The last time I saw a data/partition node error it was because I launched the erlang shell which connected to the node on the same CPU via cookies etc. Immediately after startup the shell dumped the partition error on the screen. This is terribly bothersome....
- how do I trap this exception?
- how do I repair the exception programatically? (asked in another question)
- how do I prevent this exception?
[update] I have two boxes that are running my Yaws application. The databases are replicated via Mnesia's extra_db_node feature. At some time after the servers are running I log into one of the boxes and launch 'erl' with a different sname and same cookie so that the 3 nodes can communicate. Shortly after the shell stabilizes and the shell prompt is displayed... a complex tuple is displayed on the screen indicating that there is a network partition error. This message appears to be a console dump rather than an exception that could be trapped by my yaws applications... but I want my yaws applications to detect the error and take corrective action.