DistributedR :: Issue HPdata with kerberos based hadoop environment.

[hemendray120]

Hi,

I am trying to load csv file from hadoop hdfs (Cloudera 5.8.0+Kerberos) but receiving the error.

Rscript: library(HPdata) library(distributedR) distributedR_start() Sys.setenv(DEBUG_DDC=TRUE) system("kinit -kt ") mydframe <- csv2dframe(url='hdfs:///user/<User>/Sample.csv',schema='A1:character,A2:character,A3:character', hdfsConfigurationFile='/home/<User>/hdfsconfig.json')

Error: response-parse: lexical error: invalid char in json text.

<meta http-equiv (right here) ------^

Error: basic_string::_S_construct null not valid

hdfsconfig.json file content:

{ "webhdfsPort": 50070, "hdfsPort": 8020, "hdfsHost": "<Host>", "hdfsUser": "<User>" }

Not sure whether kerberos authentication is supported in HPdata package.

I have tried by running wget command on the webhdfs url and able to get the json response after kerberos authentication.So it seems issue with kerberos authentication in HPdata package. Do we have any configuration or workaround to this.

Thanks

[hemendray120]

Is there any update on this.

Thanks,