Permalink
Commits on Jul 20, 2011
  1. ISPN-1123 - TransactionSpanningReplicatedCachesTest doesn't use the s…

    Dan Berindei
    Dan Berindei committed Jul 19, 2011
    …ame configuration on all nodes
  2. ISPN-1123 - Better formatting for cache entries

    Dan Berindei
    Dan Berindei committed Jul 20, 2011
  3. ISPN-1123 - Improve logging in BatchAsyncCacheStoreTest

    Dan Berindei Dan Berindei
    Dan Berindei authored and Dan Berindei committed Jun 26, 2011
  4. ISPN-1239 - ConcurrentOverlappingLeaveTest failing with the graceful …

    Dan Berindei authored and maniksurtani committed Jul 19, 2011
    …shutdown workaround
    
    Let's say the initial cluster members are {A, B, C, D}, numOwners = 3, and D and C leave in quick succession.
    With initial owners(k) = {B, C, D}, when D leaves B expects C to push the key to A.
    If but C dies before pushing it, on the following rehash B doesn't push the key to A.
    
    The solution is to remember the last CH for which rehashing completed successfully and base every rehash on the last successful CH.
    
    Another problem is that pending transactions should not be allowed if a rehash is interrupted until the next rehash has completed.
  5. ISPN-1239 - Graceful shutdown should be supported

    Dan Berindei authored and maniksurtani committed Jul 19, 2011
    Apply workaround in the test suite: stop all the caches before stopping the cache managers.
    Related fixes:
    * Allow unregistering the cache instance from JMX when the cache is stopped.
    * Don't log a huge stacktrace in RehashTask and in CommandAwareRpcDispatcher when the cache is shutting down.
    * Don't log "Ignoring call to stop" during normal shutdown.
    * Don't clear the data container before stopping a cache - that cache might be participating in an ongoing rehash.
  6. ISPN-1123 Avoid race conditions in state transfer tests

    galderz committed Jul 19, 2011
    * Removed unusued references to internal state transfer test classes.
    * Removed redundant log check.
Commits on Jul 19, 2011
Commits on Jul 18, 2011
  1. ISPN-1250 Upgrade TestNG dependency version only for the CDI integrat…

    kevinpollet committed Jul 18, 2011
    …ion module and fix TestNG dependency scope.
  2. ISPN-1123 Avoid @Test(timeOut=N) tests sharing same discovery instances

    galderz committed Jul 18, 2011
    * Upgraded TestNG to 6.11 as well which unfortunately does not yet
    provide a clean way to hook into the thread that executes timeOut
    enabled tests.
  3. Fix m2e warnings

    pmuir authored and galderz committed Jul 14, 2011
  4. ISPN-1232 Also scan OSGi for Infinispan modules

    pmuir authored and galderz committed Jul 14, 2011
    Note that this is a hack until 5.1 when we will
    revisit entirely how modules are implemented to
    make them more OSGi (and other modular env)
    friendly.
  5. ISPN-1123 ConcurrentStartWithReplTest should check for view completed

    galderz authored and Sanne committed Jul 18, 2011
    * Increased TEST_PING discovery timeout to avoid issues with thread
    scheduling which could have an impact on view stablishment.
  6. ISPN-1123 Make testReplaceExpiry more resilient

    galderz authored and Sanne committed Jul 15, 2011
    This should help fix random CassandraCacheStore test.
  7. ISPN-1243 - JdbcStringBasedCacheStore closes the StateTransferManager…

    Dan Berindei authored and maniksurtani committed Jul 15, 2011
    …Impl marshaller while state transfer is in progress
    
    NonStringKeyStateTransferTest was failing because
    1) The cache store was not "shared", but both cluster nodes were using the same in-memory DB URL.
    2) StateTransferManagerImpl was writing the same entries to the DB twice, once in applyInMemoryState() and then again in applyPersistentState().
  8. ISPN-1243 - JdbcStringBasedCacheStore closes the StateTransferManager…

    Dan Berindei authored and maniksurtani committed Jul 15, 2011
    …Impl marshaller while state transfer is in progress
    
    I changed StateTransferManagerImpl to use the reentrant flag for StreamingMarshaller.startObjectInput()
    I also changed FileCacheStore to not use the reentrant flag.
Commits on Jul 15, 2011
  1. ISPN-1244 Data loss possible during state transfer when numOwners == 1

    maniksurtani authored and Dan Berindei committed Jul 15, 2011
    * Add test contributed by Alex Henveld and Sanne Grinovero
    * Fix RebalanceTask's state pusher selection algorithm
  2. Improve javadoc for group api

    pmuir committed Jul 15, 2011