Thanks for your inputs.
However, my description on the issue was wrong.
Let me describe the situation again.
2 servers are in a cluster using terracotta.
Although Terracotta array server is down, 2 integration servers are working fine except clustering.
When I restart the integration servers while Terracotta server is down, it fails to bring up with following error.
jvm 1 | Jan 20, 2014 1:47:13 PM com.webMethods.sc.config.ConfigurationLogger
log
jvm 1 | WARNING: Failed to process OOB file-change event on file ‘C:\Software
AG95\profiles\IS\configuration\org.eclipse.equinox.app.manager.tmp183099735364
5242734.instance’
jvm 1 | Jan 20, 2014 1:47:13 PM com.webMethods.sc.config.ConfigurationLogger
log
jvm 1 | WARNING: Failed to process OOB file-change event on file ‘C:\Software
AG95\profiles\IS\configuration\org.eclipse.core.runtime.manager.tmp36650905438
52294218.instance’
jvm 1 | [p2] Mon Jan 20 13:47:42 KST 2014 - [System Bundle Shutdown] Unsaved
preferences when shutting down org.eclipse.equinox.internal.p2.metadata.reposito
ry.MetadataRepositoryManager
jvm 1 | OSGi framework stopped.
jvm 1 | Exception in thread “Thread-82” java.lang.IllegalStateException: Bund
leContext is no longer valid
jvm 1 | at org.eclipse.osgi.framework.internal.core.BundleContextImpl.ch
eckValid(BundleContextImpl.java:931)
jvm 1 | at org.eclipse.osgi.framework.internal.core.BundleContextImpl.ge
tServiceReference(BundleContextImpl.java:528)
jvm 1 | at com.softwareag.eda.nerv.impl.util.component.ComponentRecovery
Runnable.run(ComponentRecoveryRunnable.java:43)
jvm 1 | at java.lang.Thread.run(Thread.java:724)
jvm 1 | Bootstrap: Remote shutdown is disabled.
wrapper | ← Wrapper Stopped
If I remove watt.server.cluster.tsaURLs value, Integration Server can bring up without any error.
#webMethods#Integration-Server-and-ESB