I addressed several of these issues in Komo, although I don't know how "elegant" my efforts were. You can download the source code from the Komo project at pgFoundry, and look particularly at the ConnectionManager class. If you upgrade the code, I would definitely be interested in hearing from you back channel (MichaelMSchmidt@msn.com).
Subject: [JDBC] Handling network errors during initial connection and ongoing use
Hey all,
I'm working on fixing an application that uses jdbc so it will provide users with a better understanding of what is wrong when things go wrong. However, I cannot find any information on elegant ways to handle the 'unreliable' part of tcp/ip.
Is there some common strategy to handle network errors during setup or use of jdbc?
Specifically, I imagine all java applications using jdbc would have to handle: during initial connection no network no dns host name unknown no answer on port answer on port not postgresql bad authentication (wrong username/password if these are distinct) user doesn't have permissions to read database
during ongoing use network failure server failure
Could anyone point me toward documents/topic which cover these issues?
Thank you, adrian custer
---------------------------(end of broadcast)--------------------------- TIP 9: In versions below 8.0, the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
