General idea

Build Status:

We at @SchweizerischeBundesbahnen have a lot of projects who need changes on their projects all the time. As those settings are (and that is fine) limited to the administrator roles, we had to do a lot of manual changes like:

OpenShift:

• Creating new projects with certain attributes
• Updating projects metadata like billing information
• Updating project quotas
• Creating service-accounts

Persistent storage:

• Create gluster volumes
• Increase the size of a gluster volume
• Create PV, PVC, Gluster Service & Endpoints in OpenShift

Billing:

• Create a billing report for diffrent platforms

AWS:

• Create and manage AWS S3 Buckets

Sematext:

• Create and manage sematext logsene apps

So we built this tool which allows users to do certain things in self service. The tool checks permissions & certain conditions.

Components

• The Self-Service-Portal Backend (as container)
• The Self-Service-Portal Frontend (see https://github.com/oscp/cloud-selfservice-portal-frontend)
• The GlusterFS-API server (as a sytemd service)

Installation & Documentation

Self-Service Portal

# Create a project & a service-account
oc new-project ose-selfservice-backend
oc create serviceaccount ose-selfservice

# Add a cluster policy for the portal:
oc create -f clusterPolicy-selfservice.yml

# Add policy to service account
oc adm policy add-cluster-role-to-user ose:selfservice system:serviceaccount:ose-selfservice:ose-selfservice

# Use the token of the service account in the container

Just create a 'oc new-app' from the dockerfile.

Parameters

Param	Description	Example
GIN_MODE	Mode of the Webframework	debug/release
LDAP_URL	Your LDAP	ldap.xzw.ch
LDAP_BIND_DN	LDAP Bind	cn=root
LDAP_BIND_CRED	LDAP Credentials	secret
LDAP_SEARCH_BASE	LDAP Search Base	ou=passport-ldapauth
LDAP_FILTER	LDAP Filter	(uid=%s)
SESSION_KEY	A secret password to encrypt session information	secret
OPENSHIFT_API_URL	Your OpenShift API Url	https://master01.ch:8443
OPENSHIFT_TOKEN	The token from the service-account
MAX_QUOTA_CPU	How many CPU can a user assign to his project	30
MAX_QUOTA_MEMORY	How many GB memory can a user assign to his project	50
GLUSTER_API_URL	The URL of your Gluster-API	http://glusterserver01:80
GLUSTER_SECRET	The basic auth password you configured on the gluster api	secret
GLUSTER_IPS	IP addresses of the gluster endpoints	192.168.1.1,192.168.1.2
MAX_VOLUME_GB	How many GB storage can a user order	100
DDC_API	URL of the DDC Billing API	http://ddc-api.ch
AWS_PROD_ACCESS_KEY_ID	AWS Access Key ID to manage AWS ressources for production buckets	AKIAIOSFODNN7EXAMPLE
AWS_PROD_SECRET_ACCESS_KEY	AWS Secret Access Key to manage AWS ressources for production buckets	wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY
AWS_NONPROD_ACCESS_KEY_ID	AWS Access Key ID to manage AWS ressources for development buckets	AKIAIOSFODNN7EXAMPLE
AWS_NONPROD_SECRET_ACCESS_KEY	AWS Secret Access Key to manage AWS ressources for development buckets	wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY
AWS_S3_BUCKET_PREFIX	Prefix for all generated S3 buckets	mycompany
AWS_REGION	Region for all the aws artifacts	eu-central-1
SEMATEXT_API_TOKEN	Admin token for Sematext Logsene Apps	mytoken
SEMATEXT_BASE_URL	Base url for Sematext	for EU: https://apps.eu.sematext.com/

Route timeout

The api/aws/ec2 endpoints wait until VMs have the desired state. This can exceed the default timeout and result in a 504 error on the client. Increasing the route timeout is described here: https://docs.openshift.org/latest/architecture/networking/routes.html#route-specific-annotations

The GlusterFS api

Use/see the service unit file in ./glusterapi/install/

Parameters

glusterapi -poolName=your-pool -vgName=your-vg -basePath=/your/mount -secret=yoursecret -port=yourport

# poolName = The name of the existing LV-pool that should be used to create new logical volumes
# vgName = The name of the vg where the pool lies on
# basePath = The path where the new volumes should be mounted. E.g. /gluster/mypool
# secret = The basic auth secret you specified above in the SSP
# port = The port where the server should run
# maxGB = Optinally specify max GB a volume can be. Default is 100

Monitoring endpoints

The gluster api has two public endpoints for monitoring purposes. Call them this way:

The first endpoint returns usage statistics:

curl <yourserver>:<port>/volume/<volume-name>
{"totalKiloBytes":123520,"usedKiloBytes":5472}

The check endpoint returns if the current %-usage is below the defined threshold:

# Successful response
curl -i <yourserver>:<port>/volume/<volume-name>/check\?threshold=20
HTTP/1.1 200 OK
Content-Type: application/json; charset=utf-8
Date: Mon, 12 Jun 2017 14:23:53 GMT
Content-Length: 38

{"message":"Usage is below threshold"}

# Error response
curl -i <yourserver>:<port>/volume/<volume-name>/check\?threshold=3

HTTP/1.1 400 Bad Request
Content-Type: application/json; charset=utf-8
Date: Mon, 12 Jun 2017 14:23:37 GMT
Content-Length: 70
{"message":"Error used 4.430051813471502 is bigger than threshold: 3"}

For the other (internal) endpoints see the code (glusterapi/main.go)

Contributing

The backend can be started with Docker. All required environment variables must be set in the env_vars file.

# without proxy:
docker build -p 8080:8080 -t ssp-backend .
# with proxy:
docker build -p 8080:8080 --build-arg https_proxy=http://proxy.ch:9000 -t ssp-backend .

# env_vars must not contain export and quotes
docker run -it --rm --env-file <(sed "s/export\s//" env_vars | tr -d "'") ssp-backend

There is a small script for locally testing the API. It handles authorization (login, token etc).

go run curl.go [-X GET/POST] http://localhost:8080/api/...