Open amotamed opened 8 years ago
Just a heads up, using the latest build (1.7.2), I'm still seeing this issue
And we're connecting to hbase-0.94.6
I'm still seeing this issue on 1.8.0
still seeing this on 1.8.1
Hmm, are you using the RPC timeout feature?
When using asynchbase 1.7.1 with io.netty:netty:3.10.5.Final I'm seeing HashdWheelTimeouts continuously increasing. I'm running jmap, here's an example of the output:
/usr/lib/jvm/java-1.8.0/bin/jmap -histo 1870 | grep HashedWheelTimeout 53: 4392 245952 org.jboss.netty.util.HashedWheelTimer$HashedWheelTimeout
This number continuously grows until we run out of memory on the server, it's a very gradual increase, it takes weeks to take effect with 4G heap