nfsd Post: 302420886

Sponsored Content

Top Forums UNIX for Advanced & Expert Users nfsd Post 302420886 by methyl on Wednesday 12th of May 2010 09:10:22 PM

05-12-2010

Registered User

Consider getting a copy of the HP Tuning and Performance book.

The snapshot from "top" taken out of context doesn't mean a lot.
The percentage shown in "top" for a single process such as "nfsd" (see "man nfsd" to read about the process) is the percentage of one CPU. You have many CPUs.

If you have HP "glance" this is a much better tool for snapshots.

There are many packages for recording historical performance information including unix "sar" and commercial packages from HP. An hour-by-hour view of server performance is more valuable than a snapshot.

The "top" output posted implies a severe shortage of memory and CPU. This is only an implication and we would need better information before making a recommendation.
Whether an apparent shortage actually matters would need detailed analysis of swap statistics and CPU waits. You will need something better than "top".

One of the many advantages of unix over rival Operating Systems is that a correctly tuned server can perform well when theoretically overloaded.

This User Gave Thanks to methyl For This Post:

methyl

View Public Profile for methyl

Find all posts by methyl

6 More Discussions You Might Find Interesting

1. Solaris

nfsd won't start at boot up

Hi Inexplicably, nfsd no longer starts automatically on our Sun boxes running Solaris 9, so that 'automount' no longer functions automatically. The problem first manifested itself when we could not access files on any of the nfs automounted directories in our LAN after one of the servers (say...

2. UNIX for Dummies Questions & Answers

don't have nfsd mount point in /proc/fs/nfsd

hi guys I installed NFS server and everything started out fine but I don't have /proc/fs/nfsd entry and so I can't mount nfsd. Therefore I can't start my nfs service. Why don't I have /proc/fs/nfsd? How do I create that? Thanks

3. AIX

portmap and nfsd

Hello, what is the relation between portmap and nfsd and how communication between them looks like. Does the nfsclient contact with the portmap or nfsd first. Many thanks in advance for helping me to understand this :) BR, p

4. BSD

NFSD under OpenBSD

Hi all, I am having a following problem. Trying to run PXE boot server on my OpenBSD machine I have ended up on making NFSd daemon works. On all machines I get an error msg. nfsd : nfsd count is invalid: (null) no matter what computer I run it on. Everything works just well on FreeBSD and linux....

5. HP-UX

NFSd command utilizing more cpu

Hi, I see following 'nfsd' command is using more CPU. Could someone please comment on it's pros and cons of it? CPU TTY PID USERNAME PRI NI SIZE RES STATE TIME %WCPU %CPU COMMAND 5 ? 16890 root 152 20 34696K 12036K run 57166:48 856.13 854.64 nfsd OS -- HP-UX One...

6. Slackware

NFSd problem

Hi. Using debian 8.0 on a raspberryPI SERVER, accessing nfs from another raspberry gives quick reply. But from a slackware 14.1 SERVER on a Celeron 2Ghz dual core, is painfully slow and i cannot figure out why. Can anyone guide me?

LEARN ABOUT REDHAT

rpc.nfsd

rpc.nfsd(8)						      System Manager's Manual						       rpc.nfsd(8)

NAME

       rpc.nfsd - NFS server process

SYNOPSIS

       /usr/sbin/rpc.nfsd [-p port] nproc

DESCRIPTION

       The  rpc.nfsd program implements the user level part of the NFS service. The main functionality is handled by the nfsd.o kernel module; the
       user space program merely starts the specified number of kernel threads.

       The rpc.mountd server provides an ancially service needed to satisfy mount requests by NFS clients.

OPTIONS

       -p port
	      specify a diferent port to listen on for NFS requests. By default, rpc.nfsd will listen on port 2049.

       nproc  specify the number of NFS server threads. By default, just one thread is started. However, for optimum performance  several  threads
	      should be used. The actual figure depends on the number of and the work load created by the NFS clients, but a useful starting point
	      is 8 threads. Effects of modifying that number can be checked using the nfsstat(8) program.

SEE ALSO

       rpc.mountd(8), exportfs(8), rpc.rquotad(8), nfsstat(8).

AUTHOR

       Olaf Kirch, Bill Hawes, H. J. Lu, G. Allan Morris III, and a host of others.

								    31 May 1999 						       rpc.nfsd(8)