Can we replicated LRA Coordinator instance and use load balancing?

[barrydevp]

Hi all, Im very interesting to using LRA for Distributed Transaction base on Saga Pattern. How can I guarantee high availability characteristic in my system when using LRA? For instance I'm running LRA coordinator on K8s, can I replicated the LRA deployment? How can I setting persistent log for failure and recovering?

[mmusgrov]

Hi @barrydevp this is a useful request.

The coordinator component is an implementation piece used to implement the requirements of the specification, for example I believe that the Payara implementation does not use coordinators. However, the narayana project does use one and we have an issue for it [1] so if you would like to vote that one up then it will help to bump up its priority.

[1] https://issues.redhat.com/browse/JBTM-2997