Github user wy96f commented on the issue:
https://github.com/apache/activemq-artemis/pull/2444
> From the description it is unclear what is meant by "server crash". JVM
crash or host/machine crash. If the problem occured with unexpected JVM process
termination maybe the broker can be started in a separate JVM processes and
SIGKILLed while doing large message processing to reproduce the issue.
In our case it's a host crash. I don't think jvm crash or sending SIGKILL
can reproduce it as franz1981 said linux os would take care to flush the page
cache.
---
