Commit Graph

246 Commits

Author SHA1 Message Date
gtully 4386750fdb for mkahadb populate mdc activemq.persistenceDir such that logging for a particular instance can be identified and correlated 2016-07-04 11:19:56 +01:00
Christopher L. Shannon (cshannon) 4d6cc4b460 https://issues.apache.org/jira/browse/AMQ-6303
Properly setting typeCode value for new journal files used for ack
compaction
2016-05-23 18:42:29 +00:00
Timothy Bish 15405af2e6 https://issues.apache.org/jira/browse/AMQ-6292
allow for batch writes of forwarded acks.
2016-05-13 14:56:01 -04:00
Christopher L. Shannon (cshannon) c8a6171d04 https://issues.apache.org/jira/browse/AMQ-6288
Switching the checkpoint lock to a readlock when forwarding acks to
prevent other journal updates from being blocked.
2016-05-12 14:39:12 +00:00
Christopher L. Shannon (cshannon) e53e340262 https://issues.apache.org/jira/browse/AMQ-6288
The ack compaction task now acquires the checkpoint lock while it runs
to prevent a checkpoint from running at the same time unintentionally.
Also, getJournalLocation is now protected by a try/catch to handle
errors.
2016-05-11 13:16:57 +00:00
Christopher L. Shannon (cshannon) 91213010b3 https://issues.apache.org/jira/browse/AMQ-6285
Added an extra isShutdown() failsafe check when detecting if the
checkpoint thread should start
2016-05-09 19:35:02 +00:00
Christopher L. Shannon (cshannon) 7bdcca1bda https://issues.apache.org/jira/browse/AMQ-6285
Properly nulling out the scheduler service in MessageDatabase after it
is shutdown on a store close so that if the store is restarted the
thread will properly restart.
2016-05-09 19:06:40 +00:00
Christopher L. Shannon (cshannon) db3f8b3554 Revert "https://issues.apache.org/jira/browse/AMQ-6285"
This reverts commit 60b0c4f85a.

Inadvertantly commited a bunch of changes by mistake
2016-05-09 19:05:52 +00:00
Christopher L. Shannon (cshannon) 60b0c4f85a https://issues.apache.org/jira/browse/AMQ-6285
Properly nulling out the scheduler service in MessageDatabase after it
is shutdown on a store close so that if the store is restarted the thread will
properly restart.
2016-05-09 19:03:38 +00:00
gtully 65cef69130 https://issues.apache.org/jira/browse/AMQ-5603 - reverting default preallocatonScope to entire_journal b/c async only really works for ssd 2016-05-05 13:06:42 +01:00
gtully 1c4108545c https://issues.apache.org/jira/browse/AMQ-6277 - journal getNextLocation needs too passes to skip past if target is not initialized 2016-05-04 22:09:06 +01:00
gtully 555cd2b6bb https://issues.apache.org/jira/browse/AMQ-5603 - catch interrupted on chunked zeros case 2016-05-03 17:25:14 +01:00
gtully ba77b9f55a https://issues.apache.org/jira/browse/AMQ-6277 - tidy up logic that determines recovery location so that we don't recovery from the end of the journal in error on normal restart. This avoids suprious recovery logging 2016-05-03 12:47:49 +01:00
gtully 62bdbb0db5 https://issues.apache.org/jira/browse/AMQ-5603 - add preallocationScope=full_journal_async that will preallocate a journal in advance or use to avoid latency jitter on journal rotation. Added none option to disable preallocation 2016-04-29 16:57:28 +01:00
gtully 43d8205cf0 reduce the throughput for ci 2016-04-22 14:53:14 +01:00
gtully 930a74c696 https://issues.apache.org/jira/browse/AMQ-5578 - enable org.apache.activemq.kahaDB.files.skipMetadataUpdate for tests to validate. remove some redundant call to limit and add latencey test to validate the need for outof band or pool allocation of journals 2016-04-22 12:29:31 +01:00
Christopher L. Shannon (cshannon) 11622b3af3 https://issues.apache.org/jira/browse/AMQ-6256
Moving beforeMarshall back to the store implementations because we don't
want all store implementations to marshall (such as memory store)

This reverts commit b9b98a45ce.
2016-04-18 12:34:25 +00:00
Christopher L. Shannon (cshannon) 1ffac14f6e https://issues.apache.org/jira/browse/AMQ-6256
cleanup imports
2016-04-15 14:23:54 +00:00
Christopher L. Shannon (cshannon) b9b98a45ce https://issues.apache.org/jira/browse/AMQ-6256
Moving beforeMarshall call out of the store and into the actual
destination
2016-04-15 14:17:00 +00:00
Christopher L. Shannon (cshannon) b9f9f03829 https://issues.apache.org/jira/browse/AMQ-6256
Calling beforeMarshall on messages when they async stored before the
store task is run and before consumer dispatch to prevent two threads
from trying to mutate the message state at the same time.
2016-04-15 13:01:21 +00:00
Christopher L. Shannon (cshannon) 092c56d70c https://issues.apache.org/jira/browse/AMQ-6250
Checking for null before closing the journal in
MultiKahaDBPersistentAdapter.  Also protecting against multiple calls to
start and stop.
2016-04-14 18:39:14 +00:00
Christopher L. Shannon (cshannon) cbad8babe5 https://issues.apache.org/jira/browse/AMQ-6203
Adding a flag to enable or disable Ack Compaction
Currently defaults to enabled for 5.14.0 and above
2016-04-12 13:30:10 +00:00
Timothy Bish a9521dcebf https://issues.apache.org/jira/browse/AMQ-6203
Expose configuration on the persistence adapter level.
Double the test timeout and increase the compaction frequency to account
for very slow CI boxes.
2016-03-15 12:02:40 -04:00
Timothy Bish 946e62d702 https://issues.apache.org/jira/browse/AMQ-6203
Rewrite older acks that can be preventing GC of log files.
2016-03-14 11:04:57 -04:00
Timothy Bish b39ab7867c https://issues.apache.org/jira/browse/AMQ-6207
Perform writes / sync in predicatable order first to the recovery file
and then into the main index file.
2016-03-11 09:59:07 -05:00
gtully 13ec994939 Revert "https://issues.apache.org/jira/browse/AMQ-4495 - revisit. Reinstate check for space on pagein, so that highWaterMark is respected and full state is not reached, hense pfc is not triggered in error"
This reverts commit d8cf54b0a9.
2016-03-07 16:26:25 +00:00
Timothy Bish 5e0f49332b https://issues.apache.org/jira/browse/AMQ-5578
Remove unnecessary buffer initialization after confirmation that it is
not needed.
2016-03-03 17:35:21 -05:00
Timothy Bish edfc23ee9d https://issues.apache.org/jira/browse/AMQ-5578
Add additional preallocation strategy [CHUNKED contributed by user
jtahlborn.
2016-03-03 17:34:07 -05:00
Timothy Bish be4ad3d1c8 https://issues.apache.org/jira/browse/AMQ-5578
Replace system.out with a trace log.
2016-03-03 17:00:11 -05:00
Timothy Bish e8f44a29dd https://issues.apache.org/jira/browse/AMQ-5578
Revert to previous behavior of explicit init.
2016-03-02 17:18:30 -05:00
Timothy Bish e89b7a57f1 https://issues.apache.org/jira/browse/AMQ-5578
Remove unnecessary zero init block as the java spec requires the
allocations to do this already.  Fix some warnings.
2016-03-02 16:17:43 -05:00
gtully d8cf54b0a9 https://issues.apache.org/jira/browse/AMQ-4495 - revisit. Reinstate check for space on pagein, so that highWaterMark is respected and full state is not reached, hense pfc is not triggered in error 2016-03-01 11:44:41 +00:00
gtully 499e39e52c https://issues.apache.org/jira/browse/AMQ-6164 - allow journal write batching on a single destination 2016-02-09 12:50:27 +00:00
Christopher L. Shannon (cshannon) 25ff5699f1 https://issues.apache.org/jira/browse/AMQ-6158
Computing messageSize for a durable subscription in KahaDB now runs much
faster (n vs n^2) which is noticable when there are a large number of
pending messages for a durable subscription.
2016-02-04 15:17:49 +00:00
Timothy Bish 8c4b5f485d https://issues.apache.org/jira/browse/AMQ-6152
Ensure that when add / remove commands are colocated they don't prevent
the log from being GC'd once it is unreferenced.
2016-02-02 20:28:24 -05:00
Timothy Bish 4ecff14030 NO-JIRA Update master version to 5.14.0-SNAPSHOT to match the eventual
release version.
2016-01-20 12:40:20 -05:00
Christopher L. Shannon (cshannon) b4aa53d806 https://issues.apache.org/jira/browse/AMQ-6133
Fixing the processing of KahaUpdateMessageCommand to add the message to
the store if it doesn't exist, instead of ignoring it.  This will
prevent message loss in certain cases when the KahaDB index has to be
rebuilt.
2016-01-20 16:59:07 +00:00
Timothy Bish 193f6be687 https://issues.apache.org/jira/browse/AMQ-6131
Fix check for durable sub with no pending messages during checkpoint
cleanup.
2016-01-18 17:43:18 -05:00
Christopher L. Shannon (cshannon) 507d40a79f https://issues.apache.org/jira/browse/AMQ-6097
Fixing JournalCorruptionEofIndexRecoveryTest class

Because of the size being stored in the message order index, the test
testRecoveryAftercorruptionCheckSumExistingIndex was no longer passing
as that test was corrupting the size value in the journal, but not the
size value in the index.  The test now corrupts both values so that the
expect exception occurs again.
2015-12-21 16:46:47 +00:00
Christopher L. Shannon (cshannon) 246ccb8e04 https://issues.apache.org/jira/browse/AMQ-6097
Fixing KahaDB so that the correct marshaller is used for the message
keys inside of the message order index.  This will ensure that message
size metrics are accurate.
2015-12-18 19:26:24 +00:00
gtully 31d99b6aa9 https://issues.apache.org/jira/browse/AMQ-6083 - resolve some test regressions - AMQ2149Test - fail to start with missing data file. Ensure ack map is always saved on modification 2015-12-11 13:30:55 +00:00
gtully 768fa17085 https://issues.apache.org/jira/browse/AMQ-6005 - have plist whack it existing state on start. Start the temp store after the primary persistence adapter which does the locking 2015-12-10 14:24:03 +00:00
gtully 5db5f3e39a https://issues.apache.org/jira/browse/AMQ-6083 - ignoreMissingJournalfiles false - check for missing acks and corruption anywhere and error out so that corruption does not go unnoticed - fix and test 2015-12-09 12:30:29 +00:00
Christopher L. Shannon (cshannon) 7a7c70ad75 https://issues.apache.org/jira/browse/AMQ-6082
Propertly re-setting the storeOpenWireVersion from the BrokerService
on the KahaDB Metadata if a corrupted index is detected and the
Metadata has to be recreated.
2015-12-08 20:10:11 +00:00
Torsten Mielke 66c348b1b8 https://issues.apache.org/jira/browse/AMQ-6071 - Log info about corrupted journal records at WARN level 2015-12-04 11:29:55 +00:00
Christopher L. Shannon (cshannon) cfb4ca9f54 [maven-release-plugin] prepare for next development iteration 2015-11-30 13:48:58 +00:00
Christopher L. Shannon (cshannon) abfe038ddf [maven-release-plugin] prepare release activemq-5.13.0 2015-11-30 13:48:11 +00:00
Christopher L. Shannon (cshannon) 8f30866fdf https://issues.apache.org/jira/browse/AMQ-6063
Fixing potential NullPointerException during KahaDB index recovery.
2015-11-25 16:02:52 +00:00
Christopher L. Shannon (cshannon) e07809dec6 NO JIRA - Clear journalSize counter
The journalSize counter should be set to 0 after KahaDB is shut down.
Usually, a store is not restarted again after stopping, but if it ever
was the value would be wrong since it wasn't reset back to 0.
2015-11-24 17:53:34 +00:00
gtully fd0f71a4c8 https://issues.apache.org/jira/browse/AMQ-6033 - fix shadowing error, additional test 2015-11-02 15:08:22 +00:00
gtully dc06c8dc75 https://issues.apache.org/jira/browse/AMQ-5994 https://issues.apache.org/jira/browse/AMQ-4000 - proper fix for duplicate sub info from the store on recovery failure from AMQ2149Test. Additional test from Christopher L 2015-10-06 12:01:24 +01:00
gtully f6e1054a4b https://issues.apache.org/jira/browse/AMQ-5984 - fix sync on get for access ordered LinkedHashMap 2015-09-28 14:20:45 +01:00
gtully fcabcd282d https://issues.apache.org/jira/browse/AMQ-5960 - rework fix to reset the next sequence so that the next ack position and message reference gets cleared up in normal operation 2015-09-23 13:15:29 +01:00
Dejan Bosanac ee4672baaf https://issues.apache.org/jira/browse/AMQ-5972 - wildcard policy entries not applied in correct order 2015-09-17 11:08:21 +02:00
Dejan Bosanac f4a2543def https://issues.apache.org/jira/browse/AMQ-5960 - durable sub pendingQueueSize after force shutdown - fix npe test failure 2015-09-10 14:13:45 +02:00
Christopher L. Shannon (cshannon) d1dface01a https://issues.apache.org/jira/browse/AMQ-5923
Adding a null pointer check in PListImpl to prevent a null
pointer exception on remove
2015-09-10 11:23:27 +00:00
Dejan Bosanac 4ab55f13ae https://issues.apache.org/jira/browse/AMQ-5960 - durable sub pendingQueueSize after force shutdown 2015-09-10 11:10:37 +02:00
Christopher L. Shannon (cshannon) 734fb7dda3 https://issues.apache.org/jira/browse/AMQ-5923
Adding metrics to track the pending message size for a queue and for
subscribers.  This is useful so that not only the pending count is
known but also the total message size left to consume. Also improving
the message size store tests as well.
2015-09-09 18:12:15 +00:00
Christopher L. Shannon (cshannon) 04ee70a161 https://issues.apache.org/jira/browse/AMQ-5889
Adding support for auto detection of wire protocols over a transport.
OpenWire, AMQP, STOMP, and MQTT can all be detected and the broker
will properly handle each one over a given Transport.  Currently
auto TCP, NIO, SSL, and NIO+SSL transports can handle auto-detection
of the wire format and client but support could be added in the
future for other transports like websockets.
2015-08-11 19:39:29 +00:00
Timothy Bish 6cdd750ddc [maven-release-plugin] prepare for next development iteration 2015-08-10 11:04:26 -04:00
Timothy Bish a9eeb03520 [maven-release-plugin] prepare release activemq-5.12.0 2015-08-10 11:04:17 -04:00
Christopher L. Shannon c6542a921b AMQ-5748 - Cleanup and clearing cache on shutdown 2015-08-09 09:39:14 -04:00
Christopher L. Shannon (cshannon) b52796f794 AMQ-5748 - replace Map with ConcurrentMap for JDK7 2015-08-07 17:54:35 +00:00
Christopher L. Shannon (cshannon) de24980a62 AMQ-5748 - Fixing MessageStore cache
This fixes KahaDBStore to properly check for an existing MessageStore
in the cache before creating a new one.  This will prevent potential
issues with metrics.
2015-08-07 17:23:25 +00:00
Christopher L. Shannon (cshannon) 457dbd8b64 https://issues.apache.org/jira/browse/AMQ-5919
Converting printStackTrace statements to slf4j so that exceptions
are logged properly
2015-08-05 16:29:29 +00:00
gtully a05a9a2129 https://issues.apache.org/jira/browse/AMQ-5552 - added profile for activemq.tests=smoke to run a fast subset of tests that gives a fast verification of a full build 2015-07-09 14:10:38 +01:00
gtully 3985e7225f https://issues.apache.org/jira/browse/AMQ-5853 - fix cacheEnabled case - additional test and fix for both jdbc and kahadb stores 2015-07-09 10:30:54 +01:00
Timothy Bish 13044decce https://issues.apache.org/jira/browse/AMQ-5848
Use the latest openwire version marshallers in the KahaDB store when
starting from a clean install, drop back to the version used in the
existing store if one is found.
2015-07-08 17:29:32 -04:00
Christopher L. Shannon (cshannon) 73d1bcd7ac https://issues.apache.org/jira/browse/AMQ-5875
Resolves an issue when using mKahaDB that caused a MessageStore
being used by more than one destination to be deleted even though
there was still at least 1 other destination using the store.
2015-07-08 18:32:15 +00:00
Christopher L. Shannon (cshannon) 8871b0e496 https://issues.apache.org/jira/browse/AMQ-5748
Fixing an issue that prevented old versions of KahaDB from being
upgraded to the newest version 6
2015-07-08 16:10:11 +00:00
Christopher L. Shannon (cshannon) 785b16bf9e https://issues.apache.org/jira/browse/AMQ-5748
Added a getMessageSize method to MessageStore to support retrieving the
total message size of all stored messages for a destination.  Added a
new storeMessageSize statistic to DestinationStatistics.
2015-07-07 18:03:27 +00:00
Timothy Bish 455f1ca475 https://issues.apache.org/jira/browse/AMQ-5872
The code that adds subscriptions for retroactive subscribers did not
account for the case where the last ack position in the ackLocations
table is used as a placeholder for the next incoming message and there
may not be an element in the messageReference tracker for that index
especially after restart when the index is reloaded.  The code needed to
check this when it iterates of the existing message references to add a
value so to avoid the NPE.  

Also cleaned up the MQTT tests such that they place their data dir in
./target so that old stores get removed on 'mvn clean'
2015-07-02 17:04:35 -04:00
Christopher L. Shannon (cshannon) 12b26b641b https://issues.apache.org/jira/browse/AMQ-5745
This commit is based off a patch submitted by admin@int3solutions.com.
It fixes several locations in the code where there are potential resource leaks.
2015-06-30 13:41:48 -04:00
gtully 2c53dbcc63 https://issues.apache.org/jira/browse/AMQ-5864 - fix and test. A replayed update command did not check if already updated 2015-06-25 15:53:45 +01:00
gtully b6fea83126 https://issues.apache.org/jira/browse/AMQ-5863 - fix and test - sequenceId state in messageid needs to be cleared before doing a store 2015-06-25 12:49:52 +01:00
gtully 0a21c5f8ff better deal with journal corruption for scheduler store to report error location - intermittent failure of KahaDBSchedulerIndexRebuildTest 2015-05-27 12:26:06 +01:00
gtully 9ae5b4147b fix formatting 2015-05-20 12:50:33 +01:00
gtully 3fdf9861b2 https://issues.apache.org/jira/browse/AMQ-5783 - properly remove free topic from the store when there are no remaining subs. Issue: release of the index but remaining metadata refs to the old state. Mbean was also left dangling. test thanks to pat fox 2015-05-20 09:57:05 +01:00
gtully 27e11a388e reduce test duration for ci 2015-05-20 09:57:04 +01:00
Timothy Bish c5a1b86062 https://issues.apache.org/jira/browse/AMQ-5762
Fix for potential memory leak from creation of KahaDBTopicMessageStore.
2015-05-14 16:57:06 -04:00
Timothy Bish 9ef4259297 https://issues.apache.org/jira/browse/AMQ-5616
Use ConcurrentMap in declarations
2015-04-28 11:15:08 -04:00
gtully 4a821186a4 https://issues.apache.org/jira/browse/AMQ-5578 - preallocation could ocurr after a restart over an existing journal file! - fix and test 2015-04-20 16:04:01 +01:00
gtully 73db4d2bfd https://issues.apache.org/jira/browse/AMQ-5703 - further tests and fixes. Ensure early eof can be identified when checking for corruption and skip corruption on replay when the checksum is invalid 2015-04-16 15:25:58 +01:00
gtully a7178a46b7 https://issues.apache.org/jira/browse/AMQ-5703 - fix and test. We now skip past known corruption on a journal replay 2015-04-01 14:59:15 +01:00
Christian Posta 8e551b4c3b removed logging that snuck in there for https://issues.apache.org/jira/browse/AMQ-5578 2015-03-05 15:53:44 -07:00
gtully ecebd2413b https://issues.apache.org/jira/browse/AMQ-5626 - fix and test. on restart, next index needs to ingore priority to find the last used entry 2015-03-02 14:44:58 +00:00
gtully 260e28ecad https://issues.apache.org/jira/browse/AMQ-5620 - avoid potential deadlock on shutdown - waiting on connections to stop before stopping the pa would be an alternative but may block for ever, auto rollback ensures there is no need 2015-03-02 14:44:58 +00:00
Christian Posta 2c92c34132 Fix for https://issues.apache.org/jira/browse/AMQ-5551 The DiskBenchmark util doesn't sync (fsync) to disk for sync writes 2015-02-20 09:23:12 -07:00
gtully 2949e8dbf7 https://issues.apache.org/jira/browse/AMQ-5578 - remove timing deps from the test 2015-02-20 12:03:12 +00:00
Christian Posta 023b2ac045 Updated for https://issues.apache.org/jira/browse/AMQ-5578 adds unit tests, logging, and removes the preallocated batch stuff that snuck in there with commit 45e59e6e83 which was by accident. 2015-02-19 16:34:32 -07:00
Christian Posta 45e59e6e83 adding options for https://issues.apache.org/jira/browse/AMQ-5578 to
allow configuring the allocation strategy at finer grained controls
including zeroing out, OS copying, or sparse file
2015-02-19 10:49:40 -07:00
Timothy Bish 2b17fd8050 https://issues.apache.org/jira/browse/AMQ-5580
Use a CopyOnWriteArrayList to ensure that traversals of the adapters
list do not generate ConcurrentModificationException if a new adapter is
added during that time.
2015-02-11 17:22:26 -05:00
gtully 95f7262cb1 https://issues.apache.org/jira/browse/AMQ-5578 - implement for kahadb 2015-02-11 13:37:32 +00:00
gtully b0a1bd833c https://issues.apache.org/jira/browse/AMQ-5542 fix (via revert below) and test case applied with thanks.
Revert "resolve https://issues.apache.org/activemq/browse/AMQ-2736, logic issue in code that keeps data files with acks around pending message file gc. thanks jgenender - test case to follow"

This reverts commit dd68c61e65.

resolves: https://issues.apache.org/jira/browse/AMQ-5542 and applies test case that nicely demonstrates the defect, thanks
2015-02-02 11:34:06 +00:00
gtully ae595c95b1 [maven-release-plugin] prepare for next development iteration 2015-01-30 12:46:19 +00:00
gtully 4ba1a1689f [maven-release-plugin] prepare release activemq-5.11.0 2015-01-30 12:46:08 +00:00
gtully bf5a2997c7 [maven-release-plugin] prepare for next development iteration 2015-01-26 16:48:46 +00:00
gtully eec2955bde [maven-release-plugin] prepare release activemq-5.11.0 2015-01-26 16:48:37 +00:00
gtully 64fa706d54 Revert "[maven-release-plugin] prepare release activemq-5.11.0"
This reverts commit f47fbe4b89.
2015-01-26 16:39:35 +00:00
gtully f47fbe4b89 [maven-release-plugin] prepare release activemq-5.11.0 2015-01-26 16:17:43 +00:00