NFS Optimization: Difference between revisions

From DrewWiki
Jump to navigation Jump to search
No edit summary
No edit summary
Line 14: Line 14:
</pre>
</pre>


<pre>
echo 0 > /proc/sys/net/ipv4/tcp_sack
echo 0 > /proc/sys/net/ipv4/tcp_sack
echo 0 > /proc/sys/net/ipv4/tcp_timestamps  
echo 0 > /proc/sys/net/ipv4/tcp_timestamps  
</pre>


====TCP Segmentation offload====
====TCP Segmentation offload====

Revision as of 20:21, 5 April 2009

Server Side

/etc/exports

/mnt/raid5 192.168.15.142/32(rw,async,no_root_squash)

async - dramatic throughput increase, but dangerous if a client does not unmount cleanly..

Tuning /etc/sysctl.conf

      net.core.rmem_default = 262144
      net.core.rmem_max = 262144
      #
      # Increase the fragmented packet queue length
      net.ipv4.ipfrag_high_thresh = 524288
      net.ipv4.ipfrag_low_thresh = 393216
echo 0 > /proc/sys/net/ipv4/tcp_sack
echo 0 > /proc/sys/net/ipv4/tcp_timestamps 

TCP Segmentation offload

This will take off some of the tcp overhead if your card supports it..

# ethtool -K ethN tso on

Client Side

/etc/fstab
192.168.15.20:/mnt/raid5 /mnt/raid5 nfs rsize=32768,wsize=32768,intr,hard 0 0

[rw]size=32768 - NFSv3 maximum read write size intr - if the mount drops, you'll still be able to ^C out of whatever operation your running hard - hm?