[
https://issues.apache.org/jira/browse/FLINK-32961?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Martijn Visser updated FLINK-32961:
-----------------------------------
Fix Version/s: (was: 1.18.0)
> A new FileSystemFactory that support two high available hdfs
> ------------------------------------------------------------
>
> Key: FLINK-32961
> URL: https://issues.apache.org/jira/browse/FLINK-32961
> Project: Flink
> Issue Type: Improvement
> Components: FileSystems
> Reporter: 王茂军
> Priority: Major
> Original Estimate: 168h
> Remaining Estimate: 168h
>
> I run realtime ETL program by flink on yarn. The ETL program sink user log to
> master hdfs, and sink checkpoint to anather micro hdfs.The master hdfs and
> micro hdfs are both high available.
> By default, the ETL program can not understand the dfs.nameservices of the
> micro hdfs.
> I prepare to write a custom org.apache.flink.core.fs.FileSystemFactory that
> support two or more ha hdfs.So that , i can sink user log to master hdfs, and
> save checkpoint data to micro hdfs
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
