Cohesity Cloud Scale Technology Manual Deployment Guide for Kubernetes Clusters

Last Published:
Product(s): NetBackup (11.1.0.2)
  1. Introduction
    1. About Cloud Scale deployment
      1.  
        Decoupling of NetBackup web services from primary server
      2.  
        Decoupling of NetBackup Policy and Job Management from primary server
      3.  
        Decoupling of NetBackup Database Manager from primary server
      4.  
        Logging feature (fluentbit) in Cloud Scale
    2.  
      About NetBackup Snapshot Manager
    3.  
      Required terminology
    4.  
      User roles and permissions
  2. Section I. Configurations
    1. Prerequisites
      1.  
        Preparing the environment for NetBackup installation on Kubernetes cluster
      2.  
        Prerequisites for Snapshot Manager (AKS/EKS)
      3. Prerequisites for Kubernetes cluster configuration
        1.  
          Config-Checker utility
        2.  
          Data-Migration for AKS
        3.  
          Webhooks validation for EKS
      4. Prerequisites for Cloud Scale configuration
        1.  
          Cluster specific settings
        2.  
          Cloud specific settings
      5.  
        Prerequisites for deploying environment operators
      6.  
        Prerequisites for using private registry
    2. Recommendations and Limitations
      1.  
        Recommendations of NetBackup deployment on Kubernetes cluster
      2.  
        Limitations of NetBackup deployment on Kubernetes cluster
      3.  
        Recommendations and limitations for Cloud Scale deployment
    3. Configurations
      1.  
        Contents of the TAR file
      2.  
        Configuring the cloudscale-values.yaml file
      3.  
        Configuring an External Certificate Authority for Web UI port 443
      4. Loading docker images
        1.  
          Installing the docker images for NetBackup
        2.  
          Installing the docker images for Snapshot Manager
        3.  
          Installing the docker images and binaries for MSDP Scaleout
      5. Configuring NetBackup
        1. Primary and media server CR
          1.  
            After installing primary server CR
          2.  
            After Installing the media server CR
        2.  
          Elastic media server
    4. Configuration of key parameters in Cloud Scale deployments
      1.  
        Tuning touch files
      2.  
        Setting maximum jobs per client
      3.  
        Setting maximum jobs per media server
      4.  
        Enabling intelligent catalog archiving
      5.  
        Enabling security settings
      6.  
        Configuring email server
      7.  
        Reducing catalog storage management
      8.  
        Configuring zone redundancy
      9.  
        Enabling client-side deduplication capabilities
      10.  
        Parameters for logging (fluentbit)
      11.  
        Managing media server configurations in Web UI
  3. Section II. Deployment
    1. Deploying Cloud Scale
      1.  
        How to deploy Cloud Scale
      2.  
        Prerequisites for Cloud Scale deployment
      3.  
        Deploying the operators
      4. Deploying Cloud Scale using Helm chart
        1.  
          Installing Cloud Scale environment
        2. Single node Cloud Scale Technology deployment
          1.  
            Steps to deploy Cloud Scale in single node
      5.  
        Deploying Cloud Scale using kubectl plugin
      6.  
        Verifying Cloud Scale deployment
      7. Post Cloud Scale deployment tasks
        1.  
          Restarting Cloud Scale Technology services
  4. Section III. Monitoring and Management
    1. Monitoring NetBackup
      1.  
        Monitoring the application health
      2.  
        Telemetry reporting
      3.  
        About NetBackup operator logs
      4.  
        Monitoring Primary/Media server CRs
      5.  
        Expanding storage volumes
      6. Allocating static PV for Primary and Media pods
        1.  
          Expanding log volumes for primary pods
        2.  
          Recommendation for media server volume expansion
        3.  
          (AKS-specific) Allocating static PV for Primary and Media pods
        4.  
          (EKS-specific) Allocating static PV for Primary and Media pods
    2. Monitoring Snapshot Manager
      1.  
        Overview
      2.  
        Configuration parameters
      3.  
        Snapshot Manager manual certificate renewal in Cloud Scale
    3. Monitoring fluentbit
      1.  
        Monitoring fluentbit for logging
    4. Monitoring MSDP Scaleout
      1.  
        About MSDP Scaleout status and events
      2.  
        Monitoring with Amazon CloudWatch
      3.  
        Monitoring with Azure Container insights
      4.  
        The Kubernetes resources for MSDP Scaleout and MSDP operator
    5. Managing NetBackup
      1.  
        Managing NetBackup deployment using VxUpdate
      2.  
        Updating the Primary/Media server CRs
      3.  
        Migrating the cloud node for primary or media servers
      4.  
        Migrating cpServer controlPlane node
    6. Managing the Load Balancer service
      1.  
        About the Load Balancer service
      2.  
        Notes for Load Balancer service
      3.  
        Opening the ports from the Load Balancer service
      4.  
        Steps for upgrading Cloud Scale from multiple media load balancer to none
    7. Managing PostrgreSQL DBaaS
      1.  
        Changing database server password in DBaaS
      2.  
        Updating database certificate in DBaaS
    8. Managing logging
      1.  
        Viewing NetBackup logs
      2.  
        Extracting NetBackup logs
    9. Performing catalog backup and recovery
      1.  
        Backing up a catalog
      2. Restoring a catalog
        1.  
          Primary server corrupted
        2.  
          MSDP-X corrupted
        3.  
          MSDP-X and Primary server corrupted
  5. Section IV. Maintenance
    1. PostgreSQL DBaaS Maintenance
      1.  
        Configuring maintenance window for PostgreSQL database in AWS
      2.  
        Setting up alarms for PostgreSQL DBaaS instance
    2. Patching mechanism for primary, media servers, fluentbit pods, and postgres pods
      1.  
        Overview
      2.  
        Patching of primary containers
      3.  
        Patching of media containers
      4.  
        Patching of fluentbit collector pods
      5.  
        Update containerized PostgreSQL pod
    3. Upgrading
      1. Upgrading Cloud Scale Technology
        1.  
          Prerequisites for Cloud Scale Technology upgrade
        2.  
          Upgrade the cluster
        3. Upgrade Cloud Scale
          1.  
            Upgrade Cloud Scale using the kubectl plugin
          2.  
            Manual upgrade of Cloud Scale using the Superchart
        4.  
          Configuring NetBackup IT Analytics for NetBackup deployment
    4. Cloud Scale Disaster Recovery
      1.  
        Cluster backup
      2.  
        Environment backup
      3.  
        Cluster recovery
      4.  
        Cloud Scale recovery
      5.  
        Environment Disaster Recovery
      6.  
        DBaaS Disaster Recovery
    5. Uninstalling
      1.  
        Uninstalling Cloud Scale Technology
    6. Troubleshooting
      1. Troubleshooting AKS and EKS issues
        1.  
          View the list of operator resources
        2.  
          View the list of product resources
        3.  
          View operator logs
        4.  
          View primary logs
        5.  
          Socket connection failure
        6.  
          Resolving an issue where external IP address is not assigned to a NetBackup server's load balancer services
        7.  
          Resolving the issue where the NetBackup server pod is not scheduled for long time
        8.  
          Resolving an issue where the Storage class does not exist
        9.  
          Resolving an issue where the primary server or media server deployment does not proceed
        10.  
          Resolving an issue of failed probes
        11.  
          Resolving issues when media server PVs are deleted
        12.  
          Resolving an issue related to insufficient storage
        13.  
          Resolving an issue related to invalid nodepool
        14.  
          Resolve an issue related to KMS database
        15.  
          Resolve an issue related to pulling an image from the container registry
        16.  
          Resolving an issue related to recovery of data
        17.  
          Check primary server status
        18.  
          Pod status field shows as pending
        19.  
          Ensure that the container is running the patched image
        20.  
          Getting EEB information from an image, a running container, or persistent data
        21.  
          Resolving the certificate error issue in NetBackup operator pod logs
        22.  
          Pod restart failure due to liveness probe time-out
        23.  
          NetBackup messaging queue broker take more time to start
        24.  
          Host mapping conflict in NetBackup
        25.  
          Issue with capacity licensing reporting which takes longer time
        26.  
          Local connection is getting treated as insecure connection
        27.  
          Backing up data from Primary server's /mnt/nbdata/ directory fails with primary server as a client
        28.  
          Storage server not supporting Instant Access capability on Web UI after upgrading NetBackup
        29.  
          Taint, Toleration, and Node affinity related issues in cpServer
        30.  
          Operations performed on cpServer in cloudscale-values.yaml file are not reflected
        31.  
          Elastic media server related issues
        32.  
          Failed to register Snapshot Manager with NetBackup
        33.  
          Post Kubernetes cluster restart, flexsnap-listener pod went into CrashLoopBackoff state or pods were unable to connect to flexsnap-rabbitmq
        34.  
          Post Kubernetes cluster restart, issues observed in case of containerized Postgres deployment
        35.  
          Request router logs
        36.  
          Issues with NBPEM/NBJM
        37.  
          Issues with logging feature for Cloud Scale
        38.  
          The flexsnap-listener pod is unable to communicate with RabbitMQ
        39.  
          Job remains in queue for long time
        40.  
          Extracting logs if the nbwsapp or log-viewer pods are down
        41.  
          Helm installation failed with bundle error
        42.  
          Deployment fails with private container registry and Postgres fails to pull the images
      2. Troubleshooting AKS-specific issues
        1.  
          Data migration unsuccessful even after changing the storage class through the storage yaml file
        2.  
          Host validation failed on the target host
        3.  
          Primary pod goes in non-ready state
      3. Troubleshooting EKS-specific issues
        1.  
          Resolving the primary server connection issue
        2.  
          NetBackup Snapshot Manager deployment on EKS fails
        3.  
          Wrong EFS ID is provided in cloudscale-values.yaml file
        4.  
          Primary pod is in ContainerCreating state
        5.  
          Webhook displays an error for PV not found
        6.  
          Cluster Autoscaler initialization issue
        7.  
          Catalog backup job fails with an error (Status 9202)
      4.  
        Troubleshooting issue for bootstrapper pod
      5.  
        Troubleshooting issues for kubectl plugin
  6. Appendix A. CR template
    1.  
      Secret
    2. MSDP Scaleout CR
      1.  
        MSDP Scaleout CR template for AKS
      2.  
        MSDP Scaleout CR template for EKS
  7. Appendix B. MSDP Scaleout
    1.  
      About MSDP Scaleout
    2.  
      Prerequisites for MSDP Scaleout (AKS\EKS)
    3.  
      Limitations in MSDP Scaleout
    4. MSDP Scaleout configuration
      1.  
        Initializing the MSDP operator
      2.  
        Configuring MSDP Scaleout
      3.  
        Configuring the MSDP cloud in MSDP Scaleout
      4.  
        Using MSDP Scaleout as a single storage pool in NetBackup
      5.  
        Using S3 service in MSDP Scaleout
      6.  
        Enabling MSDP S3 service after MSDP Scaleout is deployed
    5.  
      Installing the docker images and binaries for MSDP Scaleout (without environment operators or Helm charts)
    6.  
      Deploying MSDP Scaleout
    7. Managing MSDP Scaleout
      1.  
        Adding MSDP engines
      2.  
        Adding data volumes
      3. Expanding existing data or catalog volumes
        1.  
          Manual storage expansion
      4.  
        MSDP Scaleout scaling recommendations
      5. MSDP Cloud backup and disaster recovery
        1.  
          About the reserved storage space
        2. Cloud LSU disaster recovery
          1.  
            Recovering MSDP S3 IAM configurations from cloud LSU
      6.  
        MSDP multi-domain support
      7.  
        Configuring Auto Image Replication
      8. About MSDP Scaleout logging and troubleshooting
        1.  
          Collecting the logs and the inspection information
    8. MSDP Scaleout maintenance
      1.  
        Pausing the MSDP Scaleout operator for maintenance
      2.  
        Logging in to the pods
      3.  
        Reinstalling MSDP Scaleout operator
      4.  
        Migrating the MSDP Scaleout to another node pool

Environment backup

  1. Note down the MSDP operator Namespace, NodeSelector, StorageClassName, Tolerations and Image tag as follows:

    Obtain the name of the msdp operator statefulset using the following command:

    kubectl get statefulset -n <msdp-operator-system-namespace>

    Use the following command to backup MSDP operator Image tag, Tolerations, and NodeSelector:

    kubectl get sts <msdp-operator-statefulset-name> -n <msdp-operator-sample-namespace> -o=jsonpath='{"Namespace :"}{$.metadata.namespace}{$"\nImage :"}{$.spec.template.spec.containers[0].image}{$"\nNodeSelector :"}{$.spec.template.spec.nodeSelector}{$"\nTolerations :"}{$.spec.template.spec.tolerations[2]}{$"\nStorageClassName :"}{$.spec.volumeClaimTemplates[0].spec.storageClassName}{$"\n"}'

    From the output, note down the Image tag, StorageClassName, Tolerations and NodeSelector:

    Sample Output:
    Namespace :msdp-operator-system
    Image :nbuk8sreg.azurecr.io/msdp-operator:21.0
    NodeSelector :{"agentpool":"nbuxpool"}
    Tolerations :{"key":"agentpool","operator":"Equal","value":"nbuxpool"}
    StorageClassName :nb-disk-premium

    If toleration is not provided for msdp operator, then use the following command:

    kubectl get sts <msdp-operator-statefulset-name> -n <msdp-operator-sample-namespace> -o=jsonpath='{"Namespace :"}{$.metadata.namespace}{$"\nImage :"}{$.spec.template.spec.containers[0].image}{$"\nNodeSelector :"}{$.spec.template.spec.nodeSelector}{$"\nStorageClassName :"}{$.spec.volumeClaimTemplates[0].spec.storageClassName}{$"\n"}'

    Sample Output:
    Namespace :msdp-operator-system
    Image :nbuk8sreg.azurecr.io/msdp-operator:21.0
    NodeSelector :{"agentpool":"nbuxpool"}
    StorageClassName :nb-disk-premium
  2. Backup the above msdp-operator storageClass using the following command:

    kubectl get sc <msdp-operator-storageclass-name> -o yaml > msdpopstorageclass_backup.yaml

  3. Note down the NetBackup operator Namespace, NodeSelector, Tolerations and Image tag as follows:

    Obtain the name of the NetBackup operator deployment using the following command:

    kubectl get deployment -n <netbackup-operator-system-namespace>

    Use the following command to backup NetBackup operator Image tag, Tolerations, and NodeSelector:

    kubectl get deployment <netbackup-operator-deployment-name> -n <netbackup-operator-system-namespace> -o=jsonpath='{"Namespace :"}{$.metadata.namespace}{$"\nImage :"}{$.spec.template.spec.containers[0].image}{$"\nNodeSelector :"}{$.spec.template.spec.nodeSelector}{$"\nTolerations: "}{$.spec.template.spec.tolerations}{$"\n"}'

    From the output, note down the Image tag, Tolerations and NodeSelector:

    Sample Output:
    Namespace :netbackup-operator-system
    Image :nbuk8sreg.azurecr.io/netbackup/operator:11.1.x.x.xxxx
    NodeSelector :{"agentpool":"agentpool"}
    Tolerations: [{"key":"agentpool","operator":"Equal","value":"agentpool"}]
  4. Note down the flexsnap-operator Namespace, NodeSelector, Tolerations and Image tag as follows:

    Obtain the name of the flexsnap-operator deployment using the following command:

    kubectl get deployment -n <netbackup-operator-system-namespace>

    Use the following command to backup flexsnap operator Image tag, Tolerations, and NodeSelector:

    kubectl get deployment <flexsnap-operator-deployment-name> -n <netbackup-operator-system-namespace> -o=jsonpath='{"Namespace :"}{$.metadata.namespace}{$"\nImage :"}{$.spec.template.spec.containers[0].image}{$"\nNodeSelector :"}{$.spec.template.spec.affinity.nodeAffinity.requiredDuringSchedulingIgnoredDuringExecution.nodeSelectorTerms[0].matchExpressions[0]}{$"\nTolerations :"}{$.spec.template.spec.tolerations}{$"\n"}' 

    From the output, note down the Image tag, Tolerations and NodeSelector:

    Sample Output:
    Namespace :netbackup-operator-system
    Image :nbuk8sreg.azurecr.io/veritas/flexsnap-deploy:11.1.x.x.xxxx
    NodeSelector :{"key":"agentpool","operator":"In","values":["agentpool"]}
    Tolerations :[{"effect":"NoSchedule","key":"agentpool","operator":"Equal","value":"agentpool"}]
  5. (For DBaaS) Note the FQDN of the Postgres server created.

  6. (Applicable only if unified container is created) Note the Postgres unified container image tag, containerPort:

    kubectl get statefulset.apps/nb-postgresql -n <sample-namespace> -o=jsonpath='{$"\nImage :"}{$.spec.template.spec.containers[0].image}{$"\ncontainerPort :"}{$.spec.template.spec.containers[0].ports[0].containerPort}{$"\n"}'

    Sample output:

    Image :cpautomation.azurecr.io/netbackup/postgresql:16.10.1.0-0001-DR1
     
    containerPort :13787
  7. Obtain the fluentbit image tags and nodeselector using the following command:

    kubectl get deployment.apps/nb-fluentbit-collector -n netbackup -o=jsonpath='{$"\nImage :"}{$.spec.template.spec.containers[0].image}{$"\nImage2 :"}{$.spec.template.spec.containers[1].image}{"\n"}'

    Sample output:

    Image1:cpautomation.azurecr.io/netbackup/fluentbit:11.1.x-xxxx
    Image2:cpautomation.azurecr.io/netbackup/fluentbit-log-cleanup:11.1.x-xxxx
    • Take backup of operator-values.yaml file using the following command:

      helm get values operators -n netbackup-operator-system > operator-values.yaml

      Or

      Save the operator-values.yaml file.

    • Take backup of cloudscale-values.yaml file using the following command:

      helm get values cloudscale -n netbackup > cloudscale-values.yaml

      Or

      Save the cloudscale-values.yaml file from kubectl-plugin.

  8. Note down the values of spec, cpServer, storage, log, storageClassName from the cloudscale-values.yaml file using the following command:

    kubectl get sc nb-file-premium -o yaml > CPServerLog_storageclass_backup.yaml

  9. Search the storageClassName in cloudscale-values.yaml file using the kubectl get sc nb-disk-standardssd -o yaml > storageclass_backup.yaml command and provide these storageclasses name in the following command:

    For example, nb-disk-standardssd

    kubectl get sc <storageclass-name1> <storageclass-name2> <storageclass-name3> -o yaml > storageclass_backup.yaml

  10. Note down and save the required values (names) of the secrets obtained from cloudscale-values.yaml file in the above step:

    For example, credSecretName: primary-credential-secret

    Save the secrets yaml file as follows:

    kubectl get secret <secret-name1> <secret-name2> <secret-name3> -n <sample-namespace> -o yaml > secret_backup.yaml

    For example, kubectl get secret primary-credential-secret kms-secret example-key-secret -n example-ns -o yaml > secret_backup.yaml

    Note:

    (For DBaaS) The primary-credential-secret and kms-secret key values will not be there in cloudscale-values.yaml file that have been backed up. Use the helm command to get values from the above step. By default helm would be using the values provided during the deployment.

  11. Save the secrets named as Msdp credential and drInfoSecret during creation. As the operator would delete these secrets after using it.

  12. (For DBaaS) Note the password changed during DBaaS cluster deployment:

    • (For Azure) To check the OLD_ DBADMINPASSWORD execute the following command after executing into the primary pod:

      [root@user-primary-0 mnt]# ls -a
      .  ..  atdata  .db-cert  nbdata  nbdb  .nb-kmsdb  nblogs  .nb-pgdb  nbu-primary-env  .nb-user  .passphrase-secret  .password-secret  podemptydir  .token-secret
      [root@user-primary-0 mnt]# cd .nb-pgdb/
      [root@user-primary-0 .nb-pgdb]# ls
      dbadminlogin  dbadminpassword  dbport  dbserver  pgbouncerport
      [root@user-primary-0 .nb-pgdb]# cat dbadminlogin
      dbadminlogin
      [root@user-primary-0 .nb-pgdb]# cat dbadminpassword
      5MrJrGaLSnDKxTJ0UiPvotqmbqSzqU5a
      [root@user-primary-0 .nb-pgdb]# cat dbport
      5432
      [root@user-primary-0 .nb-pgdb]# cat dbserver
      user-postgres.postgres.database.azure.com
      [root@user-primary-0 .nb-pgdb]# cat pgbouncerport
      6432
      DBADMINPASSWORD
      DBADMINPASSWORD DBADMINPASSWORD
    • (For EKS) Login to AWS UI, navigate to Secrets Manager and find adminSecret. Naming convention for admin secrets are as follows:

      admin-secret-<use cluster name remove prefix eks->

  13. Note the values (names) of the secretProviderClass.

    For example, dbSecretProviderClass: db-secret-provider-class

    Save the secretProviderClass.yaml file using the following command:

    kubectl get secretproviderclass <secretproviderclass-name> -n <sample-namespace> -o yaml> secretproviderclass_backup.yaml

    Note:

    The dbSecretProviderClass is an optional field. If it is not present in the cloudscale-values.yaml file, then skip this step.

  14. Note the required values (names)and sSave internal configmap yaml using the following command:

    kubectl get configmap nbu-media-autoscaler-configmap flexsnap-conf nbuconf cs-config -n <sample-namespace> -o yaml > internalconfigmap_backup.yaml

    Note:

    The nbu-media-autoscaler-configmap is an optional internal configmap. If it is not present in namespace, then remove nbu-media-autoscaler-configmap from the above command.

  15. Take a note of cert-manager and trust-manager as the same versions of cert-manager and trust-manager should be used when deploying them during recovery.

  16. Note the details of cloud STU used for MSDP storage, such as name of bucket, volume, credential and the respective details added through Credential management in UI.

  17. (Applicable only for DBaaS based deployment environment) Snapshot Manager backup steps:

    For AKS

    • Search the disk (PV) to which psql pvc is attached in Azure cloud portal and click on Create snapshot in the different resource group other than the cluster infra resource group and note down this resource group. Wait for the resource to be available.

      Note:

      Snapshot must be created in resource group in different availability zone to take care of the recovery in case of zone failures/corrupted.

      Save the pgsql-pv.yaml file:

      kubectl get pv | grep psql-pvc

      pvc-079b631e-a905-4586-80b5-46acc7011669 30Gi RWO Retain Bound nbu/psql-pvc managed-csi-hdd 3h10m

      kubectl describe pv <PV which is bound to psql-pvc> > pgsql-pv.yaml

      For example, kubectl describe pv pvc-079b631e-a905-4586-80b5-46acc7011669 > pgsql-pv.yaml

    • Note down the snapshot id, which would be used to create a disk from snapshot during recovery.

      Note:

      Disk Snapshot must be taken after every plugin addition as the latest database is required to recover all the plugins during Database recovery.

    For EKS

    • Describe the PV attached to psql-pvc and save the VolumeID (for example, vol-xxxxxxxxxxxxxxx), storage class name and availability zone (AZ) from the output of following command:

      kubectl get pv | grep psql-pvc

      pvc-079b631e-a905-4586-80b5-46acc7011669 30Gi RWO Retain Bound nbu/psql-pvc managed-csi-hdd 3h10m

      kubectl describe pv <PV which is bound to psql-pvc> > pgsql-pv.yaml

      For example, kubectl describe pv pvc-079b631e-a905-4586-80b5-46acc7011669 > pgsql-pv.yaml

    • Search above VolumeID in the EC2 management console > Elastic Block Store > Volumes in AWS cloud portal.

    • Create snapshot (expand the Actions drop down) from the volume and wait for the completion. Note down the snapshot id (for example, snap-xxxxxxxxxxxx)

      Note:

      Disk Snapshot must be taken after every plugin addition as the latest database is required to recover all the plugins during Database recovery.

  18. Take the backup of catalog policy at /mnt/nbdata/DrPackages on the master server.

Note:

For manual deployment using Helm charts, ensure that you save the operators-values.yaml and cloudscale-values.yaml files. These files are used at the time of recovery.