2015.6.2

The jstorm and zookeeper clusters were built on four virtual machines, and the official demo was submitted to Jstorm for successful operation. Problems encountered during the construction process and solutions:
1. The virtual machine can not start the Supervisor service, after the detection, is the current logon user does not have log write permission caused by the sudo chown user:user–r * command to modify the permissions, start the Supervisor node service success;
2. If the Jstorm configuration file Storm.yaml file modification, in time to replace the current user root ~/.jstorm folder storm.yaml file with the latest Storm.yaml file, otherwise, will cause jstorm start failure;
3. The Jstorm profile Storm.yaml must be consistent on each virtual machine, especially if the log output path Storm.zookeeper.root in the configuration file must be consistent, or even if the Jstorm service on supervisor starts successfully, it will cause jstorm The UI cannot display the Supervisor node;
4. The Jstorm service must be restarted on each supervisor node before restarting the Jstorm service on the Nimbus node, otherwise the Nimbus server boot failure may occur;
5. The jar package for a task submitted to Jstorm must contain a third-party dependent jar package (if the task contains a third-party dependent jar package), or a classnotfound error will occur even if the task is submitted successfully.

Copyright NOTICE: This article for Bo Master original article, without Bo Master permission not reproduced.

2015.6.2