Date: Thu, 19 Jul 2007 12:53:55 -0700 From: Steven Schlansker <stevenschlansker@berkeley.edu> Cc: freebsd-current@freebsd.org Subject: Re: Strange performance characteristics with ZFS Message-ID: <469FC153.4010903@berkeley.edu> In-Reply-To: <1c5c32890707170312r71f3735at958325a08db0f1c3@mail.gmail.com> References: <2978EDA9-D393-434C-B734-2DE188631761@berkeley.edu> <b41c75520707161316q5a6f9940q40e7281e1baffbfc@mail.gmail.com> <469C3900.5020703@berkeley.edu> <1c5c32890707162045u9d56cfeq2f7430ddddd55418@mail.gmail.com> <469C48F2.7000302@berkeley.edu> <1c5c32890707170312r71f3735at958325a08db0f1c3@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Brian Donnell wrote: > That was my fault, I misread what you had said your configuration was. > The problem I ran into was the same as you're seeing with the system > waiting on ZFS causing the NFS client to hang and everything becoming > unresponsive. The ls script should be running on the ZFS machine and > executing ls on the ZFS directory. If you have a smaller file to try > with first I'd recommend it. 30GB of corrupt data is just pain. To be > honest, I gave up on NFS with ZFS and started using Samba for everything > after compiling samba3 without a couple functions (check the ZFS vs > Samba debugging results thread from earlier this month) as it actually > maxes out a 100Mbit network link for the transfer. That's something I > could never get NFS with ZFS to come close to doing. > > Just to be sure, since you're using nfsd, that means you have the > sharenfs option of the zfs pool turned off? > > -- Brian > > No, I have it set up like this: [steven@universe ~]$ sudo zfs get sharenfs NAME PROPERTY VALUE SOURCE universe sharenfs -mapall=steven local universe/backup sharenfs -mapall=steven inherited from universe universe/backup/engineer sharenfs -mapall=steven inherited from universe universe/backup/zoo sharenfs -mapall=steven inherited from universe universe/data sharenfs -mapall=steven inherited from universe universe/incoming sharenfs -mapall=steven inherited from universe universe/kabamba sharenfs -mapall=steven inherited from universe universe/media sharenfs -mapall=steven inherited from universe universe/music sharenfs -mapall=steven inherited from universe universe/platypus sharenfs -maproot=root local So here's the score: Write to ZFS on local machine (dd if=/dev/zero of=file) = 36.2 MB/s This still seems a bit slow, but it's not bad so I'll let it slide :-p dd/netcat across network = 24.8 MB/s Also seems a bit slow, considering it's gigabit. Write to ZFS over nfs (dd again) = 16MB/s It does the whole ping-ponging speed (goes at like 25MB/s for a little bit, drops to 0, back to 25, etc) Any more ideas? Thanks! Steven
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?469FC153.4010903>