Code cleanup and split off responsibilities for validation and node transaction tracking

alfresco-health-processor-platform/src/main/amp/config/alfresco/module/alfresco-health-processor/alfresco-global.properties

@@ -1,10 +1,11 @@
 eu.xenit.alfresco.processor.task.cron=* * * * * ? 2099
 eu.xenit.alfresco.processor.task.delay=0
 eu.xenit.alfresco.processor.enabled=true
+eu.xenit.alfresco.processor.tracking.threadpool-size=5
 eu.xenit.alfresco.processor.tracking.enabled=true
-eu.xenit.alfresco.processor.run-once=true
-eu.xenit.alfresco.processor.scope=ALL
+eu.xenit.alfresco.processor.run-once=false
 eu.xenit.alfresco.processor.transaction.start=1
 eu.xenit.alfresco.processor.transaction.limit=1000
 eu.xenit.alfresco.processor.transaction.time.increment=15
-eu.xenit.alfresco.processor.transaction.time.start=2020-10-08
+eu.xenit.alfresco.processor.transaction.time.start=2020-11-26
+eu.xenit.alfresco.processor.validation.content.type-filter={http://www.alfresco.org/model/content/1.0}content

alfresco-health-processor-platform/src/main/amp/config/alfresco/module/alfresco-health-processor/context/job-scheduling-context.xml

@@ -9,17 +9,17 @@
         <constructor-arg name="processorService" ref="healthProcessorPlatform.proxy"/>
     </bean>
 
-    <bean id="eu.xenit.alfresco.processor.tasks.processorTaskScheduler.jobDetail"
+    <bean id="eu.xenit.alfresco.processor.tasks.processorTask.jobDetail"
             class="org.springframework.scheduling.quartz.MethodInvokingJobDetailFactoryBean">
         <property name="name" value="healthProcessorTask" />
         <property name="targetObject" ref="eu.xenit.alfresco.processor.processorTask"/>
         <property name="targetMethod" value="execute"/>
     </bean>
 
-    <bean id="eu.xenit.alfresco.processor.tasks.processorTaskScheduler.trigger"
+    <bean id="eu.xenit.alfresco.processor.tasks.processorTask.trigger"
             class="org.springframework.scheduling.quartz.CronTriggerFactoryBean">
         <property name="name" value="healthProcessorTrigger" />
-        <property name="jobDetail" ref="eu.xenit.alfresco.processor.tasks.processorTaskScheduler.jobDetail"/>
+        <property name="jobDetail" ref="eu.xenit.alfresco.processor.tasks.processorTask.jobDetail"/>
         <property name="cronExpression">
             <value>${eu.xenit.alfresco.processor.task.cron}</value>
         </property>
@@ -28,12 +28,11 @@
         </property>
     </bean>
 
-    <bean id="eu.xenit.alfresco.processor.tasks.scheduler"
-            class="org.springframework.scheduling.quartz.SchedulerAccessorBean">
+    <bean id="eu.xenit.alfresco.processor.tasks.processorTask.scheduler" class="org.springframework.scheduling.quartz.SchedulerAccessorBean">
         <property name="scheduler" ref="schedulerFactory"/>
         <property name="triggers">
             <list>
-                <ref bean="eu.xenit.alfresco.processor.tasks.processorTaskScheduler.trigger"/>
+                <ref bean="eu.xenit.alfresco.processor.tasks.processorTask.trigger"/>
             </list>
         </property>
     </bean>

alfresco-health-processor-platform/src/main/amp/config/alfresco/module/alfresco-health-processor/context/search-context.xml

@@ -0,0 +1,58 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<beans xmlns="http://www.springframework.org/schema/beans"
+       xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
+       xsi:schemaLocation="http://www.springframework.org/schema/beans
+       http://www.springframework.org/schema/beans/spring-beans.xsd">
+    <!--  Ignore indexing by node type or by node aspects  -->
+
+    <bean id="eu.xenit.alfresco.search.abstractIndexFilter" abstract="true" init-method="init" >
+        <property name="namespaceService" ref="namespaceService"/>
+        <property name="dictionaryService" ref="dictionaryService"/>
+    </bean>
+
+    <bean id="eu.xenit.alfresco.search.TypeIndexFilter"
+          class="org.alfresco.repo.search.TypeIndexFilter"
+          parent="eu.xenit.alfresco.search.abstractIndexFilter">
+        <property name="ignorePathsForSpecificTypes">
+            <value>${search.solrTrackingSupport.ignorePathsForSpecificTypes:false}</value>
+        </property>
+        <property name="typesForIgnoringPaths">
+            <list>
+                <value>{http://www.alfresco.org/model/content/1.0}person</value>
+                <value>{http://www.alfresco.org/model/application/1.0}configurations</value>
+                <value>{http://www.alfresco.org/model/content/1.0}authorityContainer</value>
+            </list>
+        </property>
+    </bean>
+
+    <bean id="eu.xenit.alfresco.search.AspectIndexFilter"
+          class="org.alfresco.repo.search.AspectIndexFilter"
+          parent="eu.xenit.alfresco.search.abstractIndexFilter">
+        <property name="ignorePathsForSpecificAspects">
+            <value>${search.solrTrackingSupport.ignorePathsForSpecificAspects:false}</value>
+        </property>
+        <property name="aspectsForIgnoringPaths">
+            <list></list>
+        </property>
+    </bean>
+
+    <bean id="eu.xenit.alfresco.search.solrTrackingComponent"
+          class="org.alfresco.repo.solr.SOLRTrackingComponentImpl"
+          init-method="init">
+        <property name="permissionService" ref="permissionService"/>
+        <property name="dictionaryService" ref="dictionaryService"/>
+        <property name="ownableService" ref="ownableService"/>
+        <property name="tenantService" ref="tenantService"/>
+        <property name="nodeDAO" ref="nodeDAO"/>
+        <property name="qnameDAO" ref="qnameDAO"/>
+        <property name="solrDAO" ref="solrDAO" />
+        <property name="dictionaryDAO" ref="dictionaryDAO" />
+        <property name="aclDAO" ref="aclDAO" />
+        <property name="enabled">
+            <value>${eu.xenit.alfresco.processor.tracking.enabled:true}</value>
+        </property>
+        <property name="typeIndexFilter" ref="eu.xenit.alfresco.search.TypeIndexFilter" />
+        <property name="aspectIndexFilter" ref="eu.xenit.alfresco.search.AspectIndexFilter" />
+        <property name="namespaceService" ref="namespaceService" />
+    </bean>
+</beans>

alfresco-health-processor-platform/src/main/amp/config/alfresco/module/alfresco-health-processor/context/service-context.xml

@@ -12,4 +12,19 @@
         <constructor-arg  name="attributeService" ref="attributeService"/>
         <constructor-arg  name="descriptorService" ref="descriptorComponent"/>
     </bean>
+    <bean id="eu.xenit.alfresco.processor.service.progressTracker"
+          class="eu.xenit.alfresco.processor.service.ProgressTracker">
+        <constructor-arg name="tracker" ref="eu.xenit.alfresco.search.solrTrackingComponent"/>
+    </bean>
+    <bean id="eu.xenit.alfresco.processor.service.nodeTxService"
+          class="eu.xenit.alfresco.processor.service.NodeTxService">
+        <constructor-arg name="tracker" ref="eu.xenit.alfresco.search.solrTrackingComponent"/>
+    </bean>
+    <bean id="eu.xenit.alfresco.processor.service.cycleService"
+          class="eu.xenit.alfresco.processor.service.CycleService">
+        <constructor-arg name="retryingTransactionHelper" ref="retryingTransactionHelper"/>
+        <constructor-arg name="progressTracker" ref="eu.xenit.alfresco.processor.service.progressTracker"/>
+        <constructor-arg name="nodeTxService" ref="eu.xenit.alfresco.processor.service.nodeTxService"/>
+        <constructor-arg name="validationService" ref="eu.xenit.alfresco.processor.service.validationService"/>
+    </bean>
 </beans>

alfresco-health-processor-platform/src/main/amp/config/alfresco/module/alfresco-health-processor/context/subsystem-context.xml

@@ -9,7 +9,7 @@
         <property name="sourceBeanName" value="healthProcessorPlatform"/>
         <property name="interfaces">
             <list>
-                <value>eu.xenit.alfresco.processor.service.ProcessorService</value>
+                <value>eu.xenit.alfresco.processor.service.IProcessorService</value>
             </list>
         </property>
     </bean>

alfresco-health-processor-platform/src/main/amp/config/alfresco/module/alfresco-health-processor/context/validation-context.xml

@@ -0,0 +1,24 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<beans xmlns="http://www.springframework.org/schema/beans"
+       xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
+       xsi:schemaLocation="http://www.springframework.org/schema/beans
+       http://www.springframework.org/schema/beans/spring-beans.xsd">
+    <bean id="eu.xenit.alfresco.processor.modules.contentExistenceValidator"
+          class="eu.xenit.alfresco.processor.modules.ContentExistenceValidator"
+        init-method="initialize">
+        <constructor-arg name="filterType">
+            <value>${eu.xenit.alfresco.processor.validation.content.type-filter}</value>
+        </constructor-arg>
+        <constructor-arg name="contentService" ref="contentService"/>
+        <constructor-arg name="nodeService" ref="nodeService"/>
+        <constructor-arg name="dictionaryService" ref="dictionaryService"/>
+    </bean>
+    <bean id="eu.xenit.alfresco.processor.service.validationService"
+          class="eu.xenit.alfresco.processor.service.ValidationService">
+        <constructor-arg name="nodeValidators">
+            <list>
+                <ref bean="eu.xenit.alfresco.processor.modules.contentExistenceValidator" />
+            </list>
+        </constructor-arg >
+    </bean>
+</beans>

alfresco-health-processor-platform/src/main/amp/config/alfresco/module/alfresco-health-processor/module-context.xml

@@ -3,6 +3,8 @@
        xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
        xsi:schemaLocation="http://www.springframework.org/schema/beans
        http://www.springframework.org/schema/beans/spring-beans.xsd">
+    <import resource="classpath:alfresco/module/alfresco-health-processor/context/validation-context.xml" />
+    <import resource="classpath:alfresco/module/alfresco-health-processor/context/search-context.xml" />
     <import resource="classpath:alfresco/module/alfresco-health-processor/context/service-context.xml" />
     <import resource="classpath:alfresco/module/alfresco-health-processor/context/subsystem-context.xml" />
     <import resource="classpath:alfresco/module/alfresco-health-processor/context/job-scheduling-context.xml" />

alfresco-health-processor-platform/src/main/amp/config/alfresco/subsystems/healthProcessor/platform/platform-healthProcessor-context.xml

@@ -6,7 +6,15 @@
     <bean id="healthProcessorPlatform"
           class="eu.xenit.alfresco.processor.service.ProcessorService">
         <constructor-arg  name="retryingTransactionHelper" ref="retryingTransactionHelper" />
-        <constructor-arg  name="configurationService" ref="eu.xenit.alfresco.processor.service.healthProcessorConfiguration"/>
+        <constructor-arg  name="configuration" ref="eu.xenit.alfresco.processor.service.healthProcessorConfiguration"/>
         <constructor-arg  name="processorAttributeService" ref="eu.xenit.alfresco.processor.service.processorAttributeService"/>
+        <constructor-arg  name="cycleService" ref="eu.xenit.alfresco.processor.service.cycleService"/>
+        <constructor-arg  name="executorService">
+            <bean class="java.util.concurrent.Executors" factory-method="newFixedThreadPool">
+                <constructor-arg>
+                    <value>${eu.xenit.alfresco.processor.tracking.threadpool-size}</value>
+                </constructor-arg>
+            </bean>
+        </constructor-arg>
     </bean>
 </beans>

alfresco-health-processor-platform/src/main/java/eu/xenit/alfresco/processor/model/Cycle.java

@@ -0,0 +1,17 @@
+package eu.xenit.alfresco.processor.model;
+
+import lombok.AllArgsConstructor;
+import lombok.Data;
+
+/**
+ * Information object containing the parameters for the tracking activity.
+ */
+@Data
+@AllArgsConstructor
+public class Cycle {
+    private int txnLimit;
+    private long firstTxn;
+    private int timeIncrementSeconds;
+    private long firstCommitTime;
+    private TrackerInfo trackerInfo;
+}

alfresco-health-processor-platform/src/main/java/eu/xenit/alfresco/processor/model/TrackerInfo.java

@@ -0,0 +1,29 @@
+package eu.xenit.alfresco.processor.model;
+
+import lombok.Data;
+
+import java.time.OffsetDateTime;
+
+/**
+ * Information object to save the state of the tracking activity. The state is persisted to allow for the correct
+ * continuation after a restart of the system.
+ */
+@Data
+public class TrackerInfo {
+    /**
+     * Timestamp of the last action
+     */
+    private OffsetDateTime timestamp;
+    /**
+     * Id of the last transaction processed by the tracker
+     */
+    private Long transactionId;
+    /**
+     * Highest commit time in ms of processed transaction
+     */
+    private Long commitTimeMs = -1L;
+    /**
+     * Name of the artifact running this tracker software
+     */
+    private String name;
+}

alfresco-health-processor-platform/src/main/java/eu/xenit/alfresco/processor/modules/ContentExistenceValidator.java

@@ -0,0 +1,65 @@
+package eu.xenit.alfresco.processor.modules;
+
+import lombok.RequiredArgsConstructor;
+import org.alfresco.model.ContentModel;
+import org.alfresco.service.cmr.dictionary.DictionaryService;
+import org.alfresco.service.cmr.repository.ContentReader;
+import org.alfresco.service.cmr.repository.ContentService;
+import org.alfresco.service.cmr.repository.NodeRef;
+import org.alfresco.service.cmr.repository.NodeService;
+import org.alfresco.service.namespace.QName;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import java.security.InvalidParameterException;
+
+@RequiredArgsConstructor
+public class ContentExistenceValidator implements NodeValidator {
+    private static final Logger logger = LoggerFactory.getLogger(ContentExistenceValidator.class);
+
+    private QName filterTypeQName = null;
+
+    protected final String filterType;
+
+    protected final ContentService contentService;
+
+    protected final NodeService nodeService;
+
+    protected final DictionaryService dictionaryService;
+
+    ContentExistenceValidator initialize() {
+        filterTypeQName = QName.createQName(filterType);
+        return this;
+    }
+
+    @Override
+    public boolean validate(NodeRef nodeRef) {
+        logger.trace(">>> Entering ContentExistenceValidator.validate()");
+        if(nodeRef == null) {
+            throw new InvalidParameterException("NodeRef cannot be null!");
+        }
+
+        if(!this.nodeService.exists(nodeRef)) {
+            logger.debug("Node does not exist: {}", nodeRef);
+            return false;
+        }
+
+        QName type = this.nodeService.getType(nodeRef);
+        if(!this.dictionaryService.isSubClass(type, filterTypeQName)) {
+            logger.debug("Node {} is of type {}, skipping because it's subclass of {}", nodeRef, type, filterTypeQName);
+            return true;
+        }
+
+        ContentReader reader = this.contentService
+                .getReader(nodeRef, ContentModel.PROP_CONTENT);
+
+        if(reader != null && reader.exists()) {
+            logger.debug("Node Content Does Exist {}", nodeRef);
+            logger.trace("<<< Exiting ContentExistenceValidator.validate()");
+            return true;
+        }
+
+        logger.error("Node Content Does Not Exist {}", nodeRef);
+        return false;
+    }
+}

alfresco-health-processor-platform/src/main/java/eu/xenit/alfresco/processor/modules/NodeValidator.java

@@ -0,0 +1,7 @@
+package eu.xenit.alfresco.processor.modules;
+
+import org.alfresco.service.cmr.repository.NodeRef;
+
+public interface NodeValidator {
+    boolean validate(NodeRef nodeRef);
+}