org.forgerock / org.forgerock.opendj

11 Dec, 2013

3 commits

  • d79bc53a   Ti amo checkstyle! ... Browse Dir » 
    git-svn-id: https://svn.forgerock.org/opendj/trunk@9961 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • fff68807   GroupManager.java: ... Browse Dir » 
    Extracted method hasGroupMembershipUpdateControl().
In loadGroup(), used StaticUtils.collectionToString().
Automatic code cleanups.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9960 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • 34637012   Code cleanups ... Browse Dir » 
    
SetPropSubCommandHandler.java:
Extracted methods parseValue() and getPropertyDefinition().

Argument.java:
Improved the toString() + javadocs.
Minor code cleanups.

ArgumentParser.java:
Removed one parseArguments() method - never used.
Extracted method indentAndWrap() to match SubCommandArgumentParser.
Minor code cleanups.

SubCommandArgumentParser.java:
Removed one parseArguments() method - never used.
Minor code cleanups.

DsconfigOptionsTestCase.java:
Added testGenerateDoc().
Minor code cleanups.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9959 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     

10 Dec, 2013

2 commits

09 Dec, 2013

2 commits

  • 0170ca45   OPENDJ-1106 fractional-replication: adding entry on a replica as DM should not exit with RC 50 ... Browse Dir » 
    Code review: Matthew Swift


Regression caused by the fix for OPENDJ-948 unauthorized disclosure of directory contents.
The code was looking for the entry matching the DN we are adding which obviously was null here.
Since it was null, the code went on to wrongly check the following conditions: op.isInternalOperation() || op.isSynchronizationOperation() || op.isInnerOperation().
The correct fix was to avoid calling AccessControlHandler.canDiscloseInformation() where it did not make sense at all, when there was no info to protect yet.


LocalBackend*Operation.java:
Do not call the setResultCodeAndMessageNoInfoDisclosure() when we fail to grab a lock, because this is not disclosing any sensitive information at that point.

LocalBackendAddOperation.java:
Do not call the setResultCodeAndMessageNoInfoDisclosure() when we fail to grab a lock, because this is not disclosing any sensitive information at that point.
In lockParent(), reversed the exception to make the code easier to read.

LocalBackendModifyDNOperation.java:
Do not call the setResultCodeAndMessageNoInfoDisclosure() when we fail to grab a lock, because this is not disclosing any sensitive information at that point.
In newDirectoryException(), removed the entryDN parameter since now this method is always called with a non null entry.


AciHandler.java:
Code cleanup


git-svn-id: https://svn.forgerock.org/opendj/trunk@9943 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • 58b89515   Fixing thread leaks in Continuous Integration. ... Browse Dir » 
    ReplicationServerDomain.java:
Fixed NPEs in stopStatusAnalyzer() and stopMonitoringPublisher().
In getEligibleCSN(), removed useless null checks.

ServerWriter.java:
Code cleanup.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9941 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     

06 Dec, 2013

8 commits

  • a3bc45ea   OPENDJ-1088 (CR-2677) Wrong error message and result code when deleting branch a… ... Browse Dir » 
    …s a user with insufficient access rights


Fixed ConcurrentModificationException introduced in r7972 for OPENDJ-475: Incorrect behaviour/result code regarding non-critical controls.
The ConcurrentModificationExcepotion was hidden via several method calls, but generally it takes such form:
1. for (Control c : Operation.getRequestControls()) - directly or via indirect calls
2. AccessControlHandler.isAllowed()
2.1. Operation.removeRequestControl()
3. ConcurrentModificationException on next loop iteration at 1.



LocalBackendWorkflowElement.java:
Replaced isControlAllowed() by removeAllDisallowedControls() that uses Iterator.remove() instead of Operation.removeRequestControl().


LocalBackend*Operation.java:
In handleRequestControls(), processControls() and processRequestControls(), called LocalBackendWorkflowElement.removeAllDisallowedControls() before the for loop + removed call to LocalBackendWorkflowElement.isControlAllowed() in the loop body.
Tiny code cleanups.

ExtendedOperationBasis.java, ECLSearchOperation.java:
Used Iterator.remove() instead of Operation.removeRequestControl().

MultimasterReplication.java:
In findDomain(), used Iterator.remove() instead of Operation.removeRequestControl() + simplified the while loop code.


Operation.java, OperationWrapper.java, PreParseOperation.java:
Removed now unused removeRequestControl().

AbstractOperation.java:
Consequence of the change to Operation.
little code tidy up.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9934 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • ef7b85b5   ArgumentParser.java: ... Browse Dir » 
    Extracted method normalizeArguments().
Made printVersion() protected.

SubCommandArgumentParser.java:
Removed normalizeArguments() and printVersion().
Extracted method indentAndWrap2().
In isUsageArgumentPresent() and isVersionArgumentPresent(), code cleanup.
In toRefSect2(), used a StringBuilder.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9933 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • 7054a1b1   CR-2675 OPENDJ-1241 Server does not start after upgrade - Part II ... Browse Dir » 
    - fixed upgrade (removed white spaces in upgrade tasks, added Filter instead of LDAPFilter, added matcher and setting schema default matching rule/syntax.)

git-svn-id: https://svn.forgerock.org/opendj/trunk@9932 41b1ffd8-f28e-4786-ab96-9950f0a78031
    violette
     
  • c2bd7532   ArgumentParser.java, SubCommandArgumentParser.java: ... Browse Dir » 
    Moved try / catch / ignore inside getUsage() methods.
Extracted method printVersion(), dashVAccepted() and normalizeArguments().
Code cleanups.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9931 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • 27078ea4   CR-2675 OPENDJ-1241 Server does not start after upgrade - Part I ... Browse Dir » 
    - fixed Rebuild Index

git-svn-id: https://svn.forgerock.org/opendj/trunk@9928 41b1ffd8-f28e-4786-ab96-9950f0a78031
    violette
     
  • 1c405d09   Temporary workaround for OPENDJ-1243: ClassCastException while accessing Environ… ... Browse Dir » 
    …mentStats appearing frequently in debug logs

* avoiding collecting buggy stats.


git-svn-id: https://svn.forgerock.org/opendj/trunk@9926 41b1ffd8-f28e-4786-ab96-9950f0a78031
    matthew
     
  • ed8dea05   Javadoc cleanup. ... Browse Dir » 
    AbstractOperation.java:
Replaced javadocs duplicated from Operation interface by {@inheritDoc}.
Put {@inheritDoc} on one line only.
Removed empty lines.
Little code formatting.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9925 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • f9d1cf76   Fixed once and for all: privilige => privilege ... Browse Dir » 
    WhoAmIExtendedOperation.java:
Initialized supportedControlOIDs once and for all.
Applied automatic refactorings

ECLSearchOperation.java:
Applied automatic refactorings

git-svn-id: https://svn.forgerock.org/opendj/trunk@9924 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     

05 Dec, 2013

4 commits

  • 54fd8d4a   OPENDJ-1174 Transfer responsibility for populating the ChangeNumberIndexDB to ChangelogDB ... Browse Dir » 
    Code review: Matthew Swift



Fixing failing unit tests, removed the following stack trace which makes test and code fail.

   [testng] Failed Test:  org.opends.server.replication.plugin.HistoricalCsnOrderingTest#buildAndPublishMissingChangesSeveralEntriesTest
   [testng] Failure Cause:  com.sleepycat.je.ThreadInterruptedException: (JE 5.0.97) Environment must be closed, caused by: com.sleepycat.je.ThreadInterruptedException: Environment invalid because of previous exception: (JE 5.0.97) <https://builds.forgerock.org/job/OpenDJ%20-%20trunk%20-%20postcommit/ws/build/unit-tests/package-instance/HistoricalCsnOrdering> Thread interrupted prior to logging the commit THREAD_INTERRUPTED: InterruptedException may cause incorrect internal state, unable to continue. Environment is invalid and must be closed.
   [testng]     com.sleepycat.je.ThreadInterruptedException.wrapSelf(ThreadInterruptedException.java:99)
   [testng]     com.sleepycat.je.dbi.EnvironmentImpl.checkIfInvalid(EnvironmentImpl.java:1594)
   [testng]     com.sleepycat.je.Database.checkEnv(Database.java:2154)
   [testng]     com.sleepycat.je.Database.getDatabaseName(Database.java:1939)
   [testng]     org.opends.server.replication.server.changelog.je.ReplicationDbEnv.closeDB(ReplicationDbEnv.java:457)
   [testng]     org.opends.server.replication.server.changelog.je.ReplicationDbEnv.shutdown(ReplicationDbEnv.java:435)
   [testng]     org.opends.server.replication.server.changelog.je.JEChangelogDB.shutdownDB(JEChangelogDB.java:409)
   [testng]     org.opends.server.replication.server.ReplicationServer.shutdown(ReplicationServer.java:702)
   [testng]     org.opends.server.replication.server.ReplicationServer.remove(ReplicationServer.java:1028)
   [testng]     org.opends.server.replication.plugin.HistoricalCsnOrderingTest.buildAndPublishMissingChangesSeveralEntriesTest(HistoricalCsnOrderingTest.java:308)



DirectoryThread.java:
In initiateShutdown(), removed the call to Thread.interrupt() added in r9911.

ChangeNumberIndexer.java:
Overrode initiateShutdown() to stop this thread on shutdown: used notify() to wake up the thread if it was waiting.
In run(), checked whether shutdown has been initiated for this thread before calling wait(): this protects against missed notifications.

JEReplicaDB.java:
In run(), removed call to Thread.currentThread().interrupt().

git-svn-id: https://svn.forgerock.org/opendj/trunk@9918 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • bc12002c   #2 Fixed a readWriteLock deadlock happening in tests on calling ChangelogDB.remo… ... Browse Dir » 
    …veDB() which then calls ReplicationDB.shutdown().
Code is blocked acquiring a writeLock on the dbCloseLock, but a readLock is never unlocked.
Debugging showed that the ChangeNumberIndexDB is holding a cursor on a ReplicationDB, said cursor holds the readLock.
Despite the ChangeNumberIndexer thread being asked to initiateShutdown(), the thread is waiting indefinitely for new changes which will never come.

DirectoryThread.java:
In initiateShutdown(), call interrupt() to ensure the thread will be interrupted if it calls Object.wait() before checking whether it needs to shutdown, or if it is already waiting.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9911 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • 74056426   ReplicationServerDomain.java: ... Browse Dir » 
    Removed change which was not intended for commit, at least not for now.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9910 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • 08fe888b   Fixed a readWriteLock deadlock happening on calling ChangelogDB.removeDB() which… ... Browse Dir » 
    … then calls ReplicationDB.shutdown().
Code is blocked acquiring a writeLock on the dbCloseLock, but a readLock is never unlocked.
Debugging showed that the ChangeNumberIndexDB is holding a cursor on a ReplicationDB, said cursor holds the readLock.
The solution is to not try to call ReplicationDB.shutdown() in ReplicationServerDomain.shutdownDomain() because it will be closed by JEChangelogDB.shutdownDB().


ReplicationServerDomain.java:
Removed the call to ReplicationDB.shutdownDomain(DN).

ReplicationDomainDB.java, JEChangelogDB.java:
Removed shutdownDomain(DN) which was never used outside of a complete shutdown of the ReplicationServer.
Inlined innerShutdownDomain().

JEReplicaDBCursor.java:
In next(), extracted a local variable to protect against a concurrent update to the cursor field.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9909 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     

04 Dec, 2013

3 commits

  • f35ae63f   ChangeNumberIndexer.java: ... Browse Dir » 
    In run(), made sure all cursors are properly closed, even in case of a RuntimeException terminating the thread.
Moved resetNextChangeForInsertDBCursor() out of removeAllCursors().
In removeAllCursors(), closed the CompositeDBCursor.

ServerHandler.java:
Made shutdownWriter volatile.

JEReplicaDBCursor.java:
Avoided a NullPointerException.



git-svn-id: https://svn.forgerock.org/opendj/trunk@9901 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • c5c06fc3   Fixing stalled builds. ... Browse Dir » 
    Fixed problem introduced in r9887: an open cursor was never released.

ReplicationDB.java:
In ctors, made sure we always call close on a cursor when it is not held. This will  permanently fix such issues in the ctors.
Other changes are just formatting changes.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9899 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • 94b852f2   OPENDJ-1107: Unexpected error when importing/exporting data as non-privileged user. ... Browse Dir » 
    CR-2657.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9898 41b1ffd8-f28e-4786-ab96-9950f0a78031
    ludo
     

03 Dec, 2013

3 commits

  • 079813d7   Fix for OPENDJ-1239 – dsreplication logs warnings for each replication server under cn=monitor. ... Browse Dir » 
    Don't check server-id for replication servers.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9891 41b1ffd8-f28e-4786-ab96-9950f0a78031
    ludo
     
  • 5d947372   Fixed a NullPointerException that is causing the following error message: ... Browse Dir » 
    (ERROR) [03/Dec/2013:12:10:55 +0100] category=CORE severity=FATAL_ERROR msgID=327820 msg=An uncaught exception during processing for thread Change number indexer has caused it to terminate abnormally. The stack trace for that exception is: NullPointerException (ChangeNumberIndexer.java:337 ChangeNumberIndexer.java:324 ChangeNumberIndexer.java:266 ChangeNumberIndexer.java:355)

ChangeNumberIndexer.java:
In getPrecedingCSN(), handled the null case.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9888 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • 4271a061   OPENDJ-1217 Servers not synchronized after a conflicting replace on multi-valued… ... Browse Dir » 
    … attribute on 2 different servers


Fixed regression introduced in r9788 by introducing a similar behaviour as what the code was doing before.

Replica 1 was trying to read a change done on itself and send it to replica 2, but failed to do so.
Hence replica 2 did not receive this UpdateMsg and as a consequence could not resolve it.

Previous code was trying to cursor to a specific CSN in the DB. It could not find it, hence tried to move after it in the DB, but could not do it because the end of the DB was reached. Previous was throwing an exception there, later catching it and then returning an empty cursor.
Newer code removed the useless catch block which let the exception fail this process.
The fix removed throwing the exception and instead returned an empty cursor.



ReplicationDB.java:
In ctor, do not throw an exception, but return an empty cursor instead (Previous code was throwing an exception, catching and ignoring it, then returning an empty cursor).

MessageHandler.java:
Prevented a NullPointerException from being thrown, then caught and ignored.


git-svn-id: https://svn.forgerock.org/opendj/trunk@9887 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     

02 Dec, 2013

9 commits

  • f1e0be4d   One liner fix to OPENDJ-1226 : Upgrade should only consider .ldif files under config/schema. ... Browse Dir » 
    Make sure we filter config/schema files like when reading it, considering only .ldif files.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9884 41b1ffd8-f28e-4786-ab96-9950f0a78031
    ludo
     
  • 64d4c04e   Fix for OPENDJ-1228, Concatenated schema may contain more than valid schema, pos… ... Browse Dir » 
    …sibly leading to further issues.
Make sure we only consider *.ldif files when handling schemas under config/schema

git-svn-id: https://svn.forgerock.org/opendj/trunk@9883 41b1ffd8-f28e-4786-ab96-9950f0a78031
    ludo
     
  • 186d2ac4   Renamed ds-cfg-compute-changenumber to ds-cfg-compute-change-number. ... Browse Dir » 
    
git-svn-id: https://svn.forgerock.org/opendj/trunk@9882 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • d91db94c   OPENDJ-1174 (CR-2631) Transfer responsibility for populating the ChangeNumberIndexDB to ChangelogDB ... Browse Dir » 
    
Hooked the ChangeNumberIndexer thread into our code base.
Managed thread lifecycle.
Added configuration to turn it on / off.
Removed code that builds the ChangeNumberIndexDB from ECLServerHandler.


ChangeNumberIndexer.java:
Added clear() + doClear field for unit tests
Used thread safe versions of MultiDomainServerState.
In run(), merged all the try blocks to factorize the catch blocks + supported doClear + fixed a bug with the mediumConsistencyRUV.
In removeCursor(), closed the cursors.
Added removeAllCursors(), getPrecedingCSN().
Renamed crossDomainDBCursor to nextChangeForInsertDBCursor and newCompositeDBCursor() to resetNextChangeForInsertDBCursor().
Changed newCursors from ConcurrentMap<CSN, DN> to ConcurrentMap<Pair<DN, Integer>, CSN> to ensure minimum memory consumption.

JEChangelogDB.java:
Removed dbDirectoryName field.
Added config, cnIndexer fields.
Changed ctor to accept ReplicationServerCfg.
In initializeDB(), shutdownDB(), clearDB(), publishUpdateMsg() took appropriate actions with the cnIndexer thread.
Added setComputeChangeNumber().
In getCursorFrom(), accept null parameter.
In setPurgeDelay(), also call it on the cnIndexDB.


MultiDomainServerState.java:
Made implementation to be fully thread safe.
Changed list from Map to ConcurrentMap + removed synchrnoized blocks.
Renamed update(DN, ServerState) to replace(DN, ServerState).
Added new version of update(DN, ServerState).

ECLServerHandler.java:
Removed assignNewChangeNumberAndStore() that builds the ChangeNumberIndexDB + adapted the code that was using this method.
Used MultiDomainServerState.replace().


ReplicationServerConfiguration.xml, ReplicationServerCfgDefn.properties:
Added ds-cfg-compute-changenumber.

ExternalChangelogDomainConfiguration.xml, ExternalChangelogDomainCfgDefn.properties:
Updated description

ReplicationServer.java:
In applyConfigurationChange(), handled computeChangenumber config.
Used MultiDomainServerState.replace().

ChangelogDB.java:
Added setComputeChangeNumber().

ReplServerFakeConfiguration.java:
Added isComputeChangenumber() and setComputeChangenumber().


ExternalChangeLogTest.java:
Adapted the code to the use of the ChangeNumberIndexer thread.

JEChangeNumberIndexDB.java:
Extracted runPurge() from run().


ReplicationDbEnv.java
Code cleanup.


git-svn-id: https://svn.forgerock.org/opendj/trunk@9881 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • 34aee712   CSN.java, CSNGenerator.java: ... Browse Dir » 
    Updated javadoc.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9880 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • 999aa775   Fix OPENDJ-1079: OpenDJ UpgradeCli adds HTTP connection handler, which breaks To… ... Browse Dir » 
    …mcat6 when OpenDJ is an embeddedDS

* refine the class loading exception handling.



git-svn-id: https://svn.forgerock.org/opendj/trunk@9878 41b1ffd8-f28e-4786-ab96-9950f0a78031
    matthew
     
  • 34091311   Fix OPENDJ-1079: OpenDJ UpgradeCli adds HTTP connection handler, which breaks To… ... Browse Dir » 
    …mcat6 when OpenDJ is an embeddedDS

* avoid initializing classes during class property value validation as this causes dependencies to be loaded which may be missing in the case where the component is disabled. The class will be fully loaded later if the component is enabled (verified with HTTP connection handler)
* prevent class loading errors from causing the server initialization to hang
* attempt to catch and propagate class loading errors up the stack. They are still not displayed in any logs (OPENDJ-1227).



git-svn-id: https://svn.forgerock.org/opendj/trunk@9876 41b1ffd8-f28e-4786-ab96-9950f0a78031
    matthew
     
  • 8f4a1e26   Fix bad comment wrapping. ... Browse Dir » 
    git-svn-id: https://svn.forgerock.org/opendj/trunk@9871 41b1ffd8-f28e-4786-ab96-9950f0a78031
    matthew
     
  • 6a2e79db   CR-2635 Second fix for OpenDJ-1196. As an issue was find by Christophe when temp… ... Browse Dir » 
    …late 02-config was deleted, added a check on template existence before copying it.
- Added i18n/path in messages.
- Added message ouput for post upgrade task.
- Moved temp file creation as it leaves .temp files in schema folder when upgrade fails.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9867 41b1ffd8-f28e-4786-ab96-9950f0a78031
    violette
     

29 Nov, 2013

3 commits

  • a9386f95   Fixing the build. ... Browse Dir » 
    LDAPReplicationDomain.java:
In parseFractionalConfig(), added back the null checks on the Iterator because some client code passes in null.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9863 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • 5acb3d03   Fixing the build. ... Browse Dir » 
    LDAPReplicationDomain.java:
Removed private modifier that was added too eagerly.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9861 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     
  • d817f92d   LDAPReplicationDomain.java: ... Browse Dir » 
    Added missing spaces in exception messages.
Extracted methods getExpandedNames() and needsExpanding() from getIncludedAttributes().

git-svn-id: https://svn.forgerock.org/opendj/trunk@9860 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     

28 Nov, 2013

1 commit

  • 625b9927   Code cleanup ... Browse Dir » 
    Added toString() for debuggging.
Added or completed javadocs.
Fixed a little bug in test code.

git-svn-id: https://svn.forgerock.org/opendj/trunk@9852 41b1ffd8-f28e-4786-ab96-9950f0a78031
    JnRouvignac
     

27 Nov, 2013

2 commits