14 Jun, 2007

5 commits


13 Jun, 2007

7 commits


12 Jun, 2007

10 commits

  • Changes since 3.2.21
    Log File On-Disk Format Changes:
    None.
    New Features:
    None.
    General Environment Changes
    
       1. Fixed a bug that causes a memory leak for applications when many Environment objects are opened and closed and the CurrentTransaction or TransactionRunner class is used. The problem was reported in this JE Forum thread [#15444]
    
       2. Fixed a bug that prevents correct recovery of data records when duplicates are configured (DatabaseConfig.setSortedDuplicates(true)) for a deferred-write database (DatabaseConfig.setDeferredWrite(true)). When a crash occurs and duplicate data records are recovered that were flushed after the last checkpoint, the wrong key is used for those records. This bug does not occur when the environment is closed normally or a checkpoint is performed after writing data records. [#15474]
    
       3. Fixed a bug that prevents Database.preload from working when duplicates are configured (DatabaseConfig.setSortedDuplicates(true)) for the database. When the preload method was called for a database for which duplicate records exist, an assertion would fire or a NullPointerException would be thrown, depending on whether assertions are enabled. [#15475]
    
       4. Enhanced SerialBinding and the Direct Persistence Layer to use the current thread's context class loader when loading application classes. This allows the JE jar file to be deployed in application servers and other containers as a shared library rather than as an application jar. [#15447]
    
       5. Fixed a bug that caused JE to incorrectly behave as if the cache was very full. The symptoms of this bug are:
              * Because the problem is a cumulative accounting error, symptoms appear after running for a long period of time.
              * JE slows down or methods fail to return. Thread dumps show that the com.sleepycat.je.evictor.Evictor class is active and looping.
              * Dumping EnvironmentStats shows that the adminBytes size (EnvironmentStats.getAdminBytes()) is unreasonably large and may be larger than the maximum cache size or the heap size.
              * The problem occurs more frequently when blocks of sequential records are deleted.
          [#15505]
    
       6. Fixed a bug where the cache usage accounting could become inaccurate if the cache size was changed while the application is active (using Environment.setMutableConfig()) and there are outstanding record locks. [#15505]
    
       7. Fixed a bug that caused incorrect query results, and possibly errors during recovery, when a custom duplicate comparison method is configured and that method does not compare all bytes of its operands. A custom comparator should always compare all bytes of its operands and should return non-zero if they are unequal. When updating existing records, JE will detect this kind of incorrect comparator in order to prevent database inconsistency. The problem was reported in this JE Forum thread. [#15527]
    
    Direct Persistence Layer (DPL)
    None.
    API Changes:
    
       1. Added the EnvironmentStats.getAdminBytes() and getLockBytes() methods, which return the amount of cache used for transactions, administrative structures such as log cleaning metadata, and locks. These values are a subset of the value returned by Environment.getCacheDataBytes() and can provide more insight into how the cache is allocated between actual data and other objects. [#15505]
    
       2. Added the EnvironmentStats.getTotalLogSize method. This new method returns the approximate total size of the JE log files in bytes. It could be used, for example, to detect that the log has exceeded an application disk space limit. [#15480]
    
    Utility Changes:
    None.
    Configuration, Documentation, Test and Build Changes:
    None.
    
    git-svn-id: https://svn.forgerock.org/opendj/trunk@2086 41b1ffd8-f28e-4786-ab96-9950f0a78031
    coulbeck
     
  • 
    git-svn-id: https://svn.forgerock.org/opendj/trunk@2085 41b1ffd8-f28e-4786-ab96-9950f0a78031
    gary_williams
     
  • git-svn-id: https://svn.forgerock.org/opendj/trunk@2084 41b1ffd8-f28e-4786-ab96-9950f0a78031
    lutoff
     
  • git-svn-id: https://svn.forgerock.org/opendj/trunk@2083 41b1ffd8-f28e-4786-ab96-9950f0a78031
    al_xipe
     
  • git-svn-id: https://svn.forgerock.org/opendj/trunk@2082 41b1ffd8-f28e-4786-ab96-9950f0a78031
    lutoff
     
  • 
    git-svn-id: https://svn.forgerock.org/opendj/trunk@2081 41b1ffd8-f28e-4786-ab96-9950f0a78031
    smaguin
     
  • 
    git-svn-id: https://svn.forgerock.org/opendj/trunk@2080 41b1ffd8-f28e-4786-ab96-9950f0a78031
    smaguin
     
  • 
    git-svn-id: https://svn.forgerock.org/opendj/trunk@2079 41b1ffd8-f28e-4786-ab96-9950f0a78031
    smaguin
     
  • 
    git-svn-id: https://svn.forgerock.org/opendj/trunk@2078 41b1ffd8-f28e-4786-ab96-9950f0a78031
    smaguin
     
  • git-svn-id: https://svn.forgerock.org/opendj/trunk@2077 41b1ffd8-f28e-4786-ab96-9950f0a78031
    lutoff
     

11 Jun, 2007

10 commits


09 Jun, 2007

3 commits


08 Jun, 2007

5 commits

  • …ways exported before top entry.
    
    Clone the database config so that we don't share it between different databases.
    
    git-svn-id: https://svn.forgerock.org/opendj/trunk@2063 41b1ffd8-f28e-4786-ab96-9950f0a78031
    coulbeck
     
  • determining the length of time the account will remain locked due to
    authentication failures (it was supposed to be reported in seconds, but the
    value returned was in milliseconds).  This only impacted components that
    reported the time until the account was unlocked and not any calculation
    determining whether an account should currently be locked.
    
    OpenDS Issue Number:  1781
    
    
    git-svn-id: https://svn.forgerock.org/opendj/trunk@2062 41b1ffd8-f28e-4786-ab96-9950f0a78031
    neil_a_wilson
     
  • - After discussing the previously noninteractive option with Brian, I changed the CLI options somewhat.  The new CLI design is to specify an interactive option (if interactivity is desired) rather than to have to specify non-interactivity.  Additionally interactivity is only supposed to apply to gathering of required information and not whether or not and application is allowed to prompt for such things as continuation following an error.  To suppress these sort of prompts you should specify the silent option.  So the common plumbing (currently not used by setup or uninstall) now operates as described above and the upgrader now has a both silent and interactive options as described.
    
    - Make BuildExtractor a Launcher.  BuildExtractor is a simple program used by the CLI for unzipping an OpenDS .zip file.   Previously is was not very smart, relying on the Upgrader to report status and make sure the unzipping went OK.  This had the unfortunate side-effect of making the scripts overly complicated resulting in issue 1772, and made it difficult to support interactivity.  So the BuildExtractor has been made smarter so that it is capable of parsing the complete command line invocation, printing usage, and performing verification that it performed properly.
    
    - Addresses issue 1772 by simplifying and correcting the upgrade and upgrade.bat scripts.
    
    
    
    git-svn-id: https://svn.forgerock.org/opendj/trunk@2061 41b1ffd8-f28e-4786-ab96-9950f0a78031
    kenneth_suter
     
  • git-svn-id: https://svn.forgerock.org/opendj/trunk@2060 41b1ffd8-f28e-4786-ab96-9950f0a78031
    kenneth_suter
     
  • …the performance of the loggers when they are disabled.
    
    Fixed for issue 1780
    
    
    git-svn-id: https://svn.forgerock.org/opendj/trunk@2059 41b1ffd8-f28e-4786-ab96-9950f0a78031
    boli