Loading...

XML

Word

Printable

Steps to Reproduce:
Hide

Either use unit test KahaDBRecoveryTest.tgz or

start AMQ 6.3

osgi:list -l -t 0 | grep mq-fabric and take note of the mq-fabric bundle id

Copy paste the following Karaf commands in one go to the Karaf shell
activemq:producer --user admin --password admin --messageCount 3000
sleep 2000
activemq:consumer --user admin --password admin --messageCount 3000
sleep 6000
restart 230

Observe broker log and check for entry similar to

Recovery replayed 223 operations from the journal in 0.019 seconds

When using unit test also check log file for above recovery note.
Show
Either use unit test KahaDBRecoveryTest.tgz or start AMQ 6.3 osgi:list -l -t 0 | grep mq-fabric and take note of the mq-fabric bundle id Copy paste the following Karaf commands in one go to the Karaf shell activemq:producer --user admin --password admin --messageCount 3000 sleep 2000 activemq:consumer --user admin --password admin --messageCount 3000 sleep 6000 restart 230 Observe broker log and check for entry similar to Recovery replayed 223 operations from the journal in 0.019 seconds When using unit test also check log file for above recovery note.

When restarting the broker gracefully KahaDB always recovers entries from the journal. The logging of the starting broker always contains a line like this

Recovery replayed 223 operations from the journal in 0.019 seconds

This even works with an idle broker.

I would expect that no journal recovery is required as index and redo file are written with the latest status when the broker shuts down gracefully.