Date: Sun, 17 Apr 2005 23:06:07 +0200 From: Willem Jan Withagen <wjw@withagen.nl> To: Kris Kennaway <kris@obsecurity.org>, stable@freebsd.org Subject: NFS defaults for read/write blocksize....(Was: Re: 5.4/amd64 console hang) Message-ID: <4262CFBF.4090709@withagen.nl> In-Reply-To: <20050416183755.GB61170@xor.obsecurity.org> References: <6eb82e05041500274172afd3@mail.gmail.com> <20050416122222.GA12385@totem.fix.no> <6eb82e0504160536572e068c@mail.gmail.com> <20050416183755.GB61170@xor.obsecurity.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Kris Kennaway wrote: >>By the way, I'm thinking that more frequently hang might related with >>large read/write block in mount_nfs -r/-w (I use 8192, original is 1024). > > > That's certainly possible since non-default settings don't get as much > testing. It would be good to get a traceback. Has it even been considered to up these values to something bigger?? Reason I ask, is since some discussions from a year ago tempted me to do some (non-scientific) NFS performance testing. (http://witahgen.dyndns.org/FreeBSD/nfs-performance/) Ever since then I'm of the opinion that sizes of read/write should be at least 8K, or bigger depending on the data. And yes, I'm aware that bonnie can hardly be considered a serious benchmark tool. But I would think that the results would warant at least this change. I no longer have some of the hardware, so redoing this test with something like 'make buildworld' is not really possible. But I'm looking into getting again a series of older boxes to run some more tests on. --WjW
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4262CFBF.4090709>