Unable to query Infinispan on JBoss 7 due to SearchFactoryIntegrator not being in the registry

Question

TL:DR I'm getting the following error and I can see no reason for it, indexing is enabled, and the SearchFactoryIntegrator is on the classpath for everything that needs to see it.

Indexing was not enabled on this cache. interface org.hibernate.search.spi.SearchFactoryIntegrator not found in registry

Long Version:

• JBoss version: 7.1.1 FINAL
• Infinispan version: 5.16 (upgraded from 5.1.2 JBoss ships with in attempts to solve issue)
• Java Version: 1.7

I am attempting to build a simple test application to run queries against an Infinispan cache.

I am successfully adding objects to the cache. The cache is defined as private org.infinispan.Cache<String, Agent> cache; with Agent being my own object.

When attempting to query the cache I get the following error from the line SearchManager searchManager = Search.getSearchManager(cache);

17:35:43,873 ERROR [org.apache.catalina.core.ContainerBase.[jboss.web].[default-host].[/SearchableCacheTest].[Faces Servlet]] (http-localhost-127.0.0.1-8080-1) Servlet.service() for servlet Faces Servlet threw exception: java.lang.IllegalArgumentException: Indexing was not enabled on this cache. interface org.hibernate.search.spi.SearchFactoryIntegrator not found in registry
    at org.infinispan.query.impl.ComponentRegistryUtils.getComponent(ComponentRegistryUtils.java:40) [infinispan-query-5.1.6.FINAL.jar:5.1.6.FINAL]
    at org.infinispan.query.SearchManagerImpl.<init>(SearchManagerImpl.java:56) [infinispan-query-5.1.6.FINAL.jar:5.1.6.FINAL]
    at org.infinispan.query.Search.getSearchManager(Search.java:39) [infinispan-query-5.1.6.FINAL.jar:5.1.6.FINAL]

The important part being

Indexing was not enabled on this cache. interface org.hibernate.search.spi.SearchFactoryIntegrator not found in registry

Prior to this error cache.getCacheConfiguration().indexing().enabled() resolves to true

The cache is a modified version of the default "cluster" cache already configured in the standalone-ha.xml of JBoss.

<cache-container name="cluster" aliases="ha-partition" default-cache="default" start="EAGER">
    <transport lock-timeout="60000"/>
    <replicated-cache name="default" mode="SYNC" batching="true" indexing="LOCAL">
        <locking isolation="REPEATABLE_READ"/>
        <store class="org.infinispan.loaders.file.FileCacheStore" preload="true" passivation="false" fetch-state="true" purge="false">
            <property name="location">
                /tempCacheIndex
            </property>
        </store>
     </replicated-cache>
</cache-container>

I see the index being created in the configured file location so it seems that enabled boolean resolving to true is a factual representation of the current state.

I have created an org.infinispan.query module in JBoss and it contains the following

• avro-1.5.1.jar
• hibernate-commons-annotations-4.0.1.Final.jar
• hibernate-search-engine-4.1.1.Final.jar
• infinispan-query-5.1.6.FINAL.jar
• jackson-core-asl-1.9.2.jar
• jackson-mapper-asl-1.9.2.jar
• lucene-core-3.5.0.jar
• paranamer-2.3.jar
• slf4j-api-1.6.1.jar
• snappy-java-1.0.4.1.jar

All of the above are defined in as resource-roots in the module.xml for that module. I have modified the module.xml for the module org.infinispan to have a dependency on the module org.infinispan.query

I have declared dependencies on both those modules in my jboss-deployment-structure.xml

Does anyone have any ideas why this might be happening and perhaps have a solution?

Note: There is a very similar question here but he was ultimately unsuccessful and chose a different route.

Answer 1

NOTE: I started with a clean JBoss 7.1.1 instance and hence reverted from Infinispan 5.1.6 to 5.1.2

For my needs I modified the existing cluster cache so it looks like this

<cache-container name="cluster" aliases="ha-partition"
default-cache="default" start="EAGER">
<transport lock-timeout="60000" />
<replicated-cache name="default" mode="SYNC" batching="true"
    indexing="LOCAL">
    <locking isolation="REPEATABLE_READ" />
    <store class="org.infinispan.loaders.file.FileCacheStore"
        preload="true" passivation="false" fetch-state="true" purge="false">
        <property name="location">
            /tempCacheIndex
        </property>
    </store>
</replicated-cache>

Then to actually be able to query the cache create the org.infinispan.query module containing the jars mentioned in the following module.xml

<module xmlns="urn:jboss:module:1.1" name="org.infinispan.query">
<resources>
    <resource-root path="avro-1.5.1.jar"/>
    <resource-root path="hibernate-commons-annotations-4.0.1.Final.jar"/>
    <resource-root path="hibernate-search-4.1.0.Beta1.jar"/>
    <resource-root path="hibernate-search-analyzers-4.1.0.Beta1.jar"/>
    <resource-root path="hibernate-search-engine-4.1.0.Beta1.jar"/>
    <resource-root path="hibernate-search-orm-4.1.0.Beta1.jar"/>
    <resource-root path="infinispan-query.jar"/>
    <resource-root path="jackson-core-asl-1.9.2.jar"/>
    <resource-root path="jackson-mapper-asl-1.9.2.jar"/>
    <resource-root path="lucene-analyzers-3.5.0.jar"/>
    <resource-root path="lucene-core-3.5.0.jar"/>
    <resource-root path="lucene-grouping-3.5.0.jar"/>
    <resource-root path="lucene-highlighter-3.5.0.jar"/>
    <resource-root path="lucene-memory-3.5.0.jar"/>
    <resource-root path="lucene-misc-3.5.0.jar"/>
    <resource-root path="lucene-smartcn-3.5.0.jar"/>
    <resource-root path="lucene-spatial-3.5.0.jar"/>
    <resource-root path="lucene-spellchecker-3.5.0.jar"/>
    <resource-root path="lucene-stempel-3.5.0.jar"/>
    <resource-root path="paranamer-2.3.jar"/>
    <resource-root path="slf4j-api-1.6.1.jar"/>
    <resource-root path="snappy-java-1.0.4.1.jar"/>
    <resource-root path="solr-analysis-extras-3.5.0.jar"/>
    <resource-root path="solr-commons-csv-3.5.0.jar"/>
    <resource-root path="solr-core-3.5.0.jar"/>
    <resource-root path="solr-solrj-3.5.0.jar"/>
</resources>

<dependencies>
    <module name="javax.api"/>
    <module name="javax.transaction.api"/>
    <module name="org.hibernate"/>
    <module name="org.infinispan" services="import" export="true"/>
    <module name="org.apache.commons.codec"/>
    <module name="org.apache.commons.io"/>
    <module name="org.apache.commons.lang"/>
    <module name="com.google.guava"/>
    <module name="org.slf4j" export="true"/>
    <module name="org.jboss.logging"/>
</dependencies>

Modify the org.jboss.as.clustering.infinspan module xml to include a dependency on the new query module but also the org.hibernate module that already exists. So add the following to that file.

<module name="org.hibernate" services="import"/>
<module name="org.infinispan.query" services="import"/>

The jboss-deployment-structure.xml of your app needs to have the following dependencies added to it

<?xml version="1.0" encoding="UTF-8"?>
<jboss-deployment-structure>
    <deployment>
         <dependencies>
              <module export="true" name="org.hibernate" services="import"/>
              <module export="true" name="org.infinispan" services="import"/>
              <module export="true" name="org.infinispan.query" services="import"/>
        </dependencies>
    </deployment>
</jboss-deployment-structure>

For annotating your objects so they can be indexed and searched I suggest you look at the samples here

EDIT: FURTHER NOTE: To get indexed objects to actually share over the cache I now had to

• Create a jar containing all the classes that you plan to put in the cache.
• Make sure these entities do not contain anything too complex it terms of imports or you'll run into class loader issues.
• Create a module from this jar, and give it a dependency on the Infinispan.query module.
• Expose this module as a global module (is this actually needed?) in the standalone.xml or domain.xml whichever is appropriate.
• Find the file modules/org/jboss/as/clustering/infinispan/main/module.xml
• Add your new module containing your cachable entities to the list of dependencies