Package: portmap Version: 5-26When rebooting or shutting down (using "init 6" or "init 0") a system that has a filesystem NFS-mounted, the /etc/rc[06].d scripts dutifully send killall5 as S20sendsigs. However, S31umountnfs.sh will pause for awhile (15+ seconds) before it times out in trying to contact the portmap server while unmounting the NFS mount(s). This produces an error on the screen briefly before the shutdown or reboot:
rpcbind: server localhost not responding, timed out...likely because the portmap server had been shutdown at S20sendsigs. It appears as though an attempt to alleviate that was implemented by placing a "S32portmap" in /etc/rc[06].d, however that sequencing places it /after/ the NFS umount script has passed. Moving the portmap restart to between the sendsigs and umountnfs.sh appears to fix the issue. I chose S21, but I suppose anything in between those two scripts will work.
It should be noted that the machine still shuts down / reboots after getting the rpcbind timeout, but with portmap running during the unmount, the shutdown occurs appx 10-15 seconds quicker.
Using Debian GNU/Linux 4.0r5.
smime.p7s
Description: S/MIME Cryptographic Signature
