4 Steps to Load Test MongoDB using YCSB workloads on Kubernetes cluster

Dawood Munavar, Arun K. Oct 06 - 2 min read

Audio : Listen to This Blog.

Introduction

In this post, we’ll be creating a Kubernetes cluster and configuring vSphere storage for kubernetes. We shall also deploy dynamic storage provisioning and configure MongoDB applications & run YCSB workloads for database simulations.

Step1: Install a Kubernetes Cluster on ESXi VMs.
Step2: Configure vSphere cloud provider, Integration of vSphere Storage for Kubernetes, Dynamic storage provisioning of PVC.
Step3: Configure custom Docker image on the Docker hub containing MongoDB, ycsb & the relevant packages to run ycsb workloads.
Step4: Load testing MongoDB using ycsb and obtaining the test results.

Step1: Installing Kubernetes cluster on ESXi VMs

Environment used:
vSphere 6.7 Update 3 infrastructure.
VM folder in vSphere to collect the kubernetes VMs.
ESXi hosts – version 6.7 u3.
Guest OS – CentOS Linux release 7.8.2003 (Core).
Cluster with HA & DRS enabled.

The following steps will run on the Master-Node.

Prepare Hostname, SELinux and Firewall

[root@localhost ~]# hostnamectl set-hostname VM1-master
[root@vm1-master ~]# setenforce 0
[root@vm1-master ~]# sed -i --follow-symlinks 's/SELINUX=enforcing/SELINUX=disabled/g' /etc/sysconfig/selinux

[root@vm1-master ~]# firewall-cmd --permanent --add-port=6443/tcp
[root@vm1-master ~]# firewall-cmd --permanent --add-port=2379-2380/tcp
[root@vm1-master ~]# firewall-cmd --permanent --add-port=10250/tcp
[root@vm1-master ~]# firewall-cmd --permanent --add-port=10251/tcp
[root@vm1-master ~]# firewall-cmd --permanent --add-port=10252/tcp
[root@vm1-master ~]# firewall-cmd --permanent --add-port=10255/tcp
[root@vm1-master ~]# firewall-cmd --reload
[root@vm1-master ~]# modprobe br_netfilter
[root@vm1-master ~]# echo '1' > /proc/sys/net/bridge/bridge-nf-call-iptables
[root@vm1-master ~]# cat /etc/hosts
10.20.176.108 VM1-master
10.20.176.110 VM2-worker
10.20.176.111 VM3-worker

Setup K8s repository

cat <<EOF > /etc/yum.repos.d/kubernetes.repo
[kubernetes]
name=Kubernetes
baseurl=https://packages.cloud.google.com/yum/repos/kubernetes-el7-x86_64
enabled=1
gpgcheck=1
repo_gpgcheck=1
gpgkey=https://packages.cloud.google.com/yum/doc/yum-key.gpg https://packages.cloud.google.com/yum/doc/rpm-package-key.gpg
EOF

Install Docker, Kubeadm

[root@vm1-master ~]# yum install kubeadm docker -y
[root@vm1-master ~]# systemctl restart docker && systemctl enable docker
[root@vm1-master ~]# systemctl restart kubelet && systemctl enable kubelet
[root@vm1-master ~]# swapoff -a

Initialize kubernetes cluster

[root@vm1-master ~]# kubeadm init
.
.
Your Kubernetes control-plane has initialized successfully!
Then you can join any number of worker nodes by running the following on each as root:

kubeadm join 10.20.176.108:6443 --token 12a5v2.fwneclcqgri0whof \
    --discovery-token-ca-cert-hash sha256:fdb98ddc292f0329cf19784e75cd359f07f560ef9aa8379f042fded7be517c7e

To use a sudo enabled user, run:

[root@vm1-master ~]# mkdir -p $HOME/.kube
[root@vm1-master ~]# sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
[root@vm1-master ~]# sudo chown $(id -u):$(id -g) $HOME/.kube/config

Deploy POD Network

[root@vm1-master ~]# export kubever=$(kubectl version | base64 | tr -d '\n')
[root@vm1-master ~]#  kubectl apply -f "https://cloud.weave.works/k8s/net?k8s-version=$kubever"

Check to see if the kubectl command is activated.

[root@vm1-master ~]# kubectl get nodes
NAME         STATUS   ROLES    AGE   VERSION
vm1-master   Ready    master   31m   v1.19.2

[root@vm1-master ~]# kubectl get pods --all-namespaces
NAMESPACE     NAME                                 READY   STATUS    RESTARTS   AGE
kube-system   coredns-f9fd979d6-2qmqt              1/1     Running   0          31m
kube-system   coredns-f9fd979d6-knl57              1/1     Running   0          31m
kube-system   etcd-vm1-master                      1/1     Running   0          31m
kube-system   kube-apiserver-vm1-master            1/1     Running   0          31m
kube-system   kube-controller-manager-vm1-master   1/1     Running   0          31m
kube-system   kube-proxy-gjqhw                     1/1     Running   0          31m
kube-system   kube-scheduler-vm1-master            1/1     Running   0          31m
kube-system   weave-net-lnhps                      2/2     Running   0          27m

The following steps will run on both the Worker-Nodes.

Prepare Hostname, SELinux and Firewall

[root@localhost ~]# hostnamectl set-hostname VM2-worker
[root@localhost ~]# setenforce 0
[root@localhost ~]# sed -i --follow-symlinks 's/SELINUX=enforcing/SELINUX=disabled/g' /etc/sysconfig/selinux
[root@localhost ~]# firewall-cmd --permanent --add-port=6783/tcp
success
[root@localhost ~]# firewall-cmd --permanent --add-port=10250/tcp
success
[root@localhost ~]# firewall-cmd --permanent --add-port=10255/tcp
success
[root@localhost ~]# firewall-cmd --permanent --add-port=30000-32767/tcp
success
[root@localhost ~]# firewall-cmd  --reload
success
[root@localhost ~]# modprobe br_netfilter
[root@localhost ~]# echo '1' > /proc/sys/net/bridge/bridge-nf-call-iptables
[root@vm2-worker ~]# cat /etc/hosts
10.20.176.108 VM1-master
10.20.176.110 VM2-worker
10.20.176.111 VM3-worker

Setup K8s repository

cat <<EOF > /etc/yum.repos.d/kubernetes.repo
[kubernetes]
name=Kubernetes
baseurl=https://packages.cloud.google.com/yum/repos/kubernetes-el7-x86_64
enabled=1
gpgcheck=1
repo_gpgcheck=1
gpgkey=https://packages.cloud.google.com/yum/doc/yum-key.gpg https://packages.cloud.google.com/yum/doc/rpm-package-key.gpg
EOF

Install Docker, Kubeadm

[root@localhost ~]# yum install kubeadm docker -y
[root@localhost ~]# systemctl restart docker && systemctl enable docker
[root@localhost ~]# swapoff -a

Join the Worker Node to the Kubernetes Cluster
(To be obtained from above master-node kubeadm init results)

[root@localhost ~]# kubeadm join 10.20.176.108:6443 --token 12a5v2.fwneclcqgri0whof --discovery-token-ca-cert-hash sha256:fdb98ddc292f0329cf19784e75cd359f07f560ef9aa8379f042fded7be517c7e

This node has joined the cluster:
* Certificate signing request was sent to apiserver and a response was received.
* The Kubelet was informed of the new secure connection details.

Run 'kubectl get nodes' on the control-plane to see this node join the cluster.

Note: As suggested go back to your master-node and check if worker-node have joined the cluster using the following command.

[root@vm1-master ~]# kubectl get nodes
NAME         STATUS   ROLES    AGE     VERSION
vm1-master   Ready    master   3h23m   v1.19.2
vm2-worker   Ready    <none>   144m    v1.19.2
vm3-worker   Ready    <none>   131m    v1.19.2

Step2: Configure vSphere cloud provider, Integration of vSphere Storage for Kubernetes, Dynamic storage provisioning of PVC

Prerequisites:

  1. All Kubernetes VMs must be placed in a vSphere VM folder.
  2. The disk.EnableUUID parameter must be set to TRUE for Kubernetes Node VMs.
  3. Configure Storage Policy-Based Management (SPBM) with VMFS - https://support.purestorage.com/Solutions/VMware_Platform_Guide/User_Guides_for_VMware_Solutions/VMware_Cloud_Native_Storage_User_Guide/Using_Storage_Policy_Based_Management_(SPBM)_with_VMFS
  4. Use above SPBM VMFS as default-datastore in the below configuration.
  5. Install jq, govc on master-node.
# Configuring the Kubernetes vSphere Cloud Provider 

On the Kubernetes Master node, Create vsphere.conf

Create the vSphere configuration file in /etc/kubernetes/vcp/vsphere.conf - you’ll need to create the folder.

[root@vm1-master ~]# cat /etc/kubernetes/vcp/vsphere.conf
[Global]
user = "[email protected]"
password = "Passw0rd"
port = "443"
insecure-flag = "1"

[VirtualCenter "10.20.176.180"]
datacenters = "munavar"

[Workspace]
server = "10.20.176.180"
datacenter = "munavar"
default-datastore = "wsame_large"
resourcepool-path = "munavar/Resources"
folder = "TKG"

[Disk]
scsicontrollertype = pvscsi

Modify the kubelet service

Append the --cloud-provider and --cloud-config flags to /var/lib/kubelet/kubeadm-flags.env

[root@vm1-master ~]# cat /var/lib/kubelet/kubeadm-flags.env
KUBELET_KUBEADM_ARGS="--network-plugin=cni --pod-infra-container-image=k8s.gcr.io/pause:3.2 --cloud-provider=vsphere --cloud-config=/etc/kubernetes/vcp/vsphere.conf"

Modify container manifests

The container manifests for the Kubernetes API server and controller-manager also need to be updated with the --cloud-provider and --cloud-config flags. These are located in the /etc/kubernetes/manifests folder.

Add the following to the spec:containers:command array, paying attention to whitespace (it is yaml, after all!)

    - --cloud-provider=vsphere
    - --cloud-config=/etc/kubernetes/vcp/vsphere.conf

We also need add a mount for the vsphere.conf file location in spec:containers:volumeMounts

    - mountPath: /etc/kubernetes/vcp
      name: vcp
      readOnly: true

And a volume in spec:volumes

  - hostPath:
      path: /etc/kubernetes/vcp
      type: DirectoryOrCreate
    name: vcp

On the Kubernetes Worker nodes
Modify the kubelet service
Append only the --cloud-provider flag to /var/lib/kubelet/kubeadm-flags.env

[root@vm2-worker ~]# cat /var/lib/kubelet/kubeadm-flags.env
KUBELET_KUBEADM_ARGS="--network-plugin=cni --pod-infra-container-image=k8s.gcr.io/pause:3.2 --cloud-provider=vsphere"

[root@vm3-worker ~]# cat /var/lib/kubelet/kubeadm-flags.env
KUBELET_KUBEADM_ARGS="--network-plugin=cni --pod-infra-container-image=k8s.gcr.io/pause:3.2 --cloud-provider=vsphere

Update the Node Provide IDs - from Master node

[root@vm1-master ~]# kubectl get nodes -o json | jq '.items[]|[.metadata.name, .spec.providerID, .status.nodeInfo.systemUUID]'
[
  "vm1-master",
  null,
  "4B531942-57E8-4A59-CFC8-72556461DAC7"
]
[
  "vm2-worker",
  null,
  "844C1942-185E-DD1A-AD54-843BEC5F6EF8"
]
[
  "vm3-worker",
  null,
  "CB251942-8386-DC5F-0863-2F5921D09456"
]

If the output contains 'null' for the providerId, then you need to set it:  the documentation provides a script - please refer to it 

https://vmware.github.io/vsphere-storage-for-kubernetes/documentation/existing.html#update-all-node-providerid-fields

Once the script is run, we may get the following output


[root@vm1-master ~]# kubectl get nodes -o json | jq '.items[]|[.metadata.name, .spec.providerID, .status.nodeInfo.systemUUID]'
[
  "vm1-master",
  "vsphere://4219534B-E857-594A-CFC8-72556461DAC7",
  "4B531942-57E8-4A59-CFC8-72556461DAC7"
]
[
  "vm2-worker",
  "vsphere://42194C84-5E18-1ADD-AD54-843BEC5F6EF8",
  "844C1942-185E-DD1A-AD54-843BEC5F6EF8"
]
[
  "vm3-worker",
  "vsphere://421925CB-8683-5FDC-0863-2F5921D09456",
  "CB251942-8386-DC5F-0863-2F5921D09456"
]

Consuming vSphere Storage - Create Storage Class

Note: Storage policies should be created on vSphere using SPBM VMFS before creating below storage class

[root@vm1-master ~]# cat sc-vsan-policy.yaml
kind: StorageClass
apiVersion: storage.k8s.io/v1
metadata:
  name: fast
provisioner: kubernetes.io/vsphere-volume
parameters:
    diskformat: thin
    storagePolicyName: kubernetes
    datastore: wsame_large

[root@vm1-master ~]# kubectl create -f sc-vsan-policy.yaml
storageclass.storage.k8s.io/fast created

[root@vm1-master ~]# kubectl describe storageclass
Name:                  fast
IsDefaultClass:        No
Annotations:           <none>
Provisioner:           kubernetes.io/vsphere-volume
Parameters:            datastore=wsame_large,diskformat=thin,storagePolicyName=kubernetes
AllowVolumeExpansion:  <unset>
MountOptions:          <none>
ReclaimPolicy:         Delete
VolumeBindingMode:     Immediate
Events:                <none>

Create a Persistent Volume Claim - Dynamic volume provisioning

[root@vm1-master ~]# cat vsphere-volume-pvcsc.yaml
kind: PersistentVolumeClaim
apiVersion: v1
metadata:
  name: pvc-volume1
  annotations:
    volume.beta.kubernetes.io/storage-class: fast
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 250Gi

[root@vm1-master ~]# kubectl create -f vsphere-volume-pvcsc.yaml
persistentvolumeclaim/pvc-volume1 created

[root@vm1-master ~]# kubectl describe pvc pvc-volume1
Name:          pvc-volume1
Namespace:     default
StorageClass:  fast
Status:        Bound
Volume:        pvc-7b58289a-709d-46f5-9e6c-585aba78f7d5
Labels:        <none>
Annotations:   pv.kubernetes.io/bind-completed: yes
               pv.kubernetes.io/bound-by-controller: yes
               volume.beta.kubernetes.io/storage-class: fast
               volume.beta.kubernetes.io/storage-provisioner: kubernetes.io/vsphere-volume
Finalizers:    [kubernetes.io/pvc-protection]
Capacity:      250Gi
Access Modes:  RWO
VolumeMode:    Filesystem
Mounted By:    <none>
Events:        <none>

[root@vm1-master ~]# kubectl describe pv
Name:            pvc-7b58289a-709d-46f5-9e6c-585aba78f7d5
Labels:          <none>
Annotations:     kubernetes.io/createdby: vsphere-volume-dynamic-provisioner
                 pv.kubernetes.io/bound-by-controller: yes
                 pv.kubernetes.io/provisioned-by: kubernetes.io/vsphere-volume
Finalizers:      [kubernetes.io/pv-protection]
StorageClass:    fast
Status:          Bound
Claim:           default/pvc-volume1
Reclaim Policy:  Delete
Access Modes:    RWO
VolumeMode:      Filesystem
Capacity:        250Gi
Node Affinity:   <none>
Message:
Source:
    Type:               vSphereVolume (a Persistent Disk resource in vSphere)
    VolumePath:         [wsame_large] kubevols/kubernetes-dynamic-pvc-7b58289a-709d-46f5-9e6c-585aba78f7d5.vmdk
    FSType:             ext4
    StoragePolicyName:  kubernetes
Events:                 <none>

Step3: Configure custom Docker image on the Docker hub containing MongoDB, ycsb, Python, maven & java to run ycsb workloads installed as part of Docker image

- Create public docker hub where we can push the custom docker image inorder to use.

  1. docker login –username= <Repository name>
  2. docker build -t <Image Name> . ==>This Dot indicates the location where Dockerfile is placed.
  3. docker tag <image name:tag> <repository name/image name :tag>

Once the Custom docker image is pushed into Docker hub, we can pull the image using
docker pull <repository name/image name :tag> for later usage.

MongoDB on Kubernetes with StatefulSets

[root@vm1-master ~]# cat mongo.yaml
apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: mongodb-replica
spec:
  serviceName: database
  replicas: 2
  selector:
    matchLabels:
      app: database
  template:
    metadata:
      labels:
        app: database
        selector: mongodb-replica
    spec:
      containers:
      - name: mongodb-stand
        image: arundockerregistry/mymongo:v2
        ports:
          - containerPort: 27017
  volumeClaimTemplates:
  - metadata:
      name: mongo-persistent-storage
      annotations:
        volume.beta.kubernetes.io/storage-class: "fast"
    spec:
      accessModes: [ "ReadWriteOnce" ]
      resources:
        requests:
          storage: 500Gi

[root@vm1-master ~]# kubectl create -f mongo.yaml
statefulset.apps/mongodb-replica created

[root@vm1-master ~]# kubectl get pods
NAME                READY   STATUS    RESTARTS   AGE
mongodb-replica-0   1/1     Running   0          6m23s
mongodb-replica-1   1/1     Running   0          6m19s

[root@vm1-master ~]# kubectl get pvc
NAME  STATUS   VOLUME  CAPACITY  ACCESS MODES   STORAGECLASS  AGE
mongo-persistent-storage-mongodb-replica-0   Bound    pvc-f974bbbc-266c-4789-ad25-8e6493f38eea   500Gi      RWO            fast           6m41s
mongo-persistent-storage-mongodb-replica-1   Bound    pvc-4bc36d34-27a7-4dcb-b36d-b503364b22a2   500Gi      RWO            fast           6m37s

[root@vm1-master ~]# kubectl exec -i -t mongodb-replica-0 -- bash
root@mongodb-replica-0:/# ls
bin   data  docker-entrypoint-initdb.d  home  lib    media  opt   root  sbin  sys  usr  ycsb-0.5.0   boot  dev   etc  js-yaml.js  lib64  mnt    proc  run   srv   tmp  var  ycsb-0.5.0.tar.gz
root@mongodb-replica-0:/# java -version
openjdk version "1.8.0_265"
OpenJDK Runtime Environment (build 1.8.0_265-8u265-b01-0ubuntu2~16.04-b01)
OpenJDK 64-Bit Server VM (build 25.265-b01, mixed mode)

Step4: Load testing MongoDB using YCSB and obtaining the test results.

YCSB is a popular Java open-source specification and program suite developed at Yahoo! to compare the relative performance of various NoSQL databases. Its workloads are used in various comparative studies of NoSQL databases

Workloads Used:
we chosen workloads A, B, C.
Workload A: Update heavy workload: 50/50% Mix of Reads/Writes
Workload B: Read mostly workload: 95/5% Mix of Reads/Writes
Workload C: Read-only: 100% reads.
The following command was used to run workload A,B & C on MongoDB pod:

/root/ycsb-0.5.0/bin/ycsb load mongodb -s -P /root/ycsb-0.5.0/workloads/workloada -p recordcount=10000000
/root/ycsb-0.5.0/bin/ycsb run mongodb -s -P /root/ycsb-0.5.0/workloads/workloada -p operationcount=10000000

/root/ycsb-0.5.0/bin/ycsb load mongodb -s -P /root/ycsb-0.5.0/workloads/workloadb -p recordcount=10000000
/root/ycsb-0.5.0/bin/ycsb run mongodb -s -P /root/ycsb-0.5.0/workloads/workloadb -p operationcount=10000000

/root/ycsb-0.5.0/bin/ycsb load mongodb -s -P /root/ycsb-0.5.0/workloads/workloadc -p recordcount=10000000
/root/ycsb-0.5.0/bin/ycsb run mongodb -s -P /root/ycsb-0.5.0/workloads/workloadc -p operationcount=10000000

Challenges Faced:

  1. Faced an issue during dynamic volume provisioning (PVC creation) fails with message - No VM found.
  2. Resolved it by following this comment - create Kubernetes master with vsphere cloud provider and add worker nodes after that with vsphere cloud provider.
    https://github.com/kubernetes/kubernetes/issues/71502

References:

  1. https://www.linuxtechi.com/install-kubernetes-1-7-centos7-rhel7/
  2. https://support.purestorage.com/Solutions/VMware_Platform_Guide/User_Guides_for_VMware_Solutions/VMware_Cloud_Native_Storage_User_Guide/Using_Storage_Policy_Based_Management_(SPBM)_with_VMFS
  3. https://www.definit.co.uk/2019/06/lab-guide-kubernetes-and-storage-with-the-vsphere-cloud-provider-step-by-step/
  4. https://vmware.github.io/vsphere-storage-for-kubernetes/documentation/existing.html#update-all-node-providerid-fields
  5. https://github.com/brianfrankcooper/YCSB/wiki/Running-a-Workload

Leave a Reply