3. Synchronize your Clusters


In order to keep the Kubernetes resource between your paired source and destination cluster in sync, you need to periodically migrate them.

For reference,

  • Source Cluster is the Kubernetes cluster where your applications are running
  • Destination Cluster is the Kubernetes cluster where the applications will be failed over, in case of a disaster in the source cluster.

Scheduling migrations

You can schedule a migration through a schedule policy. In the next sections, we will walk you through how to spec, validate, and display a schedule policy. Then, we will use our new schedule policy to schedule a migration.

Schedule policies

You can use schedule policies to specify when a specific action needs to be triggered. Schedule policies do not contain any actions themselves. Also, they are not namespaced. Storage policies are similar to storage classes where an admin is expected to create schedule policies which are then consumed by other users.

There are 4 sections in a schedule Policy spec:

  • Interval: the interval in minutes after which the action should be triggered
  • Daily: the time at which the action should be triggered every day
  • Weekly: the day of the week and the time in that day when the action should be triggered
  • Monthly: the date of the month and the time on that date when the action should be triggered

Let’s look at an example of how we could spec a policy:

apiVersion: stork.libopenstorage.org/v1alpha1
kind: SchedulePolicy
metadata:
  name: testpolicy
  namespace: mysql
policy:
  interval:
    intervalMinutes: 1
  daily:
    time: "10:14PM"
  weekly:
    day: "Thursday"
    time: "10:13PM"
  monthly:
    date: 14
    time: "8:05PM"

Validation

The following validations rules are defined:

  • The times in the policy need to follow the time.Kitchen format, example 1:02PM or 1:02pm.
  • The date of the month should be greater than 0 and less than 31. If a date doesn’t exist in a month, it will roll over to the next month. For example, if the date is specified as Feb 31, it will trigger on either 2nd or 3rd March depending on if it is a leap year.
  • The weekday can be specified in either long or short format, ie either “Sunday” or “Sun” are valid days.

Displaying a policy

To display a policy, run storkctl get with the name of the policy as a parameter:

storkctl get schedulepolicy
NAME           INTERVAL-MINUTES   DAILY     WEEKLY             MONTHLY
testpolicy     1                  10:14PM   Thursday@10:13PM   14@8:05PM

Scheduling a migration

Once a policy has been created, you can use it to schedule a migration. The MigrationSchedule object is namespaced.

NOTE: If your cluster has a DR license applied to it, you can only perform migrations in DR mode; this includes operations involving the pxctl cluster migrate command.

Continuing our previous example with testpolicy, here is how to create a MigrationSchedule object that schedules a migration:

apiVersion: stork.libopenstorage.org/v1alpha1
kind: MigrationSchedule
metadata:
  name: mysqlmigrationschedule
  namespace: migrationnamespace
spec:
  template:
    spec:
      # This should be the name of the cluster pair created above
      clusterPair: remotecluster
      # If set to false this will migrate only the Portworx volumes. No PVCs, apps, etc will be migrated
      includeResources: true
      # If set to false, the deployments and stateful set replicas will be set to 0 on the destination.
      # If set to true, the deployments and stateful sets will start running once the migration is done
      # There will be an annotation with "stork.openstorage.org/migrationReplicas" on the destinationto store the replica count from the source.
      startApplications: false
       # If set to false, the volumes will not be migrated
      includeVolumes: false
      # List of namespaces to migrate
      namespaces:
      - migrationnamespace
  schedulePolicyName: testpolicy

A few things to note:

  • The option includeVolumes is set to false because the volumes are already present on the destination cluster as there is a single storage fabric
  • The option startApplications is set to false so that the applications do not start when the resources are migrated. This is because the applications on the destination cluster should start only when the application fails over.

If the policy name is missing or invalid there will be events logged against the schedule object. Success and failures of the migrations created by the schedule will also result in events being logged against the object. These events can be seen by running a kubectl describe on the object

The output of kubectl describe will also show the status of the migrations that were triggered for each of the policies along with the start and finish times. The statuses will be maintained for the last successful migration and any Failed or InProgress migrations for each policy type.

Let’s now run kubectl describe and see how the output would look like:

kubectl describe migrationschedules.stork.libopenstorage.org -n mysql
Name:         mysqlmigrationschedule

Namespace:    migrationnamespace
Labels:       <none>
Annotations:  kubectl.kubernetes.io/last-applied-configuration:
                {"apiVersion":"stork.libopenstorage.org/v1alpha1","kind":"MigrationSchedule","metadata":{"annotations":{},"name":"mysqlmigrationschedule",...
API Version:  stork.libopenstorage.org/v1alpha1
Kind:         MigrationSchedule
Metadata:
  Creation Timestamp:  2019-02-14T04:53:58Z
  Generation:          1
  Resource Version:    30206628
  Self Link:           /apis/stork.libopenstorage.org/v1alpha1/namespaces/mysql/migrationschedules/mysqlmigrationschedule
  UID:                 8a245c1d-3014-11e9-8d3e-0214683e8447
Spec:
  Schedule Policy Name:  daily
  Template:
    Spec:
      Cluster Pair:       remotecluster
      Include Resources:  true
      Include Volumes:    false
      Namespaces:
        mysql
      Post Exec Rule:
      Pre Exec Rule:
      Selectors:           <nil>
      Start Applications:  false
Status:
  Items:
    Daily:
      Creation Timestamp:  2019-02-14T22:16:51Z
      Finish Timestamp:    2019-02-14T22:19:51Z
      Name:                mysqlmigrationschedule-daily-2019-02-14-221651
      Status:              Successful
    Interval:
      Creation Timestamp:  2019-02-16T00:40:52Z
      Finish Timestamp:    2019-02-16T00:41:52Z
      Name:                mysqlmigrationschedule-interval-2019-02-16-004052
      Status:              Successful
      Creation Timestamp:  2019-02-16T00:41:52Z
      Finish Timestamp:    <nil>
      Name:                mysqlmigrationschedule-interval-2019-02-16-004152
      Status:              InProgress
    Monthly:
      Creation Timestamp:  2019-02-14T20:05:41Z
      Finish Timestamp:    2019-02-14T20:07:41Z
      Name:                mysqlmigrationschedule-monthly-2019-02-14-200541
      Status:              Successful
    Weekly:
      Creation Timestamp:  2019-02-14T22:13:51Z
      Finish Timestamp:    2019-02-14T22:16:51Z
      Name:                mysqlmigrationschedule-weekly-2019-02-14-221351
      Status:              Successful
Events:
  Type    Reason      Age                    From   Message
  ----    ------      ----                   ----   -------
  Normal  Successful  4m55s (x53 over 164m)  stork  (combined from similar events): Scheduled migration (mysqlmigrationschedule-interval-2019-02-16-003652) completed successfully

Each migration is associated with a Migrations object. To get the most important information, type:

kubectl get migration -n migrationnamespace
NAME AGE
mysqlmigrationschedule-daily-2019-02-14-221651 1d
mysqlmigrationschedule-interval-2019-02-16-004052 5m
mysqlmigrationschedule-interval-2019-02-16-004152 4m
mysqlmigrationschedule-monthly-2019-02-14-200541 1d
mysqlmigrationschedule-weekly-2019-02-14-221351 1d

Once the MigrationSchedule object is deleted, all the associated Migration objects should also be deleted as well.



Last edited: Wednesday, Oct 9, 2019