Everything ZFS

319 readers

1 users here now

A community for the ZFS filesystem.

ZFS is an opensource COW filesystem used by enterprise and serious homelabbers for it's data safety and extensive feature set.

OpenZFS is the active branch now developed primarily for Linux with a port to it's FreeBSD roots.

This community is here to answer questions and discuss topics related to the use of ZFS in the wild.

Rules:

As always, the main rule is Don't Be a Dick. Be polite with new users asking questions that you may consider obvious. If you don't have something constructive to offer, downvote and move on.

No dirty deletes: your posts are here for posterity, perhaps the next person will get something out of it, even if it's wrong.

founded 2 years ago

MODERATORS

ikidd@lemmy.world

Needed recommendations for zpool planning (lemmy.world)

submitted 1 month ago by poinck@lemmy.world to c/zfs@lemmy.world

4 comments fedilink hide all child comments

Hi, I have an old Dell server I want to repurpose as a storage server and I decided to use zfs. I have read that gradual increases are best done with mirrored vdevs and only if I want to max out, I can choose Z2. What are your experiences in practice?

I like to go the gradual increase route because I cannot afford to maximize from the beginning. I like to start with ~10TB usable space. The server has at least 8 free drive bays. Should I go with a mirror as my first vdev or go 6x3TB with Z2? Ideally I want to be able to add more vdevs to the same pool in the future.

I am also curious whether it is a bad idea to have an uneven pool with mixed vdev types, the first being mirror and the second Z2. Will the vdevs eventually even out over time because zfs is a CoW filesystem?

you are viewing a single comment's thread
view the rest of the comments

[–] poinck@lemmy.world 1 points 1 month ago* (last edited 1 month ago) (1 children)

Thx for your answer. I think, capacity is currently more important than speed. The idea of starting with a 3-drive-z2 sounds interesting but will cost more than a mirror with the same capacity. Maybe, I start with a mirror and wait how my users will fill it up to decide whether the next increase will be a z2-vdev with 3 drives, a second mirror or even a z2 with 6 drives.

But I see the point of just using one z2-vdev and adding new drives as needed. I would have more capacity in the end, but all drives need to have the same capacity. Or, if I start with a 10 TB mirrored vdev and add a second 20 TB mirror after 1 year, and 40 TB mirror after 2 more years or replace the first mirror with larger drives after many more years.

I fear, what likely will happen is that I will start with a single disk vdev, because my users don't want to pay for redundancy. And only after some years, they see the value of it.

[–] bob_zim@infosec.exchange 2 points 1 month ago

@poinck Once you add a raidz vdev to a pool, you can’t remove the mirror. You’d be stuck with the mismatch until the whole pool is destroyed and rebuilt. This isn’t a technical problem, but it makes maintenance more painful (which two drive bays are the mirror, again?). If you want raidz, you should really start with it to save yourself future pain.

You also can’t increase the fault tolerance of a raidz vdev. That is, you can’t take a raidz1 and add a drive to make it a raidz2.

I personally don’t trust single-fault-tolerant vdevs. Resilvering takes long enough I want double fault tolerance for all spinning disks bigger than maybe 4 TB. That means raidz2 or three-member mirrors.