Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


Plex 2x2000GB JBOD. ZFS vs Ext4?
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

Plex 2x2000GB JBOD. ZFS vs Ext4?

stefemanstefeman Member
edited April 2017 in Help

As title says, would I benefit using ZFS over ext4 in performance?

System specs:

E3 1245v5
32GB DDR4 2133MHz RAM
HDD SATA disks

Comments

  • JustAMacUserJustAMacUser Member
    edited April 2017

    ZFS tends to want a lot of memory. I use it over every other file system when I can because it's quite resilient and self-healing. It works best in some form of RAID configuration (mirrored in this case since you have two disks). I have a lot of media on my Plex server and I have zero desire to re-ripped the original discs so I run four disks in RAIDZ2 using ZFS.

    I've never personally benchmarked the difference, but a quick Google search seems to suggest that ZFS has better performance.

    If you're not planning to run this in a mirror you'll lose redundency but you'll still have some benefits with ZFS since it has built-in checksums and you can "scrub" the disks regularly to find and correct some errors.

    ZFS supports snapshots, which has saved me a few times. Also has separate datasets, which kind of act like partitions but are a bit different.

    As far as I know (never experienced this myself)... Even in RAID configurations ZFS still maintains whatever data it can if the volume ends up going into a faulted state. So if you lose two out of three disks, I've read that you'll still have the data on that third disk available, which is better than nothing and better than other forms of redundency where a faulted state means the whole thing is lost.

  • WSSWSS Member

    I'm going to suggest HAMMER. ZFS is absolutely awesome; ext4 is for scrubs.

Sign In or Register to comment.