[1590] in Athena Bugs
6.0R: /usr/etc/rpc.mountd
daemon@ATHENA.MIT.EDU (Richard Basch)
Tue Jan 3 09:33:25 1989
Date: Tue, 3 Jan 89 09:33:12 EST
From: "Richard Basch" <probe@ATHENA.MIT.EDU>
To: bugs@ATHENA.MIT.EDU
Reply-To: Richard Basch <probe@ATHENA.MIT.EDU>
If the status of a mounted filesystem changes, it is not noticed by
rpc.mountd. For instance, if a server is exporting an RVD from another
server (ie. aeneas with the RT srvd packs) and it reboots, rpc*d must be
killed if a mount request occurs prior to spinning up the drive.
Sequence of events:
1) rpcd gets started
2) nfsd is started
3) all RVD's are spun up in /etc/rc.local
If anyone tries to create a mapping or tries to mount any filesystem
between steps 2 and 3, rpc.mountd will be started and not be able to
export the RVD's.
Fix, after the five minute timer expires, rpc.mountd should re-read
/etc/exports and check for valid filesystems (it keeps a cache of valid
filesystems).
-Richard
