Hi Anil,
thanks again for your advices. I think you hit the nail on the head.
the error was generated by the presence (parallel?) of a deletion task and a backup task.
let me explain better: in my bash script before proceeding with the cluster backup I went to look for the old backups to delete them with the cleanup command.
then I cyclically executed the status command and used a regular expression to go and verify that in the command output there was nothing “running”. only then I proceed with the backup. in this step something went wrong.
I split the deletion phase from the backup phase and everything seems to be ok.
it would be nice to have better tools to query the activity of the agents without evaluating the output of the “status” command as I did