Recently in the development of Dubbo services have been using the ZK platform provided by the company, because the use of too many people or no maintenance is always a problem, resulting in Dubbo services occasionally can be adjusted, occasionally adjust the situation, so take a moment to deploy a set of their own, the back of the problem is convenient to see the log troubleshooting.
1 download package upload Linux Server zookeeper-3.4.6.tar.gz
2 Decompression TAR-ZXVF zookeeper-3.4.6.tar.gz
3 Entering the zookeeper-3.4.6/conf directory
4 copies a copy of the zoo_sample.cfg renamed to Zoo.cfg (if you've ever used LTS, this lightweight task scheduling framework will also find that Tasktracker has this ZK configuration file as well)
5 Editing a configuration file
5 Create a new file myID under each/var/lib/zookeeper, the content is 1, or 2 (the second instance), or 3 (the third instance) corresponds to the SERVER.1/2/3 number above
6 do the same steps on the other two machines ( Note: The number of instances deployed must be odd because of the ZK election mechanism )
7 when the configuration is complete, start separately
After 83 instances are started, view the cluster status
9 You can try to break off one of the instances to see if the ZK switch is normal.
Dubbo Development Foreplay--zookeeper Cluster Deployment (3.4.6)