I've been trying to figure out what happens when the Kubernetes master fails in a cluster that only has one master. Do web requests still get routed to pods if this happens, or does the entire system just shut down?
According to the OpenShift 3 documentation, which is built on top of Kubernetes, (https://docs.openshift.com/enterprise/3.2/architecture/infrastructure_components/kubernetes_infrastructure.html), if a master fails, nodes continue to function properly, but the system looses its ability to manage pods. Is this the same for vanilla Kubernetes?
In typical setups, the master nodes run both the API and etcd and are either largely or fully responsible for managing the underlying cloud infrastructure. When they are offline or degraded, the API will be offline or degraded.
In the event that they, etcd, or the API are fully offline, the cluster ceases to be a cluster and is instead a bunch of ad-hoc nodes for this period. The cluster will not be able to respond to node failures, create new resources, move pods to new nodes, etc. Until both:
Enough etcd instances are back online to form a quorum and make progress (for a visual explanation of how this works and what these terms mean, see this page).
At least one API server can service requests
In a partially degraded state, the API server may be able to respond to requests that only read data.
However, in any case, life for applications will continue as normal unless nodes are rebooted, or there is a dramatic failure of some sort during this time, because TCP/ UDP services, load balancers, DNS, the dashboard, etc. Should all continue to function for at least some time. Eventually, these things will all fail on different timescales. In single master setups or complete API failure, DNS failure will probably happen first as caches expire (on the order of minutes, though the exact timing is configurable, see the coredns cache plugin documentation). This is a good reason to consider a multi-master setup–DNS and service routing can continue to function indefinitely in a degraded state, even if etcd can no longer make progress.
There are actions that you could take as an operator which would accelerate failures, especially in a fully degraded state. For instance, rebooting a node would cause DNS queries and in fact probably all pod and service networking functionality until at least one master comes back online. Restarting DNS pods or kube-proxy would also be bad.
If you'd like to test this out yourself, I recommend kubeadm-dind-cluster, kind or, for more exotic setups, kubeadm on VMs or bare metal. Note: kubectl proxy will not work during API failure, as that routes traffic through the master(s).
Kubernetes cluster without a master is like a company running without a Manager.
No one else can instruct the workers(k8s components) other than the Manager(master node)(even you, the owner of the cluster, can only instruct the Manager)
Everything works as usual. Until the work is finished or something stopped them.(because the master node died after assigning the works)
As there is no Manager to re-assign any work for them, the workers will wait and wait until the Manager comes back.
The best practice is to assign multiple managers(master) to your cluster.
Although your data plane and running applications does not immediately starts breaking but there are several scenarios where cluster admins will wish they had multi-master setup. Key to understanding the impact would be understanding which all components talk to master for what and how and more importantly when will they fail if master fails.
Although your application pods running on data plane will not get immediately impacted but imagine a very possible scenario - your traffic suddenly surged and your horizontal pod autoscaler kicked in. The autoscaling would not work as Metrics Server collects resource metrics from Kubelets and exposes them in Kubernetes apiserver through Metrics API for use by Horizontal Pod Autoscaler and vertical pod autoscaler ( but your API server is already dead).If your pod memory shoots up because of high load then it will eventually lead to getting killed by k8s OOM killer. If any of the pods die, then since controller manager and scheduler talks to API Server to watch for current state of pods so they too will fail. In short a new pod will not be scheduled and your application may stop responding.
One thing to highlight is that Kubernetes system components communicate only with the API server. They don’t
talk to each other directly and so their functionality themselves could fail I guess. Unavailable master plane can mean several things - failure of any or all of these components - API server,etcd, kube scheduler, controller manager or worst the entire node had crashed.
If API server is unavailable - no one can use kubectl as generally all commands talk to API server ( meaning you cannot connect to cluster, cannot login into any pods to check anything on container file system. You will not be able to see application logs unless you have any additional centralized log management system).
If etcd database failed or got corrupted - your entire cluster state data is gone and the admins would want to restore it from backups as early as possible.
In short - a failed single master control plane although may not immediately impact traffic serving capability but cannot be relied on for serving your traffic.
Related
Kubernetes tends to assume apps are small/lightweight/stateless microservices which can be stopped on one node and restarted on another node with no downtime.
We have a slow starting (20min) legacy (stateful) application which, once run as a set of pod should not be rescheduled without due cause. The reason being all user sessions will be killed and the users will have to login again. There is NO way to serialize the sessions and externalize them. We want 3 instances of the pod.
Can we tell k8s not to move a pod unless absolutely necessary (i.e. it dies)?
Additional information:
The app is a tomcat/java monolith
Assume for the sake of argument we would like to run it in Kubernetes
We do have a liveness test endpoint available
There is no benefit, if you tell k8s to use only one pod. That is not the "spirit" of k8s. In this case, it might be better to use a dedicated machine for your app.
But you can assign a pod to a special node - Assigning Pods to Nodes. The should be necessary only, when special hardware requirements are needed (e.g. the AI-microservice needs a GPU, which is only on node xy).
k8s don't restart your pod for fun. It will restart it, when there is a reason (node died, app died, ...) and I never noticed a "random reschedule" in a cluster. It is hard to say, without any further information (like deployment, logs, cluster) what exactly happened to you.
And for your comment: There are different types of recreation, one of them starts a fresh instance and will kill the old one, when the startup was successfully. Look here: Kubernetes deployment strategies
All points together:
Don't enforce a node to your app - k8s will "smart" select the node.
There are normally no planned reschedules in k8s.
k8s will recreate pods only, if there is a reason. Maybe your app didn't answer on the liveness-endpoint? Or someone/something deleting your pod?
I'm planning to deploy a WebRTC custom videoconference software (based on NodeJS, using websockets) with Kubernetes, but I have some doubts about scaling down this environment.
Actually, I'm planning to use cloud hosted Kubernetes (GKE, EKS, AKS or any) to be able to auto-scale nodes in the cluster to attend the demand increase and decrease. But, scaling up is not the problem, but it's about scaling down.
The cluster will scale down based on some CPU average usage metrics across the cluster, as I understand, and if it tries to remove some node, it will start to drain connections and stop receiving new connections, right? But now, imagine that there's a videoconference still running in this "pending deletion" node. There are two problems:
1 - Stopping the node before the videoconference finishes (it will drop the meeting)
2 - With the draining behaviour when it starts to scale down, it will stop receiving new connections, so if someone tries to join in this running video conference, it will receive a timeout, right?
So, which is the best strategy to scale down nodes for a video conference solution? Any ideas?
Thanks
I would say this is not a matter of resolving it on kubernetes level by some specific scaling strategy but rather application ability to handle such situations. It isn't even specific to kubernetes. Imagine that you deploy it directly on compute instances which are also subject to autoscale and you'll end up in exactly the same situation when the load decreases and one of the instances is removed from the set.
You should rather ask yourself if such application is suitable to be deployed as kubernetes workload. I can imagine that such videoconference session doesn't have to rely on the backend deployed on a single node only. You can even define some affinity or anti-affinity rules to prevent your Pods from being scheduled on the same node. So if the whole application cluster is still up and running (it's Pods are running on different nodes), eviction of a limited subset of Pods should not have a big impact.
You can actually face the same issue with any other application as vast majority of them base on some session which needs to be established between the client software and the server part. I would say it's application responsibility to be able to handle such scenarios. If some of the users unexpectedly loses the connection it should be possible to immediately redirect them to the running instance e.g. different Pod which is still able to accept new requests.
So basically if the application is designed to be highly available, scaling in (when we talk about horizontal scaling we actually talk about scaling in and scaling out) the underyling VMs, or more specifically kubernetes nodes, shouldn't affect it's high availability capabilities. From the other hand if it is not designed to be highly available, solution such as kubernetes probably won't help much.
There is no best strategy at your use case. When a cloud provider scales down, it is going to get one node randomly and kill it. It's not going to check whether this node has less resource consumption, so let's kill this one. It might end up killing the node with most pods running on it.
I would focus on how you want to schedule your pods. I would try to schedule them, if possible, on a node with running pods already (Pod inter-affinity), and would set up a Pod Disruption Budget to all Deployments/StatefulSets/etc (depending on how you want to run the pods). As a result it would only scale down when there are no pods running on a specific node, and it would kill that node, because on the other nodes there are pods; protected by a PDB.
What we're looking for is a way for an actuator health check to signal some intention like "I am limping but not dead. If there are X number of other pods claiming to be healthy, then you should restart me, otherwise, let me limp."
We have a rest service hosted in clustered Kubernetes containers that periodically call out to fetch fresh data from an external resource. Occasionally we have failures reaching those external resources, and sometimes, but not every time, a restart of the pod will resolve the issue.
The services can operate just fine on possibly stale data. Although we wouldn't want to continue operating on stale data, that's preferable to just going down entirely.
In the interim, we're planning on having a node unilaterally decide not to report any problems through actuator until X amount of time has passed since the last successful sync, but that really only delays the point at which all nodes would still report failure.
In Kubernetes you can use LivenessProbe and ReadinessProbe to let a controller to heal your service, but some situations is better handled with HTTP response codes or alternative degraded service.
LivenessPobe
Use a LivenessProbe to resolve a deadlock situation. When your pod does not respond on a LivenessProbe, it will be killed and a new pod will replace it.
ReadinessProbe
Use a ReadinessProbe when your pod is not prepared for serving requests, e.g. if your pod need to read some files or need to connect to an external service before serving requests.
Fault affecting all replicas
If you have a problem that all your replicas depends on, e.g. an external service is down, then you can not solve it by restarting your pods. You may use an OpsToogle or a circuit breaker in this situation and notifying other services that you are degraded or show a message about temporary error.
For your situations
If there are X number of other pods claiming to be healthy, then you should restart me, otherwise, let me limp.
You can not delegate that logic to Kubernetes. Your application need to understand each fault situation, e.g. if an error was a transient network error or if your error will affect all replicas.
I'm currently looking into triggering some 3D rendering from an AppEngine-based service.
The idea is that input data is submitted by an API client to this web service, which then invokes an internal Kubernetes GPU enabled application ("rendering backend") to do the hard work.
GPU-enabled clusters are relatively expensive ($$$), so I really want the cluster to be up and running on demand. I am trying to achieve that by setting the autoscaling minimum to 0 for the rendering backend.
The only pretty way of "triggering" a rendering task on such a cluster I could think of is via Pub/Sub Push. Basically, I need something like Cloud Tasks, but those seem to be aimed at long running tasks executed in AppEngine, not Kubernetes. Plus I like the way Pub/Sub decouples the web service from the rendering backend.
Google's Pub/Sub only allows pushing via HTTPS and only to a validated domain. It appears that Google is forcing me to completely "expose" my internal rendering backend by assigning a domain name to it, which feels ridiculous. I cannot just tell Pub/Sub to invoke http://loadbalancer.IP.address/handle_push.
This is making me doubt my architecture.
How would you go about building something like this on GCP?
From the GKE perspective:
You can have a cluster with a dedicated GPU-based nodepool and schedule your pods there using Taints and tolerations. Additionally, you can control the number of nodes in your nodepool using Autoscaling so that, you can use them only when your pods are to be scheduled/run.
Consider that this requires an additional default-non-GPU-based nodepool, where system pods are being run.
For triggering, as long as your default pool is running, you'd be able to deploy your application and the autoscaling should start automatically. For deploying from an App Engine application, you might want to consider talking to the Kubernetes API directly through a library.
Finally and considering the nature of your current goal (3D rendering), it might be best to use Kubernetes Jobs. With these, you can complete an sporadic computational load, allowing the nodepool to downsize once is finished.
Wrapping up, you can have a minimum cluster with a zero-sized GPU-based nodepool that will autoscale when a tainted job is requested to be run there, and once the workload is finished, it should automatically downscale. These actions can be triggered from GAE, using one of the client libraries.
If Kube proxy is down, the pods on a kubernetes node will not be able to communicate with the external world. Anything that Kubernetes does specially to guarantee the reliability of kube-proxy?
Similarly, how does Kubernetes guarantee reliability of kubelet?
It guarantees their reliability by:
Having multiple nodes: If one kubelet crashes, one node goes down. Similarly, every node runs a kube-proxy instance, which means losing one node means losing the kube-proxy instance on that node. Kubernetes is designed to handle node failures. And if you designed your app that is running on Kubernetes to be scalable, you will not be running it as single instance but rather as multiple instances - and kube-scheduler will distribute your workload across multiple nodes - which means your application will still be accessible.
Supporting a Highly-Available Setup: If you set up your Kubernetes cluster in High-Availability mode properly, there won't be one master node, but multiple. This means, you can even tolerate losing some master nodes. The managed Kubernetes offerings of the cloud providers are always highly-available.
These are the first 2 things that come to my mind. However, this is a broad question, so I can go into details if you elaborate what you mean by "reliability" a bit.