So, quick recap, Had a parity disk go belly up last week, then the flash drive died. Got the flash replaced and a new parity drive came today. Now, of course, I am seeing a bunch of pending sectors on the currently installed parity 2 disk.  My question is what is the most efficient, and safest, way to get things back to normal. I assume I just need to pre clear the new disk, add it to the array as parity 1, and start a sync to calculate parity on the new disk as well as to see if the pending sectors on the parity 2 disk go away? What am I missing?

 

Or, maybe do an extended SMART test, on parity 2, while the pre-clear is happening ?

Edited May 6, 20251 yr by ratmice
additional thought

5 hours ago, ratmice said:

Or, maybe do an extended SMART test, on parity 2, while the pre-clear is happening ?

I would do this.

6 hours ago, JorgeB said:

I would do this.

the test fails almost immediately with a read error 🤬 I guess I need to get the new disk added as parity and get another new replacement. Odd for the 2 parity disks to crap out almost simultaneously.

Edited May 6, 20251 yr by ratmice

So, circling back to my server issues. I am ready to add in the new the parity disk, but am a little worried about the Parity 2 disk (SMART errors) currently in the server. Is the best course of action to just allocate the new disk to parity, rebuild, and see what happens? I'm a bit afraid that the Parity 2 disk will be pulled out of service due to the errors. As I understand it the 2 parity disks are independent of each other so that even if that happens, I can just deallocate the Parity 2 and not have any issues with the new parity disk, or the array, aside from only having protection from one disk failure. Or, is there some other procedure I should follow to keep things efficient and safe?

 

As always, I really appreciate the help here, as I can just think myself into circles due to incomplete knowledge. 

 

p.s. the other factor here is that I will be away from home for the next couple of months, so will not be able to physically lay hands on the server to do any more disk switching, etc... Perfect timing!

If the SMART test failed, I would remove parity2 for the sync, but if you can have the array offline for the sync, do it in maintenance mode and keep parity2 intact until it's done.

6 minutes ago, JorgeB said:

If the SMART test failed, I would remove parity2 for the sync, but if you can have the array offline for the sync, do it in maintenance mode and keep parity2 intact until it's done.

Sorry, I'm a bit confused. I've never really used maintenance mode. I can have the array offline, but how do you keep parity2 intact if it's removed for the sync. There seems to be something I'm missing, maybe it's the distinction between the disk used to store parity information (parity2 disk) vs the actual information (parity2). Sorry for the endless questions, but I don't want to do something stupid because my understanding is incorrect. Thanks for your patience. 

Starting the array in maintenance mode doesn't mount the disks, so patrity2 would still be valid if, for example, there were issues with a data disk during the sync, of course, since it's failing it would likely generate some errors during an eventual rebuild, but could still be useful.

30 minutes ago, JorgeB said:

Starting the array in maintenance mode doesn't mount the disks, so patrity2 would still be valid if, for example, there were issues with a data disk during the sync, of course, since it's failing it would likely generate some errors during an eventual rebuild, but could still be useful.

Gotcha, thanks for taking time to help me out. So, I will unassign parity 2 regardless of whether I do a regular restart and sync, or start in maintenance and sync. 

 

Well, unassigning parity2 and adding the new disk as parity1  leads to an invalid configuration, so I guess I just assign the new disk, tick the maintenance mode box, and restart the array?

Edited May 8, 20251 yr by ratmice