[jira] [Created] (ARTEMIS-3277) Client connection to Artemis-cluster node lost

["Jelmer Marinus (Jira)" <ji...@apache.org>]

Jelmer Marinus created ARTEMIS-3277:
---------------------------------------

             Summary: Client connection to Artemis-cluster node lost
                 Key: ARTEMIS-3277
                 URL: https://issues.apache.org/jira/browse/ARTEMIS-3277
             Project: ActiveMQ Artemis
          Issue Type: Bug
          Components: ActiveMQ-Artemis-Native
    Affects Versions: 2.15.0
            Reporter: Jelmer Marinus
            Assignee: Clebert Suconic


We are running an Artemis-cluster with the following specs:

------

Artemis version: 2.15.0 (runs in a Docker-container)
Docker version: Docker version 20.10.3, build 48d30b5
Docker image: vromero/activemq-artemis:2.15.0 (JVM: OpenJDK Runtime Environment (build 1.8.0_232-b09))
OS-version: CentOS 7.9.2009 (x86_64)
Cluster-setup: 3 primary nodes with 3 secondary nodes

------

Sometimes we are experiencing connection issues from the clients connecting to the cluster which results in messages not beeing consumed from or produced to the cluster. This seems to be specific to a single alternating master node at the time and a restart of the node seems to solve the problem. From the client we get a "AMQ212037: Connection failure to ############################## has been detected: AMQ219015: The connection was disconnected because of server shutdown [code=DISCONNECTED]" error message.". The Artemis-console however is available so from this perspective the node seems to be reachable and working correctly.

Server-side (Artemis) we get the follow error messages which maybe related "Inconsistency during compacting: RollbackRecord ID = 82113064851 for an already rolled back transaction during compacting", "Error on reading compacting for JournalFileImpl: (activemq-data-######.amq id = #####, recordID = #####)", "Bridge Failed to ack", "Cannot find add info ######## on compactor or current records",

 

We are also experiencing the following issues which may or may not be related:

1) initial distribution/Redistribution not working correctly

2) Reset of Address-setting to default values. We use specific setting from i.e. DLA, redistributionDelay etc. but sometimes these settings seem to go back to their default values.

 



--
This message was sent by Atlassian Jira
(v8.3.4#803005)