2.1.1.4. Workspace Data Container

Warning

You are looking at documentation for an older release. Not what you want? See the current release documentation.

Each Workspace of JCR has its own persistent storage to hold workspace's items data. eXo Content Repository can be configured so that it can use one or more workspaces that are logical units of the repository content. Physical data storage mechanism is configured using the container mandatory element. The type of container is described in the class attribute = the fully qualified name of the org.exoplatform.services.jcr.storage.WorkspaceDataContainer subclass like:


<container class="org.exoplatform.services.jcr.impl.storage.jdbc.optimisation.CQJDBCWorkspaceDataContainer">

  <properties>

    <property name="source-name" value="jdbcjcr1"/>

    <property name="dialect" value="hsqldb"/>

    <property name="multi-db" value="true"/>

    <property name="max-buffer-size" value="200K"/>

    <property name="swap-directory" value="target/temp/swap/ws"/>

    <property name="lazy-node-iterator-page-size" value="50"/>

    <property name="acl-bloomfilter-false-positive-probability" value="0.1d"/>

    <property name="acl-bloomfilter-elements-number" value="1000000"/>

    <property name="check-sns-new-connection" value="false"/>

    <property name="batch-size" value="1000"/>

  </properties>

Specific parameters of Workspace Data Container

max-buffer-size: A threshold in bytes, if a value size is greater, then it will be spooled to a temporary file. The default value is 200k.
swap-directory: A location where the value will be spooled if no value storage is configured but a max-buffer-size is exceeded. The default value is the value of the "java.io.tmpdir" system property.
lazy-node-iterator-page-size: "Lazy" child nodes iterator settings. Defines the size of page, the number of nodes that are retrieved from persistent storage at once. The default value is 100.
acl-bloomfilter-false-positive-probability: ACL Bloom-filter settings. ACL Bloom-filter desired false positive probability. Range [0..1]. The default value is 0.1d.
acl-bloomfilter-elements-number: ACL Bloom-filter settings. Expected number of ACL-elements in the Bloom-filter. The default value is 1000000.
check-sns-new-connection: Defines if you need to create new connection for checking if an older same-name sibling exists. The default value is "false".
trigger-events-for-descendants-on-rename: Indicates if each descendant item must be included into the changes log in case of renaming or not.
- If its value is set to "false", the performance on the rename operations will be increased in case of a big amount of nodes under the source parent node. However, the performance will be decreased in case of a small amount of sub-nodes.
- If its value is set to "true", the performance will be better in case of a small amount of sub-nodes and worse in case of a big amount of sub-nodes.
- If this parameter is not set, the application will rely on the max-descendant-nodes-allowed-on-move parameter to add the descendant items to the changes log or not. If this parameter is not set but the trigger-events-for-descendants-on-move parameter is set, it will have the same value.
trigger-events-for-descendants-on-move: Indicates if each descendant item must be included into the changes log in case of moving or not.
- If its value is set to "false", the performance of moving operations will be increased in case of a big amount of nodes under the source parent node. However, the performance will be decreased in case of a small amount of sub-nodes.
- If its value is set to "true", the performance will be better in case of a small amount of sub-nodes and worse in case of a big amount of sub-nodes.
- If this parameter is not set, the application will rely on the max-descendant-nodes-allowed-on-move parameter to add or not the descendant items to the changes log.
max-descendant-nodes-allowed-on-move: The maximum number of descendant nodes is allowed to be included into the change log. Its value will be then automatically disabled. This allows the best performance regardless of the total amount of sub-nodes. The default value is set to "100". This parameter is used only if trigger-events-for-descendants-on-move and trigger-events-for-descendants-on-rename are not set.

Note

Bloom filters are not supported by all the cache implementations so far only the implementation for infinispan supports it. They are used to avoid read nodes that definitely do not have ACL. acl-bloomfilter-false-positive-probability and acl-bloomfilter-elements-number are used to configure such filters. You can read here for more information about Bloom filters.

Specific parameters of JDBC Workspace Data Container

eXo JCR has an RDB (JDBC) based production ready Workspace Data Container which has the following specific parameters:

source-name: JDBC data source name, registered in JDNI by InitialContextInitializer. (sourceName prior v.1.9). This property is mandatory.
dialect: Database dialect, one of ""hsqldb", "h2", "mysql", "mysql-myisam", "mysql-utf8", "mysql-myisam-utf8", "pgsql", "pgsql-scs", "oracle", "oracle-oci", "mssql", "sybase", "derby", "db2", "db2v8". The default value is "auto".
multi-db: Enables multi-database containers if the value of this parameter is "true". Otherwise, it is configured for single-database containers.
Note
This property is currently deprecated. It is advised to use db-structure-type instead.
db-structure-type: Can be set to isolated, multi, single to set corresponding configuration for data container. This property is mandatory.
db-tablename-suffix: If db-structure-type is set to isolated, tables used by repository service have the following format:
- JCR_I${db-tablename-suffix} for items.
- JCR_V${db-tablename-suffix} for values.
- JCR_R${db-tablename-suffix} for references.
  db-tablename-suffix by default equals to the workspace name, but can be set via configuration to any suitable.
batch-size: The batch size. The default value is -1 (disabled).
use-sequence-for-order-number: Indicates whether or not a sequence must be used to manage the order number. The expected value for this parameter is a boolean or "auto". By default, it is set to "auto" where the value of use-sequence will be set automatically according to your database type.
- It is enabled in case of H2, HSQLDB, PGSQL and ORACLE.
- It is disabled in case of MSSQL, MYSQL and SYBASE.

Workspace Data Container may support external storages for javax.jcr.Value (which can be the case for BLOB values for example) using the value-storages optional element. Data Container will try to read or write values using underlying value storage plugin if the filter criteria (see below) match the current property.


<value-storages>

  <value-storage id="Storage #1" class="org.exoplatform.services.jcr.impl.storage.value.fs.TreeFileValueStorage">

    <properties>

      <property name="path" value="data/values"/>

    </properties>

    <filters>

     <filter property-type="Binary" min-value-size="1M"/><!-- Values large of 1Mbyte -->

    </filters>

.........

</value-storages>

Where:

value-storage is the subclass of org.exoplatform.services.jcr.storage.value.ValueStoragePlugin and properties are optional plugin specific parameters.
filters: Each file value storage can have the filter(s) for incoming values. If there are several filter criteria, they all have to match (AND-Condition).
A filter can match values by the property type (property-type), property name (property-name), ancestor path (ancestor-path) and/or the size of values stored (min-value-size, e.g. 1M, 4.2G, 100 (bytes)).
This code sample uses a filter with property-type and min-value-size only. That means that the storage is only for binary values whose size is greater than 1Mbyte.

It is recommended to store properties with large values in a file value storage only.

Database's dialects

PostgreSQL/PostgrePlus database

PostgreSQL/PostgrePlus's dialect is set automatically. The dialect depends on the version of database. If you change default value of standard_conforming_strings parameter, you must configure one of the following dialects manually:

PgSQL : This dialect is used if standard_conforming_strings is set to 'off' which is the default value for version before 9.1.
PgSQL-SCS: This dialect is used if standard_conforming_strings is set to 'on' which is the default value for version after 9.1.

MySQL database

mysql: This dialect is used if JCR tables with InnoDB engine (by default) need to be created.
mysql-utf8: This dialect is used if JCR tables with InnoDB engine with UTF-8 encoding support need to be created.
mysql-myisam: This dialect is used if JCR tables with MyISAM engine need to be created.
mysql-myisam-utf8: This dialect is used if JCR tables with MyISAM engine with UTF-8 encoding support need to be created.
mysql-ndb: This dialect is used if JCR tables with NDB engine (mysql cluster) need to be created.
mysql-ndb-utf8: This dialect is used if JCR tables with NDB engine (mysql cluster) with UTF-8 encoding support need to be created.

Note

Since MySQL NDB engine does not support foreign keys, which may lead to improper item removal and as consequence to InvalidItemStateException. In this case, you will need to use consistency checker tool.