Recommended NFS mount options for various UNIX hosts

Recommended NFS mount options for various UNIX hosts

Note: mount options "forcedirectIO" and "noac" are only recommended on
databases.

Linux:
rw,bg,hard,nointr,rsize=32768,wsize=32768,tcp,actimeo=0,vers=3

Solaris:
rw,bg,hard,nointr,rsize=32768,wsize=32768,proto=tcp,forcedirectio,noac,vers=3

AIX:
cio,rw,bg,hard,intr,proto=tcp,vers=3,rsize=32768,wsize=32768

HPUX (11.31 or later):
rw,bg,hard,intr,rsize=32768,wsize=32768,timeo=600,noac,forcedirectio 0 0

Recommended test to collect NFS statistics


nfsstat -z (zero the NFS statistics at the client)
netstat -I (network stats before the tests at the client)
mount -o rsize=32768,wsize=32768 filerX:/vol/vol2/home /mnt/nfstest
cd /mnt/nfstest
nfsstat -m (output of the mountpoints and the mount flags)
time mkfile 1g test (write test)
time dd if=/mnt/nfstest/test /tmp/test (read test)
time cp test test1 (read and write test)
nfsstat -c (verify nfsstat output)


Check the nfsstat output against re-transmissions, timeouts and bad calls

• timeout>5%. Requests timing out before the server can answer them
• badxid~timeout. Server slow. Check nfsstat -m.
• badxid~0 and timeouts > 3%. Packets lost in the network; check netstat. If this number is the same as bad calls, the network is congested.
• retrans. May indicate network of routing problem if retransmit >5%.
• null>0. Automounter timing out. Increase the timeout parameter on the automounter configuration.

In the output of the "nfsstat -m`' command, the following parameters are critical

  • srtt. Smoothed round-trip time.
  • dev. Estimated deviation.

No comments:

Post a Comment