Ok, I have had som problem with parity errors and i might have found the disk that is causing it, but i am not completely sure.

So, my question is.

 

Can i remove the disk from the array, use the restore button (that will wipe parity from that drive)

Check parity a couple of times if it still doesnt synch without that disk in the array.

 

And if it works, can I add the old disk back and keep the (failed? data) as before.

And if it still doesnt work try other disks to see if they are the problem.

I am aware of the missing protection from parity if something else goes wrong.

 

This is a trial and error idea I want to know if I can use, or will something bad happen with the shares and so on?

 

Btw. I cant see any errors in the syslog. And smart data says:

SMART overall-health self-assessment test result: PASSED

And no (new errors) when running long smart test.

 

Can you explain what you have done and where you have seen parity errors in more detail?  A screenshot of the unRAID "Main" screen might help.  I would also suggest a syslog.  Click on the "Troubleshooting" link in my sig for instructions on getting a syslog and other info that might be helpful in troubleshooting your problem.

 

Click here for more info about parity.

Ok, I have had som problem with parity errors and i might have found the disk that is causing it, but i am not completely sure.

So, my question is.

 

Can i remove the disk from the array, use the restore button (that will wipe parity from that drive)

You must stop the array, un-assign the disk, then press the restore button.  It will then immediately invalidate parity and when you press the "Start" button it will then begin to calculate arity again, but without the removed drive.  When the parity calc is complete, you will be protected from a disk failure again.

Check parity a couple of times if it still doesnt synch without that disk in the array.

Although this might point to a drive, and you might find one that causes parity errors when read, this is *very* rare as there are internal checksums kept by the disk and sectors are re-read until either read properly, or a disk error is reported to the disk controller.     It is just as likely, (perhaps more likely) that the cause of the errors is cabling, or power to the drive.   In all the reports of random parity errors in the unRAID forum, I can remember only one that had been traced to a specific disk drive.  (So it can happen, it is just not very common)

 

And if it works, can I add the old disk back and keep the (failed? data) as before.

You'll need to stop the array, assign the disk you un-assigned, go back to the main page, press "Restore" once more, press "Start" to start re-calculating parity, wait until it is done. 

And if it still doesnt work try other disks to see if they are the problem.

As I said, if it was a marginal poor power supply, or a splitter with a poor connection it could be that removing any drive will reduce the load on the supply to where the remaining disks will work.  If you are not careful, you might think you found the issue, but find it back once you install a replacement drive.

I am aware of the missing protection from parity if something else goes wrong.

This is a trial and error idea I want to know if I can use, or will something bad happen with the shares and so on?

Should work, and it sounds like you know the risks.  Be very careful not to dislodge the cabling to the remaining drives in the array.

Btw. I cant see any errors in the syslog. And smart data says:

SMART overall-health self-assessment test result: PASSED

And no (new errors) when running long smart test.

 

Please attach to one of your next posts a full syslog, after you perform a parity check where errors are detected, but before you reboot.  Unless you are an expert in interpreting the output, it is easy to miss something.  (RobJ is one of the best in interpreting the logs)   Also post on the full smart report on your drive.   The overall "result: PASSED" is very misleading.   The drive would need to be pretty badly failing to see anything else.   

 

Joe L.

Update: Well, i bought 2 new drives instead and I am using only 1 of my old drives at the moment and that seems to work, I am not using my old PCI IDE controller anymore.

Now i will try just adding one new drive at a time and see if any parity errors occurs again or not.