Date: Sat, 23 Oct 2004 16:10:41 +0200 From: des@des.no (=?iso-8859-1?q?Dag-Erling_Sm=F8rgrav?=) To: Clifton Royston <cliftonr@lava.net> Cc: hackers@freebsd.org Subject: Re: Relative performance of swap-backed MFS vs. regular UFS? Message-ID: <xzpis91h6e6.fsf@dwp.des.no> In-Reply-To: <20041022223238.GA12502@tikitechnologies.com> (Clifton Royston's message of "Fri, 22 Oct 2004 12:32:40 -1000") References: <20041022223238.GA12502@tikitechnologies.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Clifton Royston <cliftonr@lava.net> writes: > For a large temporary file system which must hold short-lived files, > mostly small but occasionally several very large ones (e.g. 100MB+), is > it better for performance and stability if this file system: > > 1) resides on a swap-backed MFS and trusts the OS to swap out > low-priority blocks if needed under RAM pressure, or > > 2) on a regular UFS and trusts the OS to buffer as many blocks as > possible into RAM when RAM is free? the former, provided you have enough RAM. A swap-backed MFS will only swap out when it has to, while a UFS will always write out changes after a while. > I temporarily enlarged it to 256MB which is working, but as I worked > out the worst case scenario, I realized it really would need to be > nearly 1GB to handle multiple zip-bombs each hitting the 100MB size > limit. This makes me wonder if it's wise to specify a 1GB MFS on a > system with only 1GB RAM, or wiser to just revert to a regular file > system? RAM is cheap. Toss in a couple extra gig and set up a 2 GB MFS. DES --=20 Dag-Erling Sm=F8rgrav - des@des.no
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?xzpis91h6e6.fsf>