using logstash 2.2.0-1, the config works
output {
elasticsearch {
hosts => ["test-es.corp.com/elasticsearch/"]
ssl => true
cacert => ""
index => "system-%{YYYY.MM.dd}"
}
stdout { codec => rubydebug }
}
in 2.3.1
service logstash configtest
configuration OK
service logstash start
wait a few minute, it dies
logstash.log :
> {:timestamp=>"2016-04-20T12:15:27.146000-0500", :message=>"Pipeline aborted due to error", :exception=>#<LogStash::ConfigurationError: Host 'test-es.corp.com/elasticsearch' was specified, but is not valid! Use either a full URL or a hostname:port string!>,
Fine: I change the output.conf file to :
hosts => ["https://test-es.corp.com/elasticsearch/"]
restart logstash, wait a bit, dies again:
logstash.log now states:
{:timestamp=>"2016-04-20T12:29:32.963000-0500", :message=>"Pipeline aborted due to error", :exception=>#<LogStash::ConfigurationError: A path '/elasticsearch/' has been explicitly specified in the url 'https://test-es.corp.com/elasticsearch/', but you also specified a path of '/'. This is probably a mistake, please remove one setting.>
test-es.corp.com has elasticsearch on port 9200, listening only on localhost host. It has a ssl proxy which will forward request to the localhost:9200 via https://test-es.corp.com/elasticsearch/
I can hit it via a web browser to get to elasticsearch. What is logstash conf wanting.
I'm testing an upgrade before migrating our prod servers to the latest release.