Logstash silently fails on trying to import csv


#1

Hello,

I'm trying to import a csv file into ELK and it fails silently even with output set to stdout/debug.

    $ cat /vagrant/sinkhole.csv | /opt/logstash/bin/logstash -f /vagrant/logstash-sinkhole.conf  | tee log
Jun 14, 2015 8:01:51 PM org.elasticsearch.node.internal.InternalNode <init>
INFO: [logstash-elk-vagrant-32003-7948] version[1.5.1], pid[32003], build[5e38401/2015-04-09T13:41:35Z]
Jun 14, 2015 8:01:51 PM org.elasticsearch.node.internal.InternalNode <init>
INFO: [logstash-elk-vagrant-32003-7948] initializing ...
Jun 14, 2015 8:01:51 PM org.elasticsearch.plugins.PluginsService <init>
INFO: [logstash-elk-vagrant-32003-7948] loaded [], sites []
Jun 14, 2015 8:01:54 PM org.elasticsearch.node.internal.InternalNode <init>
INFO: [logstash-elk-vagrant-32003-7948] initialized
Jun 14, 2015 8:01:54 PM org.elasticsearch.node.internal.InternalNode start
INFO: [logstash-elk-vagrant-32003-7948] starting ...
Jun 14, 2015 8:01:54 PM org.elasticsearch.transport.TransportService doStart
INFO: [logstash-elk-vagrant-32003-7948] bound_address {inet[/0:0:0:0:0:0:0:0:9302]}, publish_address {inet[/10.0.2.15:9302]}
Jun 14, 2015 8:01:54 PM org.elasticsearch.discovery.DiscoveryService doStart
INFO: [logstash-elk-vagrant-32003-7948] elasticsearch/XZ4SjlIgQZWnvmrjEJ6Iag
Jun 14, 2015 8:01:57 PM org.elasticsearch.cluster.service.InternalClusterService$UpdateTask run
INFO: [logstash-elk-vagrant-32003-7948] detected_master [Grizzly][UMAeGzFgRb2TrVyY-0CjJw][elk-vagrant][inet[/127.0.0.1:9300]]{max_local_storage_nodes=1}, added {[Grizzly][UMAeGzFgRb2TrVyY-0CjJw][elk-vagrant][inet[/127.0.0.1:9300]]{max_local_storage_nodes=1},[logstash-elk-vagrant-31278-9802][pHxHzJoGSyiP03Fz_O6QFw][elk-vagrant][inet[/10.0.2.15:9301]]{client=true, data=false},}, reason: zen-disco-receive(from master [[Grizzly][UMAeGzFgRb2TrVyY-0CjJw][elk-vagrant][inet[/127.0.0.1:9300]]{max_local_storage_nodes=1}])
Jun 14, 2015 8:01:57 PM org.elasticsearch.node.internal.InternalNode start
INFO: [logstash-elk-vagrant-32003-7948] started
$

I'm using a vagrant ELK and config for sinkhole as present here https://github.com/juju4/ELK
Re-reading of file is force with since_db and start_position

Any pointer to understand why nothing is parsed and why no warning/errors?

Thanks


(Mark Walkom) #2

I've had this happen if the file doesn't contain a CR/LF at the end of it, I'd try opening it in vim etc and then adding one and see if that works.


#3

Thanks Mark. Already tried that. Checked again to be sure but it still fails.


(Mark Walkom) #4

What does your config look like?
Have you tried to change the output to just stdout to make sure data is making it through?


#5

you can see the config in the github link and yeah, I tried with and without "stdout { codec => rubydebug }"


(Christian Dahlqvist) #6

In the example output you are sending data into Logstash through stdin, although in the configuration file you have not defined stdin as an input, only a file path that is different from the one you are testing with: /home/vagrant/logs/sinkhole-*.csv. Does this location contain any files to be processed?


#7

I tried both as input file (yes log dir has one file) and stdin input and in both case, I just have logstash starting but displaying nothing else.
Elasticsearch stays empty

$ /opt/logstash/bin/logstash -f /vagrant/logstash-sinkhole.conf
$ cat file | /opt/logstash/bin/logstash -f /vagrant/logstash-sinkhole.conf
[...]
INFO: [logstash-elk-vagrant-20285-9784] started

and after nothing more


(Mark Walkom) #8

You won't get anything else as the output sends to ES, not stdout.


#9

Like input, I tried both configuration (output stdout or ES) and checked both without anythink.
only have the index properties/fields made available but no data.


(system) #10