Repository: flink Updated Branches: refs/heads/release-1.2 576cc895c -> db3c5f388
[FLINK-5894] [docs] Fix misleading HA docs This closes #3401. Project: http://git-wip-us.apache.org/repos/asf/flink/repo Commit: http://git-wip-us.apache.org/repos/asf/flink/commit/db3c5f38 Tree: http://git-wip-us.apache.org/repos/asf/flink/tree/db3c5f38 Diff: http://git-wip-us.apache.org/repos/asf/flink/diff/db3c5f38 Branch: refs/heads/release-1.2 Commit: db3c5f388b3f9acac19d3d0395df96d7a58fcdc9 Parents: 576cc89 Author: Ufuk Celebi <[email protected]> Authored: Thu Feb 23 13:30:13 2017 +0100 Committer: Ufuk Celebi <[email protected]> Committed: Thu Feb 23 13:49:44 2017 +0100 ---------------------------------------------------------------------- docs/setup/jobmanager_high_availability.md | 14 ++++---------- 1 file changed, 4 insertions(+), 10 deletions(-) ---------------------------------------------------------------------- http://git-wip-us.apache.org/repos/asf/flink/blob/db3c5f38/docs/setup/jobmanager_high_availability.md ---------------------------------------------------------------------- diff --git a/docs/setup/jobmanager_high_availability.md b/docs/setup/jobmanager_high_availability.md index aa18a4b..5949835 100644 --- a/docs/setup/jobmanager_high_availability.md +++ b/docs/setup/jobmanager_high_availability.md @@ -84,14 +84,13 @@ In order to start an HA-cluster add the following configuration keys to `conf/fl **Important**: if you are running multiple Flink HA clusters, you have to manually configure separate namespaces for each cluster. By default, the Yarn cluster and the Yarn session automatically generate namespaces based on Yarn application id. A manual configuration overrides this behaviour in Yarn. Specifying a namespace with the -z CLI option, in turn, overrides manual configuration. -- **State backend and storage directory** (required): JobManager meta data is persisted in the *state backend* and only a pointer to this state is stored in ZooKeeper. Currently, only the file system state backend is supported in HA mode. +- **Storage directory** (required): JobManager metadata is persisted in the file system *storageDir* and only a pointer to this state is stored in ZooKeeper. <pre> -high-availability.zookeeper.storageDir: hdfs:///flink/recovery</pre> -state.backend: filesystem -state.backend.fs.checkpointdir: hdfs:///flink/checkpoints +high-availability.zookeeper.storageDir: hdfs:///flink/recovery + </pre> - The `storageDir` stores all meta data needed to recover a JobManager failure. + The `storageDir` stores all metadata needed to recover a JobManager failure. After configuring the masters and the ZooKeeper quorum, you can use the provided cluster startup scripts as usual. They will start an HA-cluster. Keep in mind that the **ZooKeeper quorum has to be running** when you call the scripts and make sure to **configure a separate ZooKeeper root path** for each HA cluster you are starting. @@ -106,9 +105,6 @@ high-availability.zookeeper.path.root: /flink high-availability.zookeeper.path.namespace: /cluster_one # important: customize per cluster high-availability.zookeeper.storageDir: hdfs:///flink/recovery</pre> -state.backend: filesystem -state.backend.fs.checkpointdir: hdfs:///flink/checkpoints - 2. **Configure masters** in `conf/masters`: <pre> @@ -192,8 +188,6 @@ high-availability.zookeeper.quorum: localhost:2181 high-availability.zookeeper.storageDir: hdfs:///flink/recovery high-availability.zookeeper.path.root: /flink high-availability.zookeeper.path.namespace: /cluster_one # important: customize per cluster -state.backend: filesystem -state.backend.fs.checkpointdir: hdfs:///flink/checkpoints yarn.application-attempts: 10</pre> 3. **Configure ZooKeeper server** in `conf/zoo.cfg` (currently it's only possible to run a single ZooKeeper server per machine):
