[
https://issues.apache.org/jira/browse/YARN-895?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13830509#comment-13830509
]
Jian He commented on YARN-895:
------------------------------
Upload a patch:
- HDFS: create configs for enabling dfs client retry, and retry policy. If
enabled, dfs client will retry in case of connection failures or safe mode
exception of namnode.
- ZooKeeper: create a new config for the retry wait interval.
Test with HDFS,
- HDFS is down while RM is running or before RM starts
- NN is in safe mode when it initially starts.
- Manually control NN to enter safe mode
Test with zk:
- zk server is down before RM start or while RM is running.
All the scenarios, client is able to wait and retry
> RM crashes if it restarts while NameNode is in safe mode
> --------------------------------------------------------
>
> Key: YARN-895
> URL: https://issues.apache.org/jira/browse/YARN-895
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: resourcemanager
> Reporter: Jian He
> Assignee: Jian He
> Attachments: YARN-895.1.patch, YARN-895.2.patch, YARN-895.patch
>
>
--
This message was sent by Atlassian JIRA
(v6.1#6144)