Hy Guys,

i'm currently running the following configuration in my UnRaid server:

500GB SSD CACHE

14 Tb Parity

14 TB + 8 TB + 4 TB + 4 TB data (xfs, array)

 

Actually 8TB is empty (not cleared), and the 2x4TB are full of Media (not a super-big-deal if I loose data on that).

 

I'd like to know what's the fastest procedure to arrive to the following config:

 

Parity SWAP 14 to 18TB

2 x 14 TB data

 

I've one "big" constrain, only 6 sata port.

If needed I can evaluate to move cache to array in the meanwhile to gain one more free sata port.

 

My plan:

 

- Step 1 BACKUP important data to external destination

- Step 2 STOP ARRAY --> New Config with:

    - Actual parity As a drive

    - Actual 14TB data drive as a drive

    - New 18TB as parity

 

Let parity rebuild

 

Step 3 --> via unassigned device mount the 2x4TB and RSync to the now available and free 14TB drive (the 2 x4TB will retain data and will be readable?? I can't find this information, but I think so)

 

Does exist any way to not loose parity apart from:

-Step 1 --> clear script on 8TB --> remove without braking parity

-Step 2 --> Parity swap procedure installing 18TB

-Step 3 --> move data from 2x4TB to 14Tb mow available

-Step 4 --> clear script on 1st 4 TB -->  remove without braking parity

-Step 5 --> clear script on 2nd 4 TB -->  remove without braking parity

 

my extimation of first is a... mmm 28h parity rebuild + 6 to 8  hrs of rsyncing (from outside the array I can turbo write FAST!)

 

the second one is committed in never loose parity but will be something like mmm 

-Step 1 32hr (at around 70 Mb/s) <-- server unavailable

-Step 2 28h parity rebuild

-Step 3 16Hrs of rsync (array to array = slow)

-Step 4 16Hr <-- server unavailable

Step 5 16 Hr <-- server unavailable

 

So it would be26Hrs Vs 100+hrs which in my opinion would put A LOT of stress on the HDD.

 

Am I missing something?

 

 

Actually all HDD health is great.

 

thanks for your help

 

 

Edited June 3, 20251 yr by playmake

17 minutes ago, playmake said:

So it would be26Hrs Vs 100+hrs which in my opinion would put A LOT of stress on the HDD.

 

Not really!  These drives were designed for 24-7 operation.  There are ~8700 hours in a year and the drive should last for five years.  Do the math.  IF these are new drives and they should fail during the 100+ hours you need to get them ready for use with Unraid, you have a 'bad' drive that you don't want in your array.  If you are concerned that it will reduce the total life of the drive, what is 100 hours of shorten life compared to 43,000 hours?  Remember that the process of getting a drive installed for Unraid is not really beyond what the drive was designed to do.  Think of big data centers.  They have thousands in which hard drives read and write 24-7 for years.

 

What gets most people upset is that it takes close to a week to do everything.  (One thing I really don't like about a big drives is the time required for the monthly parity check.  About drive life of Unraid data drives.  I still have one 3TB data drive with 67,000 hours on it.  I had a couple of drives with even more hours but I have been replacing the oldest ones when I need more storage space.)  

thanks for your reply! then you suggest to follow the clear remove - clear remove ?

 

Do you know if I can clear 2 and remove in just one move? or I should follow the steps I described earlier?

I am not sure what you mean with the following statement:

3 hours ago, playmake said:

-Step 1 --> clear script on 8TB --> remove without braking parity

 

I am not aware of any 'script' that will clear a disk which will allow removal without breaking parity. (Copying all of the files off a disk does not remove the file structure and its tables.  This information is a part of the parity data stored on the parity disk!  Thus, physically removing a disk after just copying the files off the disk will break parity!!!)

 

That having been said there is the following procedure for the situation where a new array disk is larger than the current parity disk. 

 

      https://docs.unraid.net/legacy/FAQ/parity-swap-procedure/

 

That should get the present 14TB parity installed as a data disk and the new 18Tb drive installed as the parity disk.

 

For shrinking the array, there is this section in the manual:

 

     https://docs.unraid.net/legacy/FAQ/shrink-array/

 

I don't why you want to do this array shrinking.  Just use those 4Tb drives until they fail or you need more storage space (a more likely scenario).

 

Now for a bit of my personal advice.  Before you start any major changes involving disk manipulation---  Do a non-correcting parity check first.   This way you always start with a fully protected array.  (You have hear of Mr. Murphy's laws---right?)  Granted, it takes time but that is a lot better than getting started and then finding an parity issue...

 

 

Hi Frank!

 

I'd like to shrink 'cause I'm using a sata multiplier based on micron chip wich made grew too much my idle consumption. 

Here in Italy electricity is high (0,35/0,40  /kWh) so I'd like to go back to a more efficient way of storage.

 

the script to keep parity valid is under The "Clear Drive Then Remove Drive" Method procedure.

 

Thanks for the parity check tip, I will do it.

 

just to update, ad the end, I opted for the solution:

 

 

check parity

new config

take out unwanted hdd

re-write parity

 

i wasn't that confident with the clear script, after reading it needed some changes to operate correctly on 7.0.x

I already saved the script and willing to test as soon as i'll be able to build a second UnRaid server i'll use to test and to improve my CLI understanding of things, and to test out "dangerous" operation that I don't want to test on the production server

 

 

Edited May 5, 20251 yr by playmake