sostheim / krak8s Goto Github PK

View Code? Open in Web Editor NEW

1.0 1.0 5.0 6.85 MB

API Service for Kraken and Kubernetes Commands

License: Apache License 2.0

Makefile 1.96% Go 98.04%

krak8s's People

Contributors

Stargazers

Watchers

Forkers

venezia joejulian samsung-cnct

krak8s's Issues

Protect Cluster Resources from All Mutating Operations Until They Reach Steady State

See linked issue: https://github.com/reactioncommerce/launchdock/issues/80

Deleting a resource record while that resource is still being created causes the application to drop a stack trace and fails to resolve the state after the failure.

Deleted Resource not Removed Completely

A deleted project is not having it's complete configuration and resources removed.

Initially reported by @jshimko

Update to latest configuration file format for Kraken config

implement required changes to kraken config.yaml

Additional Errors in Resource Delete Operations

Resolve errors in deleting objects from the api resources.

deleting a namespace that has a cluster does not actually delete the cluster

When you delete a namespace that has a cluster associated with it, the cluster resource is deleted from krak8s as expected, but it doesn't actually delete anything from AWS.

Setup

create project
create namespace in that project
create cluster in that namespace

Reproduction

delete the namespace
note that both namespace and cluster have been deleted
go to AWS and note all servers still running

API spec inconsistency

The API spec has several inconsistently formatted inputs in the request bodies. There are underscores, camelCase, no separation (as with namespaceid for listing applications), etc. In several cases, there are a mix of these syntaxes within the same object.

Also, every resource endpoint other than clusters is a plural noun (clusters is /v1/projects/${projectId}/cluster). It also looks like you can't list clusters because that endpoint doesn't exist. Every other resource type can be listed.

After Failed Deleting Get Namespaces Returns Panic Stack Dump in HTTP Response

Initially reported by @jshimko

GET /v1/projects/5669723d/namespaces

"panic: runtime error: invalid memory address or nil pointer dereference\npanic(0x1434180, 0x200aeb0)\n\t/usr/local/go/src/runtime/panic.go:491 +0x283\nmain.MarshalNamespaceObject(0x0, 0x1)\n\t/go/src/krak8s/namespace.go:34 +0x39\nmain.(*NamespaceController).List(0xc42020e050, 0xc4201f22a0, 0xc4203d8800, 0xc420356ea0)\n\t/go/src/krak8s/namespace.go:107 +0x103\nkrak8s/app.MountNamespaceController.func4(0x2031b00, 0xc4201f2240, 0x202ac40, 0xc4203b7a40, 0xc4203d8800, 0xc4203bd3f8, 0xc4203bd400)\n\t/go/src/krak8s/app/controllers.go:335 +0xdd\nkrak8s/vendor/github.com/goadesign/goa.(*Controller).MuxHandler.func1.1(0x2031b00, 0xc4201f2240, 0x202ac40, 0xc4203b7a40, 0xc4203d8800, 0x2, 0x0)\n\t/go/src/krak8s/vendor/github.com/goadesign/goa/service.go:270 +0xaa\nkrak8s/vendor/github.com/goadesign/goa/middleware.Recover.func1.1(0x2031b00, 0xc4201f2240, 0x202ac40, 0xc4203b7a40, 0xc4203d8800, 0x0, 0x0)\n\t/go/src/krak8s/vendor/github.com/goadesign/goa/middleware/recover.go:37 +0xa0\nkrak8s/vendor/github.com/goadesign/goa/middleware.ErrorHandler.func1.1(0x2031b00, 0xc4201f2240, 0x202ac40, 0xc4203b7a40, 0xc4203d8800, 0x2, 0x2)\n\t/go/src/krak8s/vendor/github.com/goadesign/goa/middleware/error_handler.go:21 +0x90\nkrak8s/vendor/github.com/goadesign/goa/middleware.LogRequest.func1.1(0x2031b00, 0xc4201f2240, 0x202ac40, 0xc4203b7a40, 0xc4203d8800, 0xc420254080, 0x2031b00)\n\t/go/src/krak8s/vendor/github.com/goadesign/goa/middleware/log_request.go:75 +0xd48\nkrak8s/vendor/github.com/goadesign/goa/middleware.RequestIDWithHeaderAndLengthLimit.func1.1(0x2031b00, 0xc4201f21e0, 0x202ac40, 0xc4203b7a40, 0xc4203d8800, 0xc4201f2150, 0x2031b00)\n\t/go/src/krak8s/vendor/github.com/goadesign/goa/middleware/request_id.go:63 +0x144\nkrak8s/vendor/github.com/goadesign/goa.(*Controller).MuxHandler.func1(0x2030980, 0xc4201fe0e0, 0xc4203d8800, 0xc4201f2150)\n\t/go/src/krak8s/vendor/github.com/goadesign/goa/service.go:303 +0x336\nkrak8s/vendor/github.com/goadesign/goa.(*mux).Handle.func1(0x2030980, 0xc4201fe0e0, 0xc4203d8800, 0xc4201f20f0)\n\t/go/src/krak8s/vendor/github.com/goadesign/goa/mux.go:59 +0x1cb\nkrak8s/vendor/github.com/dimfeld/httptreemux.(*TreeMux).ServeLookupResult(0xc420292140, 0x2030980, 0xc4201fe0e0, 0xc4203d8800, 0xc8, 0xc42020e1b0, 0xc4201f20f0, 0x0)\n\t/go/src/krak8s/vendor/github.com/dimfeld/httptreemux/router.go:245 +0x133\nkrak8s/vendor/github.com/dimfeld/httptreemux.(*TreeMux).ServeHTTP(0xc420292140, 0x2030980, 0xc4201fe0e0, 0xc4203d8800)\n\t/go/src/krak8s/vendor/github.com/dimfeld/httptreemux/router.go:266 +0xdb\nkrak8s/vendor/github.com/goadesign/goa.(*mux).ServeHTTP(0xc420285a80, 0x2030980, 0xc4201fe0e0, 0xc4203d8800)\n\t/go/src/krak8s/vendor/github.com/goadesign/goa/mux.go:85 +0x4c\nnet/http.serverHandler.ServeHTTP(0xc42040fee0, 0x2030980, 0xc4201fe0e0, 0xc4203d8800)\n\t/usr/local/go/src/net/http/server.go:2619 +0xb4\nnet/http.(*conn).serve(0xc42018a1e0, 0x2031a40, 0xc4201b1040)\n\t/usr/local/go/src/net/http/server.go:1801 +0x71d\ncreated by net/http.(*Server).Serve\n\t/usr/local/go/src/net/http/server.go:2720 +0x288\n"

Improve Dup Checking Solution for Projects and Namespaces

Currently the system does a very simple, to simple, brute force search for duplicate terms in the kraken configuration files. If a duplicate is detected, the action is rejected.

Need to improve the communication of the status on a duplicate instance to the client
Need to improve the fidelity of checking for duplicate names to understand the object and it's context before rejecting the action
Need to add user documentation to inform client's of the API what the rules are concerning duplicate, e.g. naming requirements.

clusters not deploying on in-cluster krak8s API

As mentioned in Slack last night, the API currently does not work. Just opening an issue to track progress.

I’ve followed the readme examples for the API and I don’t seem to be able to actually create a new cluster. Here are my exact requests/responses (run from inside the Launchdock pod within the cluster). The created cluster never leaves the create_requested state and no new nodes have been created.

# create project
curl -X POST -H "Content-Type: application/json" -d '{"name":"acme"}' http://10.37.8.171:8080/v1/projects

{
  "created_at": "2017-08-29T00:03:31.043193048Z",
  "id": "f614ab6e",
  "name": "acme",
  "namespaces": null,
  "type": "project"
}


# create namespace
curl -X POST -H "Content-Type: application/json" -d '{"name":"acme-prod"}' http://10.37.8.171:8080/v1/projects/f614ab6e/namespaces

{
  "applications": null,
  "created_at": "2017-08-29T00:06:07.85514329Z",
  "id": "298f91ef",
  "name": "acme-prod",
  "resources": null,
  "type": "namespace"
}


# create cluster
curl -X POST -H "Content-Type: application/json" -d '{"namespace_id":"298f91ef", "nodePoolSize": 5}' http://10.37.8.171:8080/v1/projects/f614ab6e/cluster

{
  "created_at": "2017-08-29T01:07:28.912574145Z",
  "id": "b0ef89bd",
  "namespace_id": "298f91ef",
  "nodePoolSize": 5,
  "state": "create_requested",
  "type": "Resource",
  "updated_at": "0001-01-01T00:00:00Z"
}


# note that cluster is sitting in "create_requested" state 30 mins later
curl http://10.37.8.171:8080/v1/projects/f614ab6e/cluster/b0ef89bd

{
  "created_at": "2017-08-29T01:07:28.912574145Z",
  "id": "b0ef89bd",
  "namespace_id": "298f91ef",
  "nodePoolSize": 5,
  "state": "create_requested",
  "type": "Resource",
  "updated_at": "0001-01-01T00:00:00Z"
}

can't list applications with /v1/projects/${id}/applications

I have a project, namespace, cluster, etc. and a single application running. I can query for the single application and successfully get it, but I can't list applications in that same namespace.

# GET the single app

curl http://localhost:8080/v1/projects/77907b06/applications/bb7dd432

{
  "channel": "stable",
  "config": "",
  "created_at": "2017-09-18T13:35:53.981897503Z",
  "deployment_name": "jeremy-mongodb",
  "id": "bb7dd432",
  "json_values": "",
  "name": "mongodb-replicaset",
  "namespace_id": "02678191",
  "registry": "quay.io/samsung_cnct",
  "server": "quay.io",
  "status": {
    "deployed_at": "2017-09-18T13:35:57.832285003Z",
    "state": "DEPLOYED"
  },
  "type": "application",
  "updated_at": "2017-09-18T13:35:57.832285386Z",
  "username": "",
  "version": "latest"
}

# GET all apps in that same namespace

curl http://localhost:8080/v1/projects/77907b06/applications \
     -H "Content-Type: application/json" \
     -d '{ "namespaceid": "02678191" }'


{
  "id": "9PjU+W/E",
  "code": "invalid_request",
  "status": 400,
  "detail": "missing required payload"
}

Yes, I realize my body has namespaceid instead of namespace_id, but that’s what the spec says. Regardless, I tried both ways and got the same result.

Just to show you that the namespace definitely exists...

curl http://localhost:8080/v1/projects/77907b06/namespaces/02678191

{
  "applications": [
    {
      "oid": "bb7dd432",
      "url": "/v1/projects/77907b06/applications/bb7dd432"
    }
  ],
  "created_at": "2017-09-11T21:49:32.759664083Z",
  "id": "02678191",
  "name": "jeremy",
  "resources": {
    "oid": "e322aa1c",
    "url": "/v1/projects/77907b06/cluster/e322aa1c"
  },
  "type": "namespace"
}

"successful" application create not actually successful

I’ve been trying to get a Mongo deployment to happen via krak8s for hours and I’m still not seeing anything other than a single pending pod.

# proxying to localhost with kubectl...

curl -X "POST" "http://localhost:8080/v1/projects/77907b06/applications" \
     -H "Content-Type: application/json" \
     -d $'{
  "deployment_name": "jeremy-mongodb",
  "registry": "quay.io/samsung_cnct",
  "namespace_id": "02678191",
  "name": "mongodb-replicaset",
  "channel": "stable",
  "version": "latest"
}'

10-15 mins later…

curl http://localhost:8080/v1/projects/77907b06/applications/bb7dd432

{
  "channel": "stable",
  "config": "",
  "created_at": "2017-09-18T13:35:53.981897503Z",
  "deployment_name": "jeremy-mongodb",
  "id": "bb7dd432",
  "json_values": "",
  "name": "mongodb-replicaset",
  "namespace_id": "02678191",
  "registry": "quay.io/samsung_cnct",
  "server": "quay.io",
  "status": {
    "deployed_at": "2017-09-18T13:35:57.832285003Z",
    "state": "DEPLOYED"
  },
  "type": "application",
  "updated_at": "2017-09-18T13:35:57.832285386Z",
  "username": "",
  "version": "latest"
}

However…

kubectl get pods -n jeremy

NAME                                  READY     STATUS    RESTARTS   AGE
jeremy-mongodb-mongodb-replicaset-0   0/1       Pending   0          11m

and…

kubectl get deploy -n jeremy

No resources found.

Failed Application Deployment Shows State `DEPLOYED`

As shown in #10 - a failed application deployment is not actually being deployed (fixed in linked issue).

Still at issue is the fact that the deployment status being reported when queried is:

"status": {
    "deployed_at": "2017-09-18T13:35:57.832285003Z",
    "state": "DEPLOYED"
  }