Throughput problems with NFS
hmp at backplane.com
Sun Oct 12 06:15:50 PDT 2003
Matthew Dillon wrote:
: Ah, ok now I see. It sounds like soreserve should nominally be 2.5x
: or 3x the NFS io size in order to properly stream requests, or you
: won't be able to fill the wire.
On that note, I would like to bump up NFS_FABLKSIZE to a value
of 4096, from the current value of 512 bytes. The latter is
totally unacceptable given todays network environments ...
I propose this value, as to keep in sorta sync with Linux's
default file block size. I would like to know if Solaris
uses anything bigger than 4096.
I shall commit the change, unless there is any difference
hmp at xxxxxxxxxxxxx
More information about the Submit