Hi, I have a ReadyNAS RN426 running the latest stock firmware/OS. The RN426 is configured with 3 10T 7200 rpm disks and 2 500g SSDs. One of the 10Ts is a globalspare.The other2 are mirrored, as are the SSDs.Additionally, the 2 mirrors are tiered.
One of the mirrored 10T disks is having problems and will fail soon — but has not failed yet.I would like to replace it with the global spare.The best way to preserve redundancy during this process is to change the 2 disk mirror to a 3 disk mirror.Wait for all 3 disks to sync then remove the failing drive and change the array back to a 2 disk mirror.
I could ssh onto the device and use mdadm to fix my array.That certainly wouldn’t create the metadata that the ReadyNAS web interface needs.
My question is: is there a way to accomplish this using the web administration tools?
Model: RN426|ReadyNAS 426 – High-performance Business Data Storage - 6-Bays
No responses.... Maybe my question is too specific.
What are the best practices to preserve redundancy while replacing a failing disk?
Sure, I could pull the failing drive and the ReadyNAS will recogize the failure and start an automatic rebuild on the spare. It will take 2-3 days to rebuild during which time a failure will result in data loss.
It is really not good to expand your array if there's any impending failure, it's always best to resolve/replace the disks first before any expansion takes place.
There's no way on the Admin page to do that as well so the option is to pull out the failing disk for the global spare to take over and after it re-synced on the array then you can add another to expand or add another to set as Global spare. There's no way you can actually preserve redundancy when there's a failing/failed disk on the array.
I am not looking to expand the array. I am looking to increase redundancy from 2 to 3 mirrored copies. Even though one of the 2 disks is going to fail soon, it has not failed yet. For the moment, there are 2 copies of my data on 2 devices one of those devices just happens to be fragile. Still, I have complete redundancy.
This is just one of many use cases where mirrors of 3 or more copies are completely reasonable.
I am looking to increase redundancy from 2 to 3 mirrored copies.
The NAS web ui doesn't support creation of triple RAID-1.
The closest you can get to that is to create an independent JBOD volume of one disk, and use the built-in back up jobs to update the JBOD from the RAID-1 volume.
As a feature, this seems like an easy opportunity to bring enterprise functionality to the product. Linux already supports the underpinnings.
On that line of thought, what would happen to the rest of the ReadyNAS stack if I formatted the global spare and used mdadm to add 3rd disks to the existing mirrors.
On that line of thought, what would happen to the rest of the ReadyNAS stack if I formatted the global spare and used mdadm to add 3rd disks to the existing mirrors.
I think the point @Marc_V was trying to make is that re-syncing your array with another drive (not expanding the volume size in your case, but expanding the number of drives included) is not a good idea with a failing drive. The failing drive has a high liklihood of failing during that process, and possibly bringing down the whole RAID with it.
If you pull the bad drive, it will still have a full copy of the volume on it, since it's RAID1. So, you really do still have redundancy when you put in the new one to replace the old, just not online redundancy. If the other older drive fails durng the re-sync process, you still have an opportunity to try with the failing one.
