-
Bug
-
Resolution: Done
-
Critical
-
AMQ 7.11.7.GA
This is an active/passive AMQ 7 set-up with NFS shared store.
After an upgrade from 7.10.2 to 7.11.6, at unpredictable times, but under high production load, AMQ 7 emits this exception:
2024-06-16 03:18:26,534 WARN [org.apache.activemq.artemis.core.io.AbstractSequentialFile] null
java.nio.channels.ClosedChannelException: null
at sun.nio.ch.FileChannelImpl.ensureOpen(FileChannelImpl.java:159) ~[?:?]
at sun.nio.ch.FileChannelImpl.force(FileChannelImpl.java:459) ~[?:?]
at org.apache.activemq.artemis.core.io.nio.NIOSequentialFile.sync(NIOSequentialFile.java:322) ~[artemis-journal-2.28.0.redhat-00019.jar:2.28.0.redhat-00019]
at org.apache.activemq.artemis.core.io.AbstractSequentialFile$LocalBufferObserver.checkSync(AbstractSequentialFile.java:304) ~[artemis-journal-2.28.0.redhat-00019.jar:2.28.0.redhat-00019]
at org.apache.activemq.artemis.core.io.buffer.TimedBuffer.flushBatch(TimedBuffer.java:362) ~[artemis-journal-2.28.0.redhat-00019.jar:2.28.0.redhat-00019]
at org.apache.activemq.artemis.core.io.buffer.TimedBuffer.flush(TimedBuffer.java:306) ~[artemis-journal-2.28.0.redhat-00019.jar:2.28.0.redhat-00019]
at org.apache.activemq.artemis.core.io.buffer.TimedBuffer$CheckTimer.run(TimedBuffer.java:482) ~[artemis-journal-2.28.0.redhat-00019.jar:2.28.0.redhat-00019]
at java.lang.Thread.run(Thread.java:840) ~[?:?]
Thereafter, it continues to emit this warning, several times a second, indefinitely:
2024-06-18 12:01:24,868 WARN [org.apache.activemq.artemis.core.server] AMQ222104: IO Error completing the transaction, code = 6, message = null 2024-06-18 12:01:24,868 WARN [org.apache.activemq.artemis.core.server] AMQ222063: Error processing IOCallback code = 6 message = null
The broker may be paging on a few addresses when this happens, but that's normal in this set-up. There's nothing else in the log to indicate a problem, even going back hours. There's nothing in the log of the broker host system, or in the log of the NFS server.
Once this problem arises, the broker does seem to be able to provide a service, but with severely reduced throughput. It takes a broker restart to clear the problem.
- is cloned by
-
-
- Dev Complete
-
