<?xml version="1.0" encoding="UTF-8" standalone="no"?> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> <html xmlns="http://www.w3.org/1999/xhtml"> <head> <meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /> <title>The Locking Subsystem</title> <link rel="stylesheet" href="gettingStarted.css" type="text/css" /> <meta name="generator" content="DocBook XSL Stylesheets V1.73.2" /> <link rel="start" href="index.html" title="Getting Started with Berkeley DB Transaction Processing" /> <link rel="up" href="txnconcurrency.html" title="Chapter 4. Concurrency" /> <link rel="prev" href="blocking_deadlocks.html" title="Locks, Blocks, and Deadlocks" /> <link rel="next" href="isolation.html" title="Isolation" /> </head> <body> <div class="navheader"> <table width="100%" summary="Navigation header"> <tr> <th colspan="3" align="center">The Locking Subsystem</th> </tr> <tr> <td width="20%" align="left"><a accesskey="p" href="blocking_deadlocks.html">Prev</a> </td> <th width="60%" align="center">Chapter 4. Concurrency</th> <td width="20%" align="right"> <a accesskey="n" href="isolation.html">Next</a></td> </tr> </table> <hr /> </div> <div class="sect1" lang="en" xml:lang="en"> <div class="titlepage"> <div> <div> <h2 class="title" style="clear: both"><a id="lockingsubsystem"></a>The Locking Subsystem</h2> </div> </div> </div> <div class="toc"> <dl> <dt> <span class="sect2"> <a href="lockingsubsystem.html#configuringlock">Configuring the Locking Subsystem</a> </span> </dt> <dt> <span class="sect2"> <a href="lockingsubsystem.html#configdeadlkdetect">Configuring Deadlock Detection</a> </span> </dt> <dt> <span class="sect2"> <a href="lockingsubsystem.html#deadlockresolve">Resolving Deadlocks</a> </span> </dt> </dl> </div> <p> In order to allow concurrent operations, DB provides the locking subsystem. This subsystem provides inter- and intra- process concurrency mechanisms. It is extensively used by DB concurrent applications, but it can also be generally used for non-DB resources. </p> <p> This section describes the locking subsystem as it is used to protect DB resources. In particular, issues on configuration are examined here. For information on using the locking subsystem to manage non-DB resources, see the <em class="citetitle">Berkeley DB Programmer's Reference Guide</em>. </p> <div class="sect2" lang="en" xml:lang="en"> <div class="titlepage"> <div> <div> <h3 class="title"><a id="configuringlock"></a>Configuring the Locking Subsystem</h3> </div> </div> </div> <p> You initialize the locking subsystem by specifying <span> <code class="literal">true</code> to the <code class="methodname">EnvironmentConfig.setInitializeLocking()</code> method. </span> </p> <p> Before opening your environment, you can configure various maximum values for your locking subsystem. Note that these limits can only be configured before the environment is opened. Also, these methods configure the entire environment, not just a specific environment handle. </p> <p> Finally, each bullet below identifies the <code class="filename">DB_CONFIG</code> file parameter that can be used to specify the specific locking limit. If used, these <code class="filename">DB_CONFIG</code> file parameters override any value that you might specify using the environment handle. </p> <p> The limits that you can configure are as follows: </p> <div class="itemizedlist"> <ul type="disc"> <li> <p> The maximum number of lockers supported by the environment. This value is used by the environment when it is opened to estimate the amount of space that it should allocate for various internal data structures. By default, 1,000 lockers are supported. </p> <p> To configure this value, use the <span> <code class="methodname">EnvironmentConfig.setMaxLockers()</code> method. </span> </p> <p> As an alternative to this method, you can configure this value using the <code class="filename">DB_CONFIG</code> file's <code class="literal">set_lk_max_lockers</code> parameter. </p> </li> <li> <p> The maximum number of locks supported by the environment. By default, 1,000 locks are supported. </p> <p> To configure this value, use the <span> <code class="methodname">EnvironmentConfig.setMaxLocks()</code> method. </span> </p> <p> As an alternative to this method, you can configure this value using the <code class="filename">DB_CONFIG</code> file's <code class="literal">set_lk_max_locks</code> parameter. </p> </li> <li> <p> The maximum number of locked objects supported by the environment. By default, 1,000 objects can be locked. </p> <p> To configure this value, use the <span> <code class="methodname">EnvironmentConfig.setMaxLockObjects()</code> method. </span> </p> <p> As an alternative to this method, you can configure this value using the <code class="filename">DB_CONFIG</code> file's <code class="literal">set_lk_max_objects</code> parameter. </p> </li> </ul> </div> <p> For a definition of lockers, locks, and locked objects, see <a class="xref" href="blocking_deadlocks.html#lockresources" title="Lock Resources">Lock Resources</a>. </p> <p> For example, to configure the maximum number of locks that your environment can use: </p> <pre class="programlisting">package db.txn; import com.sleepycat.db.DatabaseException; import com.sleepycat.db.Environment; import com.sleepycat.db.EnvironmentConfig; import java.io.File; import java.io.FileNotFoundException; ... Environment myEnv = null; try { EnvironmentConfig myEnvConfig = new EnvironmentConfig(); myEnvConfig.setTransactional(true); myEnvConfig.setMaxLocks(5000); myEnv = new Environment(new File("/my/env/home"), myEnvConfig); } catch (DatabaseException de) { // Exception handling goes here } catch (FileNotFoundException fnfe) { // Exception handling goes here }</pre> </div> <div class="sect2" lang="en" xml:lang="en"> <div class="titlepage"> <div> <div> <h3 class="title"><a id="configdeadlkdetect"></a>Configuring Deadlock Detection</h3> </div> </div> </div> <p> In order for DB to know that a deadlock has occurred, some mechanism must be used to perform deadlock detection. There are three ways that deadlock detection can occur: </p> <div class="orderedlist"> <ol type="1"> <li> <p> Allow DB to internally detect deadlocks as they occur. </p> <p> To do this, you use <span><code class="methodname">EnvironmentConfig.setLockDetectMode()</code>.</span> This method causes DB to walk its internal lock table looking for a deadlock whenever a lock request is blocked. This method also identifies how DB decides which lock requests are rejected when deadlocks are detected. For example, DB can decide to reject the lock request for the transaction that has the most number of locks, the least number of locks, holds the oldest lock, holds the most number of write locks, and so forth (see the API reference documentation for a complete list of the lock detection policies). </p> <p> You can call this method at any time during your application's lifetime, but typically it is used before you open your environment. </p> <p> Note that how you want DB to decide which thread of control should break a deadlock is extremely dependent on the nature of your application. It is not unusual for some performance testing to be required in order to make this determination. That said, a transaction that is holding the maximum number of locks is usually indicative of the transaction that has performed the most amount of work. Frequently you will not want a transaction that has performed a lot of work to abandon its efforts and start all over again. It is not therefore uncommon for application developers to initially select the transaction with the <span class="emphasis"><em>minimum</em></span> number of write locks to break the deadlock. </p> <p> Using this mechanism for deadlock detection means that your application will never have to wait on a lock before discovering that a deadlock has occurred. However, walking the lock table every time a lock request is blocked can be expensive from a performance perspective. </p> </li> <li> <p> Use a dedicated thread or external process to perform deadlock detection. Note that this thread must be performing no other database operations beyond deadlock detection. </p> <p> To externally perform lock detection, you can use either the <code class="methodname">Environment.detectDeadlocks()</code> method, or use the <span class="command"><strong>db_deadlock</strong></span> command line utility. This method (or command) causes DB to walk the lock table looking for deadlocks. </p> <p> Note that like <span><code class="methodname">EnvironmentConfig.setLockDetectMode()</code>,</span> you also use this method (or command line utility) to identify which lock requests are rejected in the event that a deadlock is detected. </p> <p> Applications that perform deadlock detection in this way typically run deadlock detection between every few seconds and a minute. This means that your application may have to wait to be notified of a deadlock, but you also save the overhead of walking the lock table every time a lock request is blocked. </p> </li> <li> <p> Lock timeouts. </p> <p> You can configure your locking subsystem such that it times out any lock that is not released within a specified amount of time. To do this, use the <span><code class="methodname">EnvironmentConfig.setLockTimeout()</code></span> method. Note that lock timeouts are only checked when a lock request is blocked or when deadlock detection is otherwise performed. Therefore, a lock can have timed out and still be held for some length of time until DB has a reason to examine its locking tables. </p> <p> Be aware that extremely long-lived transactions, or operations that hold locks for a long time, may be inappropriately timed out before the transaction or operation has a chance to complete. You should therefore use this mechanism only if you know your application will hold locks for very short periods of time. </p> </li> </ol> </div> <p> For example, to configure your application such that DB checks the lock table for deadlocks every time a lock request is blocked: </p> <pre class="programlisting">package db.txn; import com.sleepycat.db.Environment; import com.sleepycat.db.EnvironmentConfig; import com.sleepycat.db.LockDetectMode; import java.io.File; ... Environment myEnv = null; try { EnvironmentConfig myEnvConfig = new EnvironmentConfig(); myEnvConfig.setTransactional(true); myEnvConfig.setInitializeCache(true); myEnvConfig.setInitializeLocking(true); myEnvConfig.setInitializeLogging(true); // Configure db to perform deadlock detection internally, and to // choose the transaction that has performed the least amount // of writing to break the deadlock in the event that one // is detected. envConfig.setLockDetectMode(LockDetectMode.MINWRITE); myEnv = new Environment(new File("/my/env/home"), myEnvConfig); // From here, you open your databases, proceed with your // database operations, and respond to deadlocks as // is normal (omitted for brevity). ...</pre> <p> Finally, the following command line call causes deadlock detection to be run against the environment contained in <code class="literal">/export/dbenv</code>. The transaction with the youngest lock is chosen to break the deadlock: </p> <pre class="programlisting">> /usr/local/db_install/bin/db_deadlock -h /export/dbenv -a y</pre> <p> For more information, see the <a class="ulink" href="http://www.oracle.com/technology/documentation/berkeley-db/db/utility/db_deadlock.html" target="_top"> <code class="literal">db_deadlock</code> reference documentation. </a> </p> </div> <div class="sect2" lang="en" xml:lang="en"> <div class="titlepage"> <div> <div> <h3 class="title"><a id="deadlockresolve"></a>Resolving Deadlocks</h3> </div> </div> </div> <p> When DB determines that a deadlock has occurred, it will select a thread of control to resolve the deadlock and then <span> throws <code class="literal">DeadlockException</code> in that thread. </span> If a deadlock is detected, the thread must: </p> <div class="orderedlist"> <ol type="1"> <li> <p> Cease all read and write operations. </p> </li> <li> <p> Close all open cursors. </p> </li> <li> <p> Abort the transaction. </p> </li> <li> <p> Optionally retry the operation. If your application retries deadlocked operations, the new attempt must be made using a new transaction. </p> </li> </ol> </div> <div class="note" style="margin-left: 0.5in; margin-right: 0.5in;"> <h3 class="title">Note</h3> <p> If a thread has deadlocked, it may not make any additional database calls using the handle that has deadlocked. </p> </div> <p> For example: </p> <pre class="programlisting">// retry_count is a counter used to identify how many times // we've retried this operation. To avoid the potential for // endless looping, we won't retry more than MAX_DEADLOCK_RETRIES // times. // txn is a transaction handle. // key and data are DatabaseEntry handles. Their usage is not shown here. while (retry_count < MAX_DEADLOCK_RETRIES) { try { txn = myEnv.beginTransaction(null, null); myDatabase.put(txn, key, data); txn.commit(); return 0; } catch (DeadlockException de) { try { // Abort the transaction and increment the // retry counter txn.abort(); retry_count++; if (retry_count >= MAX_DEADLOCK_RETRIES) { System.err.println("Exceeded retry limit. Giving up."); return -1; } } catch (DatabaseException ae) { System.err.println("txn abort failed: " + ae.toString()); return -1; } } catch (DatabaseException e) { try { // Abort the transaction. txn.abort(); } catch (DatabaseException ae) { System.err.println("txn abort failed: " + ae.toString()); return -1; } } } </pre> </div> </div> <div class="navfooter"> <hr /> <table width="100%" summary="Navigation footer"> <tr> <td width="40%" align="left"><a accesskey="p" href="blocking_deadlocks.html">Prev</a> </td> <td width="20%" align="center"> <a accesskey="u" href="txnconcurrency.html">Up</a> </td> <td width="40%" align="right"> <a accesskey="n" href="isolation.html">Next</a></td> </tr> <tr> <td width="40%" align="left" valign="top">Locks, Blocks, and Deadlocks </td> <td width="20%" align="center"> <a accesskey="h" href="index.html">Home</a> </td> <td width="40%" align="right" valign="top"> Isolation</td> </tr> </table> </div> </body> </html>