[GitHub] flink issue #4607: [FLINK-6306][connectors] Sink for eventually consistent f...

[stevenzwu]

Github user stevenzwu commented on the issue:

https://github.com/apache/flink/pull/4607
  
@aljoscha is there any doc/write-up about the reworking of BucketingSink?


---

[GitHub] flink pull request #4357: (release-1.3) [FLINK-7143, FLINK-7195] Collection ...

[stevenzwu]

Github user stevenzwu commented on a diff in the pull request:

https://github.com/apache/flink/pull/4357#discussion_r128642917
  
--- Diff: 
flink-connectors/flink-connector-kafka-base/src/main/java/org/apache/flink/streaming/connectors/kafka/FlinkKafkaConsumerBase.java
 ---
@@ -517,16 +519,13 @@ public void 
initializeState(FunctionInitializationContext context) throws Except
LOG.debug("Using the following offsets: 
{}", restoredState);
}
}
-   if (restoredState != null && restoredState.isEmpty()) {
-   restoredState = null;
-   }
} else {
LOG.info("No restore state for FlinkKafkaConsumer.");
}
}
 
@Override
-   public void snapshotState(FunctionSnapshotContext context) throws 
Exception {
+   public final void snapshotState(FunctionSnapshotContext context) throws 
Exception {
--- End diff --

```
the verrsion-specific implementations for FlinkKafkaConsumerBase may 
override that and have incorrect implementations, where as our tests would 
never realize it.
```
@tzulitai why would this be a concern for FlinkKafkaConsumerBase. if 
version-specific implementations have bugs, they should have test to catch and 
prevent bugs. We do need the capability to override the snapshot method to 
no-op. what would be your suggested alternative?


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---

[GitHub] flink pull request #4357: (release-1.3) [FLINK-7143, FLINK-7195] Collection ...

[stevenzwu]

Github user stevenzwu commented on a diff in the pull request:

https://github.com/apache/flink/pull/4357#discussion_r128642186
  
--- Diff: 
flink-connectors/flink-connector-kafka-base/src/main/java/org/apache/flink/streaming/connectors/kafka/FlinkKafkaConsumerBase.java
 ---
@@ -517,16 +519,13 @@ public void 
initializeState(FunctionInitializationContext context) throws Except
LOG.debug("Using the following offsets: 
{}", restoredState);
}
}
-   if (restoredState != null && restoredState.isEmpty()) {
-   restoredState = null;
-   }
} else {
LOG.info("No restore state for FlinkKafkaConsumer.");
}
}
 
@Override
-   public void snapshotState(FunctionSnapshotContext context) throws 
Exception {
+   public final void snapshotState(FunctionSnapshotContext context) throws 
Exception {
--- End diff --

@tzulitai looks like the behavior was changed/fix in 1.3. Here is the 
Kafka09Fetcher.java code from 1.2 that was causing the behavior I described 
earlier.

{code}
// if checkpointing is enabled, we are not automatically 
committing to Kafka.
kafkaProperties.setProperty(
ConsumerConfig.ENABLE_AUTO_COMMIT_CONFIG,
Boolean.toString(!enableCheckpointing));
{code}


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---

[GitHub] flink pull request #4357: (release-1.3) [FLINK-7143, FLINK-7195] Collection ...

[stevenzwu]

Github user stevenzwu commented on a diff in the pull request:

https://github.com/apache/flink/pull/4357#discussion_r128113797
  
--- Diff: 
flink-connectors/flink-connector-kafka-base/src/main/java/org/apache/flink/streaming/connectors/kafka/FlinkKafkaConsumerBase.java
 ---
@@ -517,16 +519,13 @@ public void 
initializeState(FunctionInitializationContext context) throws Except
LOG.debug("Using the following offsets: 
{}", restoredState);
}
}
-   if (restoredState != null && restoredState.isEmpty()) {
-   restoredState = null;
-   }
} else {
LOG.info("No restore state for FlinkKafkaConsumer.");
}
}
 
@Override
-   public void snapshotState(FunctionSnapshotContext context) throws 
Exception {
+   public final void snapshotState(FunctionSnapshotContext context) throws 
Exception {
--- End diff --

@tzulitai what's the reason to make this final? In our router use case, we 
override the snapshotState method to no-op. We disabled Flink checkpoint by 
setting checkpoint interval to Long.MAX_VALUE. we can't set Flink checkpoint to 
false, because otherwise Kafka consumer auto.commit will be hard-coded to true. 

@zhenzhongxu ^


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---