Github user hmcl commented on a diff in the pull request:
https://github.com/apache/storm/pull/2637#discussion_r185161496
--- Diff: docs/storm-kafka-client.md ---
@@ -313,4 +313,37 @@ KafkaSpoutConfig<String, String> kafkaConf =
KafkaSpoutConfig
.setTupleTrackingEnforced(true)
```
-Note: This setting has no effect with AT_LEAST_ONCE processing guarantee,
where tuple tracking is required and therefore always enabled.
\ No newline at end of file
+Note: This setting has no effect with AT_LEAST_ONCE processing guarantee,
where tuple tracking is required and therefore always enabled.
+
+# Migrating a `storm-kafka` spout to use `storm-kafka-client`
+
+This may not be an exhaustive list because the `storm-kafka` configs were
taken from Storm 0.9.6
+[SpoutConfig](https://github.com/apache/storm/blob/v0.9.6/external/storm-kafka/src/jvm/storm/kafka/SpoutConfig.java)
and
+[KafkaConfig](https://github.com/apache/storm/blob/v0.9.6/external/storm-kafka/src/jvm/storm/kafka/KafkaConfig.java).
+`storm-kafka-client` spout configurations were taken from Storm 1.0.6
+[KafkaSpoutConfig](https://github.com/apache/storm/blob/v1.0.6/external/storm-kafka-client/src/main/java/org/apache/storm/kafka/spout/KafkaSpoutConfig.java).
+
+| Storm-0.9.6 SpoutConfig | Storm-1.0.6 KafkaSpoutConfig name |
KafkaSpoutConfig usage help |
+| ------------------------- | ---------------------------- |
--------------------------- |
+| **Setting:** `startOffsetTime`<br><br> **Default:**
`EarliestTime`<br>________________________________________________ <br>
**Setting:** `forceFromStart` <br><br> **Default:** `false` <br><br>
`startOffsetTime` & `forceFromStart` together determine the starting offset.
`forceFromStart` determines whether the Zookeeper offset is ignored.
`startOffsetTime` sets the timestamp that determines the beginning offset, in
case there is no offset in Zookeeper, or the Zookeeper offset is ignored |
**Setting:**
[`FirstPollOffsetStrategy`](javadocs/org/apache/storm/kafka/spout/KafkaSpoutConfig.FirstPollOffsetStrategy.html)<br><br>
**Default:** `UNCOMMITTED_EARLIEST` <br><br> [Refer to the helper
table](#helper-table-for-setting-firstpolloffsetstrategy) for picking
`FirstPollOffsetStrategy` based on your `startOffsetTime` & `forceFromStart`
settings | **Import package:**
`org.apache.storm.kafka.spout.KafkaSpoutConfig.FirstPollOffsetStrategy.<strategy-name>`
<br><br> **Usage:** [`<KafkaSpout
Config-Builder>.setFirstPollOffsetStrategy(<strategy-name>)`](javadocs/org/apache/storm/kafka/spout/KafkaSpoutConfig.Builder.html#setFirstPollOffsetStrategy-org.apache.storm.kafka.spout.KafkaSpoutConfig.FirstPollOffsetStrategy-)|
--- End diff --
Are the "import package:" entries throughout necessary ? The ConsumerConfig
strings come all from the same Kafka package, and the KafkaSpoutConfig
configurations already need to have the package imported when the set* method
is declared in the code.
It seems most "Usage:" web links are broken. Wouldn't it be better to
simply paste the method signature and put the link for that same signature?
---
