Volume degraded after scrub, next steps

---

markwoll wrote:

 

 it looks like the 4 8tb drives are becoming another raid.

 

---

 Since your drives are now of unequal size, your data volume needs multiple RAID groups.

 

The first RAID group is 6x2TB (taking the first 2 TB of the larger drives).  The second is 4x6TB.  Those two RAID groups are then concatenated into a single volume.

 

Since you are using RAID-6, you need at least 4 disks in the group.  So the second RAID group wasn't created until you upgraded the fourth drive.  Completing this step will take a while (and upgrading the final two will actually take longer).

 

Note when you are done, you will still have two RAID groups - even though your disks will end up the same size.

 

---

markwoll wrote:

 

After this resync is done will I be able to swap the remaining 2 2tb and the unit will resync again to the final size, or will it resync twice after each swap?

 

---

The capacity rule for XRAID dual redundancy is sum the disks and subtract the two largest. 

 

So 4x8TB+2x2TB will give you a 20 TB volume. (~18.2 TiB)

 

When you add the next drive, the md127 group will be rebuilt, and then md126 group will be converted to 5x6TB.  Every block on every disk is either read or written as part of this process.  It will take longer than the current step, because you're adding another 6 TB to the array.  You'll end up with a 26 TB volume (~23.6 TiB).

 

Similarly, when you insert the last drive, the md127 group is first rebuilt, and then the md126 group is converted to 6x6TB.  Again, this step will take longer than the previous ones.  Then you'll end up with your final volume size of 32 TB (~29 TiB)

 

So there are two resync steps as you add each disk - one for each RAID group.  FWIW, the process would have been faster if you'd backed up the data and did a factory default with all the disks in place (and then restored all the data from backup).  Then the RAID array would only have been built once, instead of rebuilding it 6 times.