Better error reporting when messages "mysteriously" vanish

It is commonly the case that people try to run two instances of JBossMQ against the same
database tables. This can lead to errors like the ones in the parent task or a similar message
in the delete when doing acknowledgements.

These error messages need improving and also give a hint to the possible reasons,
e.g.
1) competing JBossMQs trying to share the same database
2) trying to use a replicating database that does not support synchronous replication (i..e there
is a delay in a row getting written to one database before it appears in the other database)

The parent issue's error message belongs in org.jboss.mq.pm.jdbc2.PersistenceManager

public SpyMessage loadFromStorage(MessageReference messageRef) throws JMSException
{
...
rs = stmt.executeQuery();
if (rs.next())
return extractMessage(rs);
+ else
throw new JMSException("Unable to load message " + messageRef + " from storage, are you trying to share database tables between JBossMQ instances? Are you using a replicating database that does not support synchronous replication?");

• return null;