[GitHub] flink issue #2822: [FLINK-5075] [kinesis] Make Kinesis consumer fail-proof t...
Github user tzulitai commented on the issue: https://github.com/apache/flink/pull/2822 Merging .. --- 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 issue #2822: [FLINK-5075] [kinesis] Make Kinesis consumer fail-proof t...
Github user StephanEwen commented on the issue: https://github.com/apache/flink/pull/2822 Makes sense. +1 to this! --- 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 issue #2822: [FLINK-5075] [kinesis] Make Kinesis consumer fail-proof t...
Github user tzulitai commented on the issue: https://github.com/apache/flink/pull/2822 Thanks for the review @StephanEwen. I'm pretty sure this doesn't affect the normal Kinesis shard discovery. I'll give it some final tests before merging (would like to get this in before the next 1.1.4 RC). Yes, using `describeStream(streamName)` only will be problematic for users with large numbers of shards, because the whole list may not be able to be returned in a single call. So that's most likely not a solution we can consider. --- 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 issue #2822: [FLINK-5075] [kinesis] Make Kinesis consumer fail-proof t...
Github user tzulitai commented on the issue: https://github.com/apache/flink/pull/2822 I'm not entirely sure whether the fix is a good workaround, or whether we should really bother with Kinesalite's incorrect behaviour at all in our Kinesis connector. I've considered the alternative is to just use the `describeStream(streamName)` variant, always fetching the whole shard list and explicitly ruling out shards we've already seen. That'll make the code clean of such "workarounds", but will be problematic for Kinesis users whose shard count exceeds the largest number of shards the API can return. --- 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. ---
