I know (and have done before) that if you enable a disk that you know works, the proper process is to remove it and re-add it and let it rebuild.

But I want to NOT rebuild and instead just re-use it.

Is there a way to do that?

 

Maybe by disabling parity, enabling the other disk and then (assuming UNRAID accepts it again as part of the array), re-enable the parity?

Or maybe by doing the "normal" process but somehow just make parity check and not write back?

 

Any idea?

 

Why I want to do that: Not only I trust the disabled disk and existing data, but actually I am afraid that other events (like temp issues with other disks) may ruin disk's content instead of building it.

You can use Tools->New Config with the option to keep all current assignments.   When you return to the Main tab you can check the Parity is Valid checkbox to avoid parity being rebuilt and then start the array.

 

This assumes that parity IS valid.   It might be a good idea after this to do a parity check to confirm this.

Yes I plan to do a parity check of course (I do every month anyway), but I am considering for once to not allow it to make fixes, just report problems.
OK thanks for this.

One more thing @itimpi please.
Assuming I do new config (retaining the assignments) and then start it with "parity is valid" and then doing a parity check without writing back corrections.

In case I DO find issues (discrepancy between parity and data), is there a way to control what corrects what? (parity to data vs data to parity)

 

41 minutes ago, NLS said:

In case I DO find issues (discrepancy between parity and data), is there a way to control what corrects what? (parity to data vs data to parity)

No.  The only option via parity check is to assume the data is valid and make the parity match.    I believe that the sequence in which parity and data drives get written means that in case of discrepancy the parity is more likely to be correct, and anyway there is not sufficient information written to determine which data disk has a problem if it is one of them.
 

If you want to know if data files are valid then you need to either use a file system (btrfs or ZFS) that has built in check-summing or if using xfs use external check summing such as provided by the Dynamix File integrity plugin.  Then when errors are detected restoring the corrupted files from backups (or perhaps re-download them).

For completeness, I am linking the original thread this discussion started in:

 

 

Yeah thought of making this its own topic, because the original had so many issues opened and closed, that even for the furure search of a user, it wouldn't help them.