| copyright | lastupdated | keywords | subcollection | ||
|---|---|---|---|---|---|
|
2024-01-05 |
containers, ha, high availability, failover, kubernetes, multi az, multi-az, szr, mzr |
containers |
{{site.data.keyword.attribute-definition-list}}
{: #ha_clusters}
Design your standard cluster for maximum availability and capacity for your app with {{site.data.keyword.containerlong}}. {: shortdesc}
Use the built-in features to make your cluster more highly available and to protect your app from downtime when a component in your cluster fails.
High availability (HA) is a core discipline in an IT infrastructure to keep your apps up and running, even after a partial or full site failure. The main purpose of high availability is to eliminate potential points of failures in an IT infrastructure. For example, you can prepare for the failure of one system by adding redundancy and setting up failover mechanisms. {: shortdesc}
What level of availability is best? : You can achieve high availability on different levels in your IT infrastructure and within different components of your cluster. The level of availability that is correct for you depends on several factors, such as your business requirements, the Service Level Agreements that you have with your customers, and the resources that you want to expend.
What level of availability does {{site.data.keyword.cloud_notm}} offer? : See How {{site.data.keyword.cloud_notm}} ensures high availability and disaster recovery.
The level of availability that you set up for your cluster impacts your coverage under the {{site.data.keyword.cloud_notm}} HA service level agreement terms. For example, to receive full HA coverage under the SLA terms, you must set up a multizone cluster with a total of at least 6 worker nodes, two worker nodes per zone that are evenly spread across three zones.
{: #fault_domains}
The {{site.data.keyword.containerlong_notm}} architecture and infrastructure is designed to ensure reliability, low processing latency, and a maximum uptime of the service. However, failures can happen. Depending on the service that you host in {{site.data.keyword.cloud_notm}}, you might not be able to tolerate failures, even if failures last for only a few minutes. {: shortdesc}
{{site.data.keyword.containerlong_notm}} provides several approaches to add more availability to your cluster by adding redundancy and anti-affinity. Review the following image to learn about potential points of failure and how to eliminate them.
{: caption="Figure 1. Overview of fault domains in a high availability cluster" caption-side="bottom"}
Potential failure point 1: Container or pod availability : Containers and pods are, by design, short-lived and can fail unexpectedly. For example, a container or pod might crash if an error occurs in your app. To make your app highly available, you must ensure that you have enough instances of your app to handle the workload plus additional instances in the case of a failure. Ideally, these instances are distributed across multiple worker nodes to protect your app from a worker node failure. For more information, see Deploying highly available apps.
Potential failure point 2: Worker node availability : A worker node is a VM that runs on physical hardware. Worker node failures include hardware outages, such as power, cooling, or networking, and issues on the VM itself. You can account for a worker node failure by setting up multiple worker nodes in your cluster. For more information, see Creating clusters with multiple worker nodes.
Worker nodes in one zone are not guaranteed to be on separate physical compute hosts. For example, you might have a cluster with three worker nodes, but all three worker nodes were created on the same physical compute host in the IBM zone. If this physical compute host goes down, all your worker nodes are down. To protect against this failure, you must set up a multizone cluster to spread worker nodes across three zones, or create multiple single zone clusters in different zones. {: note}
Potential failure point 3: Cluster availability : The Kubernetes master is the main component that keeps your cluster up and running. The master stores cluster resources and their configurations in the etcd database that serves as the single point of truth for your cluster. The Kubernetes API server is the main entry point for all cluster management requests from the worker nodes to the master, or when you want to interact with your cluster resources. To protect your cluster master from a zone failure, review the following docs. Create a cluster in a multizone location, which spreads the master across zones. Or, consider setting up a second cluster in another zone.
Potential failure point 4: Zone availability : A zone failure affects all physical compute hosts and NFS storage. Failures include power, cooling, networking, or storage outages, and natural disasters, like flooding, earthquakes, and hurricanes. To protect against a zone failure, you must have clusters in two different zones that are load balanced by an external load balancer.
Potential failure point 5: Region availability : Every region is set up with a highly available load balancer that is accessible from the region-specific API endpoint. The load balancer routes incoming and outgoing requests to clusters in the regional zones. The likelihood of a full regional failure is low. However, to account for this failure, you can set up multiple clusters in different regions and connect them by using an external load balancer. If an entire region fails, the cluster in the other region can take over the workload.
A multi-region cluster requires several Cloud resources, and depending on your app, can be complex and expensive. Check whether you need a multi-region setup or if you can accommodate a potential service disruption. If you want to set up a multi-region cluster, ensure that your app and the data can be hosted in another region, and that your app can handle global data replication. {: note}
Potential failure point 6: Storage availability : In a stateful app, data plays an important role to keep your app up and running. Make sure that your data is highly available so that you can recover from a potential failure. In {{site.data.keyword.containerlong_notm}}, you can choose from several options to persist your data. For example, you can provision NFS storage by using Kubernetes native persistent volumes, or store your data by using an {{site.data.keyword.cloud_notm}} database service. For more information, see Planning highly available data.
Your users are less likely to experience downtime when you distribute your apps across multiple worker nodes, zones, and clusters. Built-in capabilities, like load balancing and isolation, increase resiliency against potential failures with hosts, networks, or apps. Review these potential cluster setups that are ordered with increasing degrees of availability.
{: caption="Figure 1. High availability for clusters" caption-side="bottom"}
- A single zone cluster with multiple worker nodes in a worker pool.
- A multizone cluster that spreads worker nodes across zones within one region.
- Clusters with public network connectivity: Multiple clusters that are set up across zones or regions and that are connected via a global load balancer.
{: #single_zone}
To improve availability for your app and to allow failover for the case that one worker node is not available in your cluster, add additional worker nodes to your single zone cluster.
{: caption="Figure 1. High availability for clusters in a single zone" caption-side="bottom"}
When you add more worker nodes, app instances can be distributed across multiple worker nodes. If one worker node goes down, app instances on available worker nodes continue to run. Kubernetes automatically reschedules pods from unavailable worker nodes to ensure performance and capacity for your app. To ensure that your pods are evenly distributed across worker nodes, implement pod affinity{: external}.
If your cluster is created in a single zone city, the Kubernetes master of your classic cluster is highly available and includes replicas on separate physical hosts for your master API server, etcd, scheduler, and controller manager to protect against an outage such as during a master update. If your cluster resides in one of the multizone metro locations, the master is automatically deployed with three replicas and spread across the zones of the metro.
If your single zone cluster is created in one of the multizone metro locations, you can change your single zone cluster to a multizone cluster. In a multizone cluster, your workloads are distributed across worker nodes in different zones. If one zone is not available, your workloads continue to run in the remaining zones. If you prefer single zone clusters for simplified management, or if your cluster must reside in a specific single zone city that does not support multizone capabilities, you can create multiple clusters and connect them with a global load balancer.
{: #mz-clusters}
Create a multizone cluster to distribute your workloads across multiple worker nodes and zones, and protect against zone failures with hosts, networks, or apps. If resources in one zone go down, your cluster workloads continue to run in the other zones. {: shortdesc}
{: caption="Figure 1. High availability for multizone clusters" caption-side="bottom"}
In a multizone cluster, the worker nodes in your worker pools are replicated across multiple zones within one region. Multizone clusters are designed to evenly schedule pods across worker nodes and zones to assure availability and failure recovery. If worker nodes are not spread evenly across the zones or capacity is insufficient in one of the zones, the Kubernetes scheduler or {{site.data.keyword.redhat_openshift_notm}} controller might fail to schedule all requested pods. As a result, pods might go into a Pending state until enough capacity is available. If you want to change the default behavior to make Kubernetes scheduler or {{site.data.keyword.redhat_openshift_notm}} controller distribute pods across zones in a best effort distribution, use the preferredDuringSchedulingIgnoredDuringExecution pod affinity policy{: external}.
You can create a multizone cluster in one of the supported classic or VPC multizone locations only. {: note}
Distributing your workload across three zones ensures high availability for your app in case a zone becomes unavailable. You must have your worker nodes spread evenly across all three availability zones to meet the {{site.data.keyword.cloud_notm}} service level agreement (SLA) for HA configuration. This setup also makes your cluster more cost-efficient. Why is that, you ask? Here is an example.
Let's say you need a worker node with six cores to handle the workload for your app. To make your cluster more available, you have the following options:
- Duplicate your resources in another zone: This option leaves you with two worker nodes, each with six cores in each zone for a total of 12 cores.
- Distribute resources across three zones: With this option, you deploy two cores per zone, which leaves you with a total capacity of six cores. To handle your workload, two zones must be up at a time. If one zone is unavailable, the other zone can fully handle your workload. If two zones are unavailable, the two remaining cores are up to handle your parts of your workload, and you could temporarily add another worker node to that zone.
{: #mz-master-setup}
When you create a cluster in a multizone location, a highly available master is automatically deployed and three replicas are spread across the zones of the metro. For example, if the cluster is in dal10, dal12, or dal13 zones, the replicas of the master are spread across each zone in the Dallas multizone metro.
{: #mz-master-communication}
If you created a VPC multizone cluster, the subnets in each zone are automatically set up with Access Control Lists (ACLs) that allow communication between the master and the worker nodes across zones. In classic clusters, if you have multiple VLANs for your cluster, multiple subnets on the same VLAN, or a multizone classic cluster, you must enable a Virtual Router Function (VRF) for your IBM Cloud infrastructure account so your worker nodes can communicate with each other on the private network. To enable VRF, see Enabling VRF. To check whether a VRF is already enabled, use the ibmcloud account show command. If you can't or don't want to enable VRF, enable VLAN spanning. To perform this action, you need the Network > Manage Network VLAN Spanning infrastructure permission, or you can request the account owner to enable it. To check whether VLAN spanning is already enabled, use the ibmcloud ks vlan spanning get --region <region> command.
{: #convert-sz-to-mz}
To convert a single zone cluster to a multizone cluster, your cluster must be set up in one of the supported classic or VPC multizone locations. VPC clusters can be set up only in multizone metro locations, and as such can always be converted from a single zone cluster to a multizone cluster. Classic clusters that are set up in a single zone data center can't be converted to a multizone cluster. To convert a single zone cluster to a multizone cluster, see Adding worker nodes to Classic clusters or Adding worker nodes to VPC clusters.
{: #multizone-apps-faq}
It depends on how you set up the app. See Planning highly available deployments and Planning highly available persistent storage.
{: #multiple-clusters-glb}
To protect your app from a master failure or for classic clusters that must reside in one of the supported single zone locations, you can create multiple clusters in different zones within a region and connect them with a global load balancer. {: shortdesc}
To connect multiple clusters with a global load balancer, the clusters must be set up with public network connectivity. {: note}
{: caption="Figure 1. High availability for multiple clusters" caption-side="bottom"}
To balance your workload across multiple clusters, you must set up a global load balancer and add the public IP addresses of your ALBs or load balancer services to your domain. By adding these IP addresses, you can route incoming traffic between your clusters. For the global load balancer to detect if one of your clusters is unavailable, consider adding a ping-based health check to every IP address. When you set up this check, your DNS provider regularly pings the IP addresses that you added to your domain. If one IP address becomes unavailable, then traffic is not sent to this IP address anymore. However, Kubernetes does not automatically restart pods from the unavailable cluster on worker nodes in available clusters. If you want Kubernetes to automatically restart pods in available clusters, consider setting up a multizone cluster.
{: #multicluster-three-zones}
Similar to using 3 zones in multizone clusters, you can provide more availability to your app by setting up three clusters across zones. You can also reduce costs by purchasing smaller machines to handle your workload.
{: #multiple-regions-setup}
You can set up multiple clusters in different regions of one geolocation (such as US South and US East) or across geolocations (such as US South and EU Central). Both setups offer the same level of availability for your app, but also add complexity when it comes to data sharing and data replication. For most cases, staying within the same geolocation is sufficient. But if you have users across the world, it might be better to set up a cluster where your users are so that your users don't experience long waiting times when they send a request to your app.
{: #multiple-cluster-lb-options}
To load balance workloads across multiple clusters, you must make your apps available on the public network by using Application Load Balancers (ALBs) or Network Load Balancers (NLBs). The ALBs and NLBs are assigned a public IP address that you can use to access your apps.
To load balance workloads across your apps, add the public IP addresses of your ALBs and NLBs to a CIS global load balancer or your own global load balancer.
{: #cis-global-lb-setup}
- Set up Kubernetes load balancer services or the Ingress service to expose the apps in your cluster.
- Set up the CIS global load balancer by following steps 1 - 5 in Getting Started with {{site.data.keyword.cloud_notm}} Internet Services (CIS). These steps walk you through provisioning the service instance, adding your app domain, and configuring your name servers, and creating DNS records. Create a DNS record for each ALB or NLB IP address that you collected. These DNS records map your app domain to all your cluster ALBs or NLBs, and ensure that requests to your app domain are forwarded to your clusters in a round-robin cycle.
- Add health checks for the ALBs or NLBs. You can use the same health check for the ALBs or NLBs in all your clusters, or create specific health checks to use for specific clusters.
- Add an origin pool for each cluster by adding the cluster's ALB or NLB IPs. For example, if you have 3 clusters that each have two ALBs, create three origin pools that each have two ALB IP addresses. You can find the NLB or ALB IP addresses by running
kubectl get svc -n <namespace>. Add a health check to each origin pool that you create. - Add a global load balancer.
{: #byo-global-lb-setup}
- Set up Kubernetes load balancer services or set up the Ingress service to expose the apps in your cluster.
- Configure your domain to route incoming traffic to your ALB or NLB services by adding the IP addresses of all public enabled ALBs and NLB services to your domain. You can find the NLB or ALB IP addresses by running
kubectl get svc -n <namespace>. - For each IP address, enable a ping-based health check so that your DNS provider can detect unhealthy IP addresses. If an unhealthy IP address is detected, traffic is not routed to this IP address anymore.
{: #glb-private}
{{site.data.keyword.cloud_notm}} does not offer a global load balancer service on the private network. However, you can connect your cluster to a private load balancer that you host in your on-prem network by using one of the supported VPN options. Make sure to expose your apps on the private network by using Application Load Balancers (ALBs) or Network Load Balancers (NLBs), and use the private IP address in your VPN settings to connect your app to your on-prem network.