Druid.host - set to resolvable endPoint but Druid adds 'port' - corrupting end point

Hi Everyone,

I have set the druid.host, say for different nodes to a resolvable endpoint without the port. For ex.,

Properties set:

druid.host=druid-coordinator.devel.internal -> this is a resolvable end point

druid.plainTextPort=31973

Below are the traces from the historical node - thats fails to resolve the endpoint

2018-04-18T14:46:56,391 INFO [main] io.druid.discovery.DruidLeaderClient - Request[http://druid-prototype-coordinator.devel.internal:31973/druid/coordinator/v1/lookups/config/__default?detailed=true] failed with msg [Faulty channel in resource pool].

2018-04-18T14:46:56,391 WARN [HttpClient-Netty-Boss-0] org.jboss.netty.channel.SimpleChannelUpstreamHandler - EXCEPTION, please implement org.jboss.netty.handler.codec.http.HttpContentDecompressor.exceptionCaught() for proper handling. org.jboss.netty.channel.ConnectTimeoutException: connection timed out: druid-prototype-coordinator.devel.internal/10.32.119.202:31973
at org.jboss.netty.channel.socket.nio.NioClientBoss.processConnectTimeout(NioClientBoss.java:139)

Below are the traces from the co-ordinator:
2018-04-18T14:41:32,992 INFO [NodeTypeWatcher[coordinator]] io.druid.curator.discovery.CuratorDruidNodeDiscoveryProvider$NodeTypeWatcher

  • Node[druid-prototype-coordinator.devel.internal:31973:
    DiscoveryDruidNode{druidNode=DruidNode{serviceName=devel/druid-prototype/coordinator’,
    host=‘druid-prototype-coordinator.devel.internal’, port=-1, plaintextPort=31973, enablePlaintextPort=true, tlsPort=-1, enableTlsPort=false}, nodeType=‘coordinator’, services={}}] appeared.

Why does Druid append the port despite providing a resolvable endpoint through druid.host ?

Can someone suggest if I’m missing something?

Best Regards

Varaga

Hi Guys,
As per the documentation, in this link: http://druid.io/docs/latest/configuration/historical.html
“… The host for the current node. This is used to advertise the current processes location as reachable from another node and should generally be specified such that http://${druid.host}/ could actually talk to this process”
Supplying this property druid.host=druid-coordinator.devel.internal it does not seem to work.
http://druid-coordinator.devel.internal - could be accessed over the browser. However from the logs the endpoint url is constructed with the port e.g. **http://druid-coordinator.devel.internal:10090
** Has someone faced this issue? The cluster is managed/deployed through mesos/aurora !

TIA

Varaga

RESOLVED:

I guess it was the AWS ELB, that has hostname mapped to binding port. Not an issue with Druid.