ui: fix service page not rendering

[pkazmierczak]

Fixes a bug in the service detail page rendering.

Running a jobspec:

job "example" {

  group "group" {

    service {
      name     = "httpd-web"
      provider = "nomad"
    }

    task "task" {

      driver = "docker"
      user   = "www-data"

      config {
        image   = "busybox:1"
        command = "httpd"
        args    = ["-vv", "-f", "-p", "8001", "-h", "/local"]
      }

      resources {
        cpu    = 100
        memory = 100
      }

    }
  }
}
$ nomad service info http-web
Job ID   Address  Tags  Node ID   Alloc ID
example  <none>   []    2b91b659  5e117e45

In the UI I can see:

Console shows no JS/ember errors.

Resolves #27980

[tgross]

I checked out this branch to test (Chrome on Fedora Linux). I'm seeing the page rendered as I'd expect, but I see the following error in the console:

index.js:10 Uncaught (in promise) TypeError: Cannot read properties of undefined (reading 'length')
    at isIPv6 (index.js:10:321)
    at NetworkSerializer.normalize (network.js:29:1)
    at Class.deserializeSingle (fragment.js:77:1)
    at Class.<anonymous> (fragment-array.js:26:1)
    at Array.map (<anonymous>)
    at Class.deserializeFragmentArray [as deserialize] (fragment-array.js:25:1)
    at ext.js:211:1
    at has-many-d45aa09e.js:5484:1
    at Map.forEach (<anonymous>)
    at Resources.eachTransformedAttribute (has-many-d45aa09e.js:5483:1)

If I delete the service registration out from under the job, I see this error:

ember.js:308 Uncaught (in promise) Error: Assertion Failed: <model::service:_nomad-task-e6ef1eef-1f2b-d38a-2bd1-9f22b7a46337-group-web-httpd-web-www> is not a record instantiated by @ember-data/store
    at assert (ember.js:308:1)
    at recordIdentifierFor (index-cc461d33.js:2903:1)
    at lookupLegacySupport (has-many-d45aa09e.js:3650:1)
    at Service.belongsTo (has-many-d45aa09e.js:4528:1)
    at formatID (format-id.js:15:1)
    at ember.js:9589:1
    at SimpleClassicHelperManager.getValue (ember.js:9589:1)
    at ember.js:1901:1
    at ember.js:1763:1
    at track (ember.js:1654:1)

And once I hit this error, I can't tab-over to any of the other tabs in the job page without hard-reloading the page.

[pkazmierczak]

huh. ok I checked on macOS Chrome 148.0.7778.168, couldn't see any errors, but I'll do more checks with other browsers.

[pkazmierczak]

@tgross I was never able to reproduce your errors, neither on macOS nor on darwin. Tried with latest Chrome stable and FF.

@aklkv mind having a look at this? Tests are green, and in my manual testing it all works as expected.

[tgross]

@tgross I was never able to reproduce your errors, neither on macOS nor on darwin. Tried with latest Chrome stable and FF.

Including deleting the service out from under the UI? That's the one that triggered the more serious of the two bugs. I just reproduced again:

$ nomad service info -verbose httpd-web
ID           = _nomad-task-c9c0749b-b40d-4e6f-31f8-2e463039226b-group-web-httpd-web-www
Service Name = httpd-web
Namespace    = default
Job ID       = httpd
Alloc ID     = c9c0749b-b40d-4e6f-31f8-2e463039226b
Node ID      = c5cb6f23-a86d-373c-7528-cf70675557e9
Datacenter   = dc1
Address      = 192.168.1.194:26932
Tags         = []

$ nomad service delete httpd-web _nomad-task-c9c0749b-b40d-4e6f-31f8-2e463039226b-group-web-httpd-web-www
Successfully deleted service registration

[pkazmierczak]

$ nomad service delete httpd-web _nomad-task-c9c0749b-b40d-4e6f-31f8-2e463039226b-group-web-httpd-web-www
Successfully deleted service registration

a-ha! So what I did in the past was I'd update the job, removing the service. Never removed the service manually. I reproduced now!

[pkazmierczak]

ok ok. @tgross in ef0831c I fixed a few things. The services page now details the service as it should, but if a service gets deleted, it shows no allocations running a service, but the service still present in the list (since it's defined in the jobspec).

can you check on your end?

[tgross]

Works great, LGTM!

[github-actions]

I'm going to lock this pull request because it has been closed for 120 days ⏳. This helps our maintainers find and focus on the active contributions.
If you have found a problem that seems related to this change, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.