Hazelcast discovery through marathon, blacklisting all hosts.

[Andreas Joelsson]

Hi!

we have a marathon managed cluster using dcos. So I try to connect through the other instances with https://github.com/phaneesh/hazelcast-marathon-discovery

Config config = new Config();

//This is important to enable the discovery strategy

config.setProperty(GroupProperty.DISCOVERY_SPI_ENABLED, "true");

config.setProperty(GroupProperty.DISCOVERY_SPI_PUBLIC_IP_ENABLED, "true");

config.setProperty(GroupProperty.SOCKET_CLIENT_BIND_ANY, "false");

config.setProperty(GroupProperty.SOCKET_BIND_ANY, "false");

NetworkConfig networkConfig = config.getNetworkConfig();

networkConfig.setPublicAddress(System.getenv("HOST") + ":" + System.getenv("PORT_5071"));

for(Map.Entry<String, String> item : System.getenv().entrySet()) {

    log.info(item.getKey() + " => " + item.getValue());

}

JoinConfig joinConfig = networkConfig.getJoin();

joinConfig.getTcpIpConfig().setEnabled(false);

joinConfig.getMulticastConfig().setEnabled(false);

joinConfig.getAwsConfig().setEnabled(false);

DiscoveryConfig discoveryConfig = joinConfig.getDiscoveryConfig();

//Set the discovery strategy to RangerDiscoveryStrategy

DiscoveryStrategyConfig discoveryStrategyConfig = new DiscoveryStrategyConfig(new MarathonDiscoveryStrategyFactory());

//Marathon master endpoint that can provide app metadata

discoveryStrategyConfig.addProperty("marathon-endpoint", "http://XXXX:8080");

//Namespace that needs to be used by ranger for this service

discoveryStrategyConfig.addProperty("app-id", "YYYY");

//Hazelcast port index that is used in marathon configuration

discoveryStrategyConfig.addProperty("port-index", "2");

discoveryConfig.addDiscoveryStrategyConfig(discoveryStrategyConfig);

this.hzInstance = Hazelcast.newHazelcastInstance(config);

And here is the output, deploying 2 docker images towards marathon:

Node 1:

06:45:17.712 [main] INFO  com.hazelcast.instance.Node - [#1.#1.#1.#1]:8652 [dev] [3.6.4] Activating Discovery SPI Joiner
06:45:17.716 [main] INFO  com.hazelcast.core.LifecycleService - [#1.#1.#1.#1]:8652 [dev] [3.6.4] Address[#1.#1.#1.#1]:8652 is STARTING
06:45:17.787 [main] INFO  c.h.n.t.n.NonBlockingIOThreadingModel - [#1.#1.#1.#1]:8652 [dev] [3.6.4] TcpIpConnectionManager configured with Non Blocking IO-threading model: 3 input threads and 3 output threads
06:45:17.787 [main] DEBUG c.h.n.t.n.NonBlockingIOThreadingModel - [#1.#1.#1.#1]:8652 [dev] [3.6.4] IO threads selector mode is SELECT
06:45:17.803 [cached3] INFO  c.h.cluster.impl.DiscoveryJoiner - [#1.#1.#1.#1]:8652 [dev] [3.6.4] Address[#2.#2.#2.#2]:13881 is added to the blacklist.
06:45:17.803 [cached2] INFO  c.h.cluster.impl.DiscoveryJoiner - [#1.#1.#1.#1]:8652 [dev] [3.6.4] Address[#2.#2.#2.#2]:13280 is added to the blacklist.
06:45:17.803 [cached1] INFO  c.h.cluster.impl.DiscoveryJoiner - [#1.#1.#1.#1]:8652 [dev] [3.6.4] Address[#1.#1.#1.#1]:16735 is added to the blacklist.
06:45:18.801 [main] INFO  c.h.cluster.impl.DiscoveryJoiner - [#1.#1.#1.#1]:8652 [dev] [3.6.4]

Members [1] {
 Member [#1.#1.#1.#1]:8652 this
}

06:45:18.801 [main] WARN  com.hazelcast.instance.Node - [#1.#1.#1.#1]:8652 [dev] [3.6.4] Config seed port is 5701 and cluster size is 1. Some of the ports seem occupied!
06:45:18.831 [main] INFO  com.hazelcast.core.LifecycleService - [#1.#1.#1.#1]:8652 [dev] [3.6.4] Address[#1.#1.#1.#1]:8652 is STARTED

Node 2:

06:45:17.654 [cached3] INFO  c.h.cluster.impl.DiscoveryJoiner - [#2.#2.#2.#2]:13280 [dev] [3.6.4] Address[#2.#2.#2.#2]:13881 is added to the blacklist.
06:45:17.654 [cached1] INFO  c.h.cluster.impl.DiscoveryJoiner - [#2.#2.#2.#2]:13280 [dev] [3.6.4] Address[#1.#1.#1.#1]:16735 is added to the blacklist.
06:45:17.654 [cached2] INFO  c.h.cluster.impl.DiscoveryJoiner - [#2.#2.#2.#2]:13280 [dev] [3.6.4] Address[#1.#1.#1.#1]:8652 is added to the blacklist.
06:45:18.653 [main] INFO  c.h.cluster.impl.DiscoveryJoiner - [#2.#2.#2.#2]:13280 [dev] [3.6.4]

Members [1] {
 Member [#2.#2.#2.#2]:13280 this
}

06:45:18.653 [main] WARN  com.hazelcast.instance.Node - [#2.#2.#2.#2]:13280 [dev] [3.6.4] Config seed port is 5701 and cluster size is 1. Some of the ports seem occupied!
06:45:18.679 [main] INFO  com.hazelcast.core.LifecycleService - [#2.#2.#2.#2]:13280 [dev] [3.6.4] Address[#2.#2.#2.#2]:13280 is STARTED

The instances being blacklisted with port 16735 and 13881 is the nodes being replaced, however the blacklist is giving small information. I Currently have debug level active on hazel cast.

// Andreas

[Noctarius]

He Andreas,

The Marathon discovery is a community project. You’re probably better of to ask at their own github tracker if they’re not reading here.

Chris

--
You received this message because you are subscribed to the Google Groups "Hazelcast" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hazelcast+...@googlegroups.com.
To post to this group, send email to haze...@googlegroups.com.
Visit this group at https://groups.google.com/group/hazelcast.
To view this discussion on the web visit https://groups.google.com/d/msgid/hazelcast/977fb3d3-8bba-4863-98ba-521f7adc2c9b%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.