Really long events being broken up

[robgreen]

I have an event being imported with a custom source type. in that source type i have

NO_BINARY_CHECK=1
CHECK_FOR_HEADER=false
LEARN_SOURCETYPE=false
SHOULD_LINEMERGE=false

However splunk is still truncating my log lines and then generating a new event with the rest of the line (potentially broken up again) generating incorrect data. Is there a way i can tell splunk to import the whole log line into one event? The event log line can be up to 128k. I am fine with it being truncated in the display but not in the indexed data. alternatively i am fine with any one field being limited to a certain size (such as 4k) but as it stands now any fields after the really long field is missing.

thanks,
rob

[cramasta]

Add the following to what you already have in your props file

MAX_EVENTS = 10000

TRUNCATE = 0

This will cause the data to not truncate no matter how many lines you have and will break the event into a new event after 10000 lines. If you have more thank 10000 lines in a single event then increase this number accordingly

[cramasta]

Truncate=0 only stops splunk from discarding data that should be indexed after some number of characters in a single event is reached, the exact limit of characters i dont remember off the top of my head.

[cramasta]

Max_events will not do that.
Linebreaking will still occur by whatever you have defined in your config which from what it looks like is to make a new event when it detects a timestamp

MAX_EVENTS = 10000 will allow a single event to go beyond the Splunk default of 256 lines per event. This is the solution for the problem you described where the remaining part of a single event was 'overflowed' into a new event.

So for instance if your single event was 300 lines long, 256 lines will go into one event and the remaining 44 lines will be placed into a new event

[robgreen]

I think just adding TRUNCATE=0 is what i needed. I don't want to join any separate lines into the same event.