[
https://issues.apache.org/jira/browse/FLINK-4657?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15526962#comment-15526962
]
ASF GitHub Bot commented on FLINK-4657:
---------------------------------------
Github user StephanEwen commented on the issue:
https://github.com/apache/flink/pull/2550
Can we try and unify the structure / paths under which all this information
is stored in ZooKeeper?
```
/flink
+/cluster_id_1/resource_manager_lock
| |
| +/job-id-1/job_manager_lock
| | /checkpoints/latest
| | /latest-1
| | /latest-2
| |
| +/job-id-2/job_manager_lock
|
+/cluster_id_2/resource_manager_lock
|
+/job-id-1/job_manager_lock
|/checkpoints/latest
| /latest-1
|/persisted_job_graph
```
The "cluster-id" should be a generated UUID in the case of YARN/Mesos, and
should be a config value in the standalone case. In Yarn / Mesos, the UUID
should be passed via an environment variable to the Java processes with the
entry points for TaskManager / JobManager / ResourceManager.
In the Constructor, the ZooKeeper HA Services should get the "cluster-id".
> Implement HighAvailabilityServices based on zookeeper
> -----------------------------------------------------
>
> Key: FLINK-4657
> URL: https://issues.apache.org/jira/browse/FLINK-4657
> Project: Flink
> Issue Type: New Feature
> Components: Cluster Management
> Reporter: Kurt Young
> Assignee: Kurt Young
>
> For flip-6, we will have ResourceManager and every JobManager as potential
> leader contender and retriever. We should separate them by using different
> zookeeper path.
> For example, the path could be /leader/resource-manaeger for RM. And for each
> JM, the path could be /leader/job-managers/JobID
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
