Still unfinished and is not the same as MD As I can see, old data still keep the old parity, with MD all data are rewritten thus making use of all disks for performance reason
If you add 2 disks to a 4 disks raid6, you'll get 6 disks IOPS when reading, with zfs you are still limited to 4 disks IOPS for old data AFAIK the only real reason to use ZFS is for bitrot and incremental backups using shapshots. Everything else can also be made by using MD plus some additional software (in pure Unix style: one software to do one thing. ZFS it's the opposite :one software for all: raid, filesystem, snapshots, volume manager, backups,....) Il 25 feb 2018 5:37 PM, "Gilberto Nunes" <gilberto.nune...@gmail.com> ha scritto: > Take a look at > > https://twitter.com/OpenZFS/status/921042446275944448 > > > > Em dom, 25 de fev de 2018 13:36, Gandalf Corvotempesta < > gandalf.corvotempe...@gmail.com> escreveu: > > > Planning to add doesn't mean that it was added > > If I remember properly, no a single line of code was written up to > > today.... > > > > Il 25 feb 2018 5:24 PM, "Michael Bellerue" <michael.belle...@gmail.com> > ha > > scritto: > > > > > Online expansion was recently added to OpenZFS for RAID-Z. > > > > > > https://www.freebsdfoundation.org/blog/openzfs-raid-z- > > > online-expansion-project-announcement/ > > > > > > > > > On Sun, Feb 25, 2018 at 6:47 AM, Gandalf Corvotempesta < > > > gandalf.corvotempe...@gmail.com> wrote: > > > > > > > I do that quite often. > > > > Starting with a small number of expansive disks (like SSDs or NVMEs) > > and > > > > being able to expand on the fly when needed, allows me to keep > running > > > cost > > > > as low as possible > > > > > > > > With zfs this is not possible (in a simple way) thus you have to > start > > > with > > > > the proper number of disks that you plan to use for some following > > years > > > > > > > > Il 25 feb 2018 1:42 PM, "Andreas Steinel" <a.stei...@gmail.com> ha > > > > scritto: > > > > > > > > > On Sat, Feb 24, 2018 at 11:16 AM, Gandalf Corvotempesta > > > > > <gandalf.corvotempe...@gmail.com> wrote: > > > > > > In example, you can add/remove single disks > > > > > > You can change from any raid level to any raid level on the fly > > with > > > no > > > > > > downtime or loss of redundancy > > > > > > You can grow /shrink a volume > > > > > > > > > > Good points, especially the shrink that, that does indeed currently > > > > > not work in ZFS. Yet I have to say, I'd never added a single disk > in > > > > > an enterprise setup nor changed a RAID level on the fly. How often > > did > > > > > you do the latter? I just add at least one shelf to existing > storage > > > > > boxes. > > > > > _______________________________________________ > > > > > pve-devel mailing list > > > > > pve-devel@pve.proxmox.com > > > > > https://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel > > > > > > > > > _______________________________________________ > > > > pve-devel mailing list > > > > pve-devel@pve.proxmox.com > > > > https://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel > > > > > > > > > > > > > > > > -- > > > Michael Bellerue > > > michael.belle...@gmail.com > > > 651.900.3068 > > > _______________________________________________ > > > pve-devel mailing list > > > pve-devel@pve.proxmox.com > > > https://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel > > > > > _______________________________________________ > > pve-devel mailing list > > pve-devel@pve.proxmox.com > > https://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel > > > _______________________________________________ > pve-devel mailing list > pve-devel@pve.proxmox.com > https://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel > _______________________________________________ pve-devel mailing list pve-devel@pve.proxmox.com https://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
