I am looking for assistance with unwanted fields extracted automatically.
I am using a custom sourcetype that I added with a field extraction based on regex. This regex extracts four fields: "thread_name", "log_level", "event_category", and "messages". This works correctly, except for when I click the "xx more fields" link under Interesting Fields in the sidebar of search. That is where I see the unwanted fields are listed, and when I examine an event with one of those fields, I can see the field/value pairs are listed under the Event. The four fields specified in my regex are extracted correctly; I just want to suppress the extraction of the "fields" from within the SQL queries.
Following the advice from another post, I added KV_MODE=none to props.conf on the forwarder and reindexed the data, but the issue still occurred. I then added KV_MODE=none to props.conf on the indexer, and reindexed the data, but I am still seeing key/value pairs extracted from the SQL queries.
Can you please advise me of any other recommendations to stop this from happening?
I verified no data re-indexing is required. Thank you!
You need to deploy this to your Search Head tier, not the Indexers.
The deployment consists of a single server running Splunk Enterprise, and forwarders installed on various hosts. Can you confirm if need to deploy this at $SPLUNK_HOME/etc/apps/search/local/props.conf on the server ?
I deployed it to $SPLUNK_HOME/etc/apps/search/local/props.conf and it worked. Thank you!
The KV_MODE=none
is the search time field extraction setting and should be set on the Search Head. No data re-indexing is required.
Thank you!
Did you restart the Splunk service after making those props changes?
Yes. I restarted the Splunk service on the forwarder, and stopped/started Splunk on the indexer.