I messed up OES2 SP2 with NCS and FC LUNs. The systems consists of
multiple NSS ressources shared with NCS and a single one I use for NDS
Roll Forward Logs and eDirectory backup files.
Prior to joining the cluster only this LUN was available on the node. I
mounted it with fstab:
/dev/mapper/mpathd /edir xfs defaults 1 2
After joining the cluster this does not work any more:
# mount /dev/mapper/mpathd /edir
mount: /dev/mapper/mpathd already mounted or /edir busy
I checked with "multipath -l" that mpathd is still valid and the correct
device. The situation is messy as ndsd will not start properly without
access to it's Roll Forward Logs.
The problem is obviously related to how OES2 handles devices with EVMS.
Any hints to a description on how to fix the situation? Is it possible
to have multipath devices handled with and without NCS/EVMS on the same