- Mark as New
- Bookmark Message
- Subscribe to Message
- Mute Message
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Why is Splunk not ingesting .txt logfiles starting with a numeric value?
I am trying to bring in some .txt logfiles using Splunk forwarder. There are several logs in the directory, such as Log.txt, 10Log.txt, 20Log.txt, etc. These files are changed daily, and the 10, 20, etclog.txt files are written to daily. So far, I can only get Splunk to ingest the Log.txt file and nothing else. My inputs.conf file is currently as below. I have tried to monitor just *.txt with the same results. Only Log.txt is read/ingested.
[monitor://E:\Logs\CIR_Remote\*Log.txt]
disabled = false
sourcetype = LOG4NET
index = log4net
initCrcLength=1024
any input would be appreciated!
- Mark as New
- Bookmark Message
- Subscribe to Message
- Mute Message
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content

I don't know why (didn't bother to investigate it) but I noticed that sometimes specifying wildcard path to files doesn't work whereas specifying whole directory with a whitelist matching only set of files does.
Unfortunately, you can't specify two different sets of files from the same directory since each input stanza must be unique.
Anyway, check your config with
splunk btool inputs list --debug
and verify list of monitored files with
splunk list monitor
- Mark as New
- Bookmark Message
- Subscribe to Message
- Mute Message
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content

Check if first few lines/characters of all those files (Log.txt and <N>Log.txt) have some common header/characters. If they've some common header/characters, Splunk will think that they're rolled log files. To resolve it, you'll need to adjust initCrcLength value to be slighly more that the length of those common header/characters.
- Mark as New
- Bookmark Message
- Subscribe to Message
- Mute Message
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
No, the files are not the same. They start with a datetime stamp, and then include error information. They have never been ingested at all, always skipped for some reason.
