[WIP] Spike: todo example on IBM Cloud [DO NOT MERGE] #1574

hacksparrow · 2018-07-30T14:25:26Z

Add todo example on IBM Cloud - see #1446

Checklist

npm test passes on your machine
New tests added or existing tests modified to cover all changes
Code conforms with the style guide
API Documentation in code was updated
Documentation in /docs/site was updated
Affected artifact templates in packages/cli were updated
Affected example projects in examples/* were updated

bajtos · 2018-07-31T15:10:19Z

Nice start! Could you please start writing a guide explaining how to deploy your modified todo example to bluemix? For example, you can put the content into a new file examples/todo/DEPLOYING.md. That way we can review the changes in the deployment guide alongside the changes in code.

virkt25 · 2018-07-31T15:50:05Z

examples/todo/src/index.ts

@@ -8,6 +8,13 @@ import {ApplicationConfig} from '@loopback/core';
 import {RestServer} from '@loopback/rest';

 export async function main(options?: ApplicationConfig) {
+  // IBM Cloud app port is set on process.env.PORT
+  if (!options) {
+    options = {rest: {port: process.env.PORT}};


This isn't needed since the port default to process.env.PORT in the else statement below. So this can be just:

options.rest = options.rest ? options.rest : {}; options.rest.port = process.env.PORT || options.rest.port;

Where is options defined?

Ooops. For some reason I through the check was for options.rest.

bajtos

It looks like deploying LB4 apps to IBM Cloud requires only small changes - I like that.

IIUC, this deployment is based on CloudFoundry. I wish we were targeting Kubernetes based deployments, as that seems to be the new way forward, but CloudFoundry is good enough for the initial spike.

bajtos · 2018-08-13T16:57:58Z

examples/todo/src/datasources/db.datasource.json

-  "localStorage": "",
-  "file": "./data/db.json"
+  "connector": "cloudant",
+  "url": "http://admin:pass@localhost:8080",


Could you please confirm that this URL works with Cloudant running locally via Docker?

See #1446:

When developing and testing locally, the app should use local docker-based Cloudant Developer Edition.

In my experience, Docker containers cannot be accessed at localhost.

If there is a reasonably easy way how to run Cloudant locally on localhost, perhaps without Docker, then I am fine with your current approach, but we need instructions for users how to set up such Cloudant instance.

See also acceptance criteria:

A deployment guide for IBM cloud added to our docs. If the current LB4 feature-set is not sufficient to support such deployments, then create a new wiki page instead of an official doc article.

Yes, it works with Cloudant at http://admin:pass@localhost:8080 works fine.

Will push the "deployment guide" in a while.

bajtos · 2018-08-13T17:01:22Z

examples/todo/src/datasources/db.datasource.ts

+
+    // tslint:disable-next-line:no-any
+    serviceGroup.forEach((service: VcapService) => {
+      if (service.name === config.name) {


Have you considered using an npm library to handle VCAP_SERVICES processing for you?

Few examples found by my quick search:

https://www.npmjs.com/package/vcap_services

https://www.npmjs.com/package/cf-services

The latter package says in its README: Unfortunately the instance name is rarely known in advance, so you may pass it as a separate environment variable. Is this something we need to take care of?

I haven't looked at them, since I was focussed on just making the app work. Will take a look at them.

add todo example on IBM Cloud

bajtos · 2018-08-14T07:51:30Z

examples/todo/src/datasources/db.datasource.ts

+};
+
+if (shouldConfigureForIbmCloud()) {
+  const services = JSON.parse(process.env.VCAP_SERVICES!);


I realized just now: we should not be accessing process.env from inside our application code (e.g. db.datasource.ts). Any environment-specific overrides should happen as close to the top-level main function as possible.

In LoopBack 3.x, it's the job of loopback-boot to deal with environment variables.

For the purpose of this spike, we can add VCAP_SERVICES processing code to the main function as you have already did for PORT variable.

Instead of modifying the config object loaded directly from datasource JSON file, we should iterate over all datasource configurations bound to app context, match them with VCAP_SERVICES entries and update the bound value using the URL from VCAP_SERVICES.

Later on, we can extract the code processing VCAP_SERVICES and other CF-related configuration into an extension people can add to their app.

bajtos · 2018-08-14T07:51:37Z

examples/todo/src/datasources/db.datasource.ts

+
+function shouldConfigureForIbmCloud() {
+  if (
+    process.env.HOME === '/home/vcap/app' && // relatively reliable way to detect the app is running on IBM Cloud


It should not be necessary to detect IBM Cloud. If the environment provides VCAP_SERVICES variable, then we should apply it, regardless of the specific cloud we are running on. (Isn't CloudFoundry an open source project that anyone can run on their own machines?)

Agreed. If there is VCAP_SERVICES in env, it is CF env.

bajtos · 2018-08-14T07:59:43Z

examples/todo/src/datasources/db.datasource.ts

+function shouldConfigureForIbmCloud() {
+  if (
+    process.env.HOME === '/home/vcap/app' && // relatively reliable way to detect the app is running on IBM Cloud
+    config.ibmCloud && // have to explicitly config datasource to mark it as an IBM Cloud service with a matching name locally


How about changing this flag from boolean ibmCloud: true into a VCAP SERVICE name to look up (e.g. vcapServiceName: 'db'?

It looks like VCAP_SERVICES provide binding_name and instance_name properties in addition to name, maybe we should use binding_name to find the right service record? See this presentation linked from cf-services for some background info.

To be honest, my knowledge of CloudFoundry and VCAP_SERVICES is pretty minimal, please correct me if I am talking nonsense :)

Great suggestion.

bajtos · 2018-08-14T08:06:42Z

examples/todo/DEPLOYING.md

+
+### LoopBack
+
+1.  Model configuration change is drastic.


Could you please add more details? AFAICT, there is no change in model/repository configuration, it's the DataSource config that requires modification.

It is just a note about first-time experience with LB4: model-config.json in LB3 was simpler. I don't think anything has to be done about this.

bajtos · 2018-08-14T08:09:50Z

examples/todo/DEPLOYING.md

+
+1.  Model configuration change is drastic.
+2.  Repository creation and linking to a model is manual. There should be a
+    provision to do these from the `lb` commands.


Is this referring to #1588? If so, then please add a link to your document.

bajtos · 2018-08-14T08:21:11Z

examples/todo/DEPLOYING.md

+1.  Model configuration change is drastic.
+2.  Repository creation and linking to a model is manual. There should be a
+    provision to do these from the `lb` commands.
+3.  There no API Explorer to play around with the models.


Could you please explain more? We do have API explorer hosted at loopback.io, LB4 apps provide /swagger-ui endpoint to access it, e.g. http://[::1]:3000/swagger-ui. Is this perhaps an issue of a missing documentation? Or maybe we should modify scaffolded app code to print explorer (swagger-ui) URL to console at startup?

The hosted API Explorer works well in Chrome and Firefox. It does not work in Safari for me, because swagger-ui is served from HTTPS while the app is running on HTTP and Safari does not allow mixed content. I guess we have a bug to fix here - see #1603

Yes, due to mixed content the API Explorer does not load on IBM Cloud. ~~Tested on Firefox, Chrome, Safari; fails on all~~. Works on Firefox, fails on Chrome and Safari.

review 1

review 2

bajtos

I think the updated code is good enough for spike purposes now 👍

dhmlau · 2018-09-19T03:45:08Z

@hacksparrow , since your spike is done, and the docs are out, is this PR good to close?

hacksparrow · 2018-10-04T05:40:48Z

This is done - #1446

hacksparrow requested review from bajtos and raymondfeng as code owners July 30, 2018 14:25

hacksparrow force-pushed the spike/deploy-to-ibm-cloud branch from df28ab7 to 7fa075c Compare July 31, 2018 07:19

hacksparrow requested a review from virkt25 as a code owner July 31, 2018 07:19

hacksparrow self-assigned this Jul 31, 2018

hacksparrow force-pushed the spike/deploy-to-ibm-cloud branch from 7fa075c to 2a72bac Compare July 31, 2018 07:53

virkt25 reviewed Jul 31, 2018

View reviewed changes

bajtos mentioned this pull request Aug 1, 2018

Monthly Milestone - August 2018 🍉⛱ #1581

Closed

31 tasks

virkt25 added this to the August Milestone milestone Aug 1, 2018

hacksparrow force-pushed the spike/deploy-to-ibm-cloud branch 2 times, most recently from e19396f to 81e099f Compare August 9, 2018 07:33

bajtos reviewed Aug 13, 2018

View reviewed changes

hacksparrow force-pushed the spike/deploy-to-ibm-cloud branch from 81e099f to b3a1a46 Compare August 14, 2018 06:48

chore: todo example on IBM Cloud

c21fb60

add todo example on IBM Cloud

hacksparrow force-pushed the spike/deploy-to-ibm-cloud branch from b3a1a46 to c21fb60 Compare August 14, 2018 06:52

bajtos changed the title ~~[WIP] chore: todo example on IBM Cloud~~ [WIP] Spike: todo example on IBM Cloud [DO NOT MERGE] Aug 14, 2018

bajtos reviewed Aug 14, 2018

View reviewed changes

hacksparrow force-pushed the spike/deploy-to-ibm-cloud branch 3 times, most recently from af409a6 to 5bec856 Compare August 14, 2018 10:19

chore: review 1

139a702

review 1

hacksparrow force-pushed the spike/deploy-to-ibm-cloud branch from 5bec856 to 139a702 Compare August 14, 2018 13:32

chore: review 2

3f55d2d

review 2

bajtos approved these changes Aug 14, 2018

View reviewed changes

This was referenced Aug 15, 2018

Document how to deploy LB4 app to IBM Cloud via CloudFoundry #1604

Closed

[Spike] Deploy LB4 applications to IBM Cloud #1446

Closed

bajtos mentioned this pull request Aug 28, 2018

[Spike] Allow extensions to contribute custom convention for environment-specific operational configuration #1464

Open

hacksparrow closed this Oct 4, 2018

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[WIP] Spike: todo example on IBM Cloud [DO NOT MERGE] #1574

[WIP] Spike: todo example on IBM Cloud [DO NOT MERGE] #1574

hacksparrow commented Jul 30, 2018 •

edited by bajtos

Loading

bajtos commented Jul 31, 2018

virkt25 Jul 31, 2018

hacksparrow Aug 3, 2018

virkt25 Aug 7, 2018

bajtos left a comment

bajtos Aug 13, 2018

hacksparrow Aug 14, 2018 •

edited

Loading

bajtos Aug 13, 2018

hacksparrow Aug 14, 2018

bajtos Aug 14, 2018

bajtos Aug 14, 2018 •

edited

Loading

hacksparrow Aug 14, 2018

bajtos Aug 14, 2018

hacksparrow Aug 14, 2018

bajtos Aug 14, 2018

hacksparrow Aug 14, 2018

bajtos Aug 14, 2018

bajtos Aug 14, 2018

hacksparrow Aug 14, 2018 •

edited

Loading

bajtos left a comment

dhmlau commented Sep 19, 2018

hacksparrow commented Oct 4, 2018

[WIP] Spike: todo example on IBM Cloud [DO NOT MERGE] #1574

[WIP] Spike: todo example on IBM Cloud [DO NOT MERGE] #1574

Conversation

hacksparrow commented Jul 30, 2018 • edited by bajtos Loading

Checklist

bajtos commented Jul 31, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

bajtos left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

hacksparrow Aug 14, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

bajtos Aug 14, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

hacksparrow Aug 14, 2018 • edited Loading

Choose a reason for hiding this comment

bajtos left a comment

Choose a reason for hiding this comment

dhmlau commented Sep 19, 2018

hacksparrow commented Oct 4, 2018

hacksparrow commented Jul 30, 2018 •

edited by bajtos

Loading

hacksparrow Aug 14, 2018 •

edited

Loading

bajtos Aug 14, 2018 •

edited

Loading

hacksparrow Aug 14, 2018 •

edited

Loading