From: Mathieu Baudier <mbaudier@argeo.org>
Date: Sat, 28 Feb 2015 21:54:21 +0000 (+0000)
Subject: Jackrabbit search tuning
X-Git-Tag: argeo-commons-2.1.30~300
X-Git-Url: https://git.argeo.org/?a=commitdiff_plain;h=ed7a26d6f79d2091c076f1d7850a6a3550d15f14;p=lgpl%2Fargeo-commons.git

Jackrabbit search tuning

git-svn-id: https://svn.argeo.org/commons/trunk@7994 4cfe0d0a-d680-48aa-b62c-e0a02a3f76cc
---

diff --git a/demo/argeo_node_rap.properties b/demo/argeo_node_rap.properties
index 502c50e8c..520c5baf4 100644
--- a/demo/argeo_node_rap.properties
+++ b/demo/argeo_node_rap.properties
@@ -34,9 +34,16 @@ eclipse.registry.MultiLanguage=true
 log4j.configuration=file:../../log4j.properties
 
 # Tuning
+# Number of DB connections
+#argeo.node.repo.maxPoolSize=10
+# Max amount of memory available to Jackrabbit caches
 #argeo.node.repo.maxCacheMB=16
+# Persistence level cache
 #argeo.node.repo.bundleCacheMB=8
-#argeo.node.repo.maxPoolSize=10
+# Search, see http://wiki.apache.org/jackrabbit/Search
+#argeo.node.repo.extractorPoolSize=0
+#argeo.node.repo.searchCacheSize=1000
+#argeo.node.repo.maxVolatileIndexSize=1048576
 
 # DON'T CHANGE BELOW
 org.eclipse.rap.workbenchAutostart=false
\ No newline at end of file
diff --git a/org.argeo.cms/src/org/argeo/cms/internal/kernel/JackrabbitNode.java b/org.argeo.cms/src/org/argeo/cms/internal/kernel/JackrabbitNode.java
index c9c100e12..50c78edaa 100644
--- a/org.argeo.cms/src/org/argeo/cms/internal/kernel/JackrabbitNode.java
+++ b/org.argeo.cms/src/org/argeo/cms/internal/kernel/JackrabbitNode.java
@@ -125,6 +125,10 @@ class JackrabbitNode extends JackrabbitWrapper implements KernelConstants,
 		setProp(defaults, REPO_MAX_POOL_SIZE, "10");
 		// Jackrabbit default:
 		setProp(defaults, REPO_BUNDLE_CACHE_MB, "8");
+		// See http://wiki.apache.org/jackrabbit/Search
+		setProp(defaults, REPO_EXTRACTOR_POOL_SIZE, "0");
+		setProp(defaults, REPO_SEARCH_CACHE_SIZE, "1000");
+		setProp(defaults, REPO_MAX_VOLATILE_INDEX_SIZE, "1048576");
 
 		// specific
 		String dburl;
diff --git a/org.argeo.cms/src/org/argeo/cms/internal/kernel/KernelConstants.java b/org.argeo.cms/src/org/argeo/cms/internal/kernel/KernelConstants.java
index 896eca5b8..7465583f9 100644
--- a/org.argeo.cms/src/org/argeo/cms/internal/kernel/KernelConstants.java
+++ b/org.argeo.cms/src/org/argeo/cms/internal/kernel/KernelConstants.java
@@ -12,6 +12,9 @@ public interface KernelConstants {
 	final static String REPO_MAX_POOL_SIZE = "argeo.node.repo.maxPoolSize";
 	final static String REPO_MAX_CACHE_MB = "argeo.node.repo.maxCacheMB";
 	final static String REPO_BUNDLE_CACHE_MB = "argeo.node.repo.bundleCacheMB";
+	final static String REPO_EXTRACTOR_POOL_SIZE = "argeo.node.repo.extractorPoolSize";
+	final static String REPO_SEARCH_CACHE_SIZE = "argeo.node.repo.searchCacheSize";
+	final static String REPO_MAX_VOLATILE_INDEX_SIZE = "argeo.node.repo.maxVolatileIndexSize";
 
 	final static String[] DEFAULT_CNDS = { "/org/argeo/jcr/argeo.cnd",
 			"/org/argeo/cms/cms.cnd" };
diff --git a/org.argeo.cms/src/org/argeo/cms/internal/kernel/repository-h2.xml b/org.argeo.cms/src/org/argeo/cms/internal/kernel/repository-h2.xml
index 5f7d0043d..4c4875f1c 100644
--- a/org.argeo.cms/src/org/argeo/cms/internal/kernel/repository-h2.xml
+++ b/org.argeo.cms/src/org/argeo/cms/internal/kernel/repository-h2.xml
@@ -36,11 +36,13 @@
 			class="org.apache.jackrabbit.core.persistence.pool.H2PersistenceManager">
 			<param name="dataSourceName" value="dataSource" />
 			<param name="schemaObjectPrefix" value="${wsp.name}_pm_" />
-			<param name="bundleCacheSize" value="${argeo.node.repo.bundleCacheMB}"/>
+			<param name="bundleCacheSize" value="${argeo.node.repo.bundleCacheMB}" />
 		</PersistenceManager>
 		<SearchIndex class="org.apache.jackrabbit.core.query.lucene.SearchIndex">
 			<param name="path" value="${wsp.home}/index" />
-			<param name="initializeHierarchyCache" value="true" />
+			<param name="extractorPoolSize" value="${argeo.node.repo.extractorPoolSize}" />
+			<param name="cacheSize" value="${argeo.node.repo.searchCacheSize}" />
+			<param name="maxVolatileIndexSize" value="${argeo.node.repo.maxVolatileIndexSize}" />
 		</SearchIndex>
 	</Workspace>
 
@@ -55,16 +57,16 @@
 			class="org.apache.jackrabbit.core.persistence.pool.H2PersistenceManager">
 			<param name="dataSourceName" value="dataSource" />
 			<param name="schemaObjectPrefix" value="pm_ver_" />
-			<param name="bundleCacheSize" value="${argeo.node.repo.bundleCacheMB}"/>
+			<param name="bundleCacheSize" value="${argeo.node.repo.bundleCacheMB}" />
 		</PersistenceManager>
 	</Versioning>
 
 	<!-- Indexing -->
 	<SearchIndex class="org.apache.jackrabbit.core.query.lucene.SearchIndex">
 		<param name="path" value="${rep.home}/index" />
-		<param name="initializeHierarchyCache" value="true" />
-		<!-- <param name="extractorPoolSize" value="2" /> -->
-		<!-- <param name="supportHighlighting" value="true" /> -->
+		<param name="extractorPoolSize" value="${argeo.node.repo.extractorPoolSize}" />
+		<param name="cacheSize" value="${argeo.node.repo.searchCacheSize}" />
+		<param name="maxVolatileIndexSize" value="${argeo.node.repo.maxVolatileIndexSize}" />
 	</SearchIndex>
 
 	<!-- Security -->
diff --git a/org.argeo.cms/src/org/argeo/cms/internal/kernel/repository-memory.xml b/org.argeo.cms/src/org/argeo/cms/internal/kernel/repository-memory.xml
index 8181ad34d..7793c1517 100644
--- a/org.argeo.cms/src/org/argeo/cms/internal/kernel/repository-memory.xml
+++ b/org.argeo.cms/src/org/argeo/cms/internal/kernel/repository-memory.xml
@@ -12,12 +12,15 @@
 		<PersistenceManager
 			class="org.apache.jackrabbit.core.persistence.bundle.BundleFsPersistenceManager">
 			<param name="blobFSBlockSize" value="1" />
-			<param name="bundleCacheSize" value="${argeo.node.repo.bundleCacheMB}"/>
+			<param name="bundleCacheSize" value="${argeo.node.repo.bundleCacheMB}" />
 		</PersistenceManager>
 		<SearchIndex class="org.apache.jackrabbit.core.query.lucene.SearchIndex">
 			<param name="path" value="${wsp.home}/index" />
 			<param name="directoryManagerClass"
 				value="org.apache.jackrabbit.core.query.lucene.directory.RAMDirectoryManager" />
+			<param name="extractorPoolSize" value="${argeo.node.repo.extractorPoolSize}" />
+			<param name="cacheSize" value="${argeo.node.repo.searchCacheSize}" />
+			<param name="maxVolatileIndexSize" value="${argeo.node.repo.maxVolatileIndexSize}" />
 			<FileSystem class="org.apache.jackrabbit.core.fs.mem.MemoryFileSystem" />
 		</SearchIndex>
 	</Workspace>
@@ -28,7 +31,7 @@
 		<PersistenceManager
 			class="org.apache.jackrabbit.core.persistence.bundle.BundleFsPersistenceManager">
 			<param name="blobFSBlockSize" value="1" />
-			<param name="bundleCacheSize" value="${argeo.node.repo.bundleCacheMB}"/>
+			<param name="bundleCacheSize" value="${argeo.node.repo.bundleCacheMB}" />
 		</PersistenceManager>
 	</Versioning>
 
@@ -37,6 +40,9 @@
 		<param name="path" value="${rep.home}/index" />
 		<param name="directoryManagerClass"
 			value="org.apache.jackrabbit.core.query.lucene.directory.RAMDirectoryManager" />
+		<param name="extractorPoolSize" value="${argeo.node.repo.extractorPoolSize}" />
+		<param name="cacheSize" value="${argeo.node.repo.searchCacheSize}" />
+		<param name="maxVolatileIndexSize" value="${argeo.node.repo.maxVolatileIndexSize}" />
 		<FileSystem class="org.apache.jackrabbit.core.fs.mem.MemoryFileSystem" />
 	</SearchIndex>
 
diff --git a/org.argeo.cms/src/org/argeo/cms/internal/kernel/repository-postgresql-ds.xml b/org.argeo.cms/src/org/argeo/cms/internal/kernel/repository-postgresql-ds.xml
new file mode 100644
index 000000000..1e894a2e6
--- /dev/null
+++ b/org.argeo.cms/src/org/argeo/cms/internal/kernel/repository-postgresql-ds.xml
@@ -0,0 +1,82 @@
+<?xml version="1.0"?>
+<!DOCTYPE Repository PUBLIC "Jackrabbit 2.6" "http://jackrabbit.apache.org/dtd/repository-2.6.dtd">
+<Repository>
+	<!-- Shared datasource -->
+	<DataSources>
+		<DataSource name="dataSource">
+			<param name="driver" value="org.postgresql.Driver" />
+			<param name="url" value="${argeo.node.repo.dburl}" />
+			<param name="user" value="${argeo.node.repo.dbuser}" />
+			<param name="password" value="${argeo.node.repo.dbpassword}" />
+			<param name="databaseType" value="postgresql" />
+			<param name="maxPoolSize" value="${argeo.node.repo.maxPoolSize}" />
+		</DataSource>
+	</DataSources>
+
+	<!-- File system and datastore -->
+	<FileSystem class="org.apache.jackrabbit.core.fs.db.DbFileSystem">
+		<param name="dataSourceName" value="dataSource" />
+		<param name="schema" value="postgresql" />
+		<param name="schemaObjectPrefix" value="fs_" />
+	</FileSystem>
+	<DataStore class="org.apache.jackrabbit.core.data.FileDataStore">
+		<param name="path" value="${rep.home}/datastore" />
+	</DataStore>
+
+	<!-- Workspace templates -->
+	<Workspaces rootPath="${rep.home}/workspaces"
+		defaultWorkspace="${argeo.node.repo.defaultWorkspace}" />
+	<Workspace name="${wsp.name}">
+		<FileSystem class="org.apache.jackrabbit.core.fs.db.DbFileSystem">
+			<param name="dataSourceName" value="dataSource" />
+			<param name="schema" value="postgresql" />
+			<param name="schemaObjectPrefix" value="${wsp.name}_fs_" />
+		</FileSystem>
+		<PersistenceManager
+			class="org.apache.jackrabbit.core.persistence.pool.PostgreSQLPersistenceManager">
+			<param name="dataSourceName" value="dataSource" />
+			<param name="schemaObjectPrefix" value="${wsp.name}_pm_" />
+			<param name="bundleCacheSize" value="${argeo.node.repo.bundleCacheMB}" />
+		</PersistenceManager>
+		<SearchIndex class="org.apache.jackrabbit.core.query.lucene.SearchIndex">
+			<param name="path" value="${wsp.home}/index" />
+			<param name="extractorPoolSize" value="${argeo.node.repo.extractorPoolSize}" />
+			<param name="cacheSize" value="${argeo.node.repo.searchCacheSize}" />
+			<param name="maxVolatileIndexSize" value="${argeo.node.repo.maxVolatileIndexSize}" />
+		</SearchIndex>
+	</Workspace>
+
+	<!-- Versioning -->
+	<Versioning rootPath="${rep.home}/version">
+		<FileSystem class="org.apache.jackrabbit.core.fs.db.DbFileSystem">
+			<param name="dataSourceName" value="dataSource" />
+			<param name="schema" value="postgresql" />
+			<param name="schemaObjectPrefix" value="fs_ver_" />
+		</FileSystem>
+		<PersistenceManager
+			class="org.apache.jackrabbit.core.persistence.pool.PostgreSQLPersistenceManager">
+			<param name="dataSourceName" value="dataSource" />
+			<param name="schemaObjectPrefix" value="pm_ver_" />
+			<param name="bundleCacheSize" value="${argeo.node.repo.bundleCacheMB}" />
+		</PersistenceManager>
+	</Versioning>
+
+	<!-- Indexing -->
+	<SearchIndex class="org.apache.jackrabbit.core.query.lucene.SearchIndex">
+		<param name="path" value="${rep.home}/index" />
+		<param name="extractorPoolSize" value="${argeo.node.repo.extractorPoolSize}" />
+		<param name="cacheSize" value="${argeo.node.repo.searchCacheSize}" />
+		<param name="maxVolatileIndexSize" value="${argeo.node.repo.maxVolatileIndexSize}" />
+	</SearchIndex>
+
+	<!-- Security -->
+	<Security appName="Jackrabbit">
+		<SecurityManager class="org.argeo.security.jackrabbit.ArgeoSecurityManager"
+			workspaceName="security">
+		</SecurityManager>
+		<AccessManager class="org.argeo.security.jackrabbit.ArgeoAccessManager">
+		</AccessManager>
+		<LoginModule class="org.argeo.security.jackrabbit.ArgeoLoginModule">
+		</LoginModule>
+	</Security>
+</Repository>
\ No newline at end of file
diff --git a/org.argeo.cms/src/org/argeo/cms/internal/kernel/repository-postgresql.xml b/org.argeo.cms/src/org/argeo/cms/internal/kernel/repository-postgresql.xml
index 6ed23b59a..1f61ea342 100644
--- a/org.argeo.cms/src/org/argeo/cms/internal/kernel/repository-postgresql.xml
+++ b/org.argeo.cms/src/org/argeo/cms/internal/kernel/repository-postgresql.xml
@@ -19,9 +19,6 @@
 		<param name="schema" value="postgresql" />
 		<param name="schemaObjectPrefix" value="fs_" />
 	</FileSystem>
-	<DataStore class="org.apache.jackrabbit.core.data.FileDataStore">
-		<param name="path" value="${rep.home}/datastore" />
-	</DataStore>
 
 	<!-- Workspace templates -->
 	<Workspaces rootPath="${rep.home}/workspaces"
@@ -36,10 +33,13 @@
 			class="org.apache.jackrabbit.core.persistence.pool.PostgreSQLPersistenceManager">
 			<param name="dataSourceName" value="dataSource" />
 			<param name="schemaObjectPrefix" value="${wsp.name}_pm_" />
-			<param name="bundleCacheSize" value="${argeo.node.repo.bundleCacheMB}"/>
+			<param name="bundleCacheSize" value="${argeo.node.repo.bundleCacheMB}" />
 		</PersistenceManager>
 		<SearchIndex class="org.apache.jackrabbit.core.query.lucene.SearchIndex">
 			<param name="path" value="${wsp.home}/index" />
+			<param name="extractorPoolSize" value="${argeo.node.repo.extractorPoolSize}" />
+			<param name="cacheSize" value="${argeo.node.repo.searchCacheSize}" />
+			<param name="maxVolatileIndexSize" value="${argeo.node.repo.maxVolatileIndexSize}" />
 		</SearchIndex>
 	</Workspace>
 
@@ -54,15 +54,16 @@
 			class="org.apache.jackrabbit.core.persistence.pool.PostgreSQLPersistenceManager">
 			<param name="dataSourceName" value="dataSource" />
 			<param name="schemaObjectPrefix" value="pm_ver_" />
-			<param name="bundleCacheSize" value="${argeo.node.repo.bundleCacheMB}"/>
+			<param name="bundleCacheSize" value="${argeo.node.repo.bundleCacheMB}" />
 		</PersistenceManager>
 	</Versioning>
 
 	<!-- Indexing -->
 	<SearchIndex class="org.apache.jackrabbit.core.query.lucene.SearchIndex">
 		<param name="path" value="${rep.home}/index" />
-		<param name="extractorPoolSize" value="2" />
-		<param name="supportHighlighting" value="true" />
+		<param name="extractorPoolSize" value="${argeo.node.repo.extractorPoolSize}" />
+		<param name="cacheSize" value="${argeo.node.repo.searchCacheSize}" />
+		<param name="maxVolatileIndexSize" value="${argeo.node.repo.maxVolatileIndexSize}" />
 	</SearchIndex>
 
 	<!-- Security -->