We're using 'batch' stanza on our Splunk forwarders so they delete the log files once they've been indexed. Obviously, this means we've lost Splunk's capability of detecting duplicates. We're trying to set something up manually so we can detect if Splunk has already indexed the file before submitting that file to indexers. For a duplicate, file name and file data is exactly the same. Do you have any suggestions/pointers on how to set this up? Thanks.
If you are using batch and it is working correctly, it should be impossible for any files's data to be indexed more than once UNLESS the same file appears on more than 1 forwarder OR the same data appears in more than 1 file. Splunk has built-in protections to disallow the latter case (see crcSalt) so the most likely situation is that you have the same file being sent to more than 1 forwarder. For your duplicates, do the have the same host as well (or is host overridden)? Share some more detail about why you think you have duplicate events (I am highly skeptical).