Hi,

I am running into an issue when using the JMSQueueAppender on Weblogic application server, where application code talks to SLF4j with Logback as the underlying logging framework (core+classic).

Below my logback.xml configuration. In this scenario, SLF4j is being initialized before the JNDI context is available from Weblogic, thus resulting an error like:

1:01:19,348 |-ERROR in ch.qos.logback.classic.net.JMSQueueAppender[JMSQueue] - Error while activating options for appender named [JMSQueue]. javax.naming.NoInitialContextException: JNDI subsystem is not ready for use

at javax.naming.NoInitialContextException: JNDI subsystem is not ready for use

at weblogic.jndi.WLInitialContextFactoryDelegate.newLocalContext(WLInitialContextFactoryDelegate.java:569)

... stacktrace omitted ...

Is there any reconnection/retrycount property that can be set on the JMSQueueAppender (can't find documentation about it online manual)? Or any other way of re-initialization?

(N.b. a workaround would be: define scan="true" scanPeriod="10 seconds", perform a unix 'touch' on logback.xml manually, which will reconfigure the appender. But by then we might have missed a lot of critical logging not being sent to the JMS queue).

Copy of JMS appender definition in logback.xml:

weblogic.jndi.WLInitialContextFactory

</InitialContextFactoryName>

<ProviderURL>t3://localhost:7001</ProviderURL>

weblogic.jms.ConnectionFactory

</QueueConnectionFactoryBindingName>

<QueueBindingName>jms/foo/barQueue</QueueBindingName>

</appender>

....

<appender-ref ref="JMSQueue"/>

</logger>

Thanks in advance!

Kind regards,

Rutger