From: Matthew Dillon Date: Mon, 22 Feb 2010 16:23:04 +0000 (-0800) Subject: swapcache - documentation update (data caching, read/write concurrency) X-Git-Tag: v2.7.0~141 X-Git-Url: https://gitweb.dragonflybsd.org/dragonfly.git/commitdiff_plain/2dc854bc5f54a4874701b2189fb6eda9a3d77076 swapcache - documentation update (data caching, read/write concurrency) * Update the manual page to expand on issues when caching bulk data, including read/write concurrency issues with SSDs. --- diff --git a/share/man/man8/swapcache.8 b/share/man/man8/swapcache.8 index 92d63d9ec8..b7145ecc4d 100644 --- a/share/man/man8/swapcache.8 +++ b/share/man/man8/swapcache.8 @@ -153,19 +153,26 @@ memory (KVM) and cannot reliably support more than around 100,000 active vnodes. 64-bit kernels can support 300,000+ active vnodes. .Pp -Data caching is definitely more wasteful of SSD write bandwidth -than meta-data caching. It doesn't hurt performance per se, -but may cause the -.Nm -to exhaust its burst and smack against the long term average -bandwidth limit, causing the SSD to wear out at the maximum rate you -programmed. Data caching is far less wasteful and more efficient +Data caching is definitely more wasteful of the SSD's write durability +than meta-data caching. +The swapcache may exhaust its burst and smack against the long term +average bandwidth limit, causing the SSD to wear out at the maximum rate +you programmed. Data caching is far less wasteful and more efficient if (on a 64-bit system only) you provide a sufficiently large SSD and increase .Cd kern.maxvnodes to cover the entire directory topology being served. Each vnode requires about 1K of physical ram. .Pp +Due to the higher SSD write rate you may want to use a +medium-sized SSD with good write performance to reduce interference +between reading and writing. +Write durability also scales with larger SSDs. +For example, an Intel X25-V only has 40MB/s in write performance +and burst writing by swapcache will seriously interfere with +concurrent read operation on the SSD. The 80G X25-M on the +otherhand has double the write performance. +.Pp When data caching is turned on you generally want to use .Xr chflags 1 with the