Skip to main content

Migration to Managed Kubernetes FAQ

How to migrate your Kubernetes cluster to Selectel?

Use the instructions:

Can I create a fault-tolerant cluster?

Yes, when creating a cluster, you can select the Fault-Tolerant or Multi-Zone Fault-Tolerant cluster type.

We recommend creating Multi-Zone Fault-Tolerant clusters because in such clusters master nodes are located in different segments of a multi-zone pool — the nodes will be hosted in different data centers that are geographically remote from each other. A single point of failure for a multi-zone pool is failures across all data centers where cluster nodes are hosted.

For more information, see the Cluster types subsection of the Managed Kubernetes product overview guide.

What to do if an error occurs when creating or modifying the configuration of a Managed Kubernetes cluster or node group?

Errors when creating or modifying the configuration of a Managed Kubernetes cluster or node groups can occur for various reasons, including insufficient resources in the location.

We recommend following these steps:

  1. Check the error description and follow the troubleshooting recommendations provided in the event log in the control panel: in the top menu, click Products ⟶ Managed Kubernetes ⟶ section Clusters ⟶ tab Event log → in the event row, click .

  2. If the cause of the error is not listed in the event log, check the up-to-date information on Managed Kubernetes operation and resource availability in locations in the control panel or on the status page.

  3. If there is no information about an incident or resource unavailability, try creating or modifying the cluster or node group configuration again. Try changing the configuration and replacing local disks with network disks.

  4. If there is information about resource unavailability, try creating a cluster in another location.

  5. If the previous steps did not help and you could not determine the cause on your own, create a ticket. In the ticket, specify:

    • UUID of the Managed Kubernetes cluster. You can copy it in the control panel: in the top menu, click Products → Managed Kubernetes → cluster page → copy the UUID under the cluster name, next to the region and pool;
    • Project ID. You can copy it in the control panel: in the top menu, click IAM → Projects → in the project row, in the ID column, click ;
    • pool where the cluster was created.

How do I delete a Managed Kubernetes cluster stuck in the ERROR or PENDING_CREATE status if deletion via the control panel or API does not work?

  1. If the cluster is in the ERROR status, check the error description and follow the troubleshooting recommendations provided in the event log in the control panel: in the top menu, click Products ⟶ Managed Kubernetes ⟶ section Clusters ⟶ tab Event log → in the event row, click .

  2. If the cause of the error is unknown or you still cannot delete the cluster via the control panel or API, create a ticket and specify in it:

    • UUID of the Managed Kubernetes cluster. You can copy it in the control panel: in the top menu, click Products → Managed Kubernetes → cluster page → copy the UUID under the cluster name, next to the region and pool;
    • Project ID. You can copy it in the control panel: in the top menu, click IAM → Projects → in the project row, in the ID column, click ;
    • pool where the cluster was created.

Why does a load balancer in Managed Kubernetes switch to ERROR status during creation, and how do I fix it?

Errors when creating a load balancer can occur for various reasons:

  • due to platform-side incidents (Octavia/Amphora failures);
  • insufficient resources to create instances (amphora);
  • or temporary overloads.

We recommend following these steps:

  1. Check the up-to-date information on service status and resource availability on the status page.

  2. If there is no information about an incident or resource unavailability, try creating it again. To keep the public IP address when recreating the load balancer, add the loadbalancer.openstack.org/keep-floatingip: "true" annotation to the manifest.

  3. If the creation error recurs, check for available space in the subnet and your load balancer quotas. If there are not enough quotas, increase them.

  4. If there is information about resource unavailability in a specific location or the creation error recurs in that location, try creating a load balancer in another location.

  5. If the previous steps did not help and you could not determine the cause on your own, create a ticket. In the ticket, specify:

    • Load balancer ID. You can copy it in the control panel: in the top menu, click Products → Cloud Servers → Load Balancers → load balancer page → under the load balancer name, click ;
    • Project ID. You can copy it in the control panel: in the top menu, click IAM → Projects → in the project row, in the ID column, click ;
    • pool where the load balancer was created.

How do I increase project quotas and limits?

Each project has quotas and limits set:

  • limits — the maximum amount of resources that can be created in a project. Limits are set automatically upon project creation and may vary across pools and pool segments;
  • quotas — constraints within limits used to control costs. When a project is created, quotas equal to the limits are allocated for each resource. Once resource usage reaches the quota value, you will not be able to create new resources.

If there are not enough quotas to create resources, you can change them yourself. The quota value cannot exceed the project limit.

If you need more quotas to create resources than defined by the project limit, increase the limits. To do this, create a ticket and specify in it:

  • Project ID. You can copy it in the control panel: in the top menu, click IAM → Projects → in the project row, in the ID column, click ;
  • pool where the limit needs to be increased;
  • the new limit value.

After a Selectel specialist increases the limit, change the quotas yourself in the control panel.

Are there availability guarantees for a Managed Kubernetes cluster?

Selectel guarantees the availability of Managed Kubernetes clusters only for High Availability and Multizone High Availability types. For these, the guaranteed availability is 99.98%. For more information, see Terms of Use of Specific Services: Kubernetes Cluster .

How is a Managed Kubernetes cluster backed up?

In a Managed Kubernetes cluster, you can use volume snapshots (VolumeSnapshots) — point-in-time copies of persistent volume (PersistentVolume) content — to back up data. You can create a snapshot and then restore a persistent volume from it.

Snapshots can also be used to transfer data to another resource without the need to create a new disk.

How do I work with Managed Kubernetes after migration?

For more information about how Managed Kubernetes works, its versions, and limitations, see the Managed Kubernetes product overview. Instructions on working with the product can be found in the Managed Kubernetes section.

We also recommend taking the Managed Kubernetes course Product Dive: Managed Kubernetes on the Selectel.Stack learning platform.