Troubleshooting mapr-loopbacknfs Service Issues

Describes solutions for mapr-loopbacknfs issues.

To debug authentication issues, follow these steps:

  1. If you receive a standard error (stderr):
    • Make sure rpcinfo/portmap is installed and/or run service portmap start.
    • Run service rpcbind restart.
  2. Examine the log files for error messages:
    /usr/local/mapr-loopbacknfs/logs/loopbacknfs.log
    /usr/local/mapr-loopbacknfs/logs/mount_local_fs.log
    Error messages in loopbacknfs.log file:
    Error Message Solution
    Refresh User tickets failed as security layer could not be initialized with user ticket /tmp/maprticket_0 Unset MAPR_TICKETFILE_LOCATION in initscripts/mapr-loopbacknfs.
    exiting: license only allows 10 NFS/mfs server(s), currently alive=10 If you have multiple clusters listed in the mapr-clusters.conf file on the client, make sure the first one listed is a MapR 4.0.2 or later cluster. If that is not the problem, you will probably need to purchase additional licenses, or reduce number of installations of the mapr-loopbacknfs service.
    mount.nfs: Protocol not supported The NFS directory with mapr-loopbacknfs is already mounted.
  3. Verify that settings in configuration files are correct.
    • For all clusters: /usr/local/mapr-loopbacknfs/conf/mapr-clusters.conf
    • For secure clusters: MAPR_TICKETFILE_LOCATION
  4. Check for “stale” mounts.
    • The mount_local_fs.pl script can cause the initscript wrapper to force unmounts of the mounted file systems.
    • Always check for stale mounts after stopping the service:
      • df -k should return instantly.
      • Use umount -f <mount_point> to force the unmount.
      • Use ps -ef | grep mount_local to confirm that the script is not stuck.