Where Are TLS Handshake Errors In APISERVER Logs Coming From? - kubernetes

I have a cluster provisioned using KubeSpray on AWS. It has two bastions, one controller, one worker, and one etcd server.
I am seeing endless messages in the APISERVER logs:
http: TLS handshake error from 10.250.227.53:47302: EOF
They come from two IP addresses, 10.250.227.53 and 10.250.250.158. The port numbers change every time.
None of the cluster nodes correspond to those two IP addresses. The subnet cidr ranges are shown below.
The cluster seems stable. This behavior does not seem to have any negative affect. But I don't like having random HTTPS requests.
How can I debug this issue?

They're from the health check configured on the AWS ELB; you can stop those messages by changing the health check configuration to be HTTPS:6443/healthz instead of the likely TCP one it is using now
How can I debug this issue?
Aside from just generally being cognizant of how your cluster was installed, and then observing that those connections come at regular intervals, I would further bet that those two IP addresses belong to the two ENIs that are allocated to the ELB in each public subnet (they'll show up in the Network Interfaces list on the console as "owner: elasticloadbalancer" or something similar)

Related

How to automatically update the Service `spec.externalIPs` when a Kubernetes worker is drained/down?

I'm hosting a Kubernetes cluster on VMs/VPS from a random cloud provider not providing any Kubernetes things at all, meaning with a dedicated public IP address and to allow the trafic coming to the worker nodes, I'm defining my Service with the spec.externalIPs with the fixed list of IP addresses.
I'm looking for a way to get that list updated when a node is drained/down automatically.
I had a look at the existing operators from https://operatorhub.io/ but I haven't found any that seem to cover my use case.
The idea would that when the event of a node passing to NotReady is emitted, the Service is updated with the Nodes being Ready.
Is there any operator that could allow doing that?
After some time working on this, I finally figured out that this is not possible, at least today, there's no known operator or what so ever that could update the field with the IP addresses.
And even if it was the case, there would be delays to update the DNS records.
What I've done instead is to buy another VPS, installing HAproxy in order to proxy the Kubernetes API trafic to the master nodes, and the web trafic (both 80 and 443) to the Kubernetes worker nodes.
HAproxy monitors the nodes, and add/remove nodes automagically and in a very quick way.
With this, you just need one DNS record, pointing to the Load Balancer (or VIP of the Load Balancers in order to avoid SPOF), and HAproxy will do the rest!

DNS problem on AWS EKS when running in private subnets

I have an EKS cluster setup in a VPC. The worker nodes are launched in private subnets. I can successfully deploy pods and services.
However, I'm not able to perform DNS resolution from within the pods. (It works fine on the worker nodes, outside the container.)
Troubleshooting using https://kubernetes.io/docs/tasks/administer-cluster/dns-debugging-resolution/ results in the following from nslookup (timeout after a minute or so):
Server: 172.20.0.10
Address 1: 172.20.0.10
nslookup: can't resolve 'kubernetes.default'
When I launch the cluster in an all-public VPC, I don't have this problem. Am I missing any necessary steps for DNS resolution from within a private subnet?
Many thanks,
Daniel
I feel like I have to give this a proper answer because coming upon this question was the answer to 10 straight hours of debugging for me. As #Daniel said in his comment, the issue I found was with my ACL blocking outbound traffic on UDP port 53 which apparently kubernetes uses to resolve DNS records.
The process was especially confusing for me because one of my pods worked actually worked the entire time since (I think?) it happened to be in the same zone as the kubernetes DNS resolver.
To elaborate on the comment from #Daniel, you need:
an ingress rule for UDP port 53
an ingress rule for UDP on ephemeral ports (e.g. 1025–65535)
I hadn't added (2) and was seeing CoreDNS receiving requests and trying to respond, but the response wasn't getting back to the requester.
Some tips for others dealing with these kinds of issues, turn on CoreDNS logging by adding the log configuration to the configmap, which I was able to do with kubectl edit configmap -n kube-system coredns. See CoreDNS docs on this https://github.com/coredns/coredns/blob/master/README.md#examples This can help you figure out whether the issue is CoreDNS receiving queries or sending the response back.
I ran into this as well. I have multiple node groups, and each one was created from a CloudFormation template. The CloudFormation template created a security group for each node group that allowed the nodes in that group to communicate with each other.
The DNS error resulted from Pods running in separate node groups from the CoreDNS Pods, so the Pods were unable to reach CoreDNS (network communications were only permitted withing node groups). I will make a new CloudFormation template for the node security group so that all my node groups in my cluster can share the same security group.
I resolved the issue for now by allowing inbound UDP traffic on port 53 for each of my node group security groups.
So I been struggling for a couple of hours i think, lost track of time, with this issue as well.
Since i am using the default VPC but with the worker nodes inside the private subnet, it wasn't working.
I went through the amazon-vpc-cni-k8s and found the solution.
We have to sff the environment variable of the aws-node daemonset AWS_VPC_K8S_CNI_EXTERNALSNAT=true.
You can either get the new yaml and apply or just fix it through the dashboard. However for it to work you have to restart the worker node instance so the ip route tables are refreshed.
issue link is here
thankz
Re: AWS EKS Kube Cluster and Route53 internal/private Route53 queries from pods
Just wanted to post a note on what we needed to do to resolve our issues. Noting that YMMV and everyone has different environments and resolutions, etc.
Disclaimer:
We're using the community terraform eks module to deploy/manage vpcs and the eks clusters. We didn't need to modify any security groups. We are working with multiple clusters, regions, and VPC's.
ref:
Terraform EKS module
CoreDNS Changes:
We have a DNS relay for private internal, so we needed to modify coredns configmap and add in the dns-relay IP address
...
ec2.internal:53 {
errors
cache 30
forward . 10.1.1.245
}
foo.dev.com:53 {
errors
cache 30
forward . 10.1.1.245
}
foo.stage.com:53 {
errors
cache 30
forward . 10.1.1.245
}
...
VPC DHCP option sets:
Update with the IP of the above relay server if applicable--requires regeneration of the option set as they cannot be modified.
Our DHCP options set looks like this:
["AmazonProvidedDNS", "10.1.1.245", "169.254.169.253"]
ref: AWS DHCP Option Sets
Route-53 Updates:
Associate every route53 zone with the VPC-ID that you need to associate it with (where our kube cluster resides and the pods will make queries from).
there is also a terraform module for that:
https://www.terraform.io/docs/providers/aws/r/route53_zone_association.html
We had run into a similar issue where DNS resolution times out on some of the pods, but re-creating the pod couple of times resolves the problem. Also its not every pod on a given node showing issues, only some pods.
It turned out to be due to a bug in version 1.5.4 of Amazon VPC CNI, more details here -- https://github.com/aws/amazon-vpc-cni-k8s/issues/641.
Quick solution is to revert to the recommended version 1.5.3 - https://docs.aws.amazon.com/eks/latest/userguide/update-cluster.html
As many others, I've been struggling with this bug a few hours.
In my case the issue was this bug https://github.com/awslabs/amazon-eks-ami/issues/636 that basically sets up an incorrect DNS when you specify endpoint and certificate but not certificate.
To confirm, check
That you have connectivity (NACL and security groups) allowing DNS on TCP and UDP. For me the better way was to ssh into the cluster and see if it resolves (nslookup). If it doesn't resolve (most likely it is either NACL or SG), but check that the DNS nameserver in the node is well configured.
If you can get name resolution in the node, but not inside the pod, check that the nameserver in /etc/resolv.conf points to an IP in your service network (if you see 172.20.0.10, your service network should be 172.20.0.0/24 or so)

Can't get to GCE instance from k8s pods on the same subnet

I have a cluster with container range 10.101.64.0/19 on a net A and subnet SA with ranges 10.101.0.0/18. On the same subnet, there is VM in GCE with IP 10.101.0.4 and it can be pinged just fine from within the cluster, e.g. from a node with 10.101.0.3. However, if I go to a pod on this node which got address 10.101.67.191 (which is expected - this node assigns addresses 10.101.67.0/24 or something), I don't get meaningful answer from that VM I want to access from this pod. Using tcpdump on icmp, I can see that when I ping that VM machine from the pod, the ping gets there but I don't receive ACK in the pod. Seems like VM is just throwing it away.
Any idea how to resolve it? Some routes or firewalls? I am using the same topology in the default subnet created by kubernetes where this work but I cannot find anything relevant which could explain this (there are some routes and firewall rules which could influence it but I wasn't successful when trying to mimic them in my subnet)
I think it is a firewall issue.
Here I've already provided the solution on Stakoverflow.
It may help to solve your case.

How to setup fail2ban with Rancher Load Balancer (HAProxy)

I want to setup fail2ban on my Rancher agents.
I have a Cattle environment running a managed network where each Rancher agent is running a HAProxy used as a web-server + load balancer.
I want to ban users based on different criteria (too many failed logins, too many requests, etc) from HTTP and HTTPs ports.
Currently I have fail2ban setup with regex that is working in dummy logs but fail2ban is not banning any IPs.
fail2ban-client status <my-jail> shows that the jail has been started but has 0 bans even when I do incorrect requests.
After running fail2ban in debug mode and investigating IP tables I have found the problem.
The problem occurs because of 3 reasons:
1) HAProxy is running a docker image with server time in UTC and your servers might be in a different time zone.
2) HAProxy is running in a docker container in a Cattle managed network which means that the incoming packets are Forward packets and not Input packets for iptables.
3) The way Cattle handles forwarding is a bit ugly and hence does not allow custom fail2ban rules
In my case as HAProxy is in a docker image with a different timezone, fail2ban was ignoring ban attempts as the time was a few hours off. Changing the server time fixed the first problem.
I could now see that IPs were indeed being banned when using:
fail2ban-client status <my-jail>
But the problem still remained cause even though I could see the correct IPs being banned, I could still access the server completely fine.
This is because of the way Rancher sets up iptables. To fix this problem I changed my /etc/fail2ban/jail.local from:
[DEFAULT]
...
chain = INPUT
...
To:
[DEFAULT]
...
chain = CATTLE_FORWARD
...
Now the users are correctly banned because the timezone matches and are then put into a jail in the Forward chain hence dropping requests from banned users.

Getting ZooKeeper to run on Google's Compute Engine using external IPs

I have been trying to setup a ZooKeeper cluster on the Google Compute Engine and have run into some issues when using the external IPs of the machines. My cluster consists of 3 nodes on their own separate instances on GCE.
Now, when I configure each node to use the external IP of the instance they seem to be unable to communicate with each other.
zoo.cfg
tickTime=2000
dataDir=/var/lib/zookeeper
clientPort=2181
initLimit=5
syncLimit=2
server.1=externalIp1:2888:3888
server.2=externalIp2:2888:3888
server.3=externalIp3:2888:3888
If I configure them with their internal IP, however, everything works perfectly fine. My guess is that when ZooKeeper starts up, it binds itself to the internal IP of the instance regardless of the configurations. Because of this, when each node tries to look for the other 2 using the external IPs that they were configured, they're unable to find them.
So my question is, is there any way to make it so that ZooKeeper uses the external IP of the machine instead of the internal one? I'm relatively new to the Google Cloud Platform and to setting up hardware in general, so I'm not really sure if something like ip forwarding, firewall rules, or something else would achieve what I'm trying to do (assuming it's even possible).
According to the Zookeeper 3.4.5 docs, you need to specify the following option:
clientPortAddress
New in 3.3.0: the address (ipv4, ipv6 or hostname) to listen for client connections; that is, the address that clients attempt to connect to. This is optional, by default we bind in such a way that any connection to the clientPort for any address/interface/nic on the server will be accepted.
Although it appears that by default, it will bind to all available IPs on the server, so theoretically, it should have worked as you have set it up.
Important note: if Zookeeper instances talk to each other using external IPs rather than internal IPs, you will be charged for data egress whereas if all communication is over internal network (using internal IPs) within the same zone, you won't.