We have an issue when we are trying to collect syslogdata from a filesystem on a syslog server with help from a deployment-app.
The folder structure looks like this, all data is being written to /var/log/netlogs/"IP-adress of the device"/messages.log
Here is the inputs.conf file we are using right now (we have experimented back and forward with crcSalt, initCrcLength)
sourcetype = syslog
whitelist = .log$
SHOULD_LINEMERGE = false
crcSalt = <string>
initCrcLength = 6250
ignoreOlderThan = 10m
index = network
Any ides what we need to do in order to index every single messages.conf in this filestructure
How about just this
[monitor:///var/log/netlogs/*/messages.log] sourcetype = syslog SHOULD_LINEMERGE = false crcSalt = <string> initCrcLength = 6250 ignoreOlderThan = 10m index = network
Have you looked at https://wiki.splunk.com/Community:Troubleshooting_Monitor_Inputs in particular https://your-splunk-server:8089/services/admin/inputstatus/TailingProcessor:FileStatus ? Can be quite helpful in seeing why files aren't being processed.
My guess would be "ignoreOlderThan = 10m" is the issue. This means that Splunk ignores any files that is older than 10minutes when it first sees the file (on startup) even if it is written to later on while Splunk is still running. The only way to get Splunk to process the files once it has 'tagged' them as old, is to restart Splunk.