https://bz.apache.org/bugzilla/show_bug.cgi?id=57700

            Bug ID: 57700
           Summary: Exception in a LifecycleListener on Context makes
                    webapp unreachable in spite of subsequent successful
                    startups
           Product: Tomcat 7
           Version: 7.0.59
          Hardware: PC
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Catalina
          Assignee: dev@tomcat.apache.org
          Reporter: knst.koli...@gmail.com

Using Tomcat 7.0.59.
I have an implementation of LifecycleListener that is configured in
META-INF/context.xml of a web application and performs some configuration work
on CONFIGURE_START_EVENT.

Roughly:
[[[
import org.apache.catalina.Context;
import org.apache.catalina.Lifecycle;
import org.apache.catalina.LifecycleEvent;
import org.apache.catalina.LifecycleListener;

public class FooListener implements LifecycleListener {
  @Override
  public void lifecycleEvent(LifecycleEvent event) {
    if (event.getLifecycle() instanceof Context) {
      if (Lifecycle.CONFIGURE_START_EVENT.equals(event.getType())) {
        if (loader == null) {
          Context context = (Context) event.getLifecycle();
          try {
             // <Do some work...>
          } catch (Exception ex) {
            final String msg = "Configuration failed for web application \"" +
context.getName()
                + "\".";
            throw new ConfigurationException(msg, ex);
          }
        }
      }
    }
  }
}
]]]

The work of the listener can fail due to some external condition.

Steps to reproduce:
All the following is performed via HTML UI of the Manager web application.
1. Undeploy application foo.
2. Upload foo.war.
3. The external condition is not configured.
The following is observed:
- The application fails to start.
- A org.apache.catalina.LifecycleException happens and is logged into
catalina.dddd-yy-mm.log
- The Manager app shows the application as deployed and is in "stopped" state.

This all is as expected.

4. Fix the external condition and start the web application via Manager.
Observed behaviour:
- The application is started. The manager webapp shows that it is started and
running.

Expected behaviour:
- Being able to access /foo

Actual behaviour:
- Error 404. The access log shows that the request was served by the ROOT web
application.
- No errors or warnings are logged

If I stop/start the foo application, it stops and starts successfully, but
nothing changes.
If I stop/start the ROOT application, it stops and starts successfully, but
nothing changes.

If I undeploy and re-deploy the foo application, the situation resolves and the
application can be accessed successfully.


Thoughts:
----------
1. Maybe the error is in Mapper. Maybe the error is handling of FailedContext. 

The initial error may be a deployment error (resulting in use of
FailedContext), but in this case it is odd that starting the application (after
fixing the external condition) is performed successfully and it is shown as
running.

If the error is in the Mapper, why the error is fixed when the application is
undeployed?

2. We already have test cases for failed startups,
org.apache.catalina.core.TestStandardContext
-> testBug46243()
-> testWebappLoaderStartFail()
-> testWebappListenerConfigureFail()

The testWebappListenerConfigureFail() test was designed to reflect this
situation. I guess there is a difference in deployment scenario via Manager
webapp vs. the one used in the test.

3. I wonder, if the Manager webapp had a "[Redeploy]" button for a webapp, like
it has a "[Reload]" one, would pressing it be able to resolve this issue. I
will file an enhancement request for that.

-- 
You are receiving this mail because:
You are the assignee for the bug.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org
For additional commands, e-mail: dev-h...@tomcat.apache.org

Reply via email to