Nix App netstat.sh line breaking on some servers

[glitchcowboy]

I've got the nix app running on my search head, and 300 or so nix Forwarders sending data in. On some of the forwarders, the netstat.sh output gets broken into one event per line in a search, but the others have a single event per netstat.sh run.

This is consistent whether the OS is AIX or RHEL.

Has anyone ever seen this before?

[glitchcowboy · Accepted solution]

It took me several days to find it (with the expert leading of mlanghor).

Turns out that the unix app was disabled on the indexers, so they didn't know about the props.conf settings used to parse the netstat.sh input.

The key troubleshooting tip was:

$SPLUNK_HOME/bin/splunk cmd btool props list --debug netstat

On a test indexer I built just for troubleshooting this, it returns data about the netstat props config. On my production indexer, I got a blank stare (no output).