showmount

[Stefan Schulz]

When i use showmount serverA i get a list with a lot different machines. Some of this machines are long gone.

Although the server has recently rebooted showmount shows those old entrys. How do i cleanup this list ?

[CHRIS_ANORUO]

Remove the client entries in /etc/rmtab and reboot the server. You can then do umount -a or reboot the server.

[Stefan Schulz]

I was afraid you'd say to reboot the server. I forgot to mention that i already edited the /etc/rmtab but this didn't have any effect (as you already know).

Rebooting is no option for the next weeks as this is a production system.

After reading the manpages again i couldnt find any disadvantages in this outdated list. Am i right? Is this list just for informational purposes or has this some meaning regarding performance?

But if this list is not reliable, of what use is it then?

Thanks for your reply.

[Rainer_1 · Accepted solution]

as long as rpc.mountd runs the showmount will not change it's output even if /etc/rmtab is changed.
To get a new fresh showmount output you have to edit /etc/rmtab and then stop/restart rpc.mountd

Find the pid with

ps -ef | grep rpc.mountd

Then kill the pid an restart rpc.mountd with

/usr/sbin/rpc.mountd

(i've tested it on a running system and the established NFS connections were not affected)

[Dan Hetzel]

Hi Stefan,

Rainer is right! As rpc.mountd is the mount request server, it won't break your present NFS mounts.
You can safely issue that command on a production box as long as you restart it immediately.

After the restart, showmount should give you the right info.

Best regards,

Dan

[Stefan Schulz]

Dan, Rainer thanks for your help. I have been afraid of killing the mountd, but with your assurence that this is ok i solved this issue.

I'm still not sure if this was a problem or just cosmetics but it is solved now.

Thanks a lot.

Stefan

[Dan Hetzel]

Glad it helped !

Dan