Rahul
2009-09-24 20:20:36 UTC
I ran "nfsstat -s" on our NFS server and get this result quoted below. I am
not much experienced with this so wanted any comments that fellow users
might have. Any red flags? Any opportunities for tuning?
[Apologies for the cross-post; I posted before on comp.os.linux.misc but
that is probably too generic a group for this question]
I had read that " for writes > 10%: Write caching (either array-based or
host-based, such as a Prestoserv card) would speed up operation." True?
Any other such rules of thumb?
------------------
Server rpc stats:
calls badcalls badauth badclnt xdrcall
349230699 0 0 0 0
Server nfs v3:
null getattr setattr lookup access readlink
26 0% 4715292 1% 232194 0% 179819 0% 2321884 0% 461
0%
read write create mkdir symlink mknod
191503923 54% 150048993 42% 23005 0% 35 0% 23 0% 0
0%
remove rmdir rename link readdir
readdirplus
5595 0% 0 0% 125 0% 0 0% 0 0% 2986
0%
fsstat fsinfo pathconf commit
8 0% 43 0% 0 0% 174135 0%
--------------------------------
not much experienced with this so wanted any comments that fellow users
might have. Any red flags? Any opportunities for tuning?
[Apologies for the cross-post; I posted before on comp.os.linux.misc but
that is probably too generic a group for this question]
I had read that " for writes > 10%: Write caching (either array-based or
host-based, such as a Prestoserv card) would speed up operation." True?
Any other such rules of thumb?
------------------
Server rpc stats:
calls badcalls badauth badclnt xdrcall
349230699 0 0 0 0
Server nfs v3:
null getattr setattr lookup access readlink
26 0% 4715292 1% 232194 0% 179819 0% 2321884 0% 461
0%
read write create mkdir symlink mknod
191503923 54% 150048993 42% 23005 0% 35 0% 23 0% 0
0%
remove rmdir rename link readdir
readdirplus
5595 0% 0 0% 125 0% 0 0% 0 0% 2986
0%
fsstat fsinfo pathconf commit
8 0% 43 0% 0 0% 174135 0%
--------------------------------
--
Rahul
Rahul