[Bug] Samples seemingly not executing workflow

[backhand]

What are you really trying to do?

I'm trying to run the samples here, and getting a feel for how the whole architecture of temporal works. Part of that is getting a basic workflow up and running, and being able to observe how it's scheduled, executed etc.

Describe the bug

On running the samples I tried (hello-world, hello-world-js, activities-examples) I spun up a dev cluster via docker-compose and followed the instructions for getting it started. The cluster spun up fine and seems active, UI is working, worker starts and connects, but when I run the actual workflow it seems to just silently exit out of the client.workflow.start call. All logs or attempts to write to a file works up to that point, but after that nothing happens. Not in the client.ts functions nor those in activities.ts. I'm new to this so perhaps there's a step I'm missing or I misunderstood something.

Minimal Reproduction

• Clone temporal docker repo (https://github.com/temporalio/docker-compose)
• Spin up cluster: docker-compose up
• Clone samples repo (https://github.com/temporalio/samples-typescript)
• cd samples-typescript/hello-world-js
• npm install
• npm run start.watch
• (In another shell) npm run workflow

Output from worker:

% npm run start                                                                                                                             samples-typescript (main)

> [email protected] start
> node src/worker.js

2023-01-26T09:50:35.033Z [INFO] Creating worker {
  options: {
    namespace: 'default',
    identity: '86321@xyz-MBP',
    shutdownGraceTime: '10s',
    maxConcurrentActivityTaskExecutions: 100,
    maxConcurrentLocalActivityExecutions: 100,
    enableNonLocalActivities: true,
    maxConcurrentWorkflowTaskExecutions: 100,
    stickyQueueScheduleToStartTimeout: '10s',
    maxHeartbeatThrottleInterval: '60s',
    defaultHeartbeatThrottleInterval: '30s',
    isolateExecutionTimeout: '5s',
    workflowThreadPoolSize: 8,
    maxCachedWorkflows: 761,
    enableSDKTracing: false,
    showStackTraceSources: false,
    debugMode: false,
    interceptors: { activityInbound: [Array], workflowModules: [Array] },
    sinks: { defaultWorkerLogger: [Object] },
    workflowsPath: '/Users/xyz/Documents/workspace/samples-typescript/hello-world-js/src/workflows.js',
    activities: [Module: null prototype] { greet: [AsyncFunction: greet] },
    taskQueue: 'hello-javascript',
    shutdownGraceTimeMs: 10000,
    stickyQueueScheduleToStartTimeoutMs: 10000,
    isolateExecutionTimeoutMs: 5000,
    maxHeartbeatThrottleIntervalMs: 60000,
    defaultHeartbeatThrottleIntervalMs: 30000,
    loadedDataConverter: {
      payloadConverter: [DefaultPayloadConverter],
      failureConverter: [DefaultFailureConverter],
      payloadCodecs: []
    }
  }
}
2023-01-26T09:50:36.338Z [INFO] asset workflow-bundle-55f159d4e77301e74160.js 700 KiB [emitted] [immutable] (name: main)
2023-01-26T09:50:36.338Z [INFO] runtime modules 670 bytes 3 modules
2023-01-26T09:50:36.338Z [INFO] modules by path ./node_modules/@temporalio/ 164 KiB
2023-01-26T09:50:36.338Z [INFO]   modules by path ./node_modules/@temporalio/common/lib/ 69 KiB 18 modules
2023-01-26T09:50:36.338Z [INFO]   modules by path ./node_modules/@temporalio/workflow/ 93.1 KiB
2023-01-26T09:50:36.338Z [INFO]     ./node_modules/@temporalio/workflow/lib/worker-interface.js 9.81 KiB [built] [code generated]
2023-01-26T09:50:36.339Z [INFO]     + 12 modules
2023-01-26T09:50:36.339Z [INFO]   ./node_modules/@temporalio/worker/lib/workflow-log-interceptor.js 2.27 KiB [built] [code generated]
2023-01-26T09:50:36.339Z [INFO] modules by path ./src/ 1000 bytes
2023-01-26T09:50:36.339Z [INFO]   ./src/workflows-autogenerated-entrypoint.cjs 591 bytes [built] [code generated]
2023-01-26T09:50:36.339Z [INFO]   ./src/workflows.js 413 bytes [built] [code generated]
2023-01-26T09:50:36.339Z [INFO] __temporal_custom_payload_converter (ignored) 15 bytes [built] [code generated]
2023-01-26T09:50:36.339Z [INFO] __temporal_custom_failure_converter (ignored) 15 bytes [built] [code generated]
2023-01-26T09:50:36.339Z [INFO] ./node_modules/long/umd/index.js 43.1 KiB [built] [code generated]
2023-01-26T09:50:36.339Z [INFO] ./node_modules/ms/index.js 2.95 KiB [built] [code generated]
2023-01-26T09:50:36.339Z [INFO] webpack 5.75.0 compiled successfully in 590 ms
2023-01-26T09:50:36.342Z [INFO] Workflow bundle created { size: '0.68MB' }
2023-01-26T09:50:36.768622Z  INFO temporal_sdk_core::worker: Initializing worker task_queue=hello-javascript namespace=default
2023-01-26T09:50:36.769Z [INFO] Worker state changed { state: 'RUNNING' }

Output from workflow:

 % npm run workflow                                                                                                                          samples-typescript (main)

> [email protected] workflow
> node src/client.js

%

Expected:

Hello, Temporal!

Environment/Versions

MacOS Ventura 13.1, Intel core i7
Node.js v19.4.0
Temporal version: 1.19.1 (also tried 1.18.5)
Run via docker-compose from the aforementioned docker-compose repo

Additional context

[bergundy]

It's likely the grpc-js version issue reported here: temporalio/sdk-typescript#1033

[backhand]

It's likely the grpc-js version issue reported here: temporalio/sdk-typescript#1033

That worked! Thanks for the quick response.

[diervo]

I'm running into a similar issue (even after upgrading to latest and greatest grpc-js)
Server never returns with any response and I get a timeout...
(I'm running the helm/k8s version in case has something to do with that)

ServiceError: Failed to start Workflow
[...]
 code: 14,
    details: 'last connection error: failed to receive server preface within timeout',
    metadata: Metadata { internalRepr: [Map], options: {} },
    progress: 'PROCESSED'

Is anyone experiencing the sam issue?

For the record, I'm evaluating the technology, so the fact that I have problems setting up the hello world, doesn't give me much confidence at the moment... (at least on the hosted version)

[bergundy]

Can you connect to a local server using temporalite or our docker compose setup?
If that doesn’t work, can you try with a fresh installation and SDK version 1.6.0?

[diervo]

I'm using SDK 1.6.0 and I tried to override the grpc-js with "@grpc/grpc-js": "1.8.7" just in case, but no luck.
The weird thing is that I do see Handling connection for 7233 on the frontend-headless

The docker-compose version works just fine, so it has to be an internal misconfiguration of one of the services deployed within the cluster...

I filed a bug in the helm repo, since the sdk seems to be just fine invoked when standing up the services via docker.

[diervo]

@bergundy I tried temporalite just to triple check:

If I run temporalite without the frontend ./temporalite start --headless --namespace default
Everything works just fine.

If I run temporalite with the UI ./temporalite start --namespace default, I get the following errors and the whole process enters and endless loop when the worker or the client are unable to run.

{"level":"info","ts":"2023-02-01T21:02:10.665-0800","msg":"matching client encountered error","service":"history","error":"Not enough hosts to serve the request","service-error-type":"serviceerror.Unavailable","logging-call-at":"metric_client.go:220"}

Seems like worker and client logic seem to be fine, but the services are not well configured neither for kubernetes or temporalite?
🤷‍♂️

[bergundy]

Apart from that error log with temporalite, is the SDK sample working for you?

For the k8s deployment, I suggest you get support on our community slack: https://t.mp/slack.

I’m closing this issue since it is no longer relevant since the last sdk release.