Platform: CDH5.0.0 + Splunk 6.1 and Hunk. We have configured virtual indexes. We are successfully issuing queries against Splunk and getting data back from MR jobs. When MR jobs run, should they produce splunk index files, so next time Splunk makes a query against this data it doesn't have to run an MR job again? I see result files with names like 1401212205.1082/0/part-m-00043 (1 per MR task, I think). Each file contains what looks like a hash and a path to a different file. Example 81a76d523faab948b16fcd2eb6dd56a0 /user/splunk/scratch/dispatch/1401212205.1082/0/splunk-m-0000043 But the splunk-m-0000043 file (or any like them) do not exist at that directory location. Are they supposed to exist? ... View more