Hi,

 

I do a non-correcting parity check once every 3 months. Because it takes almost 1,5 days I don't really want to do it more often.

 

It seems that I always get a couple of errors, about 5, at the beginning of the parity check. In the first hour or so. What I have been doing is, I let the check complete, then run a correcting check after and cancel that check when it's corrected those 5 errors.

 

Is this an OK thing to do? Or is there a better way to deal with this? I suppose running the correcting check completely and then run another non-correcting check would be the best, but that would take about 5 days.

 

Any advice would be much appreciated. Thanks.

Edited May 7, 20251 yr by Aemstel

6 minutes ago, Aemstel said:

It seems that I always get a couple of errors, about 5, at the beginning of the parity check.

Are you having unclean shutdowns?

 

 

1 minute ago, trurl said:

Are you having unclean shutdowns?

 

 

No, I don't think so. It's nearly always on. I reboot sometimes when there's an Unraid update, but those reboots seem clean to me.

You would get a parity check if you were. But a few sync errors at the beginning is typical for unclean shutdowns.

14 minutes ago, Aemstel said:

I suppose running the correcting check completely and then run another non-correcting check

That would at least confirm that the sync errors had been corrected. If not, then further investigation required.

If I run the correcting check for an hour, have it fix those 5 errors, then run another non-correcting check, also for an hour and see it finds errors. Could I do that? Or do I really need to run the 2nd and 3rd checks completely as well?

That should work.

Great, thanks for your help! I'll report back here with the result.

So here's what happened:

 

Check 1 - non-correcting
5 errors found within first couple hours. Not sure when exactly, I didn't keep an eye on it much. Let the entire check complete. Took 36 hours. Did not find any more errors.

 

Check 2 - correcting

5 errors corrected in first 20 minutes. Cancelled the check.

 

Check 3 - non-correcting
Let it run for 1 hour and it did not find any errors. Cancelled the check.

 

So I guess I'm good? So very few errors is not really something to worry about I guess, and it seems they got corrected just fine. It's still kind of weird why this happens every time. Always like 2 to 10 errors at the start of the parity check. Maybe when only running the check once every 3 months (I said 2 months before, but that was a typo), a couple of errors sneaking in is to be expected? I do use the array quite a lot.

Parity errors not expected. I have over 2.5 years without any. Those older ones I had were probably unclean shutdown.

Could the error come when the drives are going from down to spinning up during the parity check?   I had some instability in my array, and the big brains on this forum suggested a SATA or power cable problem.   I think I had too many disks on a single PSU line.   I added another line and kept the disks going 24/7 -  the array has been solid for months through a number of parity checks. 

Are errors only fond after a reboot, or even without rebooting?

16 hours ago, demanding-chief3698 said:

Could the error come when the drives are going from down to spinning up during the parity check?   I had some instability in my array, and the big brains on this forum suggested a SATA or power cable problem.   I think I had too many disks on a single PSU line.   I added another line and kept the disks going 24/7 -  the array has been solid for months through a number of parity checks. 

I don't know. I do have many disks and they do spin down after 3 hours.

6 hours ago, JorgeB said:

Are errors only fond after a reboot, or even without rebooting?

Don't know. I just do the check every 3 months. It's unlikely I didn't reboot at least once in that time to do an update.

Edited May 8, 20251 yr by Aemstel

20 minutes ago, Aemstel said:

I do have many disks

Ideally no more than 4 disks per PSU cable.

1 hour ago, Aemstel said:

Don't know. I just do the check every 3 months. It's unlikely I didn't reboot at least once in that time to do an update.

Start a new check, if errors are found, run another one without rebooting, if no errors are found, reboot and run a another one to see if they come up again, if it only happens after a reboot, it's a common issue with some controllers/hardware combinations

I wanted to mark JorgeB's answer as a solution too, but can only mark one. Thanks to all who helped.