"Exception - com.splunk.mr.JobStartException: Failed to start MapReduce job."

[dannili]

I just moved my whole dashboard to production environment but when I tried to test using a search string, following exception and error occur:

2 errors occurred while the search was executing. Therefore, search results might be incomplete. Hide errors.   
•[hdfsprovider] Error while running external process, return_code=255. See search.log for more info 
•[hdfsprovider] Exception - com.splunk.mr.JobStartException: Failed to start MapReduce job. Please consult search.log for more information. Message: Failed to create Hunk directories due to permission denied. Please make sure you have write permission to the directory specified by vix.splunk.home.hdfs in your provider 

**********Search

 index="my_index" | stats count(eval( OverallAvgNetworkMOS < 3.5 OR Stream_1_JitterInterArrival > 30 OR Stream_2_JitterInterArrival > 30 OR Stream_1_RoundTrip > 500 OR Stream_2_RoundTrip > 500 OR Stream_1_PacketLossRate>0.1 OR Stream_2_PacketLossRate>0.1)) AS bad_call count(eval(OverallAvgNetworkMOS >= 3.5)) AS good_call | transpose

I checked other posts related to error but could not find a suitable solution. Is this knida search related to config?

Does anyone know how these two can happen? How do I fix this? Thanks a lot!

[mwdbhyat · Accepted solution]

Are you using Hunk ? To me it looks like its a permissions issue with virtual index - check this out https://docs.splunk.com/Documentation/Hunk/6.4.10/Hunk/Setupavirtualindex

Have you checked search.log ?

Does it only do it when you run that specific search, or any search ?

Are your DEV and PROD environments using the same architecture/setup ?

[dannili]

thanks for ur response. Turns out it's the setup issue. System team helped fix it.