hopefully your backup processes are good enough that if a drive fails, or if the filesystem becomes corrupt enough, that you can easily just recover from backup, rather than try to repair the disk.
But it happens... that data that isn't quite important enough to worry about, but is worth a little effort to try to recover.
Starting with TestDisk
Sometimes the hardware does fail and in that case, check if the product is still under warranty most drives have a 3 year warranty, and many drives fail within that time, if they're going to fail.
haven't been using software level RAID systems as much these days. Recovery is not straight forward.
Started looking into it here, then abandoned the effort:
requires mdadm tools.
then: assemble the raid:
mdadm --assemble /dev/<disk1> /dev/<disk2> /dev/<disk3> /dev/<disk4> mdadm -E -s
If you have a proprietary file system (e.g. NTFS, HFS+) and something goes wrong -- if there is important information on the drive it can be worth the cost of reputable commercial recovery software
Get Data Back
in the past have used: Get Data Back software on Windows (usually seems to be NTFS filesystems that go bad most often)
on HFS+, mac formatted drives... If drive has been synchronized and meta folders have been deleted / corrupted, when first connecting drive to an actual OS X system, give it time to rebuild indexes before attempting a First Aid / Repair with Disk Utility. I have seen a perfectly fine drive report "Cannot recover, back up all data". Then I let the indexer run, and a few hours later no errors were reported with Disk Utility. Give it time.