[
https://issues.apache.org/jira/browse/SPARK-26019?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16693664#comment-16693664
]
Imran Rashid commented on SPARK-26019:
--------------------------------------
Yeah I agree with [~viirya]'s analysis, my suggestion was from just a quick
glance at the code. I don't think swapping those lines is likely to help at
all ... but I can't come up with any other explanation for how it does happen.
From SPARK-26113, it doesn't seem particular to the cloudera distribution, but
we'll poke at it a bit. SPARK-26113 also makes it sound like a race as it
works after the initial failure ...
[~Tagar] are you running a pyspark shell, or with spark-submit? the token
generation is different in those two cases, so that might matter (though I
don't see how yet ...)
[~hyukjin.kwon] for errors which appear to be from a race, I don't think we
should close immediately because we can't reproduce it, as it can be tricky to
reproduce and involve something about the user environment that we dont'
immediately understand, that doesn't mean its not a real issue. (I absolutely
agree that if it appears to be related to a specific distribution, it doesn't
belong as an issue here).
> pyspark/accumulators.py: "TypeError: object of type 'NoneType' has no len()"
> in authenticate_and_accum_updates()
> ----------------------------------------------------------------------------------------------------------------
>
> Key: SPARK-26019
> URL: https://issues.apache.org/jira/browse/SPARK-26019
> Project: Spark
> Issue Type: Bug
> Components: PySpark
> Affects Versions: 2.3.2, 2.4.0
> Reporter: Ruslan Dautkhanov
> Priority: Major
>
> Started happening after 2.3.1 -> 2.3.2 upgrade.
>
> {code:python}
> Exception happened during processing of request from ('127.0.0.1', 43418)
> ----------------------------------------
> Traceback (most recent call last):
> File "/opt/cloudera/parcels/Anaconda/lib/python2.7/SocketServer.py", line
> 290, in _handle_request_noblock
> self.process_request(request, client_address)
> File "/opt/cloudera/parcels/Anaconda/lib/python2.7/SocketServer.py", line
> 318, in process_request
> self.finish_request(request, client_address)
> File "/opt/cloudera/parcels/Anaconda/lib/python2.7/SocketServer.py", line
> 331, in finish_request
> self.RequestHandlerClass(request, client_address, self)
> File "/opt/cloudera/parcels/Anaconda/lib/python2.7/SocketServer.py", line
> 652, in __init__
> self.handle()
> File
> "/opt/cloudera/parcels/SPARK2-2.3.0.cloudera4-1.cdh5.13.3.p0.611179/lib/spark2/python/lib/pyspark.zip/pyspark/accumulators.py",
> line 263, in handle
> poll(authenticate_and_accum_updates)
> File
> "/opt/cloudera/parcels/SPARK2-2.3.0.cloudera4-1.cdh5.13.3.p0.611179/lib/spark2/python/lib/pyspark.zip/pyspark/accumulators.py",
> line 238, in poll
> if func():
> File
> "/opt/cloudera/parcels/SPARK2-2.3.0.cloudera4-1.cdh5.13.3.p0.611179/lib/spark2/python/lib/pyspark.zip/pyspark/accumulators.py",
> line 251, in authenticate_and_accum_updates
> received_token = self.rfile.read(len(auth_token))
> TypeError: object of type 'NoneType' has no len()
>
> {code}
>
> Error happens here:
> https://github.com/apache/spark/blob/cb90617f894fd51a092710271823ec7d1cd3a668/python/pyspark/accumulators.py#L254
> The PySpark code was just running a simple pipeline of
> binary_rdd = sc.binaryRecords(full_file_path, record_length).map(lambda .. )
> and then converting it to a dataframe and running a count on it.
> It seems error is flaky - on next rerun it didn't happen.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
