The messages on nsrwatch on both the server and the client should be the same. Possibly your nsrwatch on the client is connecting to some other machine with the NW server package installed. In any case, if the cause of your backup abort was the message you have mentioned, your backup of filesystems would fail as well.

You should be able to get the log files for the NMO from the applogs directory on your client machine. On Unix / Linux the directory would be /nsr/applogs. On Windows it would be something like C:\Program Files\Legato\nsr\applogs (this would depend on where you installed your NW binaries on Windows). Let us know if there are any errors in the log files.

Actually if you type an nsrwatch on the client it returns the message I said before, but if you type and nsrwatch -s nsrserver it works fine. Apparently the client is trying to use himself as a Networker Server. The /etc/hosts is set correctly, all the aliases of the nsrserver are set in the /nsr/res/servers and, right now, it just seems to be well configured. The backups were working fine until yesterday, but something happens since then.

If I try to do a save command on the client I get this:
storage@mir:/nsr/applogs>save
save: Using mir as server
Warning: Could not determine job id: Could not connect to server.. Continuing ...
save: RAP error: system error
save: Cannot open save session with mir

mir is the name of the client. If I use the command save -s nsrserver it works ok. Don't know in which configuration file or network related config I'm getting the problem.

You need to use save -s -b . If your client doesn't know what is your server (eg. servers file) then you must use -s. That's normal thing.

But let's go back to RMAN backup. Give us whole output of rman execution when it fails for you...