I am monitoring with a forwarder logs file that are being written.
And sometimes the events indexed are broken in multiple events in the middle of the line :
example :
Thu Sep 19 17:13:32 PDT 2013 myhost my event and the end of my line
becomes
event1 :
Thu Sep 19 17:13:32 PDT 2013 myhost my event and
event2
the end of my line
If the event is but in the middle of the line, it can be a write buffer issue.
by example if the End of Line is reached when Splunk is reading the file.
To confirm, you can index the same log file separately in a test index, and see if the events are properly indexed with a static file. If this is confirmed, the option is to increase the time_before_close on the inputs.conf
see time_before_close =
* Modtime delta required before Splunk can close a file on EOF.
* Tells the system not to close files that have been updated in past
* Defaults to 3.
http://docs.splunk.com/Documentation/Splunk/latest/admin/Inputsconf
If the event is but in the middle of the line, it can be a write buffer issue.
by example if the End of Line is reached when Splunk is reading the file.
To confirm, you can index the same log file separately in a test index, and see if the events are properly indexed with a static file. If this is confirmed, the option is to increase the time_before_close on the inputs.conf
see time_before_close =
* Modtime delta required before Splunk can close a file on EOF.
* Tells the system not to close files that have been updated in past
* Defaults to 3.
http://docs.splunk.com/Documentation/Splunk/latest/admin/Inputsconf
I confirm the application generating the log file is writing by chunks.
It doesn't happen if I index the file after.