Date: Thu, 31 Aug 2017 15:57:14 +0100 From: Rafal Lukawiecki <raf@rafal.net> To: Frank Leonhardt <frank2@fjl.co.uk> Cc: freebsd-questions@freebsd.org Subject: Re: Does NFS client cache? Message-ID: <FE17480E-ED05-412A-9C63-74EC4F32BF23@rafal.net> In-Reply-To: <59A81E95.1060300@fjl.co.uk> References: <206BE5FE-7A80-4CCA-8107-F3BBD3FC00FA@rafal.net> <59A81E95.1060300@fjl.co.uk>
next in thread | previous in thread | raw e-mail | index | archive | help
Thank you, Frank. I will benchmark performance first in the hope that = FreeBSD NFS client is somehow faster than what I have experienced with = Linuxes. Otherwise the copy to local strategy makes a lot of sense. = It=E2=80=99s just so much easier to think and manage with a caching = system=E2=80=A6 Good luck with AWS. And thank you, for helping me understand FreeBSD. Rafal -- Rafal Lukawiecki Data Scientist=20 Project Botticelli Ltd > On 31 Aug 2017, at 15:35, Frank Leonhardt <frank2@fjl.co.uk> wrote: >=20 > Sometimes :-) NFS caching is a problem, especially when you consider = the effects of fsync() and being POSIX compliant in this respect. If you = cache a file, even R/O, then an fsync() operation on one host must = invalidate the cache on ALL other hosts before returning. Strictly = speaking. In practice there are a lot of kernel tuneables to ignore this = (some undocumented IME). You can also comment out the code that deals = with the fsync() requests. >=20 > Best strategy depends on exactly what you're trying to do. I'd = probably copy files to local storage as a cron job as an early option. >=20 > (Thanks for the info on AWS) >=20 > Regards, Frank. >=20 > _______________________________________________ > freebsd-questions@freebsd.org mailing list > https://lists.freebsd.org/mailman/listinfo/freebsd-questions > To unsubscribe, send any mail to = "freebsd-questions-unsubscribe@freebsd.org"
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?FE17480E-ED05-412A-9C63-74EC4F32BF23>