- 29 11月, 2017 14 次提交
-
-
由 gyao 提交于
-
由 gyao 提交于
-
由 gyao 提交于
Migrate logic from org.apache.flink.runtime.rest.handler.legacy.metrics.SubtaskMetricsHandler to new handler. Add new handler to DispatcherRestEndpoint. [FLINK-8143][flip6] Assert that SubtaskIndexPathParameter is mandatory [FLINK-8143][flip6] Use path parameter constants in SubtaskMetricsHandlerTest This closes #5082.
-
由 gyao 提交于
-
由 gyao 提交于
Migrate logic in org.apache.flink.runtime.rest.handler.legacy.metrics.JobManagerMetricsHandler to new handler and add new handler to DispatcherRestEndpoint. [FLINK-7716][Javadoc] Deprecate method MetricStore#getJobManager(). There is a semantically equivalent method in MetricStore. This closes #5083.
-
由 gyao 提交于
Migrate logic in org.apache.flink.runtime.rest.handler.legacy.metrics.TaskManagerMetricsHandler to new handler, and add new handler to DispatcherRestEndpoint. [FLINK-7717][flip6] Use taskmanagerid constant in TaskManagerMetricsHandlerTest This closes #5081.
-
由 gyao 提交于
-
由 Till Rohrmann 提交于
Before, the TaskExecutor's InstanceID was exposed as TaskExecutor id. This was wrong since the InstanceID is bound the registration of a TaskExecutor whereas the ResourceID is bound to the lifetime of the TaskExecutor. Thus, it is better to identify the TaskExecutor by its ResourceID which does not change. This commit changes the behaviour accordingly on the ResourceManager and the TaskManagerDetailsHandler. This closes #5093.
-
由 yew1eb 提交于
This closes #5001.
-
由 Aljoscha Krettek 提交于
-
由 Aljoscha Krettek 提交于
-
由 Till Rohrmann 提交于
-
由 zjureel 提交于
This closes #4574
-
由 Greg Hogan 提交于
When printing the job plan the operator description is typically formatted as the operator name followed by the user given or generated function name in parenthesis. For example, "Reduce (My Function)". This normalizes the node names to include a space between the operator and function names. This closes #4383
-
- 28 11月, 2017 6 次提交
-
-
由 Till Rohrmann 提交于
Add "Remote connection to [null] failed with java.nio.channels.NotYetConnectedException" to the list of whitelisted log statements in YarnTestBase. This logging statement seems to appear since we moved from Flakka to Akka 2.4.0. This closes #5085.
-
由 Piotr Nowojski 提交于
This closes #4748
-
由 zhangminglei 提交于
This closes #4946
-
由 yew1eb 提交于
This closes #5034
-
由 yew1eb 提交于
This closes #5067
-
由 Greg Hogan 提交于
This closes #5061
-
- 27 11月, 2017 5 次提交
-
-
由 Aljoscha Krettek 提交于
-
由 Aljoscha Krettek 提交于
-
由 Raycee 提交于
This closes #5077.
-
由 zentol 提交于
This closes #5049.
-
由 Piotr Nowojski 提交于
Increasing timeout for reading the records from 30s to 60s seems to solve the issue for failing at-least-one tests.
-
- 25 11月, 2017 1 次提交
-
-
由 gyao 提交于
Migrate logic in org.apache.flink.runtime.rest.handler.legacy.metrics.JobVertexMetricsHandler to new handler and add new handler to DispatcherRestEndpoint. Add common classes for remaining implementations of org.apache.flink.runtime.rest.handler.legacy.metrics.AbstractMetricsHandler, which require migration as well. [FLINK-7718] [flip6] Clean up JobVertexMetricsHandlerHeaders [FLINK-7718] [flip6] Assert that HTTP code is 404 if metric is unknown [FLINK-7718] [flip6] Minor fixes in Javadocs [FLINK-7718] [flip6] Add unit test for AbstractMetricsHandlerHeaders [FLINK-7718] [flip6] Let unit tests inherit from TestLogger [FLINK-7718] [flip6] Re-format Metric constructor [FLINK-7718] [flip6] Fix mistake in Javadoc of AbstractMetricsHandlerHeaders [FLINK-7718] [flip6] Rename AbstractMetricsHandlerHeaders to AbstractMetricsHeaders Strip the term Handler from the Header class. Also rename its subclasses. [FLINK-7718] [flip6] No longer return HTTP 404 if metric is unknown [FLINK-7718] [flip6] Make JobVertexMetricsHeaders class final [FLINK-7718] [flip6] Introduce MetricsHandlerTestBase for future MetricHandlers [FLINK-7718] [flip6] Always return same MessageParameter objects in JobVertexMetricsMessageParameters This closes #5055.
-
- 24 11月, 2017 5 次提交
-
-
由 Aljoscha Krettek 提交于
-
由 Stephan Ewen 提交于
-
由 Stephan Ewen 提交于
-
由 Nico Kruber 提交于
This is another step at using or own (off-heap) buffers for network communication that we pass through netty in order to avoid unnecessary buffer copies. This closes #4481.
-
由 Aljoscha Krettek 提交于
This was causing the snapshot deployment to fail.
-
- 23 11月, 2017 9 次提交
-
-
由 Stefan Richter 提交于
[FLINK-5465] [streaming] Wait for pending timer threads to finish or to exceed a time limit in exceptional stream task shutdown. This closes #5058.
-
由 Piotr Nowojski 提交于
-
由 twalthr 提交于
-
由 Xingcan Cui 提交于
This closes #5056.
-
由 Piotr Nowojski 提交于
-
由 Piotr Nowojski 提交于
Previously faulty scenario with producer pool of 2. 1. started transaction 1 with producerA, written record 42 2. checkpoint 1 triggered, pre committing txn1, started txn2 with producerB, written record 43 3. checkpoint 1 completed, committing txn1, returning producerA to the pool 4. checkpoint 2 triggered , committing txn2, started txn3 with producerA, written record 44 5. crash.... 6. recover to checkpoint 1, txn1 from producerA found to "pendingCommitTransactions", attempting to recoverAndCommit(txn1) 7. unfortunately txn1 and txn3 from the same producers are identical from KafkaBroker perspective and thus txn3 is being committed result is that both records 42 and 44 are committed. With this fix, after re-initialization txn3 will have different producerId/epoch counters compared to txn1.
-
由 Piotr Nowojski 提交于
-
由 Bowen Li 提交于
-
由 Fabian Hueske 提交于
This closes #5043.
-