Re: Thoughts and obesrvations on Samza

2015-07-06 Thread Martin Kleppmann
Hi all, Lots of good thoughts here. I agree with the general philosophy of tying Samza more firmly to Kafka. After I spent a while looking at integrating other message brokers (e.g. Kinesis) with SystemConsumer, I came to the conclusion that SystemConsumer tacitly assumes a model so much like

Re: Thoughts and obesrvations on Samza

2015-07-06 Thread Jay Kreps
Hey Martin, For the YARN/Mesos/etc decoupling I actually don't think it ties our hands at all, all it does is refactor things. The division of responsibility is that Samza core is responsible for task lifecycle, state, and partition management (using the Kafka co-ordinator) but it is NOT

Review Request 36224: SAMZA-728: Samza job fails due to null pointer in JobCoordinator refreshJobModel

2015-07-06 Thread Navina Ramesh
--- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/36224/ --- Review request for samza, Yan Fang, Naveen Somasundaram, and Yi Pan (Data

Re: Thoughts and obesrvations on Samza

2015-07-06 Thread Yi Pan
@Jay, you got my point. {quote} I think the question is whether for the as a service you are proposing actually trying to build some layer over YARN/Mesos/AWS that abstracts these away? {quote} I am not very strong on this, but I do see that as an option. The reason I had is: if all we need from

RE: Thoughts and obesrvations on Samza

2015-07-06 Thread Ken Krugler
Hi Martin, As a lurker here, this has been a very interesting thread. I would suggest talking to one of the Solr committers about their experience in merging with Lucene, as that's got many similarities to what you're discussing here…though Solr was more mature when that happened. From what