2.2.2. Lock Manager configuration

Warning

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

What LockManager does?

In general, LockManager stores Lock objects, so it can give a Lock object or can release it. Also, LockManager is responsible for removing Locks that live too long. This parameter may be configured with "time-out" property.

JCR provides one basic implementations of LockManager:


org.exoplatform.services.jcr.impl.core.lock.infinispan.ISPNCacheableLockManagerImpl

CacheableLockManagerImpl

ISPNCacheableLockManagerImpl stores Lock objects in Infinispan, so Locks are replicable and affect on cluster, not only a single node. Also, Infinispan has a JdbcStringBasedStore, so Locks will be stored to the database.

You can enable LockManager by adding lock-manager-configuration to workspace-configuration.

For example:


<workspace name="ws">

   ...

   <lock-manager class="org.exoplatform.services.jcr.impl.core.lock.infinispan.ISPNCacheableLockManagerImpl">

      <properties>

         <property name="time-out" value="15m" />

         ...

      </properties>

   </lock-manager>               

   ...

</workspace>

Where time-out parameter represents interval to remove Expired Locks. LockRemover separates threads, that periodically ask LockManager to remove Locks that live so long.

Configuration:

The configuration uses the template Infinispan configuration for all LockManagers.

The lock template configuration:

test-infinispan-lock.xml:




<infinispan xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"

            xmlns="urn:infinispan:config:8.2"

            xsi:schemaLocation="urn:infinispan:config:8.2 http://www.infinispan.org/schemas/infinispan-config-8.2.xsd">

  <threads>

    <thread-factory name="infinispan-factory" group-name="infinispan" thread-name-pattern="%G %i" priority="5"/>



    <!-- listener-executor -->

    <blocking-bounded-queue-thread-pool name="infinispan-listener" thread-factory="infinispan-factory" core-threads="1" 

                                        max-threads="5" queue-length="0" keepalive-time="0"/>

  </threads>



  <jgroups transport="org.infinispan.remoting.transport.jgroups.JGroupsTransport">

    <stack-file name="stack" path="${exo.jcr.cluster.jgroups.config}"/>

  </jgroups>



  <cache-container name="lock-manager"  default-cache="default"  listener-executor="infinispan-listener"

                   statistics="true">

    <jmx duplicate-domains="true" domain="jcr.ispn.cache" mbean-server-lookup="org.infinispan.jmx.PlatformMBeanServerLookup"/>

    <transport cluster="${exo.cluster.partition.name}-jcr-lock" stack="stack" lock-timeout="240000"/>

    <replicated-cache-configuration mode="SYNC" name="default" statistics="true" remote-timeout="${exo.jcr.cluster.lock.sync.repltimeout:240000}">

      <locking isolation="READ_COMMITTED" concurrency-level="500" striping="false" write-skew="false" 

                acquire-timeout="${exo.jcr.lock.lockacquisitiontimeout:180000}"/>

      <transaction transaction-manager-lookup="org.exoplatform.services.transaction.infinispan.JBossStandaloneJTAManagerLookup" mode="NON_XA"/>

      <state-transfer enabled="${exo.jcr.cluster.lock.statetransfer.fetchinmemorystate:false}" 

                        timeout="${exo.jcr.cluster.lock.statetransfer.timeout:240000}"/>

      <eviction strategy="NONE" />

      <expiration lifespan="-1" />

      <persistence passivation="false">

        <string-keyed-jdbc-store xmlns="urn:infinispan:config:store:jdbc:8.0" shared="true" fetch-state="true" 

                                read-only="false" purge="false" preload="true">

          <string-keyed-table drop-on-exit="${infinispan-cl-cache.jdbc.table.drop}" create-on-start="${infinispan-cl-cache.jdbc.table.create}"

                              prefix="${infinispan-cl-cache.jdbc.table.name}">

            <id-column name="${infinispan-cl-cache.jdbc.id.column}" type="${infinispan-cl-cache.jdbc.id.type}" />

            <data-column name="${infinispan-cl-cache.jdbc.data.column}" type="${infinispan-cl-cache.jdbc.data.type}" />

            <timestamp-column name="${infinispan-cl-cache.jdbc.timestamp.column}" type="${infinispan-cl-cache.jdbc.timestamp.type}" />

          </string-keyed-table>

        </string-keyed-jdbc-store>

      </persistence>

    </replicated-cache-configuration>

  </cache-container>

</infinispan>

Note

To prevent any consistency issue regarding the lock data, please ensure that your cache loader is org.infinispan.persistence.jdbc.stringbased.JdbcStringBasedStore and that your database engine is transactional.

For more information about JdbcStringBasedStore, refer to this link.

As you see, all configurable parameters are filled by templates and will be replaced by LockManagers configuration parameters:


<lock-manager class="org.exoplatform.services.jcr.impl.core.lock.infinispan.ISPNCacheableLockManagerImpl">

   <properties>

      <property name="time-out" value="15m" />

      <property name="infinispan-configuration" value="conf/standalone/cluster/test-infinispan-lock.xml" />

      <property name="jgroups-configuration" value="udp-mux.xml" />

      <property name="infinispan-cluster-name" value="JCR-cluster" />

      <property name="infinispan-cl-cache.jdbc.table.name" value="lk" />

      <property name="infinispan-cl-cache.jdbc.table.create" value="true" />

      <property name="infinispan-cl-cache.jdbc.table.drop" value="false" />

      <property name="infinispan-cl-cache.jdbc.id.column" value="id" />

      <property name="infinispan-cl-cache.jdbc.data.column" value="data" />

      <property name="infinispan-cl-cache.jdbc.timestamp.column" value="timestamp" />

      <property name="infinispan-cl-cache.jdbc.datasource" value="jdbcjcr" />

      <property name="infinispan-cl-cache.jdbc.dialect" value="${dialect}" />

      <property name="infinispan-cl-cache.jdbc.connectionFactory" value="org.exoplatform.services.jcr.infinispan.ManagedConnectionFactory" />

   </properties>

</lock-manager>

Configuration requirements:

infinispan-cl-cache.jdbc.id.type, infinispan-cl-cache.jdbc.data.typeand infinispan-cl-cache.jdbc.timestamp.type are injected in the Infinispan configuration into the property respectively idColumnType, dataColumnType and timestampColumnType. You can set those data types according to your database type or set it as AUTO (or do not set at all) and data type will be detected automatically.
jgroups-configuration is moved to separate the configuration file udp-mux.xml. In this case, the udp-mux.xml file is a common JGroup configuration for all components (QueryHandler, Cache, LockManager), but we can still create our own configuration.

our udp-mux.xml:


<config>

    <UDP

         singleton_name="JCR-cluster" 

         mcast_addr="${jgroups.udp.mcast_addr:228.10.10.10}"

         mcast_port="${jgroups.udp.mcast_port:45588}"

         tos="8" 

         ucast_recv_buf_size="20000000"

         ucast_send_buf_size="640000" 

         mcast_recv_buf_size="25000000" 

         mcast_send_buf_size="640000" 

         loopback="false"

         discard_incompatible_packets="true" 

         max_bundle_size="64000" 

         max_bundle_timeout="30"

         use_incoming_packet_handler="true" 

         ip_ttl="${jgroups.udp.ip_ttl:2}"

         enable_bundling="false" 

         enable_diagnostics="true"

         thread_naming_pattern="cl" 



         use_concurrent_stack="true" 



         thread_pool.enabled="true" 

         thread_pool.min_threads="2"

         thread_pool.max_threads="8" 

         thread_pool.keep_alive_time="5000" 

         thread_pool.queue_enabled="true"

         thread_pool.queue_max_size="1000"

         thread_pool.rejection_policy="discard"



         oob_thread_pool.enabled="true"

         oob_thread_pool.min_threads="1"

         oob_thread_pool.max_threads="8"

         oob_thread_pool.keep_alive_time="5000"

         oob_thread_pool.queue_enabled="false" 

         oob_thread_pool.queue_max_size="100" 

         oob_thread_pool.rejection_policy="Run" />



    <PING timeout="2000"<config xmlns="urn:org:jgroups"

        xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"

        xsi:schemaLocation="urn:org:jgroups http://www.jgroups.org/schema/JGroups-3.2.xsd">

    <UDP

         singleton_name="JCR-cluster" 

         mcast_port="${jgroups.udp.mcast_port:45588}"

         tos="8"

         ucast_recv_buf_size="20M"

         ucast_send_buf_size="640K"

         mcast_recv_buf_size="25M"

         mcast_send_buf_size="640K"

         loopback="true"

         max_bundle_size="64K"

         max_bundle_timeout="30"

         ip_ttl="${jgroups.udp.ip_ttl:8}"

         enable_bundling="true"

         enable_diagnostics="true"

         thread_naming_pattern="cl"



         timer_type="old"

         timer.min_threads="4"

         timer.max_threads="10"

         timer.keep_alive_time="3000"

         timer.queue_max_size="500"



         thread_pool.enabled="true"

         thread_pool.min_threads="2"

         thread_pool.max_threads="8"

         thread_pool.keep_alive_time="5000"

         thread_pool.queue_enabled="true"

         thread_pool.queue_max_size="10000"

         thread_pool.rejection_policy="discard"



         oob_thread_pool.enabled="true"

         oob_thread_pool.min_threads="1"

         oob_thread_pool.max_threads="8"

         oob_thread_pool.keep_alive_time="5000"

         oob_thread_pool.queue_enabled="false"

         oob_thread_pool.queue_max_size="100"

         oob_thread_pool.rejection_policy="Run"/>



    <PING timeout="2000"

            num_initial_members="20"/>

    <MERGE2 max_interval="30000"

            min_interval="10000"/>

    <FD_SOCK/>

    <FD_ALL/>

    <VERIFY_SUSPECT timeout="1500"  />

    <BARRIER />

    <pbcast.NAKACK2 xmit_interval="1000"

                    xmit_table_num_rows="100"

                    xmit_table_msgs_per_row="2000"

                    xmit_table_max_compaction_time="30000"

                    max_msg_batch_size="500"

                    use_mcast_xmit="false"

                    discard_delivered_msgs="true"/>

    <UNICAST  xmit_interval="2000"

              xmit_table_num_rows="100"

              xmit_table_msgs_per_row="2000"

              xmit_table_max_compaction_time="60000"

              conn_expiry_timeout="60000"

              max_msg_batch_size="500"/>

    <pbcast.STABLE stability_delay="1000" desired_avg_gossip="50000"

                   max_bytes="4M"/>

    <pbcast.GMS print_local_addr="true" join_timeout="3000"

                view_bundling="true"/>

    <UFC max_credits="2M"

         min_threshold="0.4"/>

    <MFC max_credits="2M"

         min_threshold="0.4"/>

    <FRAG2 frag_size="60K"  />

    <RSVP resend_interval="2000" timeout="10000"/>

    <pbcast.STATE_TRANSFER />

    <!-- pbcast.FLUSH  /-->

</config>

Data, id and timestamp type in different databases:

DataBase name	Data type	Id type	Timestamp type
HSQL	VARBINARY(65535)	VARCHAR(512)	BIGINT
MySQL	LONGBLOB	VARCHAR(512)	BIGINT
ORACLE	BLOB	VARCHAR2(512)	NUMBER(19, 0)
PostgreSQL/PostgrePlus	bytea	VARCHAR(512)	BIGINT
MSSQL	VARBINARY(MAX)	VARCHAR(512)	BIGINT