Re:Linux 5.10 BTRFS performance regression
by kefyras on Wednesday December 23, 2020 @10:26AM
(#25518514)
> but ZFS doesn't like living on a partitioned disk nearly as much.
Don't know how true is this, since it's not even possible to create a zpool on the whole unpartitioned device on linux. It automatically creates GPT label with zfs and a small efi partition.
Re:Linux 5.10 BTRFS performance regression
by zaarn on Wednesday December 23, 2020 @10:32AM
(#25518592)
That is true, but I suspect that only happens to protect the zpool against tools like the Windows Repair Install (which likes to attempt repairs on disks without GPT label). Using ZFS on a partition inside a GPT disk tends to lead to a few quirky behaviours, like for example ZFS not always reporting correct SMART data of the device.
Re:Linux 5.10 BTRFS performance regression
by fomine3 on Thursday December 24, 2020 @02:26AM
(#25525695)
IIRC ZoL `zpool create /dev/disk/by-id/ata-XXX` makes gpt partition table and zfs partition inside it.
Re:Linux 5.10 BTRFS performance regression
by zaarn on Thursday December 24, 2020 @06:08AM
(#25526554)
A stub table.
Re:Linux 5.10 BTRFS performance regression
by rleigh on Thursday December 24, 2020 @03:54AM
(#25526021)
It's nonsense. A ZFS pool is created from block devices, and ZFS doesn't care if those are whole discs or partitions.
Every ZFS pool I have is composed of GPT partitions, and this is also recommended in ZFS books as good practice to make it easier to identify a failed drive. Since you see the GPT partition name in the "zpool status" and other tools' output, it's handy.
Re:Linux 5.10 BTRFS performance regression
by projct on Wednesday December 23, 2020 @05:38PM
(#25523104)
On ZFS you can have multiple vdevs of different sizes in the same pool, and you can also set copies=N to have multiple copies of the data available for extra redundancy.
You can't unfortunately tell it that different disks have different speeds.
It also works just fine w/ a small (1GB or so) ARC if you want, just at an obvious performance penalty compared to having a larger cache.
I'm currently on Bcachefs, which has erasure coding with similar promises and works better for me.
ZFS is more rigid, I have to have all matched disks for best performance, plus RAM unless I like bad performance (I don't). This could be solved by striping disks into 1TB partitions, but ZFS doesn't like living on a partitioned disk nearly as much.
Mixed disk has a bunch of cool properties for homelab users. Bcachefs also solves another problem; caching and speed. I could designate my 1TB NVMe as fast and my SATA 2TB SSDs as slow. I could also define that /tmp requiers no Erasure Coding and everything else requires 2 replicas. Then I would be able to combine the performance of my NVMe with the capacity of my SATA SSDs while gaining a simple redundance solution. My NAS has a dozen of mixed disks that I could combine more easily than ever.
The issue with ZFS is not that it's not ready to use. The issue is that it's born from enterprise and requires a costly enterprise setup to run, rather than a cheap homelab setup.