Seems to be a cluster problem - if i take one node down, slowdown doesn't
happen. What i also noticed that sometimes cpu usage would rise out of the
normal range (on both brokers) and would only go down
when i stopped the network bridge (or restarted one of the brokers).
Network connector setting
Brokers were slow at the time. Jvm max heap is set at 2048MB, starting the
brokers clean memory usage usually starts from 200MB and slowly starts
rising as more threads are being used. Usual thread usage will go up to
about 300-400 on brokers together. Right now broker b is running 246 threads
and
okerA.txt>
Threaddumps from both brokers
Regards
Priit
--
View this message in context:
http://activemq.2283324.n4.nabble.com/Connecting-to-broker-slows-down-tp4667017p4667042.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.
Hey
I have a problem with the broker itself slowing down over a few days of
uptime. Cpu and memory utilization are both ok. But establishing a
connection to the broker takes longer and longer. I'm using tcp+ssl
connector. Same issue with nio and regular tcp.
I have a setup of two brokers, both o