Avoid run away parity checks

We have had a number of situations over the years where a user will be trying to recover from a failure and run a parity check only to have it encounter huge numbers of sync errors on virtually every block. UnRaid faithfully updates parity (at least some of it).before the mistake is noticed which renders parity useless to do a recovery.

Would be nice to offer some protection.

For example:

Run the beginning of the parity check in non-correcting mode, and if all looks reasonable, restart it in correcting mode.

Or

Run a quick abbreviated read only check, that samples data at several locations on the disk before kicking off the correcting check. (This could also be run on starting the array to detect extreme out of sync parity situations.)