Hadoop Data Roll cannot create a directory - because it sees it as a file?

[derick_kotze]

I am trying to configure Hadoop Data Roll to archive data to a S3-compatible data store.

I can confirm that I can access the data store via s3cmd (https://s3tools.org/usage) as well as Hadoop on the local machine.

Below is the error that I'm seeing after the | archivebuckets saved search ran.

2019-11-19 18:17:09.064 +0200 ERROR SearchOutputStream - java.lang.RuntimeException: {exception=org.apache.hadoop.fs.FileAlreadyExistsException, message=Can't make directory for path 's3a://sam-splunk-archive/_internal_archive/tmp' since it is a file., stack=[org.apache.hadoop.fs.s3a.S3AFileSystem.innerMkdirs(S3AFileSystem.java:1731), org.apache.hadoop.fs.s3a.S3AFileSystem.mkdirs(S3AFileSystem.java:1681), org.apache.hadoop.fs.FileSystem.mkdirs(FileSystem.java:2216), com.splunk.roll.Transactor.prepare(Transactor.java:84), com.splunk.roll.Roller.doCopyBucket(Roller.java:158), com.splunk.roll.Roller.copy(Roller.java:87), com.splunk.roll.RollHandler.doCopyFromSplunkIndex(RollHandler.java:182), com.splunk.roll.RollHandler.executeCopyFromSplunkIndex(RollHandler.java:145), com.splunk.roll.RollHandler.executeCopyToVix(RollHandler.java:130), com.splunk.roll.RollHandler.execute(RollHandler.java:68), com.splunk.mr.SplunkMR.runImpl(SplunkMR.java:1802), com.splunk.mr.SplunkMR.run(SplunkMR.java:1551), org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:76), org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:90), com.splunk.mr.SplunkMR.main(SplunkMR.java:1839)]}
java.lang.Exception: java.lang.RuntimeException: {exception=org.apache.hadoop.fs.FileAlreadyExistsException, message=Can't make directory for path 's3a://sam-splunk-archive/_internal_archive/tmp' since it is a file., stack=[org.apache.hadoop.fs.s3a.S3AFileSystem.innerMkdirs(S3AFileSystem.java:1731), org.apache.hadoop.fs.s3a.S3AFileSystem.mkdirs(S3AFileSystem.java:1681), org.apache.hadoop.fs.FileSystem.mkdirs(FileSystem.java:2216), com.splunk.roll.Transactor.prepare(Transactor.java:84), com.splunk.roll.Roller.doCopyBucket(Roller.java:158), com.splunk.roll.Roller.copy(Roller.java:87), com.splunk.roll.RollHandler.doCopyFromSplunkIndex(RollHandler.java:182), com.splunk.roll.RollHandler.executeCopyFromSplunkIndex(RollHandler.java:145), com.splunk.roll.RollHandler.executeCopyToVix(RollHandler.java:130), com.splunk.roll.RollHandler.execute(RollHandler.java:68), com.splunk.mr.SplunkMR.runImpl(SplunkMR.java:1802), com.splunk.mr.SplunkMR.run(SplunkMR.java:1551), org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:76), org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:90), com.splunk.mr.SplunkMR.main(SplunkMR.java:1839)]}
            at com.splunk.mr.SplunkMR.run(SplunkMR.java:1567)
            at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:76)
            at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:90)
            at com.splunk.mr.SplunkMR.main(SplunkMR.java:1839)
Caused by: java.lang.RuntimeException: {exception=org.apache.hadoop.fs.FileAlreadyExistsException, message=Can't make directory for path 's3a://sam-splunk-archive/_internal_archive/tmp' since it is a file., stack=[org.apache.hadoop.fs.s3a.S3AFileSystem.innerMkdirs(S3AFileSystem.java:1731), org.apache.hadoop.fs.s3a.S3AFileSystem.mkdirs(S3AFileSystem.java:1681), org.apache.hadoop.fs.FileSystem.mkdirs(FileSystem.java:2216), com.splunk.roll.Transactor.prepare(Transactor.java:84), com.splunk.roll.Roller.doCopyBucket(Roller.java:158), com.splunk.roll.Roller.copy(Roller.java:87), com.splunk.roll.RollHandler.doCopyFromSplunkIndex(RollHandler.java:182), com.splunk.roll.RollHandler.executeCopyFromSplunkIndex(RollHandler.java:145), com.splunk.roll.RollHandler.executeCopyToVix(RollHandler.java:130), com.splunk.roll.RollHandler.execute(RollHandler.java:68), com.splunk.mr.SplunkMR.runImpl(SplunkMR.java:1802), com.splunk.mr.SplunkMR.run(SplunkMR.java:1551), org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:76), org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:90), com.splunk.mr.SplunkMR.main(SplunkMR.java:1839)]}
            at com.splunk.roll.RollHandler.execute(RollHandler.java:82)
            at com.splunk.mr.SplunkMR.runImpl(SplunkMR.java:1802)
            at com.splunk.mr.SplunkMR.run(SplunkMR.java:1551)
            ... 3 more

The following message also appears - which ties back to the error in the message above.

I can see that directories have been created in the bucket - but no actual data has been transferred.

Any ideas on how to resolve the issue? I'm pretty sure that I'm missing something obvious ... so any help will be appreciated.