Thanks, I think your adding volatile may be the key. I don’t think the thread.interrupt does anything as the socket methods don’t declare that they throw InterruptedException. After I added the System.println statements I was able to get the release build to work. I am still in the process of finishing all the tasks.
Ralph > On Nov 19, 2017, at 9:07 AM, Mikael Ståldal <mi...@apache.org> wrote: > > I made an attempt to make that test more reliable. > > On 2017-11-19 00:57, Ralph Goers wrote: >> I am preparing to do the release build and am getting the following error: >> Tests run: 5, Failures: 0, Errors: 1, Skipped: 0, Time elapsed: 1.25 s <<< >> FAILURE! - in org.apache.logging.log4j.core.appender.SyslogAppenderTest >> testTCPAppender(org.apache.logging.log4j.core.appender.SyslogAppenderTest) >> Time elapsed: 0.039 s <<< ERROR! >> java.net.BindException: Address already in use >> at >> org.apache.logging.log4j.core.appender.SyslogAppenderTest.initTCPTestEnvironment(SyslogAppenderTest.java:100) >> at >> org.apache.logging.log4j.core.appender.SyslogAppenderTest.testTCPAppender(SyslogAppenderTest.java:54) >> I thought I fixed this by adding a thread.join to wait for shutdown to >> complete but I got it again. This doesn’t seem to happen on every build. >> Before today I have never had this error before. >> Any unit test failures during the build fail the release. Each release >> attempt takes about 45 minutes so I really need a build without failing >> tests. If I can’t figure it out I may have to disable the test. >
