Hi everyone! I'm using Unraid 6.12.10.

On Monday, I was doing a check-read on all the hard drives to set a couple of them as parity. Apparently, everything went well. I also did a check in maintenance mode, and I didn't get any errors. And neither have SMART errors.

 

 

 

Today, I set a couple of them as parity hard drives, and I started getting a lot of read errors.

 

 

Apparently, the content seems to be fine on those drives, and I randomly checked some things, and there don't appear to be any read errors. If there had been errors, the check-read would have returned some error.

 

What could be happening? Should I let the parity process finish with so many errors? I am attaching the diagnostic file in case someone can help me.

 

Kind regards.

 

 

 

tower-diagnostics-20240627-2114.zip

It's not logged as a disk problem and SMART looks OK, so most likely a power/connection issue.

Thanks for your reply JorgeB!

 

It seems that I have solved it, for now, by reversing the order of the hard drive cables. That is, the last cable is now on the first hard drive, and so on with all of them. Honestly, I don't understand anything, and what happened is very strange. I'll come back when the parity process is finished if finished ok, to tell you.

 

 

1 hour ago, dravenar said:

It seems that I have solved it, for now, by reversing the order of the hard drive cables

 

That suggests it could have been a connection issue and re-ordering the cables meant they ended up being better seated.   The SATA connector seems to be quite frequently prone to making bad connections.

Last night, after the process passed 50%, one of the drives started giving me constant errors, so I canceled the process. I'll try again next week again.

It's very strange because if the cable or connection was truly faulty, it should also give me read errors while I'm using the array, right?

 

15 minutes ago, dravenar said:

It's very strange because if the cable or connection was truly faulty, it should also give me read errors while I'm using the array

Connection issues can often be intermittent.   The DATA connector in particular seems to be prone to this.    Vibration and/of thermal issues can mean a connection that was working suddenly starts producing errors for no apparent reason.

I don't know why, but I've tried changing the cables, the order,... and when I try to do the parity process, at a certain point (it's not always the same %), it starts giving errors on one or two drives. The problem is that it doesn't always give errors on the same drive; sometimes it changes. This seems very strange, because the server has been running for a month without any read problems; they only appear during the parity process.

 

Could it be that the controller gets overloaded at some point? If the parity process finishes with errors, does it mean that the parity is incorrect and useless?

Another possibility might be that the PSU is struggling as parity check is one time all drives are simultaneously in use.