An error is reported when hue executes impala query: internalexception: memory limit exceeded: error occurred on backend
The error is caused by too many impala connections, which can be released by modifying the number of connections or other connections
When using impala query, an error will be reported when using cast function to convert data to decimal type:
Bad status for request TFetchResultsReq(fetchType=0, operationHandle=TOperationHandle(hasResultSet=True,
modifiedRowCount=None, operationType=0, operationId=THandleIdentifier(secret='\xc8\xc1j\xfe\xd6.B\xa2\x9e\xc8\x1a63\xfb\x15\xf2'
, guid='\xf8\x82.\x81^\xabE\x7f\x00\x00\x00\x00\xda\xf8\xa9\xa1')), orientation=4, maxRows=100)
: TFetchResultsResp(status=TStatus(errorCode=None, errorMessage='UDF ERROR: Decimal expression
overflowed\n', sqlState='HY000', infoMessages=None, statusCode=3), results=None, hasMoreRows=None)
After checking a wave, I found that it was a bug in impala
bug link: https://issues.apache.org/jira/browse/IMPALA-6405
It seems that the bug still exists. Here is a screenshot of my test:
Solution: do not convert to decimal, or convert in other ways.