You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@beam.apache.org by amir bahmanyari <am...@yahoo.com> on 2016/08/12 00:16:40 UTC

Appropriate Kafka & Hadoop versions...

Hi Colleagues,I am rebuilding my Beam lab servers with the latest clusters Flink & eventually Spark.Whats the appropriate versions of Kafka & Hadoop to install so there wont be any potential short ckt?
ThanksAmir-

Re: Appropriate Kafka & Hadoop versions...

Posted by Emanuele Cesena <em...@shopkick.com>.
Hi,

The last part (the jira issue) is about Flink.
(Flink itself should have support for all Kafka versions, but I don’t think that’s the recommended way to go with Beam.)

Best,


> On Aug 11, 2016, at 11:19 PM, Amir Bahmanyari <am...@yahoo.com> wrote:
> 
> Thanks Emanuele
> Does this all apply to Flink as well?
> Cheers 
> 
> 
> Sent from my iPhone
> 
>> On Aug 11, 2016, at 10:48 PM, Emanuele Cesena <em...@shopkick.com> wrote:
>> 
>> Hi Amir,
>> 
>> Which version of Spark are you using, 1.6 or 2?
>> 
>> On hadoop I don’t think there’s any issue and I’d go with the latest 2.7 (or 2.6 if you have any special requirements).
>> 
>> To the best of my knowledge:
>> - Spark 2 works with Kafka 0.9+
>> - Spark 1.6 requires Kafka 0.8.x — although some distro can have different versions, for instance we’re using CDH 5.7 that ships Spark 1.6 working with Kafka 0.9.
>> 
>> If you use/plan to use Beam’s KafkaIO then it requires Kafka 0.9+ (I’d personally go with 0.10), but the last time I tried it with Flink it wasn’t fully working because of this:
>> https://issues.apache.org/jira/browse/BEAM-102
>> (Aljoscha told me he was working on it and we can expect a patch soon)
>> 
>> This is all I know.
>> 
>> Best,
>> 
>> 
>>> On Aug 11, 2016, at 5:16 PM, amir bahmanyari <am...@yahoo.com> wrote:
>>> 
>>> Hi Colleagues,
>>> I am rebuilding my Beam lab servers with the latest clusters Flink & eventually Spark.
>>> Whats the appropriate versions of Kafka & Hadoop to install so there wont be any potential short ckt?
>>> 
>>> Thanks
>>> Amir-
>> 
>> -- 
>> Emanuele Cesena, Data Eng.
>> http://www.shopkick.com
>> 
>> Il corpo non ha ideali
>> 
>> 
>> 
>> 
> 

-- 
Emanuele Cesena, Data Eng.
http://www.shopkick.com

Il corpo non ha ideali





Re: Appropriate Kafka & Hadoop versions...

Posted by Amir Bahmanyari <am...@yahoo.com>.
Thanks Emanuele
Does this all apply to Flink as well?
Cheers 


Sent from my iPhone

> On Aug 11, 2016, at 10:48 PM, Emanuele Cesena <em...@shopkick.com> wrote:
> 
> Hi Amir,
> 
> Which version of Spark are you using, 1.6 or 2?
> 
> On hadoop I don’t think there’s any issue and I’d go with the latest 2.7 (or 2.6 if you have any special requirements).
> 
> To the best of my knowledge:
> - Spark 2 works with Kafka 0.9+
> - Spark 1.6 requires Kafka 0.8.x — although some distro can have different versions, for instance we’re using CDH 5.7 that ships Spark 1.6 working with Kafka 0.9.
> 
> If you use/plan to use Beam’s KafkaIO then it requires Kafka 0.9+ (I’d personally go with 0.10), but the last time I tried it with Flink it wasn’t fully working because of this:
> https://issues.apache.org/jira/browse/BEAM-102
> (Aljoscha told me he was working on it and we can expect a patch soon)
> 
> This is all I know.
> 
> Best,
> 
> 
>> On Aug 11, 2016, at 5:16 PM, amir bahmanyari <am...@yahoo.com> wrote:
>> 
>> Hi Colleagues,
>> I am rebuilding my Beam lab servers with the latest clusters Flink & eventually Spark.
>> Whats the appropriate versions of Kafka & Hadoop to install so there wont be any potential short ckt?
>> 
>> Thanks
>> Amir-
> 
> -- 
> Emanuele Cesena, Data Eng.
> http://www.shopkick.com
> 
> Il corpo non ha ideali
> 
> 
> 
> 


Re: Appropriate Kafka & Hadoop versions...

Posted by Emanuele Cesena <em...@shopkick.com>.
Hi Amir,

Which version of Spark are you using, 1.6 or 2?

On hadoop I don’t think there’s any issue and I’d go with the latest 2.7 (or 2.6 if you have any special requirements).

To the best of my knowledge:
- Spark 2 works with Kafka 0.9+
- Spark 1.6 requires Kafka 0.8.x — although some distro can have different versions, for instance we’re using CDH 5.7 that ships Spark 1.6 working with Kafka 0.9.

If you use/plan to use Beam’s KafkaIO then it requires Kafka 0.9+ (I’d personally go with 0.10), but the last time I tried it with Flink it wasn’t fully working because of this:
https://issues.apache.org/jira/browse/BEAM-102
(Aljoscha told me he was working on it and we can expect a patch soon)

This is all I know.

Best,


> On Aug 11, 2016, at 5:16 PM, amir bahmanyari <am...@yahoo.com> wrote:
> 
> Hi Colleagues,
> I am rebuilding my Beam lab servers with the latest clusters Flink & eventually Spark.
> Whats the appropriate versions of Kafka & Hadoop to install so there wont be any potential short ckt?
> 
> Thanks
> Amir-

-- 
Emanuele Cesena, Data Eng.
http://www.shopkick.com

Il corpo non ha ideali