We are running Red Hat Linux release 6.1, kernel 2.2.12-20smp on i686. The rpc.mountd dameon begins to use up our system's memory until it is beyond 100MB of ram used. Usually at this point the external machines trying to mount the machine are unable to connect. Our temporary fix was to kill the daemon every 3 hours and restart it. The memory still grows but the kill/restart allows for the remote machines to mount and umount cleanly. Some of test machines go up and down 10-15 times and hour. It seems that the rpc.mountd is not cleaning up memory properly. Thanks Bill Snyder Software Engineer Intel Corporation
We have observed similar behaviour with our Red Hat 6.2 LAN (using the out-of-the-box kernel-2.2.14-5.0), we have a i686 with SMP on 550 MHz PIII.
assigned to johnsonm
Unable to reproduce in current release