[AMBARI-24273] hadoop-env is not regenerated when OneFS is used as a FileSystem

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.7.0
• Fix Version/s: 2.7.1
• Component/s: ambari-server
• Labels: pull-request-available
• Assignee: Attila Magyar
• Reporter: Attila Magyar
• Created: 10/Jul/18 12:04
• Updated: 19/Jul/18 09:21
• Resolved: 19/Jul/18 08:30

Description

The before-ANY/shared_initialization.py only regenerates hadop_env if there is a namenode or dfs_type is set to HCFS

  def hook(self, env):
    import params
    env.set_params(params)

    setup_users()
    if params.has_namenode or params.dfs_type == 'HCFS':
      setup_hadoop_env()
    setup_java()

This is no longer true because in the latest ambari-server we set dfs_type as follows:

    Map<String, ServiceInfo> serviceInfos = ambariMetaInfo.getServices(stackId.getStackName(), stackId.getStackVersion());
    for (ServiceInfo serviceInfoInstance : serviceInfos.values()) {
      if (serviceInfoInstance.getServiceType() != null) {
        LOG.debug("Adding {} to command parameters for {}", serviceInfoInstance.getServiceType(),
            serviceInfoInstance.getName());

        clusterLevelParams.put(DFS_TYPE, serviceInfoInstance.getServiceType());
        break;
      }
    }

This iterates over all of the stack service which will find HDFS first, so that the dfs_type will be HDFS instead of HCFS.

Issue Links

• links to: GitHub Pull Request #1773
• links to: GitHub Pull Request #1788