- 27 4月, 2020 32 次提交
-
-
由 Aljoscha Krettek 提交于
This reverts commit b7ce6119. Our per-push CI does not cover Scala 2.12 so we didn't notice that ScalaShellITCase does not actually pass.
-
由 Terry Wang 提交于
[FLINK-17313] Fix type validation error when sink table ddl contains columns with precision of decimal/varchar (#11848)
-
由 Jeff Zhang 提交于
-
由 Robert Metzger 提交于
-
由 Prem Santosh 提交于
-
由 Andrey Zagrebin 提交于
This closes #11863.
-
由 Andrey Zagrebin 提交于
-
由 Andrey Zagrebin 提交于
Deprecate legacy heap options: jobmanager.heap.size (and update deprecation of jobmanager.heap.mb) to use jobmanager.memory.heap.size for standalone and jobmanager.memory.process.size for containerized (Yarn/K8s) environments. Expose the new JM memory options of FLIP-116 in user docs. This closes #11787.
-
由 Andrey Zagrebin 提交于
The resulting default JVM heap size for standalone deployment is 1024Mb as before FLIP-116. The requested container size (Yarn and Kubernetes) will increase by default from 1024Mb to 1600Mb.
-
由 Andrey Zagrebin 提交于
[hotfix] Fix ProcessMemoryUtilsTestBase#testExceptionShouldContainRequiredConfigOptions to run also for TM
-
由 Andrey Zagrebin 提交于
[hotfix] Remove 'TaskManager' from MemoryBackwardsCompatibilityUtils#getLegacyTaskManagerHeapMemoryIfExplicitlyConfigured
-
由 Andrey Zagrebin 提交于
-
由 Andrey Zagrebin 提交于
This closes #11675.
-
由 Andrey Zagrebin 提交于
-
由 Andrey Zagrebin 提交于
-
由 Xintong Song 提交于
[FLINK-16742][runtime] Refactor BashJavaUtils, move to separated package and extract configuration loading logics.
-
由 Xintong Song 提交于
-
由 Xintong Song 提交于
This closes #11545.
-
由 Xintong Song 提交于
-
由 Xintong Song 提交于
-
由 Xintong Song 提交于
-
由 Till Rohrmann 提交于
We need to configure BIND_HOST to localhost, otherwise Netty will use the wildcard addresss to bind to an IPv6 address whose port is not blocked. This closes #11864.
-
由 Till Rohrmann 提交于
The problem is that we used Mockito to mock a RpcService. Due to this, some of the fields which are later used are being null.
-
由 Till Rohrmann 提交于
-
由 Till Rohrmann 提交于
-
由 Till Rohrmann 提交于
-
由 kevin.cyj 提交于
This commit is the first part of implementation to solve the dead lock problem when reducing the exclusive buffer of receiver side to 0. Reducing the number of exclusive buffers of receiver side to 0 can bring several advantages (may at the cost of some performance regression). One is that memory can be saved from the reduced network buffer usage. Another important benefit is that the in-flight data can be reduced so we can speed up checkpoint in cases of back pressure. However, for the current implementation, reducing the exclusive buffer of receiver side can incur deadlock problem because all the floating buffers might be requested away by some blocked input channels and never recycled until barrier alignment. To solve the problem, this commit mainly makes the following changes: 1. At sender side, after sending a checkpoint barrier when aligned exactly-once checkpoint mode is used, the outgoing channel will be blocked and no data will be sent out until the channel is unblocked. 2. At receiver side, no buffer will be stored in BufferStorage any more and after a checkpoint is completed or canceled, the receiver side will resume data consumption and unblock the upstream by sending a special event to the sender side. Note that after this patch we still can't set the exclusive buffer of receiver side to 0 because there is still deadlock problem which will be totally solved in the following up patches.
-
由 kevin.cyj 提交于
This closes #11351.
-
由 kevin.cyj 提交于
For unaligned checkpoint, no buffer is stored in BufferStorage so CheckpointBarrierUnaligner#processEndOfPartition should always return false. This closes #11351.
-
由 Jiangjie (Becket) Qin 提交于
[FLINK-17268][connector/common] Fix the shutdown sequence in SplitFetcher to avoid blocking on shutdown. Fix flaky SourceReaderTestBase#testAvailableOnEmptyQueue().
-
由 Jiangjie (Becket) Qin 提交于
-
由 Jiangjie (Becket) Qin 提交于
[FLINk-17268][connector/common] Fix the bug of missing records during SplitFetcher wakeup. Added unit test for SplitFetcher.
-
- 26 4月, 2020 8 次提交
-
-
由 Kurt Young 提交于
-
由 Kurt Young 提交于
-
由 Kurt Young 提交于
-
由 Kurt Young 提交于
-
由 Kurt Young 提交于
-
由 Kurt Young 提交于
This closes #11910
-
由 Aljoscha Krettek 提交于
This partially reverts commit 0650e4b6. Only the tests are reverted because the fix is still good. The added tests don't work because of a race condition: if the query finishes before the test cancels them the job status will be FINISHED, and not CANCELLED, as the text expects.
-
由 Jingsong Lee 提交于
This closes #11853
-