Closed imcom closed 8 years ago
Oh, I see... That was introduced in storm 0.8.1 but we never extended support for this option. Thank you for pointing this out. :)
However, you based your change on #134, so, before you can merge, you either need to remove #134's changes from your pull request (try out the pip.conf
trick I mentioned in my last comment on that pull request) or to wait for #134 to be merged into master and then rebase.
hi @poros I tried pip.conf but got no luck... did not really dig into it though. I prefer to wait #134 to be merged first. Thanks for checking out my PR.
seems like the #134 will be revoked, so I will remove the code from #134 here
+1 Thanks for contributing!
Thank you for adding more features! Please fix the documentation issue, so we can go ahead and merge.
hi @poros I've updated the doc and added more features to Pyleus. Enjoy!
Let me know if there are more needs
Thanks
Just noticed that this implements what has been asked in #88. Close #88
@ecanzonieri do you want to weigh in? It's +1 for me
@howardx you are on the right track, under spout options section you have zk_hosts
property, this is where the Kafka lies. Kafka uses ZK for registration so when you set the topic, Storm will use that topic and the ZK servers to find corresponding Kafka brokers. The only problem is with Storm you can not separate Kafka with Storm, they have to share the same ZK cluster .. but with different (better) chroot
sleep spout wait strategy time is quite useful when use kafka spout or similar. When the source queue is empty, we do not want storm cluster run into the crazy looping