Date: Wed, 20 Nov 2024 17:22:21 +0000 From: void <void@f-m.fm> To: freebsd-fs <freebsd-fs@freebsd.org> Subject: Re: improving nfs client & server performance Message-ID: <7c108c57-e005-47d4-b9e2-947502d951da@app.fastmail.com> In-Reply-To: <CABXB=RQuJana-GHCszZdWZpuw5CfVxFeVHFugmG9-cbhEG6R_w@mail.gmail.com> References: <ZxZbPmv_11tS5pxZ@int21h> <CAM5tNy4scNutJXdOL=UmK_NhObcfbwpnUpL1dFqe3JVeJVWvcQ@mail.gmail.com> <Zxai_n-LhohogSEY@int21h> <CAM5tNy5j6RknwbvcXjCDs02r1JEO4_De3mY8JHrxYNb0nKhNrw@mail.gmail.com> <ZyAhC8GGA5zEbK4P@int21h> <CABXB=RQuJana-GHCszZdWZpuw5CfVxFeVHFugmG9-cbhEG6R_w@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi, On Fri, 15 Nov 2024, at 16:30, J David wrote: > Perhaps I'm overstating the case, but I believe that using ZFS on SMR > disks is strongly discouraged. I haven't tried myself, mainly due to > the horror stories I've read. Stories that sound a lot like yours. I've heard about SMR drives too [0]. But when I was offered one at a Great Price [1] I installed it. I'll admit i didn't hear about SMR+ZFS combination being inadvisable, though. it *was* working ok for the purpose I allocated it - namely, more like a tape drive than a hard drive, in terms of write speed. But read speed *was* acceptable. Ish. It's this which has deteriorated. Everything that was on that drive that was important is now on other non-SMR based hardware or backed up offsite. [0] in the context of archiving not-often-accessed data [1] a gift [2] theoretically, with eg copies=3 it should have been self-healing errors [3] a failure mode of slowing down is preferable to completely giving up. Now, with all the data off it, it'll be junked as too slow for purpose.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?7c108c57-e005-47d4-b9e2-947502d951da>