We just implemented search-head pooling. I have a lookup that gets generated via wget. How do people handle situations like this? I don't want to tie the scripts to a specific server, but I also want to make sure it runs when one of the servers is down. Is there a way to have Splunk run it, but not generate indexed data?
One option: create a Python search command that updates your lookup. Then, schedule a search that invokes that search command on a desired interval (e.g. every night at midnight). The search heads in your pool will coordinate such that only one instance runs the scheduled search "at a time" (i.e. only one instance in the pool will run the scheduled search, per interval).