search

uber / uReplicator

Improvement of Apache Kafka Mirrormaker
Apache License 2.0
917 stars 199 forks source link

Failover & failback documentation #135

Open chowell5 opened 6 years ago

chowell5 commented 6 years ago

Hello,

I'm currently evaluating uReplicator. There is some documentation about how to set up a test environment for uReplicator (https://github.com/uber/uReplicator), but documentation on running a production environment seems a little lacking.

For instance, I can find no interest on what happens if a datacentre fails, and you need failover/failback. It would be great to document this.

Thank you.

xhl1988 commented 6 years ago

Hi, @chowell5 , thanks for the suggestion to document failover/failback. For your reference, a typical all-active setup we use is like in this video, around 24:00 (https://www.youtube.com/watch?v=K-fI2BeTLkk). Producer produces to regional Kafka cluster and consumer consumes from aggregation Kafka cluster to achieve all-active.

chowell5 commented 6 years ago

Hi, thanks for your reply @xhl1988 - that helps although I still don't feel that I have the understanding of uReplicator that I need for a successful trial. It would be great to see more, example configurations, etc etc.

Many thanks.