<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>topic Re: NFS in Operating System - HP-UX</title>
    <link>https://community.hpe.com/t5/operating-system-hp-ux/nfs/m-p/2454892#M770307</link>
    <description>Distance might not be the problem, you should also consider:&lt;BR /&gt;&lt;BR /&gt;client-server ratio, has it increased recently&lt;BR /&gt;what the shared apps are&lt;BR /&gt;what the nfs server spec is and what else is it doing&lt;BR /&gt;what is the network link load/reliability&lt;BR /&gt;&lt;BR /&gt;If the problem is due to distance it might be due to the NFS process.  The client has to ask the server what port mountd is running on it reponds; client then asks mountd for the filehandle for a specific file you've requested; server responds; then with this filehandle the client asks the portmapper what nfsd to use; server responds; client then queries the nfsd with the filehandle and it provides a filehandle for the desired object.  Although the whole process isn't repeated every time, the stateless nature of NFS means a fair amount of network traffic which is obviously exacerbated by the distance in your scenario.&lt;BR /&gt;&lt;BR /&gt;You could upgrade your NFS server&lt;BR /&gt;Create an aditional server to balance out the load&lt;BR /&gt;Add more NFS daemons (nfsd) on the server&lt;BR /&gt;&lt;BR /&gt;Monitor it with nfsstat, especially look at the write rates.  Other fields to watch are:&lt;BR /&gt;&lt;BR /&gt;badcalls&lt;BR /&gt;nullrecv&lt;BR /&gt;badlen&lt;BR /&gt;retrans&lt;BR /&gt;badxid&lt;BR /&gt;timeout&lt;BR /&gt;&lt;BR /&gt;A nonzero in any of these fields could mean trouble.&lt;BR /&gt;&lt;BR /&gt;Hope this helps, Paul</description>
    <pubDate>Wed, 18 Oct 2000 11:03:30 GMT</pubDate>
    <dc:creator>Paul McCleary</dc:creator>
    <dc:date>2000-10-18T11:03:30Z</dc:date>
    <item>
      <title>NFS</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/nfs/m-p/2454889#M770304</link>
      <description>Why does a large distance(200 miles) between an NFS server and client cause so many performance problems ? &lt;BR /&gt;</description>
      <pubDate>Wed, 18 Oct 2000 09:37:57 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/nfs/m-p/2454889#M770304</guid>
      <dc:creator>Tony Lowrie</dc:creator>
      <dc:date>2000-10-18T09:37:57Z</dc:date>
    </item>
    <item>
      <title>Re: NFS</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/nfs/m-p/2454890#M770305</link>
      <description>The distance should not affect the NFS performance, it is the nature/quality of your link between the servers that you have to put into cosiderations.</description>
      <pubDate>Wed, 18 Oct 2000 10:43:59 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/nfs/m-p/2454890#M770305</guid>
      <dc:creator>CHRIS_ANORUO</dc:creator>
      <dc:date>2000-10-18T10:43:59Z</dc:date>
    </item>
    <item>
      <title>Re: NFS</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/nfs/m-p/2454891#M770306</link>
      <description>Hi,&lt;BR /&gt;Without more information about your particular environment the standard answer would have to be that there must be some WAN links involved over a distance of 200 miles and that the performance of NFS over this distance will be impacted most by the slowest WAN link in place betweens the sites involved. Traffic prioritisation may be a factor as well, i.e. routers may have been configured to place a lower priority on NFS traffic in relation to more interactive or OLTP related traffic, eg telnet etc.&lt;BR /&gt;&lt;BR /&gt;Maybe NFS is not the best method for achieving you data transfer requirements.&lt;BR /&gt;&lt;BR /&gt;Regards, Trevor</description>
      <pubDate>Wed, 18 Oct 2000 10:46:07 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/nfs/m-p/2454891#M770306</guid>
      <dc:creator>Trevor Dyson</dc:creator>
      <dc:date>2000-10-18T10:46:07Z</dc:date>
    </item>
    <item>
      <title>Re: NFS</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/nfs/m-p/2454892#M770307</link>
      <description>Distance might not be the problem, you should also consider:&lt;BR /&gt;&lt;BR /&gt;client-server ratio, has it increased recently&lt;BR /&gt;what the shared apps are&lt;BR /&gt;what the nfs server spec is and what else is it doing&lt;BR /&gt;what is the network link load/reliability&lt;BR /&gt;&lt;BR /&gt;If the problem is due to distance it might be due to the NFS process.  The client has to ask the server what port mountd is running on it reponds; client then asks mountd for the filehandle for a specific file you've requested; server responds; then with this filehandle the client asks the portmapper what nfsd to use; server responds; client then queries the nfsd with the filehandle and it provides a filehandle for the desired object.  Although the whole process isn't repeated every time, the stateless nature of NFS means a fair amount of network traffic which is obviously exacerbated by the distance in your scenario.&lt;BR /&gt;&lt;BR /&gt;You could upgrade your NFS server&lt;BR /&gt;Create an aditional server to balance out the load&lt;BR /&gt;Add more NFS daemons (nfsd) on the server&lt;BR /&gt;&lt;BR /&gt;Monitor it with nfsstat, especially look at the write rates.  Other fields to watch are:&lt;BR /&gt;&lt;BR /&gt;badcalls&lt;BR /&gt;nullrecv&lt;BR /&gt;badlen&lt;BR /&gt;retrans&lt;BR /&gt;badxid&lt;BR /&gt;timeout&lt;BR /&gt;&lt;BR /&gt;A nonzero in any of these fields could mean trouble.&lt;BR /&gt;&lt;BR /&gt;Hope this helps, Paul</description>
      <pubDate>Wed, 18 Oct 2000 11:03:30 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/nfs/m-p/2454892#M770307</guid>
      <dc:creator>Paul McCleary</dc:creator>
      <dc:date>2000-10-18T11:03:30Z</dc:date>
    </item>
    <item>
      <title>Re: NFS</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/nfs/m-p/2454893#M770308</link>
      <description>The WAN speed is very important.  I've seen a 10ms latency going 3000 miles on an ATM line with new equipment, and I've seen a 200ms latency going across town with some older stuff.  The distance is not important in of itself.  Traffic on your WAN connection can slow these figures down significantly when things get busy.  The bottom line is that NFS can be used successfully over a WAN, but requires that some network engineering be performed to make sure you get the performance you are looking for.</description>
      <pubDate>Wed, 18 Oct 2000 11:11:39 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/nfs/m-p/2454893#M770308</guid>
      <dc:creator>Tim Malnati</dc:creator>
      <dc:date>2000-10-18T11:11:39Z</dc:date>
    </item>
    <item>
      <title>Re: NFS</title>
      <link>https://community.hpe.com/t5/operating-system-hp-ux/nfs/m-p/2454894#M770309</link>
      <description>Part of the problem may be that NFS uses UDP by default, which may lead to premature timeouts and poor performance.  Patch #PHNE_20344 was specifically designed to allow the use of NFS over TCP to improve performance over WANs.  After installing the patch, look at the release notes in &lt;BR /&gt;/usr/share/doc/RelNotesNFS_TCP.txt for more information.&lt;BR /&gt;&lt;BR /&gt;Darren</description>
      <pubDate>Wed, 18 Oct 2000 18:26:36 GMT</pubDate>
      <guid>https://community.hpe.com/t5/operating-system-hp-ux/nfs/m-p/2454894#M770309</guid>
      <dc:creator>Darren Miller</dc:creator>
      <dc:date>2000-10-18T18:26:36Z</dc:date>
    </item>
  </channel>
</rss>

