James Netherton created CAMEL-22784:
---------------------------------------
Summary: Failover in FileLockClusterService is unreliable when
running multiple JVMs
Key: CAMEL-22784
URL: https://issues.apache.org/jira/browse/CAMEL-22784
Project: Camel
Issue Type: Bug
Components: camel-file
Reporter: James Netherton
Assignee: James Netherton
Based on feedback gained from the changes in CAMEL-22430, failover in
FileLockClusterService is unreliable when running multiple JVMs.
The current method of recording heartbeat timestamps is flawed. It uses
System.nanoTime, which is inappropriate for coordination across multiple JVMs.
The simplest fix to is to switch to System.currentTimeMillis. It is safe so
long as clocks are in sync across cluster nodes. Which is a general requirement
for most cluster failover algorithms anyway.
There's also some additional potential split-brain scenarios that I also plan
to fix as part of this.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
