[
https://issues.apache.org/jira/browse/FLINK-13758?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16981411#comment-16981411
]
Yang Wang commented on FLINK-13758:
-----------------------------------
I go over the current code base and find that it is actually a bug. Because
`StreamExecutionEnvironment#registerCachedFile` allows a remote distributed
path(for example, hdfs://host:port/and/path). However, the remote path could
not properly be processed by `RestClusterClient#submitJob`. It does not
differentiate the local and remote path.
I think we need to check whether the artifacts is a remote path and skip to
upload in `RestClusterClient#submitJob`.
This bug has been introduced since RestClusterClient is used to submit job.
> failed to submit JobGraph when registered hdfs file in DistributedCache
> ------------------------------------------------------------------------
>
> Key: FLINK-13758
> URL: https://issues.apache.org/jira/browse/FLINK-13758
> Project: Flink
> Issue Type: Bug
> Components: Command Line Client
> Affects Versions: 1.6.3, 1.6.4, 1.7.2, 1.8.0, 1.8.1, 1.9.0, 1.10.0, 1.9.1
> Reporter: luoguohao
> Priority: Major
> Labels: pull-request-available
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> when using HDFS files for DistributedCache, it would failed to submit
> jobGraph, we can see exceptions stack traces in log file after a while, but
> if DistributedCache file is a local file, every thing goes fine.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
