Dont use the full disk during pool/vdev creation

Problem/Justification
SSD getting rejected as replacement for being too small. Turns out 1TB label in case of SSDs is not universal in terms of actual storage space.

Impact
The only negative i can see ATM is a loss of storage space, maybe it could be an optional/opt-out thing during pool creation. But it would help greatly if we do not have to always dig around if the ssd we want to buy as replacement is the correct size or not.

User Story
As i said before it could be a webui option (either optional or opt-out, idk which would be better) and could be used to improve compatibility for the pool avoiding situations of confusion when the user gets a 1 tb drive just to get a cold shower when zfs rejects it.

Isn’t that what the 2 GiB partition is for? (As the first partition / beginning of the device.)

Previously it was used for swap creation and to buffer a small amount of room to prevent issues that you are experiencing.

Now with the latest iteration of SCALE, I thought they still default to this 2 GiB buffer space, even in the absence of any swap activation?

There is no swap partition on the SSDs, some of them dont even have a partition just the bare device.

Yeah, there’s a bug in SCALE (again).

It shouldn’t be doing that.

One issue (RAM / ARC / Swap) was fixed, and I think this new issue came out of that, unintentionally.

1 Like

Well bug or not a 2g partition wouldnt help when one 1TB drive is 931,51GB the other on the other hand is 953,87…
(And so far what i have seen there is no clear indication before buying which case the drive falls into.)
In my case i have a mixed pool because of this, it just so happened in the case that made me make this request the failed drive is the bigger one and the replacement unfortunately the smaller.

I believe it’s not a bug but a deliberate choice.