Questions about product deployment setup

Community Support Server Deployment
1

Hi, I have a problem related to product deployment.

  1. my temporal clustering is running at container A with ip a.a.a.a
  2. my application (workflow definitions, activities) and workers are running at container B with ip b.b.b.b
  3. my client(WorkflowClient) is running at container C with ip c.c.c.c;

so, can this work? Is setTarget when creating WorkflowServiceStubs in container C enough?

WorkflowServiceStubsOptions workflowServiceStubsOptions =
                WorkflowServiceStubsOptions.newBuilder()
                        .setTarget("a.a.a.a")
                        .build();
WorkflowClientOptions clientOptions = WorkflowClientOptions.newBuilder()
                .setNamespace("default")
                .build();

WorkflowServiceStubs workflowServiceStubs = WorkflowServiceStubs.newServiceStubs(workflowServiceStubsOptions);        
return WorkflowClient.newInstance(workflowServiceStubs, clientOptions);

Or, I have to put client and application at same place?
Thanks.

2

Hi,

They can be in different host, the only requirement is that the Temporal Server should be reachable from clients and workers,

Are you facing any issues?

Antonio

3

Thanks for replying.

I do face some problem when deploying. In my product env, i cannot use k8s or docker compose. All i can do is start each services one by one on different hosts and containers using docker run below

docker run
    -e SERVICES=history \                      -- Spin up one or more: history, matching, worker, frontend
    ....
    -e BIND_ON_IP 0.0.0.0 \
    -e BROAD_CAST_ADDR {container ip} \ 
    ...
    temporalio/server:<tag>

how can i know the container ip before docker run? I cannot use docker network neither btw.
Is there any other way to deploy the cluster? (or by changing some env value or config file?)

Also, if 4 services are in different host or container. which ip is the ip that my external client, workers, … should connect to?

Thanks.