If user omits port # in database connection string, the test still "goes green" - but the install will fail with a particularly terse error message. It appears that the test is assuming a port of 5432, but the install itself does not. If user enters an otherwise invalid port #, such as "999", test will appropriately fail. It only seems to assume a port of 5432 if none is specified. Repro steps: 1. begin installation of server with the assumption that you'll be using an external postgres db., i.e., a db that is not on this box. 2. Replace the default Database Connection URL ("jdbc:postgresql://127.0.0.1:5432/rhq") with a string akin to "jdbc:postgresql://foobar.example.com/rhq". Note that you have "forgotten" to replace the port # 3. Click the "Test Connection" button. 4. Noting that the test returns green, Choose to install (and/or overwrite/keep data, if applicable) 5. Submit. Current results: * Test goes green * Install fails with a cryptic message: Could not save the settings for some reason. Error Could not save the settings for some reason. Expected results: * Test should fail - test should use exact data as entered in the fields from install. * Error message should be useful.
what's the server log say? probably has more specific error messages. Not putting a port in the URL is probably OK from the JDBC driver point of view - the JDBC driver itself will assume a default (in the postgres driver, it probably defaults to 5432) - in the similar manner that "http://host" will force browsers to default to port 80.
This bug was previously known as http://jira.rhq-project.org/browse/RHQ-2238
Temporarily adding the keyword "SubBug" so we can be sure we have accounted for all the bugs. keyword: new = Tracking + FutureFeature + SubBug
making sure we're not missing any bugs in rhq_triage