran a non-correcting parity check and the kernal crashed



, and ran another correcting parity check and got 3 errors)


When running a diagnostics on my windows machine firefox uses all my ram and then crashes (I do get a chuckle that there is a running theme with ram here lol).

I guess it doesn't like 32gb of DDR3 ram for some reason? I will downgrade to 16gb for the time being (tbh, I only upgraded b/c we were running a palworld server at the time and it had a ram leak. 16gb has been fine without that)  and I will look into a system upgrade in the future.