Hi All,

We are running multiple trident topologies on a storm cluster. Quite often
we are seeing transaction node errors in Zookeeper.

 Node already deleted
or directory not empty.

 Which results in a runtime exception and worker crash. Have anyone faced
similar issues with storm zookeeper? Any reference to how storm/zookeeper
tags state data from different topologies will be helpful too.

Thanks,
Sunil

Reply via email to