Hey, welcome to the forums!
dRAID vdevs cannot be expanded in any form after creation, the only option would be to create a new vdev, preferably with the same number of drives and dRAID layout as the first vdev, and add it to the pool.
dRAID uses precomputed permutation maps to map how data and distributed spare content is distributed across the vdev. Due to this, it is unlikely that dRAID will support expansion as the permutation map would no longer properly map to the drive arrangement. There is currently no work towards enabling dRAID expansion.
2 Likes
No - dRAID is designed for very large storage servers which will typically add extra storage by the 10s or 100s of disks at a time. As such it has several down-sides to compensate for the upside of faster resilvering - including inflexibility.
Thanks so much for the quick and detailed reply!
That makes complete sense — I really appreciate the clarification about dRAID’s internal structure and limitations.
Just to be sure:
Which types of vdevs do support expansion by adding individual drives over time?
I understand dRAID doesn’t, but is RAID-Z1 (or other RAID-Z levels) more flexible in this regard?
Thanks again!
All RAIDz levels can up expanded a single disk at a time, however there are some caveats regarding capacity calculation and efficiency. We have a calculator to give you an idea about efficiency here. And there have been plenty of discussions on the forums regarding this as well. ZFS capacity calculations are hard.
1 Like
And mirrors have always been totally flexible…
… but inefficient in redundancy space usage and so more costly unless you need the IOPS for small random reads (eg. for virtual disks/zVols/iSCSI or database files).
1 Like
Can’t have your cake and eat it, can you? 