foundationdb

FoundationDB

According to the foundationDB documentation FoundationDB gives us the power of ACID transactions in a distributed database and has below properties

• Multi-model data store
• Easily scalable and fault tolerant
• Industry-leading performance

Note These steps are taken from the steps that are mentioned in the official repository of foundationDB operator. Since they don't have any any way to install the operator without building the operator image, we have followed the the upstream Makefile to create this README that can be used to create a simple foundationDB cluster.

Prerequisite

• Install GO on your machine, see the Getting Started guide for more information, and set GOPATH.
• Install KubeBuilder and its dependencies on your machine, see The KubeBuilder Book for more information.
• You should have Kustomize installed on you cluster.
• Kubernetes 1.9+ with Beta APIs enabled.
• PV support on the underlying infrastructure.
• Kanister version 0.109.0 with profiles.cr.kanister.io CRD installed.
• Docker CLI installed
• A docker image containing the required tools to back up FoundationDB. The Dockerfile for the image can be found here. To build and push the docker image to your registry, execute these steps.

Build docker image

Execute the commands below to build and push the foundationdb docker image to a registry.

# Clone Kanister Github repo locally
$ git clone https://github.com/kanisterio/kanister.git <path_to_kanister>

# Build FoundationDB docker image
$ docker build -t <registry>/<repository>/foundationdb:<tag_name> <path_to_kanister>/docker/foundationdb
$ docker push <registry>/<repository>/foundationdb:<tag_name>

Installation

We don't have the foundationDB helm chart yet to get the foundationDB installed on our Kubernetes cluster. So we will be manually installing foundationDB on our Kubernetes cluster. FoundationDB team recently started working on an operator, and we wil be using operator to install foundationDB cluster on our Kubernetes cluster.

To get to know about how Kubernetes operators, please follow this link.

Steps that are mentioned below can be followed to install the foundationDB operator

• Change your current directory to $GOPATH/src/github.com using the command cd $GOPATH/src/github.com and run mkdir foundationdb to create the directory foundationdb.
• CD into newly created directory and clone this github repo inside the created directory that is foundationdb using below command git clone https://github.com/FoundationDB/fdb-kubernetes-operator.git.
• Run sed -i '/IMG ?= fdb-kubernetes-operator:latest/c\IMG ?= ghcr.io/kanisterio/fdb-kubernetes-operator:latest' Makefile to correct the name of the operator image.
• Create Secrets to set up a secret with self-signed test certs using the command ./config/test-certs/generate_secrets.bash
• Run make rebuild-operator to install the operator. Please make sure this completes successfully.

Create the cluster

Once we have the CRD FoundationDBCluster (this will be taken care by make rebuild-operator) created we can go ahead with creating the CR for the already created CRD that will spin up a foundationDB cluster.

Please follow below command to create the CR, please make sure to use the local_cluster.yaml file from this repo.

NOTE:

Replace <registry>, <repository> and <tag_name> for the imageName value in ./local_cluster.yaml before running the following command.

$ kubectl create -f local_cluster.yaml

If you now go ahead and try to list all the pods, you will be able to see that there are some pods running for foundationDB cluster. By default we are running this cluster in double redundancy mode, because we faced some issues while running it as single redundancy mode, and have discussion going on about that, here.

Once we have the database pods up and running let's try to exec into a pod and then try to insert some key value pairs in the database

# exec into the foundation db pod
$ kubectl exec -it foundationdbcluster-sample-1 -c foundationdb bash

# once you are in the foundation DB pod get the fdbcli to run the command
$ fdbcli

fdb> writemode on
fdb> set name Tom
fdb> set lastname Manville

Protect the application

Once we have the foundationDB cluster up and running we can go ahead and integrate this with Kanister in order to take the backup and restore that backup. To achieve that we will have to create below mentioned Kanister resources

Create profile

kanctl create profile s3compliant --access-key <access-key>                 \
                --secret-key <secret-key>                                   \
                --bucket infracloud.kanister.io --region <region-name>      \
                --namespace default
profile <profile-name> created

The command will configure a location where artifacts resulting from Kanister data operations such as backup should go. This is stored as a profiles.cr.kanister.io CustomResource (CR) which is then referenced in Kanister ActionSets. Every ActionSet requires a Profile reference to complete the action. This CR (profiles.cr.kanister.io) can be shared between Kanister-enabled application instances.

Create blueprint

NOTE: v2 Blueprints are experimental and are not supported with standalone Kanister.

In order to perform backup, restore, and delete operations on the running foundationDB, we need to create a Blueprint Kanister resource. You can create the Blueprint using the command below.

# replace kanister-op-ns with the namespace where kanister is installed
kubectl create -f foundationdb-blueprint.yaml -n <kanister-op-ns>

Once we have created the Blueprint let's go ahead and insert some data into the foundationDB database.

Insert some records

To insert some records into the database we will have to EXEC into the foundationDB pod and then run the database command using the fdbcli utility

# EXEC into the pod
$ kubectl exec -it foundationdbcluster-sample-1 -c foundationdb bash
# get the fdbcli
$ fdbcli
fdb> writemode on
fdb> set name Tom
fdb> set companyname Kasten
fdb> set lasltname Manville

Create ClusterRoleBinding

Since the Kanister resources need access to the the foundationDB cluster that has just been created, using the CR, we will have create a ClusterRoleBinding to enable that access. You can create that ClusterRoleBinding resource using below command

$ kubectl create -f kanister-clusteradmin-binding.yaml

Create backup actionset

To take the backup of the data tha we have just inserted into the database, we will have to create Actionset Kanister resource. Please follow below command to create the Actionset

$ kanctl create actionset --action backup --namespace <kanister-op-ns> --blueprint foundationdb-blueprint  \
    --profile default/<profile-name>  \
    --objects apps.foundationdb.org/v1beta1/foundationdbclusters/default/foundationdbcluster-sample

actionset backup-jx2d2 created

Once you have created the Actionset, you can check the status of the Actionset by describing it to make sure the backup is completed. Please follow below command to check the status of the Actionset

$ kubectl describe actionset -n <kanister-op-ns> backup-jx2d2

Disaster strikes!

Once the backup is completed we can go ahead and delete the data manually from the database to imitate disaster.

$ kubectl exec -it foundationdbcluster-sample-1 -c foundationdb bash
$ fdbcli
fdb> writemode on
fdb> clearrange '' \xFF

Once we have cleared all the keys, we can go ahead and try get the value of any key that we have inserted and we should not get the value of that key because we have deleted the data.

$ kubectl exec -it foundationdbcluster-sample-1 -c foundationdb bash
$ fdbcli
fdb> get name
`name': not found

Once we have deleted the data from the database, to imitate the disaster, we will restore the backup that we have already created. To restore the backup we will have to create another Actionset, with restore action.

# Please replace kanister-op-ns with namespace where your kanister is installed
$ kanctl --namespace <kanister-op-ns> create actionset --action restore --from "backup-jx2d2"
actionset <restore-actionset> created

Once you have created the restore actionset, you can make sure that the actionset is completed by describing the actionset. You can describe the actionset using below command

$ kubectl describe actionset -n <kanister-operator-namespace> <restore-actionset-name>

If you have verified that the restore actionset is completed we can EXEC into the foundationDB pod once again to verify if the data has been restored back.

Delete the artifacts

The artifacts created by the backup action can be cleaned up using the following command:

# Replace kanister-op-ns with the namespace where your kanister operator is installed
$ kanctl --namespace <kanister-op-ns> create actionset --action delete --from "backup-jx2d2"
actionset "<delete-actionset-name>" created

# View the status of the ActionSet
# Replace kanister-op-ns with the namespace where your kanister operator is installed
$ kubectl --namespace <kanister-op-ns> describe actionset <delete-actionset-name>

Troubleshooting

If you run into any issues with the above commands, you can check the logs of the controller using:

# Replace kanister-op-ns with the namespace where your kanister operator is installed
$ kubectl --namespace <kanister-op-ns> logs -l app=kanister-operator

you can also check events of the actionset

$ kubectl describe actionset <actionset-name> -n kanister

Delete Blueprint and Profile CR

# Replace kanister-op-ns with the namespace where your kanister operator is installed
$ kubectl delete blueprints.cr.kanister.io foundationdb-blueprint -n <kanister-op-ns>

$ kubectl get profiles.cr.kanister.io -n default
NAME               AGE
s3-profile-sph7s   2h

$ kubectl delete profiles.cr.kanister.io s3-profile-sph7s -n default