Community Articles
Find and share helpful community-sourced technical articles
Announcements
Alert: Welcome to the Unified Cloudera Community. Former HCC members be sure to read and learn how to activate your account here.
Labels (1)
Community Manager

Question

The documentation for Kafka states that Kafka is being distributed as a parcel.

 

Documentation:

http://www.cloudera.com/content/cloudera/en/documentation/cloudera-kafka/latest/topics/kafka_install...

 

I already have a cluster based on packages.

 

Is it possible to install Kafka without Cloudera Manager? By using packages?

 

Here are a few concerns around a shift to a parcel based distribution.

 

1. If a local repository of parcels is used for a cluster connected to the internet,
will couldera manager still be able to detect if new parcels are available? Is there an option to not upgrade?

 

2. Can a cluster using be built from older versions of parcels from a local repository that have been downloaded previously?

Answer

 

We haven't released the package distribution of Kafka yet, it is only parcels for now.

 

1. If a local repository of parcels is used for a cluster connected to the internet,
will couldera manager still be able to detect if new parcels are available? Is there an option to not upgrade?

 

Yes the parcels would just appear as available parcels, you pick what to download

 

The overview is here: http://www.cloudera.com/content/cloudera/en/documentation/core/latest/topics/cm_ig_parcels.html

 

This goes into how to create a local parcel repository 

 

http://www.cloudera.com/content/cloudera/en/documentation/core/latest/topics/cm_ig_create_local_parc...

 

 

2. Can a cluster using be built from older versions of parcels from a local repository that have been downloaded previously?

 

You could rebuild a cluster from scratch and then import data in yes, you could not reliably back-rev a deployment into a previous parcel version and expect things to work.  But that is an issue with package based installations as well and not unique to parcels.

 

1,687 Views
0 Kudos
Comments
New Contributor
Hi,
 I am setting up kafka in Bluedata environment
I have created a topic “test” with 4 partitions with replication factor 1
 

Topic: test Partition: 0 Leader: 1 Replicas: 1 Isr: 1

Topic: test Partition: 1 Leader: 0 Replicas: 0 Isr: 0

Topic: test Partition: 2 Leader: 1 Replicas: 1 Isr: 1

Topic: test Partition: 3 Leader: 0 Replicas: 0 Isr: 0

 

 

When I run Console-Producer

(bin/kafka-console-producer.sh --broker-list localhost:9092 --topic test) 

and enter some message, Its giving error like

 

 

WARN Error while fetching metadata with correlation id 0 : {test=LEADER_NOT_AVAILABLE} 

 

WARN Error while fetching metadata with correlation id 1 : {test=LEADER_NOT_AVAILABLE} 

 

WARN Error while fetching metadata with correlation id 2 : {test=LEADER_NOT_AVAILABLE} 

So on…

 

Can any one help me in resolving this issue