That's what I get for becoming our internal package maintainer for pmacct
by default - updating it now! :D
One quick sanity check regarding output to Kafka - would setting
kafka_refresh_time to 1 add significant overhead to nfacctd, e.g. related
to buffer flushing process/etc?
My goal is to produce something closer to a live stream, but with
micro-flows aggregated to 1s granularity, then use Spark to do additional
transformations/aggregations against local data along with basic anomaly
On Fri, Feb 17, 2017 at 5:58 AM, Paolo Lucente <pa...@pmacct.net> wrote:
> Hi Aaron,
> The feature is post 1.6.1. Can you please switch to master code on GitHub?
> On Thu, Feb 16, 2017 at 10:44:23AM -0800, Aaron Finney wrote:
> > Hi Paolo/all,
> > I've been unable to get nfacctd to send the Avro schema to a Kafka topic
> > I receive the following message when starting nfacctd:
> > WARN: [/etc/nfacctd.conf:14] Unknown key: kafka_avro_schema_topic.
> > WARN: [/etc/nfacctd.conf:15] Unknown key: kafka_avro_schema_refresh_
> > Ignored.
> > Kafka/Avro are otherwise working; the process successfully writes the
> > schema to a local file specified in avro_schema_output_file, but I'd like
> > to have it send to a schema topic in the cluster itself if possible.
> > Version info is below.
> > Thanks!
> > Aaron
> > INFO ( default/core ): NetFlow Accounting Daemon, nfacctd 1.6.1
> > (20161001-00+c5)
> > INFO ( default/core ): '--enable-kafka' '--enable-jansson'
> > 'KAFKA_LIBS=-L/usr/lib64/ -lrdkafka' 'AVRO_CFLAGS=-I/usr/include/avro'
> > 'AVRO_LIBS=-L/usr/lib -lavro'
> > _______________________________________________
> > pmacct-discussion mailing list
> > http://www.pmacct.net/#mailinglists
> pmacct-discussion mailing list
pmacct-discussion mailing list