Commit Graph

159 Commits

Author SHA1 Message Date
Justin Bertram b4f8aada3c ARTEMIS-3092 inconsistent JDBC config logging 2021-02-04 15:32:51 -05:00
Clebert Suconic c47122ea15 ARTEMIS-3084 Avoid pending IO to close files 2021-02-03 16:56:07 -05:00
Clebert Suconic 755947ee0b ARTEMIS-3084 Deal with async close and double close
Since the libaio.close is now async
there might be a situation with more than one close called during a server.stop();

This should deal with that scenario
2021-02-01 15:29:39 -05:00
Clebert Suconic 03b0fcd2c2 ARTEMIS-3084 Issue a warning instead of blocking forever in case of not completing closes 2021-01-29 13:24:57 -05:00
Clebert Suconic b05bea7db9 ARTEMIS-3084 Issue a warning instead of blocking forever in case of not completing closes 2021-01-29 12:44:34 -05:00
Clebert Suconic 7ce5315d7a ARTEMIS-3084 Fix on checkstyle
sorry for breaking it in the first place.
I messed up into some gitfu sending an older version by accident.
2021-01-29 10:50:47 -05:00
Clebert Suconic 873c2bcc18 ARTEMIS-3084 Avoiding Race condition on async close 2021-01-29 10:44:55 -05:00
Clebert Suconic 4b5c29939c ARTEMIS-3084 Fixing intermittent failures on testsuite. file.close() should block
This fixed a failure I saw once at GroupingFailoverReplicationTest.testGroupingLocalHandlerFailsMultipleGroups
2021-01-28 14:24:12 -05:00
Clebert Suconic 03f8f83935 ARTEMIS-3084 Small tweak. Log a Thread dump if a server.stop cannot get a completion of pending closes 2021-01-28 11:31:58 -05:00
Clebert Suconic c019218c4e ARTEMIS-3084 Eliminate Block on moving to next file on libaio 2021-01-28 11:10:40 -05:00
Tomas Hofman b1835fa2d8 ARTEMIS-3037 JournalImpl#checkKnownRecordID() implementation can leave a thread hanging in WAITING state 2020-12-18 12:44:18 -05:00
Emmanuel Hugonnet fdfc58171b [ARTEMIS-2939]: Artemis should not delete corrupt log files.
* Moving corrupted journal files to the attic folder.

Jira: https://issues.apache.org/jira/browse/ARTEMIS-2939
2020-10-20 18:39:06 -04:00
franz1981 a680f7d52e ARTEMIS-2928 blocking CallbackCache can be replaced with a JCTools lock-free queue 2020-10-06 07:14:24 +02:00
franz1981 a6bf7d0e04 ARTEMIS-2837 Bursts of open files under high load
This would prevent to push a new opened file if there
is already one available to be consumed
2020-08-17 13:01:41 -04:00
franz1981 851aef1172 Revert "ARTEMIS-2837 Avoiding bursts on writes and pending callbacks"
This reverts commit 1761f763
2020-08-17 13:01:41 -04:00
franz1981 7902dd1757 ARTEMIS-2849 Optimize zeroing of buffers while writing the ASYNCIO journal 2020-07-23 15:44:52 -04:00
Clebert Suconic 1761f76308 ARTEMIS-2837 Avoiding bursts on writes and pending callbacks 2020-07-08 23:08:27 -04:00
Robbie Gemmell d2c46762e8 ARTEMIS-2109: fix javadoc errors, remove doclint config for JDK8, remove superfluous plugin versions, update to 3.0.1 javadoc plugin
Using 3.0.1 like the current apache parent pom [not being used], and also as newer ones break build.
2020-06-11 18:50:01 +01:00
Francesco Nigro 4e1af3aea3 ARTEMIS-2762 JournalRecord is using too much memory to track JournalFile updates 2020-05-15 09:21:33 -04:00
Clebert Suconic 3541ebf218 ARTEMIS-2732 Cleanup on testsuite logging 2020-04-22 23:08:46 -04:00
Clebert Suconic af796d5ce4 ARTEMIS-2701 Improving DLQ/check over previously removed records 2020-04-06 12:25:35 -04:00
Clebert Suconic 03fb630f73 ARTEMIS-1975 Fixing LargeMessage encoding for replication 2020-03-25 16:16:05 -04:00
Clebert Suconic 31c945f8b0 ARTEMIS-1975 Fixing encodesize cached on AMQP Large Message
Encoding ahead is broken with AMQP Large Message
This is still part of ARTEMIS-1975
2020-03-24 15:51:46 -04:00
Clebert Suconic ddd8ed4402 ARTEMIS-1975 Real Large Message support into AMQP
This is a Large commit where I am refactoring largeMessage Body out of CoreMessage
which is now reused with AMQP.

I had also to fix Reference Counting to fix how Large Messages are Acked

And I also had to make sure Large Messages are transversing correctly when in cluster.
2020-02-25 15:25:01 +01:00
Francesco Nigro 5897909dc9 ARTEMIS-2617 use core pools to reduce GC on journal loading 2020-02-12 13:29:51 -05:00
Francesco Nigro fa0c187ae6 ARTEMIS-2602 Reduce number of copies for non JDBC Journal 2020-01-18 18:45:12 +01:00
Francesco Nigro b10d765139 ARTEMIS-2602 load surviving records into SparseArrayLinkedList 2020-01-17 14:31:52 -05:00
Francesco Nigro 4cc6464ddd ARTEMIS-2602 Improve Journal loading heap usage 2020-01-17 14:31:52 -05:00
Francesco Nigro f51c799ac0 ARTEMIS-1811 NIO Seq File should use RandomAccessFile with heap buffers
It use RandomAccessFile to allow using heap buffers without additional
copies and/or leaks of direct buffers, as performed by FileChannel JDK
implementation (see https://bugs.openjdk.java.net/browse/JDK-8147468)
2019-09-23 10:55:20 -04:00
Clebert Suconic aa5d76e1bb ARTEMIS-2414 AIOSequentialFile was ignoring sync and leaking files 2019-07-30 11:03:06 -04:00
Wei Yang edace8845e ARTEMIS-2414 Sync before closing file in case data loss 2019-07-29 14:44:26 -04:00
Clebert Suconic ebbc500b82 NO-JIRA Duplicate Call on getEncodeSize 2019-07-16 15:15:43 -04:00
yang wei dd62be1602 ARTEMIS-2414 Sync before closing file in case data loss 2019-07-15 23:53:44 +01:00
Francesco Nigro 4445261f34 ARTEMIS-2382 Reclaimer doesn't need to be instantiatable 2019-06-17 22:31:30 -04:00
brusdev 9b52547ff7 ARTEMIS-2371 Message with huge header shuts broker down
Add max record size check before adding a record to prevent that the
broker shuts down, when there is one really large header sent with the
message. Add message size check before allocating large message resource
if it can't be stored.
2019-06-10 17:36:18 -04:00
Jiri Danek f7a36300ef ARTEMIS-2320 Fix IntLongMath errorprone warning
Expression of type int may overflow before being assigned to a long
    (see https://errorprone.info/bugpattern/IntLongMath)
2019-04-29 11:27:29 +01:00
Clebert Suconic 828a4856da ARTEMIS-2274 Fix on Journal buffer overflow with almost large messages 2019-03-13 14:56:46 -04:00
Francesco Nigro 89c02f1cc0 ARTEMIS-2211 Refactor ByteBuffer pooling, alignment and zeroing
Refactored thread local ByteBuffer pooling, alignment
and zeroing in order to avoid duplicate code and
improve code coverage with tests.
In addition are being provided faster branchless
alignment operations and optional zeroing of
pooled ByteBuffers for both ASYNCIO and
NIO/MAPPED journal types.
2019-03-12 18:32:53 -04:00
Clebert Suconic 061fb2787e ARTEMIS-1977 Stripping activemq-artemis as a separated proejct 2019-03-07 11:34:15 -05:00
Francesco Nigro 4da9d84311 ARTEMIS-2239 Zero-copy NIO/MAPPED TimedBuffer
NIO/MAPPED journal types can use directly the buffer of TimedBuffer
to perform file writes, avoiding an expensive copy + zeroing.
2019-02-14 00:20:51 +00:00
Clebert Suconic 05d893036e ARTEMIS-2200 NPE fix
This was a typo on my last commit around ARTEMIS-2200
2019-01-17 21:51:30 -05:00
Justin Bertram 57aacf784c NO-JIRA fix lgtm.com warnings
Warnings enumerated at
https://lgtm.com/projects/g/apache/activemq-artemis/alerts/?mode=tree&severity=warning
2019-01-17 22:30:28 +00:00
Clebert Suconic b3f0a87e0d ARTEMIS-2200 NPE when calling journal.delete from Multiple Threads 2019-01-17 15:38:51 -05:00
Francesco Nigro 702f445205 ARTEMIS-2200 NPE while dropping/failing large messages on paging
Large messages pendingRecordID is not accessed atomically, leading
to races that would lead to records that cannot been found on the
journal for deletion: it would lead to cause NPE that won't clean
the pending tasks on the current OperationContextImpl.
Adding a cleanup on error of those tasks and avoiding the race
to happen by adding proper synchronization will both enforce
correct clean up when something bad happen and avoid NPE.
2019-01-17 14:35:25 -05:00
Francesco Nigro 1c0ef5d7f7 ARTEMIS-2196 Avoid creating RandomAccessFile when FileChannel is needed 2018-12-14 15:10:30 -06:00
feuillemorte a3001fd9bd ARTEMIS-2201 Added tests on DEFAULT_JOURNAL_FILE_OPEN_TIMEOUT value 2018-12-14 11:32:13 +01:00
Francesco Nigro 2967df6a99 ARTEMIS-1996 MappedSequentialFileFactory may cause DirectByteBuffer off-heap memory leaks
Compaction is now reusing direct ByteBuffers on both
reading and writing with explicit and deterministic
release to avoid high peak of native memory utilisation
after compaction.
2018-10-22 09:16:15 +02:00
Francesco Nigro cae253d136 ARTEMIS-2131 Error compacting journal
Compaction cannot free a sliced view of a ByteBuffer on Java >=9:
the fix is using the original ByteBuffer instead of the slice
to perform a file write and allow it to be correctly released by
the Cleaner.
2018-10-17 20:49:50 -04:00
Clebert Suconic 6b1abd1aad ARTEMIS-2053 avoiding data loss after compacting 2018-08-26 18:17:29 -04:00
Clebert Suconic 63e6cd98f8 ARTEMIS-2046 Fixing issues with JournalStorageManager.stop in replication, JDBC and shared storage 2018-08-20 21:25:57 -04:00