Parity Check always returns 5 errors

[Ceco]

Hi,

 

Since about 3/4 months every time my monthly Parity-Check runs it returns exactly 5 errors. I have parity corrections enabled so I don't think they should be the same errors (though i have no idea how to check that) but it still seems like an odd coincidence.

I never had a problem with parity check before which always returned 0 errors for over a year priori to that.

 

Not sure how to diagnose the potential problem or if i should even be worried about it.

 

System diagnostics are below.

unraid-diagnostics-20240601-0941.zip

[JorgeB]

Run another parity check without rebooting and post new diags if error are found.

[Ceco]

Started a new parity check, will send the diagnostics in about a day when it's finished

[Ceco]

Here are the new diagnostics after the second parity check, which returned 0 errors this time.

 

unraid-diagnostics-20240602-1342.zip

[JorgeB]

Now try rebooting and run another check, in some cases the persistent errors only happen after a reboot.

[Ceco]

Thanks, took me a minute but I ran another parity check after rebooting and got 5 errors again.

 

Diagnostics are attached.

unraid-diagnostics-20240609-1951.zip

[JorgeB]

When the same 5 errors happen, but only after a reboot, it's usually an add-on controller, in this case the Asmedia.

[Ceco]

Interesting, didn't think that would be the problem. Is this something that can be safely ignored or is there something i can/should do about it ?

[JonathanM]

18 hours ago, Ceco said:

Is this something that can be safely ignored or is there something i can/should do about it ?

Depends. If the errors are always in the same addresses, and those addresses aren't critical to any of the files or filesystems, then you can ignore them.

If not, then you need to replace parts until the parity check is always clean.

 

Another consideration is normalizing parity errors. Typically you would alert on any parity errors, currently you have to examine the logs every time to see if all the errors are "expected" or you have new ones which need attention.

 

Personally I wouldn't feel comfortable running like that.

[Ceco]

19 minutes ago, JonathanM said:

If the errors are always in the same addresses, and those addresses aren't critical to any of the files or filesystems, then you can ignore them.

Pardon the noob question but how can i check that ? I might not be looking in the right place but i can't find anything in the documenation for the parity or diagnostics tool explaining how to read logs or get more information about the parity errors. Not even sure what log i should be reading...

 

I went for a SATA Card with the Asmedia chip cause many users reported it running well in Unraid. Looking back, the start of the errors do coincide with the card's installation so is this kind of errors commonplace with such cards or could i possibly have a faulty one ?

[JonathanM]

Parity error addresses are in the syslog, but there is no easy way to correlate them to files or what they might corrupt. Determining whether you can safely ignore them is a very technical process, I gave the 10,000ft overview. If you aren't comfortable with that explanation, no, you can't safely ignore them.

 

It's not the Asmedia chip that seems to be the issue, more the implementation on the specific model of card. I would start by replacing the card with a different brand version of the same chip.

[Ceco]

That's a bummer, I specifically went for a SilverStone card hoping it would be more reliable than generic alternatives as it had positive feedback.

Anyway, i'll find a cheap replacement card and see what happens.

 

Thanks for the help.

[JonathanM]

33 minutes ago, Ceco said:

That's a bummer, I specifically went for a SilverStone card hoping it would be more reliable than generic alternatives as it had positive feedback.

Anyway, i'll find a cheap replacement card and see what happens.

 

Thanks for the help.

Have you checked to see if there is an update available for the card? IIRC silverstone is one of the few vendors that offer updated firmware for their cards.

[Ceco]

They do ! Unfortunately I already updated the firmware when i received the card and checking their website there hasn't been an update, so it doesn't look like this will be my solution.