I've read the H810 manual, the OpenMange Help pages, and a Knowledge Base article about this. I'm still confused. This is a RAID 6.
A drive with a predicted failure is still online. So if you want to deal with it you can:
1.) Force it offline. This immediately causes a rebuild on the hot spare. One caveat mentioned only in the OpenManage Help pages is you are urged to do a consistency check first because they say in some circumstances forcing the drive offline will cause loss of data if the consistency check isn't done first. Wouldn't this caution also extend to a drive failure? Can't do a consistency check before a drive failure :-).
2.) If you have "Allow Revertible Hot Spare & Replace Member" and "Auto Replace Member on Predictive Failure" enabled then (I think) the controller tries to just copy the predicted failure disk to the hot spare. This is not a rebuild, just a copy so it's faster. You may be able to manually do the replace member command as well to do the same thing if "Auto Replace Member on Predictive Failure" is not enabled. I don't know.
In any case there are two approaches listed above for handling a disk with predicted failure. The Knowledge Base article only mentions number 1 (without the check consistency caveat).
What is the best way to do this?