My project has been using activemq as the central part of our integration platform for a while now. Recently we have made a few changes to our broker setup, hoping to improve throughput and efficiency in the broker. We also noticed some problems on our brokers that we thought could be related to concurrency, so we decided to go with the nio transport and -Dorg.apache.activemq.UseDedicatedTaskRunner=false in the hopes of reducing the amount of threads in the broker. We then put our broker into our test environment and let it run for a few weeks.
Our current setup seems to be more stable than the one we had previously (which pretty much left everything at the defaults), except for a few recurring log messages. At this point I have given up trying to decipher whether they are serious situations or just something I should disregard, so I had hoped for some help deciding what to do with them. We are running activemq-5.6.0. The tracebacks and our broker setup file can be found here: https://gist.github.com/3748262 We are running a total of 4 brokers, which are in pairwise networks of brokers, such that host1 is paired with host 9 and host 2 is paired with host 10. Most of the consumers and producers run on host 1 or host 2 and connect to tcp://localhost:30500. My question is basically what these warnings mean. They did not appear before we started using the nio transport, so we know we can make them disappear again. We would however like to keep our current setup if this is just aggressive logging. Does anyone know? -- Kind regards, Robin Kåveland Hansen
