[
https://issues.apache.org/jira/browse/FLINK-4228?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16249213#comment-16249213
]
ASF GitHub Bot commented on FLINK-4228:
---------------------------------------
Github user tillrohrmann commented on the issue:
https://github.com/apache/flink/pull/4939
I think excluding dependencies without declaring them explicitly has the
disadvantage that this relies implicitly on another dependency to pull in the
right versions. Once this dependency is removed or its version bumped the whole
build might fail because it no longer pulls in the dependency.
> YARN artifact upload does not work with S3AFileSystem
> -----------------------------------------------------
>
> Key: FLINK-4228
> URL: https://issues.apache.org/jira/browse/FLINK-4228
> Project: Flink
> Issue Type: Bug
> Components: State Backends, Checkpointing
> Reporter: Ufuk Celebi
> Assignee: Nico Kruber
> Priority: Blocker
> Fix For: 1.4.0
>
>
> The issue now is exclusive to running on YARN with s3a:// as your configured
> FileSystem. If so, the Flink session will fail on staging itself because it
> tries to copy the flink/lib directory to S3 and the S3aFileSystem does not
> support recursive copy.
> h2. Old Issue
> Using the {{RocksDBStateBackend}} with semi-async snapshots (current default)
> leads to an Exception when uploading the snapshot to S3 when using the
> {{S3AFileSystem}}.
> {code}
> AsynchronousException{com.amazonaws.AmazonClientException: Unable to
> calculate MD5 hash:
> /var/folders/_c/5tc5q5q55qjcjtqwlwvwd1m00000gn/T/flink-io-5640e9f1-3ea4-4a0f-b4d9-3ce9fbd98d8a/7c6e745df2dddc6eb70def1240779e44/StreamFlatMap_3_0/dummy_state/47daaf2a-150c-4208-aa4b-409927e9e5b7/local-chk-2886
> (Is a directory)}
> at
> org.apache.flink.streaming.runtime.tasks.StreamTask$AsyncCheckpointThread.run(StreamTask.java:870)
> Caused by: com.amazonaws.AmazonClientException: Unable to calculate MD5 hash:
> /var/folders/_c/5tc5q5q55qjcjtqwlwvwd1m00000gn/T/flink-io-5640e9f1-3ea4-4a0f-b4d9-3ce9fbd98d8a/7c6e745df2dddc6eb70def1240779e44/StreamFlatMap_3_0/dummy_state/47daaf2a-150c-4208-aa4b-409927e9e5b7/local-chk-2886
> (Is a directory)
> at
> com.amazonaws.services.s3.AmazonS3Client.putObject(AmazonS3Client.java:1298)
> at
> com.amazonaws.services.s3.transfer.internal.UploadCallable.uploadInOneChunk(UploadCallable.java:108)
> at
> com.amazonaws.services.s3.transfer.internal.UploadCallable.call(UploadCallable.java:100)
> at
> com.amazonaws.services.s3.transfer.internal.UploadMonitor.upload(UploadMonitor.java:192)
> at
> com.amazonaws.services.s3.transfer.internal.UploadMonitor.call(UploadMonitor.java:150)
> at
> com.amazonaws.services.s3.transfer.internal.UploadMonitor.call(UploadMonitor.java:50)
> at java.util.concurrent.FutureTask.run(FutureTask.java:266)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
> at java.lang.Thread.run(Thread.java:745)
> Caused by: java.io.FileNotFoundException:
> /var/folders/_c/5tc5q5q55qjcjtqwlwvwd1m00000gn/T/flink-io-5640e9f1-3ea4-4a0f-b4d9-3ce9fbd98d8a/7c6e745df2dddc6eb70def1240779e44/StreamFlatMap_3_0/dummy_state/47daaf2a-150c-4208-aa4b-409927e9e5b7/local-chk-2886
> (Is a directory)
> at java.io.FileInputStream.open0(Native Method)
> at java.io.FileInputStream.open(FileInputStream.java:195)
> at java.io.FileInputStream.<init>(FileInputStream.java:138)
> at
> com.amazonaws.services.s3.AmazonS3Client.putObject(AmazonS3Client.java:1294)
> ... 9 more
> {code}
> Running with S3NFileSystem, the error does not occur. The problem might be
> due to {{HDFSCopyToLocal}} assuming that sub-folders are going to be created
> automatically. We might need to manually create folders and copy only actual
> files for {{S3AFileSystem}}. More investigation is required.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
