Pulled active drive by accident, triggering error state, now what?

[heffe2001]

I have a few empty drive sleds in a Compellant SC200, and went to pull one to zero out a 6tb SAS drive.  I didn't pay close enough attention, and pulled an active drive in the array (the light on that sled is extremely dim, not bright like the rest), and the drive entered an error state (not for any write errors, but 1 read error).  Upon reboot, it's got he drive marked as bad, and being emulated, even though it's not ACTUALLY bad.

 

What's the best way to handle this and get the drive back into the array?  Pull it, reboot, then reinstall and run a preclear is the process I've run in the past, or would I just be better off dropping the parity drive, readding it, and just rebuilding parity?  I'm 99.999% certain there's no data errors (mover wasn't running, and I don't have any processes that write directly to the array disks).  Either way I'm going to have to preclear and I'm guessing either way is going to require one a drive rebuild, but parity is larger (10tb), where this was an 8tb, so it'd (i'm guessing) be faster to go that route.  Just wondering what other folks process for bone-head moves like this, lol.

 

[trurl]

8 minutes ago, heffe2001 said:

Either way I'm going to have to preclear

Can't imagine why you think preclear is needed.

 

If you New Config/Trust Parity, you will still need a parity check, so might as well rebuild the disk or rebuild parity. As you say, rebuilding the smaller disk will be faster.

[heffe2001]

In the past when I've tried to just re-add a drive back in that had been triggered bad, I'd had to wipe it to get it to accept it again. 

 

What would be the procedure to NOT have to do that in this instance?  Drop the disk from the array, add it back and then rebuild, or will it allow me to do that?

 

In the past when I've had drives drop due to issues like this, I've usually precleared it again just to make sure the smart attributes and all hadn't been incremented or anything like that, but in this instance, I know EXACTLY what happened, lol.

 

Went ahead and dropped it from the array, started, stopped, added it back, and now it's rebuilding.  Like I said, I've always run the clear again in the past to 'test' the drive, but I don't think it's necessary in this case.

 

 

Edited August 23, 2023 by heffe2001