Smaller drves in new pool

ok - so ml 350 g6 with 4 x 2TB drives in the external drive bays hooked up to an old lsi SAS raid controller in channel A port

i’ve put some 4 x 500GB drivers in half of the sff drive cage in port B on the LSI card

rockstor recognises the drives - but here’s the thing

how do i use those drives?

the 4 x 2TB drives on channel A are in RAID 10

if i get a second card and plug all 8 slots in the drive cage into the second LSI card
can i configure the 8 x 500GB drives as 2 x 2TB drives and add them to the existing RAID10 pool


@Greg_Simpson Hello again and welcome back by the way.

You can use them whoever you like. They can form a new pool of their own or multiple pools, or you can add them to the existing pool. Btrfs does not require that all pool members are the same size. There will be limitations on what space can be used with this arrangement but given the raid in btrfs is chunk based, as long as, for raid1 or 10, there is space in existing chunks or space for new chunk across the minimum number of devices to mee the raid level, all is good. So to same new data on raid1 there must be 2 drives with enough fee space to save the data, 4 drives for raid10. But it’s any 2 or 4 drives. Plus you can always ReRaid at a later date if that makes sense.

Yes. And you can continue to use the pool while you are at it. The process may take several hours thought as it’s quite a lot of work to move all the data around. Performance of the pool/system will be degraded during the period but it should still work for accessing the data.

Hope that helps.

1 Like

@Greg_Simpson just another note on this.

We have recently updated our docs in the area so it may be worth taking another look if you haven’t seen them for a bit:

Specifically in your case the following subsection:

Hope that helps.

EDIT: You didn’t mention if you were running our newer “Built on openSUSE” variant. That’s quite an important element here as the kernel and btrfs is much improved in our new OS base. I only mention this as we no longer release updates to the older CentOS variant and we failed to keep it’s kernel updated also. But we can now use our new upstream’s kernel which is nice as that was a step too far for us previously. Plus the actively support btrfs which was the core reason for our move from CentOS in the first place.

1 Like