A Kubernetes Operator that automates the packaging, provisioning, and managing of operational tasks for Kubernetes containerized applications. By default the NuoDB Kubernetes Operator deploys the NuoDB with Community Edition (CE) capability in the following tested and verified Kubernetes distributions:
- Red Hat OpenShift 3.11, 4.x
- On-prem or OpenShift supported public cloud platforms
- Google Cloud Platform (GCP)
- GKE managed Kubernetes
- Anthos GKE (on-prem) managed Kubernetes
- Open source Kubernetes
- Amazon Web Services (AWS)
- EKS managed Kubernetes
- Open source kubernetes
- Rancher Kubernetes Manager
- Rancher RKE and Rancher supported Kubernetes (e.g. EKS, AKS) on Rancher supported cloud platforms
Note: The NuoDB Operator is ideal for development, test and for product evaluation purposes in single cluster Kubernetes environments. The NuoDB Operator is not yet available for production use. Coming soon in 1H'20 is automated LEVEL 2 (rolling upgrade) and LEVEL 3 (backup and recovery) operational support. Multi-cluster support is also on the roadmap.
The NuoDB Operator also supports deploying NuoDB with either ephemeral or persistent storage options with configurations to run NuoDB Insights, a visual database monitoring Web UI, and start a sample application (ycsb) to quickly generate a user-configurable SQL workload against the database.
For a list of supported NuoDB Operator releases and where to download, click the Releases tab above. To enable automated notification of new releases, click the Watch button above and subscribe to the Releases Only selection.
Note: The master branch may contain changes not fully tested and represents a work-in-progress view of the next planned release.
The NuoDB Community Edition (CE) capability is a full featured, but limits the database to one Storage Manager (SM) and three Transaction Engine (TE) processes. The Community Edition is free of charge and allows you to self-evaluate NuoDB at your own pace. The NuoDB Community Edition (CE) will allow first time users to experience all the benefits and value points of NuoDB including:
- Ease of scale-out to meet changing application throughput requirements
- Continuous availability even in the event of common network, hardware, and software failures
- NuoDB database and workload visual monitoring with NuoDB Insights
- ANSI SQL
- ACID transactions
To effectively evaluate the NuoDB Community Edition (CE) we recommend creating a Kubernetes cluster of at least three nodes. To fully demonstrate transactional scale-out and database continuous availability we recommend four or five nodes.
As you proceed through the steps outlined on this page, we would like your self-guided NuoDB CE operator and database experience in Kubernetes to be a positive one! Please reach out to us at support@nuodb.com with any questions or comments you may have. We would be glad to learn more about your specific use case and provide assistance if needed!
To trial or run a PoC of the NuoDB Enterprise Edition (EE) which also allows users to scale the Storage Manager (SM) database process, contact NuoDB Sales at sales@nuodb.com for a PoC time-based enterprise edition license. For more information about NuoDB, see the NuoDB Website.
This page is organized in the following sections:
Configure NuoDB Insights Visual Monitor
Connect to your Database using nuosql
NuoDB Features and Benefits Evaluation Steps
Configuration Parameter Descriptions
Note: The instructions on this page use the Kubernetes kubectl command (for portability reasons across Kubernetes environments). For environments, the kubectl command is an alias that points to the OpenShift client program oc.
Create a Kubernetes cluster and connect to the cluster. In our verification tests, we regularly verify the samples workloads outlined on this page using the following minimal configuration:
- 4 worker nodes, each with with 4 CPUs and 16 GB of RAM (e.g. AWS m5.xlarge instance type)
- 5 GB disk for Admin pods
- 20 GB disk for Storage Manager(SM) pods
Please use this as a guideline for a minimal configuration when you create your cluster. To run larger SQL workloads using the included YCSB sample application, adjust node CPU and Memory upwards as required. To determine resources used, monitor your NuoDB database process resource consumption using the NuoDB Insights visual montioring tool.
export OPERATOR_NAMESPACE=nuodb
export NUODB_OPERATOR_VERSION=2.0.3 --confirm you set the correction NuoDB Operator version here.
3. Clone a copy of the NuoDB Operator release branch or download from the Releases tab a release zip file and unpack
For example, in your home or working directory, run:
git clone -b "rel-2.0.3" https://github.com/nuodb/nuodb-operator
kubectl create namespace $OPERATOR_NAMESPACE
NuoDB supports cloud platform storage (e.g. AWS EBS, GCP PD, Azure Disk, etc), 3rd-party CSI storage (e.g. Portworx, OpenEBS, Linbit, etc.), and the use of local storage via Hostpath. If planning to use any of these storage types, configure them prior to deploying your database.
Configure the local storage permissions on each cluster node to enable hosting storage for either the NuoDB Admin or the Storage Manager (SM) pods. Note: When using the local disk storage option only 1 Admin pod is supported.
sudo mkdir -p /mnt/local-storage/disk0
sudo chmod -R 777 /mnt/local-storage/
sudo chcon -R unconfined_u:object_r:svirt_sandbox_file_t:s0 /mnt/local-storage
sudo chown -R root:root /mnt/local-storage
Create the Kubernetes storage class "local-disk" and persistent volume
kubectl create -f nuodb-operator/build/etc/charts/nuodb-helm/local-disk-class.yaml
Label the cluster nodes you want to use to run NuoDB pods.
kubectl label node <node name> nuodb.com/zone=nuodb
Note: The label value, in this example "nuodb", can be any value.
Next, label one of these nodes as your storage node that will host the NuoDB Storage Manager (SM) pod. If using Local storage, ensure there is sufficient disk space on this node. To create this label run:
kubectl label node <yourStorageNodeDNSName> nuodb.com/node-type=storage
Once your cluster nodes are labeled for NuoDB use, run the following kubectl get nodes command to confirm nodes are labeled properly. The display output should look similar to the below
kubectl get nodes -l nuodb.com/zone -L nuodb.com/zone,nuodb.com/node-type
NAME STATUS ROLES AGE VERSION ZONE NODE-TYPE
ip-10-0-141-113.ec2.internal Ready worker 15d v1.13.4+cb455d664 nuodb storage
ip-10-0-152-147.ec2.internal Ready worker 15d v1.13.4+cb455d664 nuodb
ip-10-0-162-73.ec2.internal Ready worker 15d v1.13.4+cb455d664 nuodb
ip-10-0-184-233.ec2.internal Ready worker 15d v1.13.4+cb455d664 nuodb
ip-10-0-206-8.ec2.internal Ready worker 15d v1.13.4+cb455d664 nuodb
Each time a NuoDB Admin pod starts it will load a Kubernetes configmap that contains the current NuoDB license level information and places its contents in the /etc/nuodb/nuodb.lic file. When a request is made to either start a NuoDB Transaction Engine (TE) or Storage Manager (SM) process, the NuoDB Admin will check the license file contents to ensure the process request is authorized.
To apply a NuoDB Communiity Edition (CE) license file, run
kubectl create configmap nuodb-lic-configmap -n $OPERATOR_NAMESPACE --from-literal=nuodb.lic=""
To apply a NuoDB Enterprise Edition (EE) license file to a running NuoDB database system that is using a CE license,
obtain your EE license file from your NuoDB Sales or Support representative and copy the file to nuodb.lic, then run
kubectl delete configmap nuodb-lic-configmap -n $OPERATOR_NAMESPACE
kubectl create configmap nuodb-lic-configmap -n $OPERATOR_NAMESPACE --from-file=nuodb.lic
Then, delete a NuoDB Admin pod, and once the Admin pod has been restarted, connect to the new pod and run,
nuocmd set license --license-file /etc/nuodb/nuodb.lic
This command will propagate the new NuoDB EE license throughout the Admin tier (remaining pods).
Note: The filename specified in the above commands must be nuodb.lic
To check the effective NuoDB license and confirm license level, run
nuocmd --show-json get effective-license
To permit the pulling of the NuoDB database and operator container images, create the Kubernetes image pull secret
This secret will be used to pull the NuoDB Operator and NuoDB container images from the Red Hat Container Catalog (RHCC). Enter your Red Hat login credentials for the --docker-username and --docker-password values.
kubectl create secret docker-registry pull-secret \
--docker-username='yourUserName' --docker-password='yourPassword' \
--docker-email='yourEmailAddr' --docker-server='registry.connect.redhat.com'
Note: If using Quay.io (or other supported public repo) to pull the NuoDB container images, a Kubernetes secret is not required because the NuoDB repository is public. For example, to pull the image from quay.io, run at the command prompt, docker pull quay.io/nuodb/nuodb-operator.
Disable Linux Transparent Huge Pages (THP). Run the following required command to create a security context constraint which will allow the Operator to disable THP during Operator deployment.
kubectl create -n $OPERATOR_NAMESPACE -f nuodb-operator/deploy/thp-scc.yaml
oc adm policy add-scc-to-user privileged system:serviceaccount:$OPERATOR_NAMESPACE:nuodb-operator
oc adm policy add-scc-to-user privileged system:serviceaccount:elastic-system:elastic-operator
oc adm policy add-scc-to-user privileged system:serviceaccount:$OPERATOR_NAMESPACE:insights-server-release-logstash
Optionally deploy the NuoDB Insights visual monitoring tool (recommended). NuoDB Insights is a powerful database monitoring tool that can greatly aid in visualizing database workload and resource consumption. For more information about the benefits of Insights, and data privacy when using HOSTED, please refer to the NuoDB Insights Webpage.
Before deploying NuoDB, to enable NuoDB Insights you will need to choose one of the available deployment methods:
-
LOCAL: Deploy Insights locally on your Kubernetes cluster. With this option, all performance data is privately stored and managed locally on your cluster by starting local elasticsearch, logstash, kibana, and grafana components that are utilized by the Insights on-cluster monitoring solution. To enable this option:
- Apply the
nuodbinsightsserver_crd.yamlduring Operator deployment andnuodbinsightsserver_cr.yamlduring database deployment.
- Apply the
-
HOSTED in AWS: This option streams your NuoDB Insights performance data to the NuoDB hosted Insights data portal on the Amazon AWS public cloud. To access your secure performance data using this option, you will use a private Subscriber ID that will be provided to you once the Insights collection agent starts. Note: Before selecting this option, you must agree to our NuoDB Terms of Use and Data Privacy Policy located here: Terms of Use and Privacy Policy. If you agree to these terms, following the below instructions to enable HOSTED NuoDB Insights. This option can also be enabled at a later time if you choose. To enable this option:
- Set
insightsEnabledparamater totruein yournuodb-cr.yamlfile. By setting this option, this is your acknowledgement and acceptance of the NuoDB terms of use and data Privacy policy.
- Set
To install the NuoDB Operator into your Kubernetes cluster, follow the steps indicated for the appropriate Kubernetes Distribution you are using.
In OpenShift 4.x, the NuoDB Operator is available to install directly from the OpenShift OperatorHub, an integrated service catalog, accessible from within the OpenShift 4 Web UI which creates a seamless - single click experience - that allows users to install the NuoDB Operator from catalog-to-cluster in seconds.
Prerequisite: Run the following yaml in your OpenShift cluster to authorize the NuoDB Operator service account before installing the NuoDB Operator.
kubectl create -f nuodb-operator/deploy/cluster-op-admin.yaml
Steps:
- Select
Projectsfrom the OpenShift 4 left toolbar and click theNuoDBproject to make it your current project. - Select the
OperatorHubunder theCatalogsection in the OCP 4 left toolbar. - Select the
Databasefilter and scroll down to the NuoDB Application tile and click the tile. - In the right-hand corner of the NuoDB Operator page, click the
Installbutton. - On the "Create Operator Subscription" page, select the radio group option "A specific namespace on the cluster"
and enter the project/namespace in the pull-down field that you would like to install the NuoDB Operator,
then select
Subscribeto subscribe to the NuoDB Operator. - In less than a minute, on the page that displays should indicate the NuoDB Operator has been installed, see "1 installed" message.
- To verify the NuoDB Operator installed correctly, select
Installed Operatorsfrom the left toolbar. The STATUS column should show "Install Succeeded". - Select
Statusunder theProjectson the left toolbar to view your running Operator.
The following video provides a full walk-thru of how to deploy the NuoDB Operator and database in OpenShift 4.x.
Note: The same benefit areas can also be demonstrated in any NuoDB supported Kubernetes managed environment.
- Using the GCP Marketplace, locate the NuoDB Operator. Click the
Configurebotton and follow the on screen instructions to deploy the NuoDB Operator. During this step: - Select a GCP Project.
- Either create a GKE cluster or choose an existing one from the list
- Create a namespace called
nuodbin which to install the NuoDB Operator - Take defaults for
App instance nameandCluster Admin Service Account - Click the
Deploybutton. Your NuoDB Operator will deploy in several minutes. You can use the GCP Kubernetes Engine Web UI "Workloads" selection to monitor progress.
kubectl apply -f https://github.com/operator-framework/operator-lifecycle-manager/releases/download/0.10.1/crds.yaml
kubectl apply -f https://github.com/operator-framework/operator-lifecycle-manager/releases/download/0.10.1/olm.yaml
Clone the Operator marketplace repository
https://github.com/operator-framework/operator-marketplace
Note: If you experience the following error when running the catalogSource.yaml file in the next section, then you can install the Operator Marketplace to resolve this error. However, the error can also be ignored. The NuoDB Operator will install successfully without the Operator Marketplace.
error: unable to recognize "catalogSource.yaml": no matches for kind "OperatorSource" in version "operators.coreos.com/v1"
# Set the environment context to the namespace you will deploy the NuoDB Operator
kubectl config set-context --current --namespace=$OPERATOR_NAMESPACE
kubectl create -f nuodb-operator/deploy/catalogSource.yaml
kubectl create -n $OPERATOR_NAMESPACE -f nuodb-operator/deploy/operatorGroup.yaml
kubectl create -n $OPERATOR_NAMESPACE -f nuodb-operator/deploy/cluster_role.yaml
kubectl create -f nuodb-operator/deploy/cluster_role_binding.yaml
kubectl create -f nuodb-operator/deploy/cluster-op-admin.yaml
kubectl create -n $OPERATOR_NAMESPACE -f nuodb-operator/deploy/role.yaml
kubectl create -n $OPERATOR_NAMESPACE -f nuodb-operator/deploy/role_binding.yaml
kubectl create -n $OPERATOR_NAMESPACE -f nuodb-operator/deploy/service_account.yaml
## add NuoDB, Insights, and ycsb sample SQL app CRDs
kubectl create -f nuodb-operator/deploy/crds/nuodb_v2alpha1_nuodbadmin_crd.yaml
kubectl create -f nuodb-operator/deploy/crds/nuodb_v2alpha1_nuodb_crd.yaml
kubectl create -f nuodb-operator/deploy/crds/nuodb_v2alpha1_nuodbinsightsserver_crd.yaml
kubectl create -f nuodb-operator/deploy/crds/nuodb_v2alpha1_nuodbycsbwl_crd.yaml
# create a local copy of the nuodb-csv.yaml file
sed "s/placeholder/$OPERATOR_NAMESPACE/" nuodb-operator/deploy/olm-catalog/nuodb-operator/$NUODB_OPERATOR_VERSION/nuodb-operator.v$NUODB_OPERATOR_VERSION.clusterserviceversion.yaml > nuodb-csv.yaml
# You may pull the NuoDB Operator from quay.io which is the default. Optionally, to replace quay.io as the default location to pull the NuoDB Operator image, follow these examples:
# Create a new nuodb-csv.yaml flie that pulls from the Red Hat Container Catalog, run
# sed "s/quay.io/registry.connect.redhat.com/" nuodb-csv.yaml > nuodb-csv-rhcc.yaml
# Create a new nuodb-csv.yaml file that pulls from the Google Marketplace, run
# sed "s/quay.io/marketplace.gcr.io/" nuodb-csv.yaml > nuodb-csv-gcp.yaml
# To pull from the AWS Marketplace, make a copy of nuodb-csv.yaml and name it nuodb-csv-aws.yaml.
# Replace in the nuodb-csv-aws.yaml file the two image references with the following image pull value:
# 117940112483.dkr.ecr.us-east-1.amazonaws.com/d893f8e5-fe12-4e43-b792-8cb98ffc11c0/cg-756769224/quay.io/nuodb/nuodb-operator:2.0.3-1-latest
# If appliable, copy your customized nuodb-csv-xxx.yaml file to nuodb-csv.yaml and run,
kubectl create -n $OPERATOR_NAMESPACE -f nuodb-csv.yaml
# Check deployment rollout status every 5 seconds (max 10 minutes) until complete.
ATTEMPTS=0
ROLLOUT_STATUS_CMD="kubectl rollout status deployment/nuodb-operator -n $OPERATOR_NAMESPACE"
until $ROLLOUT_STATUS_CMD || [ $ATTEMPTS -eq 60 ]; do
ATTEMPTS=$((attempts + 1))
echo ""
kubectl get pods -n $OPERATOR_NAMESPACE
sleep 5
done
The nuodb-operator/deploy/crds directory includes sample Custom Resources yaml files that can be used to deploy a NuoDB database with on-cluster (locally deployed) NuoDB Insights, and a ycsb sample workload.
nuodb_v2alpha1_nuodbadmin_cr.yaml
nuodb_v2alpha1_nuodb_cr.yaml
nuodb_v2alpha1_nuodbinsightsserver_cr.yaml
nuodb_v2alpha1_nuodbycsbwl_cr.yaml`
Optionally, you can add or modify any of the below parameters values to your own customer resource files to customize your deployment. Each parameter is described in the Optional Database Parameters section. Sample deployment files are provided below.
This sample file uses persistent storage, uses AWS general purpose EBS storage (gp2), disables hosted Insights monitoring and starts three NuoDB Admin pods,
spec:
storageMode: persistent
insightsEnabled: false
adminCount: 3
adminStorageSize: 5Gi
adminStorageClass: gp2
apiServer: https://domain:8888
container: nuodb/nuodb-ce:latest
Note: If your Kubernetes cluster has less than three worker nodes then modify your adminCount value to equal the number of worker nodes in your cluster. This will deploy one NuoDB Admin process per worker node.
This sample file starts a database named test, uses persistent storage, uses AWS general purpose EBS storage (gp2), and includes various others configurations like controlling the number of desired TE pods, CPU, and memory used per NuoDB process type. Running multiple nuodb_cr.yaml files will start multiple databases within the same NuoDB domain.
spec:
storageMode: persistent
dbName: test
dbUser: dba
dbPassword: <changeMe>
smCount: 1
smMemory: 4Gi
smCpu: "2"
smStorageSize: 20G
smStorageClass: gp2
engineOptions: ""
teCount: 1
teMemory: 4Gi
teCpu: "2"
container: nuodb/nuodb-ce:latest
We recommend replacing the database password dbPassword value with one of your choice.
This sample file starts on-cluster NuoDB Insights and uses AWS general purpose EBS storage (gp2) to persist monitoring data.
spec:
elasticVersion: 7.3.0
elasticNodeCount: 1
kibanaVersion: 7.3.0
kibanaNodeCount: 1
storageClass: gp2
Note: For parameters adminStorageClass, smStorageClass, and storageClass enter the Kubernetes storage class value you wish to use. For example,
| Public Cloud | Kubernetes Storage Class |
|---|---|
| AWS | gp2 |
| GCP | standard |
| AZURE | standard_lrs or premium_lrs |
| User Config'd | Kubernetes Storage Class |
|---|---|
| Local Hostpath | local-disk |
| 3rd Party | vendor specific value |
spec:
ycsbLoadName: ycsb-load
ycsbWorkload: b
ycsbLbPolicy: ""
ycsbNoOfProcesses: 5
ycsbNoOfRows: 10000
ycsbNoOfIterations: 0
ycsbOpsPerIteration: 10000
ycsbMaxDelay: 240000
ycsbDbSchema: User1
ycsbContainer: nuodb/ycsb:latest
This sample deploys a NuoDB database using "on-cluster" NuoDB Insight visual monitoring and start a sample SQL application
# Set the environment context to the namespace you will deploy the NuoDB Operator
kubectl config set-context --current --namespace=$OPERATOR_NAMESPACE`
# add cluster-admin permissions to the nuodb-operator service account
kubectl create -f nuodb-operator/deploy/cluster-op-admin.yaml
# Modify / customize your NuoDB cr yaml files and run, (see samples below in next section)
kubectl create -n $OPERATOR_NAMESPACE -f nuodb-operator/deploy/crds/nuodb_v2alpha1_nuodbadmin-cr.yaml
kubectl create -n $OPERATOR_NAMESPACE -f nuodb-operator/deploy/crds/nuodb_v2alpha1_nuodb-cr.yaml
kubectl create -n $OPERATOR_NAMESPACE -f nuodb-operator/deploy/crds/nuodb_v2alpha1_nuodbinsightsserver_cr.yaml
kubectl create -n $OPERATOR_NAMESPACE -f nuodb-operator/deploy/crds/nuodb_v2alpha1_nuodb-ycsbwl-cr.yaml
#Wait for nuodb to be logstash instance to be ready
# Check deployment rollout status every 10 seconds (max 10 minutes) until complete.
ATTEMPTS=0
ROLLOUT_STATUS_CMD="kubectl rollout status sts/insights-server-release-logstash -n $OPERATOR_NAMESPACE"
until $ROLLOUT_STATUS_CMD || [ $ATTEMPTS -eq 60 ]; do
ATTEMPTS=$((attempts + 1))
echo ""
kubectl get pods -n $OPERATOR_NAMESPACE
sleep 5
done
# create the Insights client
kubectl create -f nuodb-operator/build/etc/insights-server/insights-client.yaml
echo "Obtain your NuoDB Insights Dashboard URL:"
echo "For Red Hat OpenShift, go to URL:"
echo " https://$(kubectl get route grafana-route --output=jsonpath={.spec.host})/d/000000002/system-overview?orgId=1&refresh=10s"
echo "For Google GKE, go to URL:"
echo " http://$(kubectl get ingress grafana-ingress --output=jsonpath={.status.loadBalancer.ingress[0].ip})/d/000000002/system-overview?orgId=1&refresh=10s"
echo "For EKS or open source K8S,"
echo "Run the following command in a terminal window suitable for logging output commands:"
echo " $ kubectl port-forward service/grafana-service 3000 &"
echo "Go to URL:"
echo " localhost:3000/d/000000002/system-overview?orgId=1&refresh=10s"
To obtain your on-cluster NuoDB Insights URL,
For Red Hat OpenShift, go to URL:
echo " https://$(kubectl get route grafana-route --output=jsonpath={.spec.host})/d/000000002/system-overview?orgId=1&refresh=10s"
For Google GKE, go to URL:
echo " http://$(kubectl get ingress grafana-ingress --output=jsonpath={.status.loadBalancer.ingress[0].ip})/d/000000002/system-overview?orgId=1&refresh=10s"
For EKS or open source K8S, run the following command in a terminal window suitable for logging output commands:
$ kubectl port-forward service/grafana-service 3000 &"
Go to URL:
localhost:3000/d/000000002/system-overview?orgId=1&refresh=10s
Note: It may take several minutes for the NuoDB Insights URL to become available.
Optionally, you can choose to send your performance data to the NuoDB publicly hosted Insights portal. Your performance data remains private and is only accessible by using your private Subscriber ID. With this option, you can find your NuoDB Insights SubcriberID by locating the "nuodb-insights" pod in your Kubernetes dashboard, go to the Logs tab, and find the line that indicates your Subscriber ID.
Insights Subscriber ID: yourSubID#
Note: When using the open source Kubernetes dashboard:** A current Kubernetes dashboard Web UI issue doesn't allow users to retrieve their Insights Subscription ID using the dashboard to inspect the nuodb-insights log file. Instead run,
kubectl logs nuodb-insights -n nuodb -c insights
To connect to NuoDB Insights, open a Web browser using the following URL
https://insights.nuodb.com/yourSubID#
To check the status of hosted NuoDB Insights visual monitoring tool, run
oc exec -it nuodb-insights -c insights -- nuoca check insights
Once your database is running, you can connect and run SQL using the NuoDB nuosql command tool. See sample below:
kubectl exec -it <te-pod-name> bash
nuosql <db-name> --user dba --password secret
SQL>
For more information on how to run SQL, see Using NuoDB SQL Command Line and NuoDB SQL Development online documenation.
The NuoDB Operator includes a sample SQL application that will allow you to get started quickly running SQL statements against your NuoDB database. The sample workload uses YCSB (the Yahoo Cloud Servicing Benchmark). The cr.yaml includes YCSB parameters that will allow you to configure the SQL workload to your preferences.
To start a SQL Workload (if your nuodb-ycsbwl-cr.yaml wasn't configured to start one by default) locate the ycsb Replication Controller in your Kubernetes dashboard and scale it to your desired number of pods to create your desired SQL application workload. Once the YCSB application is running the resulting SQL workload will be viewable from the NuoDB Insights visual monitoring WebUI.
Once your NuoDB database and Insights visual monitor are running, here are a few steps to try out to quickly realize the benefits of running a NuoDB SQL database
-
Demonstrate Transactional Scale-out
To easily scale out your NuoDB Transaction Engines (TEs) to meet an increased SQL application workload, from the CLI, edit the teCount value by running,
kubectl edit nuodbs.nuodb.comUsing NuoDB Insights, monitor the SQL application during the time of increased application workload and the scale-out of TEs and observe the increase in transactional throughput, Transactions Per Second (TPS)
-
Demonstrate Continuous Availability
Deploy your NuoDB system with three NuoDB Admins and three Transaction Engines (TEs). Using either your Kubernetes dashboard or the
kubectl delete podCLI command, forcibly delete either a NuoDB Admin or a TE pod. Note: when using the Community Edition, only 1 Storage Manager (SM) is availabe.Using NuoDB Insights, monitor the SQL application during the time of the forced pod deletes and observe that SQL application availability continues unimpacted while the Kubernetes system and NuoDB recovery from failure events in seconds.
kubectl delete -n $OPERATOR_NAMESPACE configmap nuodb-lic-configmap
kubectl delete pod/insights-client
kubectl delete -f nuodb-operator/deploy/crds/nuodb_v2alpha1_nuodbadmin-cr.yaml
kubectl delete -f nuodb-operator/deploy/crds/nuodb_v2alpha1_nuodb-cr.yaml
kubectl delete -f nuodb-operator/deploy/crds/nuodb_v2alpha1_nuodb-insights-cr.yaml
kubectl delete -f nuodb-operator/deploy/crds/nuodb_v2alpha1_nuodb-ycsbwl-cr.yaml
# Delete the NuoDB persistent storage volumes claims
kubectl delete -n $OPERATOR_NAMESPACE pvc --all
If the local-disk storage class was used, then delete the NuoDB Storage Manager(SM) disk storage and storage class
ssh -i ~/Documents/cluster.pem $JUMP_HOST
ssh -i ~/.ssh/cluster.pem core@ip-n-n-n-n.ec2.internal 'rm -rf /mnt/local-storage/disk0/*'
kubectl delete -f local-disk-class.yaml
From the OpenShift WebUI, locate the OperatorHub under the Catalog left-bar selection. Select the NuoDB Operator and click the Uninstall button.
Run the following commands
kubectl delete pod/insights-client
kubectl delete nuodbinsightsservers/insightsserver
kubectl delete nuodbs/nuodb-db
kubectl delete nuodbycsbwls/nuodbycsbwl
kubectl delete pvc --all
kubectl delete pv --all
kubectl delete -f nuodb-operator/deploy/catalogSource.yaml
kubectl delete -n $OPERATOR_NAMESPACE -f nuodb-operator/deploy/operatorGroup.yaml
kubectl delete -n $OPERATOR_NAMESPACE -f nuodb-operator/deploy/cluster_role.yaml
kubectl delete -f nuodb-operator/deploy/cluster_role_binding.yaml
kubectl delete -f nuodb-operator/deploy/cluster-op-admin.yaml
kubectl delete -n $OPERATOR_NAMESPACE -f nuodb-operator/deploy/role.yaml
kubectl delete -n $OPERATOR_NAMESPACE -f nuodb-operator/deploy/role_binding.yaml
kubectl delete -n $OPERATOR_NAMESPACE -f nuodb-operator/deploy/service_account.yaml
kubectl delete -f nuodb-operator/deploy/crds/nuodb_v2alpha1_nuodbadmin_crd.yaml
kubectl delete -f nuodb-operator/deploy/crds/nuodb_v2alpha1_nuodb_crd.yaml
kubectl delete -f nuodb-operator/deploy/crds/nuodb_v2alpha1_nuodbinsightsserver_crd.yaml
kubectl delete -f nuodb-operator/deploy/crds/nuodb_v2alpha1_nuodbycsbwl_crd.yaml
kubectl delete -f nuodb-csv.yaml
kubectl delete clusterrolebinding nuodb-op-admin
# For OPENSHIFT only, delete the thp security context constraint
kubectl delete scc thp-scc
kubectl delete namespace $OPERATOR_NAMESPACE
Verify cleanup
kubectl get grafanas
kubectl get grafanadatasources
kubectl get grafanadashboards
kubectl get kibanas
kubectl get elasticsearches
kubectl get sa/grafana-operator
kubectl get secrets | grep grafana-operator
kubectl get rolebindings/grafana-operator
kubectl get role/grafana-operator
storageMode - Run NuoDB using persistent durable storage "persistent" or volatile storage "ephemeral". Must be set to one of those two values.
storageMode: persistent
insightsEnabled - Use to control NuoDB Insights Opt In. NuoDB Insights provides database monitoring and visualization. Set to "true" to activate or "false" to deactivate.
insightsEnabled: false
adminCount - Number of admin service pods. Requires 1 cluster node available for each Admin Service
adminCount: 3
adminStorageSize - Admin service log volume size (GB)
adminStorageSize: 5G
adminStorageClass - Admin persistent storage class name
adminStorageClass: glusterfs-storage
apiServer - Load balancer service URL. hostname:port (or LB address) for nuocmd and nuodocker process to connect to.
apiServer: https://domain:8888
storageMode - Run NuoDB using persistent durable storage "persistent" or volatile storage "ephemeral". Must be set to one of those two values.
storageMode: persistent
dbName - NuoDB Database name. must consist of lowercase alphanumeric characters '[a-z0-9]+'
dbName: test
dbUser - Name of Database user
dbUser: dba
dbPassword - Database password
dbPassword: secret
smMemory - SM memory (in GB)
smMemory: 2Gi
smCpu - SM CPU cores to request
smCpu: "1"
smStorageSize - Storage manager (SM) volume size (GB)
smStorageSize: 20G
smStorageClass - SM persistent storage class name
smStorageClass: local-disk
engineOptions - Additional "nuodb" engine options Format: …
engineOptions: ""
teCount - Number of transaction engines (TE) nodes. Limit is 3 in CE version of NuoDB
teCount: 1
teMemory - TE memory (in GB)
teMemory: 2Gi
teCpu - TE CPU cores to request
teCpu: "1"
container - NuoDB fully qualified image name (FQIN) for the Docker image to use
Below are examples that pull the NuoDB container image from Red Hat (RHCC), Google Cloud Platform Marketplace, AWS Marketplace, and DockerHub.
container: registry.connect.redhat.com/nuodb/nuodb-ce:latest
container: marketplace.gcr.io/nuodb/nuodb:latest
container: 117940112483.dkr.ecr.us-east-1.amazonaws.com/d893f8e5-fe12-4e43-b792-8cb98ffc11c0/cg-756769224/docker.io/nuodb/nuodb-ce:2.0.3-1-latest
container: nuodb/nuodb-ce:latest
elasticVersion - Version of ElasticSearch
elasticVersion: 7.3.0
elasticNodeCount - Number of nodes in the ElasticSearch Cluster
elasticNodeCount: 1
kibanaVersion - Version of Kibana
kibanaVersion: 7.3.0
kibanaNodeCount - Version of Kibana
kibanaNodeCount: 1
storageClass - Kubernetes Persistent Storage Class
storageClass: ""
ycsbLoadName - YCSB workload pod name
ycsbLoadName: ycsb-load
ycsbWorkload - Sample SQL activity workload. Valid values are a-f. Each letter determines a different mix of read and update workload percentage generated. a= 50/50, b=95/5, c=100 read. Refer to YCSB documentation for more detail.
ycsbWorkload: b
ycsbLbPolicy - YCSB load-balancer policy. Name of an existing load-balancer policy, that has already been created using the 'nuocmd set load-balancer' command.
ycsbLbPolicy: ""
ycsbNoOfProcesses - Number of YCSB processes. Number of concurrent YCSB processes that will be started in each YCSB pod. Each YCSB process makes a connection to the Database.
ycsbNoOfProcesses: 2
ycsbNoOfRows - YCSB number of rows inserted into the ycsb table
ycsbNoOfRows: 10000
ycsbNoOfIterations - YCSB number of iterations. After the last iteration finishes the benchmark will stop. A value of 0 will run the benchmark continuously.
ycsbNoOfIterations: 0
ycsbOpsPerIteration - Number of YCSB SQL operations to perform in each iteration. This value controls the number of SQL operations performed in each benchmark iteration. Increasing this value increases the run-time of each iteration, and also reduces the frequency at which new connections are made during the sample workload run period.
ycsbOpsPerIteration: 10000
ycsbMaxDelay - YCSB maximum workload delay in milliseconds (Default is 4 minutes)
ycsbMaxDelay: 240000
ycsbDbSchema - YCSB Database schema. Default schema to use to resolve tables, views, etc.
ycsbDbSchema: User1
ycsbContainer - YCSB fully qualified image name (FQIN) for the ycsb docker image to use. See examples below pulling the image from dockerhub and the AWS Marketplace.
ycsbContainer: nuodb/ycsb:latest
ycsbContainer: 117940112483.dkr.ecr.us-east-1.amazonaws.com/d893f8e5-fe12-4e43-b792-8cb98ffc11c0/cg-756769224/docker.io/nuodb/ycsb:2.0.3-1-latest
The NuoDB Operator is built using the K8s Operator SDK. The CRDs (deploy/crds/*_crd.yaml, OpenAPI (zz_generated.openapi.go), and deep copy functions (zz_generated.deepcopy.go) are automatically generated using the Operator SDK commands:
operator-sdk generate k8s
operator-sdk generate openapi