something really not good. Usually, perhaps mostly, occurs when database runs in ARCHIVELOG mode and space is running out in archive destination directory. Oracle archives the redo logs timely to a specified archive location (timely/on commit...) so that recovery is guaranteed at later time. But we should keep a check on space utilization in destination directory. If we miss, Oracle may most likely stop new any external connections until DBA clears the space on dest directory.

Fix to this was briefly explained by Donald Burleson in his pages below ...

http://www.dba-oracle.com/sf_ora_00257_archiver_error_connect_internal_only_until_freed.htm

We do this problem recently in one of our production environments, after which only i started finding some details about this error. Respective 10g database was recently moved to a new box and so the error as backup maintenance cron scripts from old box were not migrated to new box. As there was not job to backup the archive logs on schedule basis space on file system ran out and finally database hung up. What is surprising me most is why would we still need a cron to manage few of the Oracle jobs. We may likely loose the control when things are scattered. Moment I realized crons were being used for archive back up, I felt Oracle should have already had a better way to handle these tasks. I searched for it and found there is a way...OEM can do this for us. Below link explains that...