In response to Tim Uckun <timuckun@gmail.com>:
> Today the database shut down unexpectedly. I have included the log file
> that shows the shutdown. Can anybody tell me why this happened and how I can
> make sure it doesn't happen again.
>
> The only thing I can think of that I did was to specify a password for the
> postgres user in the operating system.
Not likely to cause the DB to restart ... at least not in any OS
configuration that I'm aware of. However, you don't mention what
OS you're running ... that might be important.
> Here is the log file. Very strange.
>
> 2009-03-25 00:02:01 GMT LOG: incomplete startup packet
> 2009-03-25 00:30:05 GMT LOG: could not receive data from client: Connection
> timed out
> 2009-03-25 00:30:05 GMT LOG: unexpected EOF on client connection
> 2009-03-25 00:30:05 GMT LOG: could not receive data from client: Connection
> timed out
> 2009-03-25 00:30:05 GMT LOG: unexpected EOF on client connection
> 2009-03-25 00:32:15 GMT LOG: could not receive data from client: Connection
> timed out
> 2009-03-25 00:32:15 GMT LOG: unexpected EOF on client connection
> 2009-03-25 02:41:57 GMT LOG: incomplete startup packet
> 2009-03-25 02:41:57 GMT LOG: received smart shutdown request
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Sure looks like someone intentionally shut the database down.
Who has login access to this system with enough privs to do that? I'd
start by questioning them on their activities at that time. If that
doesn't answer your question, then implement OS auditing so you can
catch the culprit next time. It's quite possible that someone is sending
signals to the DB system without knowledge of how those signals are
interpreted.
--
Bill Moran
http://www.potentialtech.com
http://people.collaborativefusion.com/~wmoran/