Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Flume >> mail # user >> help configuring flume scribe sink


Copy link to this message
-
help configuring flume scribe sink
I am working on a POC in which I am trying to have flume emulate scribe output by using a flume scribe sink that I found online.  There was no docs with it so I have poked through code to try and figure out how to configure this properly.  I have gotten the it configured to the point where a scribe server receives the data but does not recognize any scribe categories.  I have included my config file below.  In this config file I'm simply doing a tail -f on some bogus log files that I generate for the input and trying to create 3 streams of output similar to what a scribe intake server might be doing in receiving multiple streams of scribe data and funneling it to a scribe integration server.

Has anyone used this flume scribe sink successfully and, if so, can you help me get my config file setup properly?  Thank you.

Regards,
Dwight Marzolf

# Name the components on this agent
agent1.sources = exec-tail-ds1 exec-tail-ds2 exec-tail-ds3
agent1.sinks = scribe-forward-sink1 scribe-forward-sink2 scribe-forward-sink3
agent1.channels = file-channel-1

##########################
# SOURCE
##########################

# Configure source 1
agent1.sources.exec-tail-ds1.type = exec
agent1.sources.exec-tail-ds1.command = tail -F /tmp/dataset1
agent1.sources.exec-tail-ds1.channels = file-channel-1

# Configure source 2
agent1.sources.exec-tail-ds2.type = exec
agent1.sources.exec-tail-ds2.command = tail -F /tmp/dataset2
agent1.sources.exec-tail-ds2.channels = file-channel-1

# Configure source 3
agent1.sources.exec-tail-ds3.type = exec
agent1.sources.exec-tail-ds3.command = tail -F /tmp/dataset3
agent1.sources.exec-tail-ds3.channels = file-channel-1

##########################
# SINK
##########################

# Config for scribe sink 1
agent1.sinks.scribe-forward-sink1.type = org.apache.flume.sink.scribe.ScribeSink
agent1.sinks.scribe-forward-sink1.scribe.sink.name = ds1
agent1.sinks.scribe-forward-sink1.scribe.port = 1463
agent1.sinks.scribe-forward-sink1.scribe.host = 10.48.140.8
agent1.sinks.scribe-forward-sink1.scribe.category = dataset1
agent1.sinks.scribe-forward-sink1.scribe.category.header = category
agent1.sinks.scribe-forward-sink1.channel = file-channel-1

# Config for scribe sink 2
agent1.sinks.scribe-forward-sink2.type = org.apache.flume.sink.scribe.ScribeSink
agent1.sinks.scribe-forward-sink2.scribe.sink.name = ds2
agent1.sinks.scribe-forward-sink2.scribe.port = 1463
agent1.sinks.scribe-forward-sink2.scribe.host = 10.48.140.8
agent1.sinks.scribe-forward-sink2.scribe.category = dataset2
agent1.sinks.scribe-forward-sink2.scribe.category.header = category
agent1.sinks.scribe-forward-sink2.channel = file-channel-1

# Config for scribe sink 3
agent1.sinks.scribe-forward-sink3.type = org.apache.flume.sink.scribe.ScribeSink
agent1.sinks.scribe-forward-sink3.scribe.sink.name = ds3
agent1.sinks.scribe-forward-sink3.scribe.port = 1463
agent1.sinks.scribe-forward-sink3.scribe.host = 10.48.140.8
agent1.sinks.scribe-forward-sink3.scribe.category.header= category
agent1.sinks.scribe-forward-sink3.scribe.category = dataset3
agent1.sinks.scribe-forward-sink3.channel = file-channel-1

##########################
# CHANNEL
##########################

# Channel file buffer 1
agent1.channels.file-channel-1.type = file
agent1.channels.file-channel-1.checkpointDir = /scribe/channel/checkpoint1
agent1.channels.file-channel-1.dataDirs = /scribe/channel/data1

##########################
# BINDING
##########################

# Bind the source and sink to the channel
agent1.sources.exec-tail-ds1.channels = file-channel-1
agent1.sinks.scribe-forward-sink1.channel = file-channel-1
[Description: cid:B446C73C-E88B-4947-BD2A-586ECE21AB0D]
Dwight Marzolf
Sr. Engineer
Data & Analytics
Phone: +1 781 430 5228

HERE Burlington
5 Wayside Rd
42° 29' 7" N. 71° 11' 32" W
HERE, a Nokia business