The database takes an OS-level file lock by default - it looks like the NFS server believes the file is still locked even though the process has died.
This should clear the NFS lock (since it will create new inodes) and allow you to start - once started you can remove/archive /nexus-data/db.tmp
If you don’t want to copy the whole DB directory then you can also do this once for each file mentioned during startup:
1. mv /nexus-data/db/config/database.ocf /nexus-data/db/config/database.ocf.tmp
2. cp /nexus-data/db/config/database.ocf.tmp /nexus-data/db/config/database.ocf
The alternative is to talk to your NFS administrator to get the file locks cleared for that client.
To stop this happening again in the future you can turn off the database file lock by adding this property to your etc/system.properties file:
file.lock=false
Nexus has it’s own higher-level lock to stop multiple Nexus instances from using the same data directory, so this should be safe..
--
Cheers, Stuart