Information related tp Advanced Partitioning

[quaint bronze]

Hi,

I understand the point of doing root-data partitions to avoid using entirely a disk for the system.
But what is the main purpose of doing root-data-data over root-data ?

What are the main differences between both?

Regards,

[grizzled raft]

Better utilization of the media.

SSD/NVMe media are solid state, not spinning. They can be split across two nodes for data aggregates. The RAID groups can be larger with solid state and therefore instead of using two parity drives per node, you use two parity drive partitions per node (effectively two drives instead of four)

Parity tax is reduced

[brave pasture]

It's not something you can decide/select. HDDs always use Root-Data partitioning, while SSDs always use Root-Data-Data partitioning.
The idea is that 2 nodes accessing the same HDD at the same time is a bad idea (the r/w head needs to move physically between the two partitions each time), while for SSDs it is no issue at all (reading/writing is independent from the "actual location" of the bits)

[quaint bronze]

apparently this depends from the documentation. if you have an aggregate with both HDD and SSD then adpv2 is not supported

and for the 2nd line it makes sense ! thanks for your answer

thanks both for your answers, I see now better the reasons

[brave pasture]

yes because that is a FlashPool Aggregate which is technically still an HDD aggregate 🙂

[quaint bronze]

with SSD caching data so