Solr container doesn't work with cloud (cluster) Solr client

[acsbendi]

It is not possible to use the Solr container with the cloud Solr client (CloudSolrClient, this client can be used for accessing Solr clusters), because the following exception is thrown when trying to send any request:

Tried fetching cluster state using the node names we knew of, i.e. [localhost:8983_solr]. However, succeeded in obtaining the cluster state from none of them.If you think your Solr cluster is up and is accessible, you could try re-creating a new CloudSolrClient using working solrUrl(s) or zkHost(s).
java.lang.RuntimeException: Tried fetching cluster state using the node names we knew of, i.e. [localhost:8983_solr]. However, succeeded in obtaining the cluster state from none of them.If you think your Solr cluster is up and is accessible, you could try re-creating a new CloudSolrClient using working solrUrl(s) or zkHost(s).
	at org.apache.solr.client.solrj.impl.BaseHttpClusterStateProvider.getState(BaseHttpClusterStateProvider.java:105)
	at org.apache.solr.client.solrj.impl.BaseCloudSolrClient.resolveAliases(BaseCloudSolrClient.java:1182)
	at org.apache.solr.client.solrj.impl.BaseCloudSolrClient.requestWithRetryOnStaleState(BaseCloudSolrClient.java:894)
	at org.apache.solr.client.solrj.impl.BaseCloudSolrClient.request(BaseCloudSolrClient.java:866)
	at org.apache.solr.client.solrj.SolrRequest.process(SolrRequest.java:214)
	at org.apache.solr.client.solrj.SolrClient.add(SolrClient.java:106)
	at org.apache.solr.client.solrj.SolrClient.add(SolrClient.java:71)
	at com.example.SolrQueryTest.setUp(SolrQueryTest.java:37)
...

You can see this exception after changing this line to
solrClient = new CloudSolrClient.Builder(Collections.singletonList("http://" + solrContainer.getContainerIpAddress() + ":" + solrContainer.getSolrPort() + "/solr")).build(); and running any test in SolrQueryTest.java.

The underlying issue is that the cluster will report 8983 as the port used by Solr, but the container uses a different host port, so localhost:8983 will not be available to the client.

[kiview]

We traditionally had similar challenges in KafkaContainer due to ZooKepper. In the latest iteration, those were solved by updating the advertised connection after the container has started and the mapped port is known:

testcontainers-java/modules/kafka/src/main/java/org/testcontainers/containers/KafkaContainer.java

Lines 109 to 119 in 4df9b90

	protected void containerIsStarted(InspectContainerResponse containerInfo) {
	String brokerAdvertisedListener = brokerAdvertisedListener(containerInfo);
	ExecResult result = execInContainer(
	"kafka-configs",
	"--alter",
	"--bootstrap-server", brokerAdvertisedListener,
	"--entity-type", "brokers",
	"--entity-name", getEnvMap().get("KAFKA_BROKER_ID"),
	"--add-config",
	"advertised.listeners=[" + String.join(",", getBootstrapServers(), brokerAdvertisedListener) + "]"
	);

I lack the knowledge of Solr to understand if such an approach would be feasible here as well.

The other alternative is starting Solr with the connection URLs as part of the configure() callback.

[acsbendi]

Thanks for the fast response. As far as I know, it's not possible to change the port of a Solr node after it has been registered. Also, there is no concept of advertised URL or port and the client simply uses the name/ID of the node as the URL to connect to it.

The other alternative works, in fact, that is how we test our code that uses CloudSolrClient. More specifically, we find 2 free ports for Solr and Zookeeper, have the container expose them, and then set the SOLR_PORT environment variable. Of course, finding free ports in our own code is less nice than having Docker do it for us, but that's the only way we manage to make it work currently.