linkedin/brooklin

Give me a specific example for system.destination.identityPartitioningEnabled ,thanks

Bruce2jiang opened this issue · 4 comments

Give me a specific example for system.destination.identityPartitioningEnabled ,thanks

I set it to go like this : destination.identityPartitioningEnabled =true ,but have no effect

Don't know how to add it, in this place :
Datastream.pdsc
{
"name": "metadata",
"doc": "Generic metadata for Datastream (e.g. owner, expiration, etc). Metadata is stored as user defined name/value pair.",
"type": {
"type": "map",
"values": "string"
},
"optional": true
}

You can set this option in datastream metadata, e.g.

bin/brooklin-rest-client.sh -o CREATE -m '{"system.destination.identityPartitioningEnabled":"true","owner":"test-user","system.reuseExistingDestination":"false"}' -u http://localhost:32311/ -n first-mirroring-stream -s "kafka://localhost:9093/^(first|second)-topic$" -c kafkaMirroringConnector -t kafkaTransportProvider 2>/dev/null

My test steps are as follows: first, create a topic with three partitions in the source kafka and execute it as you do:"./brooklin-rest-client.sh -o CREATE -u http://localhost:32311/ -n first-mirroring-stream-nine -s "kafka://10.124.201.51:9092,10.124.201.52:9092,10.124.201.53:9092/bigdata_test_topic_nine" -c kafkaMirroringConnector -t kafkaTransportProvider -m '{"system.destination.identityPartitioningEnabled":"true","owner":"test-user","system.reuseExistingDestination":"false"}' 2>/dev/null
",

but the topic of the target kafka has only one partition.
Should three partitions be created in advance in the destination kafka?

Yes, topic needs to be created with correct partition count in advance.