runaway server processes

[John Dominic T. Binuya]

good day, here is my problem,our system experienced multiple runaway processes trying to connect to the database,we use a two-clustered system 16X875MHz CPU, 64G ram each.
But due to an error in the env_file for the process(it was using the env_file for node1), it just hangs until it is started again (the process runs every 10 mins via cron). According to our DBA, the process does not exist in the DB but it is seen on the OS level. The quick solution was that the DBA killed the process. After a few minutes it started again, it reached about 3000 processes until the database crashed and had to be opened again. The problem is why the process still persists when it is not connected to the database. It just doesn't terminate normally. The problem occured on the 2nd node.
Sorry for the long story.