kernel hanging problem - nfs?...
jason at rtfmconsult.com
Wed Jan 8 22:53:01 CST 2003
On Wed, 8 Jan 2003, nathan r. hruby wrote:
> > that they recommend increasing the socket read/write buffer size when
> > increasing the number of nfs daemons. Did you do that on your
> > server(s)?
yes, i believe so. as per a redhat recommendation for heavy use servers:
echo 262143 > /proc/sys/net/core/rmem_default
echo 262143 > /proc/sys/net/core/rmem_max
echo 262143 > /proc/sys/net/core/wmem_default
echo 262143 > /proc/sys/net/core/wmem_max
daemon rpc.nfsd $RPCNFSDCOUNT
echo 65536 > /proc/sys/net/core/rmem_default
echo 131072 > /proc/sys/net/core/rmem_max
echo 65536 > /proc/sys/net/core/wmem_default
echo 131072 > /proc/sys/net/core/wmem_max
> that's so friggin' odd. I've been using the 2.4.19 kernel for a while as
> a NFS server with wr=8192 and it's just been super happy. jason, are you
> using the Redhat kernels or vanilla kernels?
i am using the redhat kernels.
it hasn't really mattered as i have had this issue on every kernel since
redhat 7.0 (2.2.X) and i have never been able to successfully use a rsize
of > 1024. i am now using rsize=1024 for most production things and
using tcp with a rsize of 4096 for servers that are more lightly loaded.
More information about the Linux-PowerEdge