Kubernetes: Difference between revisions
| Line 286: | Line 286: | ||
k exec -it $i -- bash -c "set | grep ENVVARX" | k exec -it $i -- bash -c "set | grep ENVVARX" | ||
== Monitoroing == | |||
=== Prometheus queries of note: === | |||
Cluster wide CPU usage percent: | |||
sum (rate (container_cpu_usage_seconds_total{id="/"}[5m])) / sum (machine_cpu_cores{}) * 100 | |||
Cluster wide memory usage: | |||
sum (container_memory_working_set_bytes{id="/"}) / sum (machine_memory_bytes{}) * 100 | |||
Cluster wide cpu and memory request v available as a percent ( keep this below 100% ) | |||
CPU: | |||
sum(kube_pod_container_resource_requests_cpu_cores) / sum(kube_node_status_capacity_cpu_cores) * 100 | |||
Memory: | |||
sum(kube_pod_container_resource_requests_memory_bytes) / sum(kube_node_status_capacity_memory_bytes) * 100 | |||
== Troubleshooting == | == Troubleshooting == | ||
Revision as of 18:15, 10 February 2020
Useful
alias:
alias k="kubectl"
alias ks="kubectl --namespace kube-system" # Kubernetes Events
alias ke="kubectl get events --sort-by='{.lastTimestamp}'" # Kubernetes System stuff
alias kse="kubectl --namespace kube-system get events --sort-by='{.lastTimestamp}'" # Kubernetes Systems Events
For moving around namespaces fast
alias kX="kubectl --name-sapce nameX"
dump all :
kubectl get all --export=true -o yaml
( namespace kube-system not dumped )
list form:
k get pods k get rs # replica set k get rc # replication controller
what are all the things ?
kubectl api-resources
kubectl api-versions
event sorted by time
kubectl get events --sort-by=.metadata.creationTimestamp
what storage classes does my cluster support?
k get storageclass
how are pod spread out over nodes:
k describe node | grep -E '(^Non-t|m |^Name)' | more
( doesn't scale well, no indication of )
what deployments should be on what instance groups:
kubectl get deploy --all-namespaces -o=jsonpath='{range .items[*]}{.metadata.name}{"\t"}{..nodeSelector}{"\n"}{end}' |sort
or the other way round ( ig first )
kubectl get deploy --all-namespaces -o=jsonpath='{range .items[*]}{..nodeSelector}{"\t"}{.metadata.name}{"\n"}{end}' |sort
if you used kops to deploy the cluster then nodes are labes with their instance groups, you can be more specific like this:
k describe node -l kops.k8s.io/instancegroup=<instance group name> | grep -E '(^Non-t|m |^Name)' | more
how many nods in each instance group? ( tested under kops )
for i in `kops get ig 2>/dev/null| grep -v NAME | awk '{print $1}'`
do
echo $i
kubectl get nodes -l kops.k8s.io/instancegroup=$i
done
how many pods per node:
k get pod -o wide | grep -v NAME | awk '{print $8}' | sort | uniq -c | sort -rn
k get pod --all-namespaces -o wide | grep -v NAME | awk '{print $8}' | sort | uniq -c | sort -rn
audit: who tried to do what?
ks get pod | grep kube-apiserver-ip
ks logs $podname
who tried to scale unsuccessfully?
ks logs $podname | grep scale | grep cloud | awk '$8!=200{print $0}'
Where is the service account token that I gave this pod?
It's in here: /var/run/secrets/kubernetes.io/serviceaccount/token
Scripting Scaling
Manually edit the replicas of a deployment from within the same namespace, but in a different pod:
- give the actor pod a service account ( possibly via it's deployment ).
- create a Role as below.
- create the RoleBinding to connect the ServiceAccount to the Role.
Now you have: Pod -> Deployment -> ServiceAccount -> RoleBinding -> Role
Now the Pod has permission to do what it needs. Very similar to AWS's "IAM Role" where you give an instance a role that has the permissions that it needs to operate.
Note that in this case "ClusterRole" and ClusterRoleBinding are not required. It's all namespaced to the namespace that your deployment is in. In this case: "default".
export API_URL="https://${KUBERNETES_SERVICE_HOST}:${KUBERNETES_SERVICE_PORT}/${KUBE_ENDPOINT}"
export TOKEN=`cat /var/run/secrets/kubernetes.io/serviceaccount/token`
export CURL_CA_BUNDLE=/var/run/secrets/kubernetes.io/serviceaccount/ca.crt
curl \
-H 'Accept: application/json' \
-H "Authorization: Bearer $TOKEN" \
$API_URL \
> scale.json
# edit scale.json, set replicas to 4
curl -X PUT \
-d@scale.json \
-H 'Content-Type: application/json' \
-H "Authorization: Bearer $TOKEN" \
$API_URL
CURL_CA_BUNDLE - kubenerets is it's own CA, and presennts to each pod a ca bundle that makes ssl "in" the cluster valid.
This was the role that did it. FIXME: pare it down
apiVersion: rbac.authorization.k8s.io/v1
kind: Role
metadata:
name: kube-cloudwatch-autoscaler
labels:
app: kube-cloudwatch-autoscaler
rules:
- apiGroups:
- ""
resources:
- nodes
verbs:
- list
- apiGroups:
- apps
resources:
- deployments
- deployments.apps
- deployments.apps/scale
- "*/scale"
verbs:
- get
- update
- patch
- put
- apiGroups:
- ""
resources:
- configmaps
verbs:
- get
- create
On patching
There are a couple of way to change an object.
export TOKEN=`cat /var/run/secrets/kubernetes.io/serviceaccount/token` export CURL_CA_BUNDLE=/var/run/secrets/kubernetes.io/serviceaccount/ca.crt
1. dump whole "thing" , make change post object back ( as above ) GET -> PUT
curl \ -v \ -H 'Accept: application/json' \ -H "Authorization: Bearer $TOKEN" \ $API_URL \ > scale.json # edit scale.json, set replicas to 4 curl -X PUT \ -d@scale.json \ -H 'Content-Type: application/json' \ -H "Authorization: Bearer $TOKEN" \ $API_URL
2. terse PATCH
curl -sS \
-X 'PATCH' \
-H "Authorization: Bearer ${TOKEN}" \
-H 'Content-Type: application/merge-patch+json' \
$API_URL \
-d '{"spec": {"replicas": 1}}'
3. old / full PATCH ?
reference: https://stackoverflow.com/questions/41792851/manage-replicas-count-for-deployment-using-kubernetes-api ( 1 year 8 months old at tie of _this_ writing )
Careful, compare:
BORKEN!
PAYLOAD='[{"op":"replace","path":"/spec/replicas","value":"3"}]'
curl \
-X PATCH \
-d ${PAYLOAD} \
-H 'Content-Type: application/json-patch+json' \
-H "Authorization: Bearer ${TOKEN}" \
$API_URL
WERKS!
curl \
-X PATCH \
-d '[{"op":"replace","path":"/spec/replicas","value":3}]' \
-H 'Content-Type: application/json-patch+json' \
-H "Authorization: Bearer ${TOKEN}" \
$API_URL
Closely:
-d '[{"op":"replace","path":"/spec/replicas","value":"3"}]' <- broken
-d '[{"op":"replace","path":"/spec/replicas","value":3}]' <- works
Template Examples
list images by pod:
kubectl get pods --all-namespaces -o=jsonpath='{range .items[*]}{"\n"}{.metadata.name}{":\t"}{range .spec.containers[*]}{.image}{", "}{end}{end}{"\n"}'
list images by deploy
kubectl get deploy -o=jsonpath='{range .items[*]}{"\n"}{.metadata.name}{":\t"}{range .spec.template.spec.containers[*]}{.image}{", "}{end}{end}{"\n"}'
list all deployment cpu and mem request:
kubectl get deploy -o=jsonpath='{range .items[*]}{"\n"}{.metadata.name}{":\t"}{range .spec.template.spec.containers[*]}{.resources.requests.cpu}{", "}{.resources.requests.memory}{end}{end}{"\n"}'
list nodeslectors for all deploys:
kubectl get deploy -o=jsonpath='{range .items[*]}{"\n"}{.metadata.name}{":\t"}{.spec.template.spec.nodeSelector}{end}{"\n"}'
all nodes by their condition statuses:
kubectl get nodes -o=jsonpath='{range .items[*]}{@.metadata.name}:{"\n"}{range @.status.conditions[*]}{"\t"}{@.type}={@.status};{"\n"}{end}{end}{"\n"}'
Note the double loop.
examine the resource requesnt and limit for all deploys:
kubectl get deploy -o=jsonpath='{range .items[*]}{@.metadata.name}{range @.spec.template.spec.containers[*]}{@.resources.request.cpu}{@.resources.request.memory}{@.resources.limits.cpu}{@.resources.limits.memory}{"\n"}{end}{end}{"\n"}'
ContainersIDs by host: ( accounts for multiple containers per pod )
for i in `kns describe node NODENAME | grep "m (" | grep -v cpu | awk '{print $2}'`
do
echo -n $i " "
ns=$(kns get pod -A | grep $i | awk '{print $1}')
echo -n $ns " "
# k get pod $i -n $ns -o jsonpath='{.metadata.name}{" "}{.status.containerStatuses[*].containerID}{"\n"}' ; done
k get pod $i -n $ns -o jsonpath='{"\n"}{"\t"}{range .status.containerStatuses[*]}{.name}{" "}{.containerID}{"\n"}{end}' ; done
done
request for every replicaset ,across all name spaces
kubectl get rs -A -o=jsonpath='{range .items[*]}{"\n"}{.metadata.namespace}{","}{.metadata.name}{","}{.status.readyReplicas}{","}{range .spec.template.spec.containers[*]}{.resources.requests.cpu}{","}{.resources.requests.memory}{end}{end}{"\n"}' | sort -n -r -k 3 -t , | head
of the form:
namespace,name,replicasready,cpu,mem
metrics
wget "$(kubectl config view -o jsonpath='{range .clusters[*]}{@.cluster.server}{"\n"}{end}')"
also:
k top nodes k top pods
exec notes
k exec -it $i -- bash -c "set | grep ENVVARX"
Monitoroing
Prometheus queries of note:
Cluster wide CPU usage percent:
sum (rate (container_cpu_usage_seconds_total{id="/"}[5m])) / sum (machine_cpu_cores{}) * 100
Cluster wide memory usage:
sum (container_memory_working_set_bytes{id="/"}) / sum (machine_memory_bytes{}) * 100
Cluster wide cpu and memory request v available as a percent ( keep this below 100% )
CPU:
sum(kube_pod_container_resource_requests_cpu_cores) / sum(kube_node_status_capacity_cpu_cores) * 100
Memory:
sum(kube_pod_container_resource_requests_memory_bytes) / sum(kube_node_status_capacity_memory_bytes) * 100
Troubleshooting
Pod
pod starts and dies too quick, inthe deployment, stateuleset, or deamonset overrid the command and args with this:
look at logs:
put this in a deployment so that you can exec into a failing pod and see what it's upto:
command: ["/bin/sh"]
args: ["-c", "while true; do date; sleep 10;done"]
the pod will come up and stay up long enugh for you to get in and look around.
this only works with a pod that has enough of an operating system for you to do that . Like bash ps ls cd and such tools. Some very slim containers do not have those tools.
kube Api server
You want to see what is up with the api server. If there is a problem with the cluster the api server is going to give you a better view.
by default -v ( verbose ) is not set.
in this example we are using KOPS to make the kubenertes cluster, so to change the api server config we need to edit the cluster_spec.
kops edit cluster
add a section
apiServer: LogLevel: X
This adds --v=X to the api server command line.
I had something pounding the server with a bad token , ut the log only said: "Bad token"... who ?
Had to turn loglevel up to 20 to get to the bottom of it... it was a bad fluentd token ( wish fluentd had some sort of back-off. )
Note that a kops update showed "no changed required".
I could force it, but instead, I used the "small change" trick Kops#Tricks_for_making_small_updates_to_a_host
Practices and Guidelines
https://medium.com/devopslinks/security-problems-of-kops-default-deployments-2819c157bc90
- Do not use replication controllers, instead use replica sets
- when changing the shape of the cluster , number and type of instance groups , you will use kops edit ig <ig name> , but don't for get to update the cluster-autoscaler config ( ks edit deploy cluster-autoscaler )
- Stay up to date, read release notes, just like you do for all the other stuff you manage right ? https://github.com/kubernetes/kubernetes/blob/master/CHANGELOG-1.14.md
ConfigMaps
All things configmaps:
https://kubernetes.io/docs/tasks/configure-pod-container/configure-pod-configmap/
Cgroup / slice errors
https://github.com/kubernetes/kubernetes/issues/56850
log message:
Sep 18 21:32:37 ip-10-10-37-50 kubelet[1681]: E0918 21:32:37.901058 1681 summary.go:92] Failed to get system container stats for "/system.slice/docker.service": failed to get cgroup stats for "/system.slice/docker.service": failed to get container info for "/system.slice/docker.service": unknown container "/system.slice/docker.service"
MAAS ubuntu
https://stripe.com/blog/operating-kubernetes
https://medium.com/@adriaandejonge/moving-from-docker-to-rkt-310dc9aec938
https://coreos.com/rkt/docs/latest/rkt-vs-other-projects.html#rkt-vs-docker
Security
Todo / read:
- https://github.com/aquasecurity/kube-hunter/blob/master/README.md
- https://www.arctiq.ca/events/2018/10/5/building-a-secure-container-strategy-with-aqua-security-microsoft-azure-and-hashicorp-vault/
- https://koudingspawn.de/secure-kubernetes-with-vault/
References and Reading
- Replica set versus Replication controller
- https://www.mirantis.com/blog/kubernetes-replication-controller-replica-set-and-deployments-understanding-replication-options/
- Publishing services - service types
- https://kubernetes.io/docs/concepts/services-networking/service/#publishing-services-service-types
- Kuberenetes the hard way
- https://github.com/kelseyhightower/kubernetes-the-hard-way
- Hadolint - A smarter Dockerfile linter that helps you build best practice Docker images.
- https://github.com/hadolint/hadolint
HPA broken
Blue is test
Blue env:
Client Version: v1.12.2 Server Version: v1.10.6
Prod env:
Client Version: v1.12.2 Server Version: v1.9.8
In prod HPAs work. When I ask for them I see:
NAME REFERENCE TARGETS MINPODS MAXPODS REPLICAS AGE adjust Deployment/adjust 0%/70% 1 5 1 1d web-admin Deployment/web-admin 0%/70% 1 3 1 2h
In blue env they don't work, I see:
NAME REFERENCE TARGETS MINPODS MAXPODS REPLICAS AGE adjust Deployment/adjust <unknown>/70% 1 5 1 1d web-admin Deployment/web-admin <unknown>/70% 1 3 1 2h
in Kubernetes events we see:
HorizontalPodAutoscaler Warning FailedGetResourceMetric horizontal-pod-autoscaler unable to get metrics for resource cpu: no metrics returned from resource metrics API
Note that the metrics server is running in kube-system, but there are no repo files for that in /third-party" in prod.
In blue we store all metrics-server related files in /thirdpary/metrics-server ( taken from git@github.com:kubernetes-incubator/metrics-server.git )
In prod the deployment has:
- command:
- /metrics-server
- --source=kubernetes.summary_api:''
In blue this seemed to do the trick
- /metrics-server
- --kubelet-preferred-address-types=InternalIP
- --kubelet-insecure-tls
Cluster Scaling
ks get configmap cluster-autoscaler-status -o yaml
https://github.com/kubernetes/autoscaler/blob/master/cluster-autoscaler/FAQ.md
Steps to move hardware around
In this case we are removing the last node from an instance group and then removing the instance group.
Reference: https://kubernetes.io/docs/concepts/architecture/nodes/
1. Cordon the node
k cordon ip-xx-xx-xx-xx.region.compute.internal
No new pods will be deployed here.
2. drain ( move pods here to somewhere else )
k drain ip-xx-xx-xx-xx.region.compute.internal
You may need to add "--ignore-daemonsets" if you have daemonsets running ( data dog , localredis )
You may need to "--delete-local-data" if you have a metrics server on this node. BE CAREFUL. You will loose metrics, but probably you have an "out of cluster" place where metrics are stored ( datadog, elastic search, etc )
3. remove the nodegroup from the autoscaler:
ks edit deploy cluster-autoscaler
4. tell kops to delete the instance group.
kops delete ig myig
at this point the vms will be shut down.
k get nodes
Downing nodes
kubectl drain <node name> --delete-local-data --force --ignore-daemonsets kubectl delete node <node name>
Kubeadm way
1. light up some instances:
if you are using amzn linux you can cloud-init like this:
yum_repos:
# The name of the repository
kubernetes:
# Any repository configuration options
# See: man yum.conf
#
# This one is required!
baseurl: https://packages.cloud.google.com/yum/repos/kubernetes-el7-x86_64
enabled: true
gpgcheck: true
gpgkey:
- https://packages.cloud.google.com/yum/doc/yum-key.gpg
- https://packages.cloud.google.com/yum/doc/rpm-package-key.gpg
name: kubernetes
packages:
- curl
- git
- iproute-tc
- jq
- kubeadm
- kubectl
- kubelet
- lsof
- mlocate
- ntp
- screen
- strace
- sysstat
- tcpdump
- telnet
- traceroute
- tree
- unzip
- wget
runcmd:
- [ /usr/bin/updatedb ]
- [ 'amazon-linux-extras', 'install', 'docker', '-y' ]
- [ 'setenforce', '0']
- [ 'systemctl', 'enable', 'docker']
- [ 'systemctl', 'start', 'docker' ]
- [ 'systemctl', 'enable', 'kubelet']
- [ 'systemctl', 'start', 'kubelet' ]
write_files:
- content: |
net.bridge.bridge-nf-call-ip6tables = 1
net.bridge.bridge-nf-call-iptables = 1
path: /etc/sysctl.d/k8s.conf
permissions: '0755'
owner: root:root
- content: |
SERVER=$(/usr/bin/aws ec2 describe-tags --region us-east-1 --filters "Name=resource-id,Values=$(wget -q -O - http://169.254.169.254/latest/meta-data/instance-id)" "Name=key,Values=Name" --query 'Tags[*].Value' --output text)
PRIVATE_IP=$(curl http://169.254.169.254/latest/meta-data/local-ipv4)
# if hostname was set this would work, but hostname is not set
# sed -i "s/^\(HOSTNAME\s*=\s*\).*$/\1$SERVER/" /etc/sysconfig/network
echo "HOSTNAME=$SERVER" >> /etc/sysconfig/network
echo "$PRIVATE_IP $SERVER" >> /etc/hosts
echo "$SERVER" > /etc/hostname
hostname $SERVER
path: /root/sethostname.sh
permissions: '0755'
owner: root:root
on the first instance, do a
kubeadm init
and save the output.
run that output on the other instances.
boom! a kubernetes cluster...
what is it missing ?
- your app
- logging
- monitoring
- dashboard
Custer status
kubectl get componentstatuses
The helm way
Setup
prereq: k cluster is up already.
step 1. install local helm 3 binary.
https://github.com/helm/helm/releases/tag/v3.0.0-alpha.1
MacOS
cd ~/work mkdir helm cd helm wget https://get.helm.sh/helm-v3.0.0-alpha.1-darwin-amd64.tar.gz tar darwin-amd64.tar.gz cp darwin-amd64/helm ~/bin
Loonix
cd ~/work mkdir helm cd helm wget https://get.helm.sh/helm-v3.0.0-alpha.1-linux-amd64.tar.gz tar zxvf linux-amd64.tar.gz cp linux-amd64/helm ~/bin/
helm can use different kubectl contexts, but we just use on concext so simple is fine.
to get the kubernetes side of helm setup you do this:
blue@kubernetescluster:~$ helm init --debug Creating /home/blue/.helm Creating /home/blue/.helm/repository Creating /home/blue/.helm/repository/cache Creating /home/blue/.helm/plugins Creating /home/blue/.helm/starters Creating /home/blue/.helm/cache/archive Creating /home/blue/.helm/repository/repositories.yaml Adding stable repo with URL: https://kubernetes-charts.storage.googleapis.com $HELM_HOME has been configured at /home/blue/.helm. Happy Helming! blue@kubernetescluster:~$
note that "--dry-run" worked in helm 2 but doesn't owrk in helm 3.
taints
Still learning about this .
kubectl get nodes -o json | jq .items[].spec.taints
Kubernetes dashboard
get the token for the service account to loggin to the web ui:
ks get secret `ks get sa kubernetes-dashboard -o=jsonpath='{.secrets[0].name}'` -o=jsonpath='{.data.token}' | base64 -d ; echo
Kubernetes Quiz Links
- Pods
https://kodekloud.com/p/practice-test-kubernetes-ckad-pods
- ReplicaSets
https://kodekloud.com/p/practice-test-kubernetes-ckad-replicasets
- Deployments
https://kodekloud.com/p/practice-test-kubernetes-ckad-deployments
- Namespaces
https://kodekloud.com/p/practice-test-kubernetes-ckad-namespaces
- Commands and Arguments
https://kodekloud.com/p/practice-test-kubernetes-ckad-commands-and-arguments
- ConfigMaps
https://kodekloud.com/p/practice-test-kubernetes-ckad-configmaps
- Secrets
https://kodekloud.com/p/practice-test-kubernetes-ckad-secrets
- Security Contexts
https://kodekloud.com/p/practice-test-kubernetes-ckad-security-contexts
- Service Accounts
https://kodekloud.com/p/practice-test-kubernetes-ckad-service-account
- Taints and Tolerations
https://kodekloud.com/p/practice-test-kubernetes-ckad-taints-tolerations
- Node Affinity
https://kodekloud.com/p/practice-test-kubernetes-ckad-node-affinity
- Multi-Container Pods
https://kodekloud.com/p/practice-test-kubernetes-ckad-multicontainer-pods
- Readiness and Liveness Probes
https://kodekloud.com/p/practice-test-kubernetes-ckad-readiness-probes
- Container Logging
https://kodekloud.com/p/practice-test-kubernetes-ckad-logging
- Monitoring
https://kodekloud.com/p/practice-test-kubernetes-ckad-monitoring
- Labels & Selectors
https://kodekloud.com/p/practice-test-kubernetes-ckad-labels-and-selectors
- Rolling Updates And Rollbacks
https://kodekloud.com/p/practice-test-kubernetes-ckad-rolling-updates-and-rollbacks
- Services
https://kodekloud.com/p/kubernetes-for-beginners-services-493847859
Datadog
what version of data dog am I running ?
do this:
kubectl get pods -l app=datadog-agent -o=jsonpath='{range .items[*]}{"\n"}{.metadata.name}{":\t"}{range .spec.containers[*]}{.image}{", "}{end}{end}{"\n"}'
and get "latest"! except what does that mean?
instead ask the agent itself:
for i in `kubectl get pods -l app=datadog-agent | awk '{print $1}' | grep -v NAME `; do echo $i; k exec -it $i -- /opt/datadog-agent/bin/agent/agent version; done
datadog-agent-XXXX
Agent X.X.X - Commit: XXX - Serialization version: X.X.X
datadog-agent-YYYY
Agent X.X.Y - Commit: XXX - Serialization version: X.X.X
ah ah! inconsistant versions! can be fixed with a ds delete -> k apply, or even just a pod kill.
TCPDump a container
or pod.
Reference:
https://community.pivotal.io/s/article/How-to-get-tcpdump-for-containers-inside-Kubernetes-pods
- Get the container ID and host.
k get pod XXX -o=jsonpath='{.status.containerStatuses[0].containerID}{"\n"}{.status.hostIP}{"\n"}'
docker://YYYYYYYYYYYYYYYYYYYYYY
Z.Z.Z.Z
- get the interface index
docker exec XXX cat /sys/class/net/eth0/iflink <NUMBER>
- find the interface on the host
ip link |grep ^<NUMBER>:
- On those dump that interface
tcpdump -i veth235ab8ff
Disk usage of container
kns get pod -l app=myapp,env=datacenter-production -o=jsonpath='{range .items[*]}{.status.containerStatuses[0].containerID}{"\t"}{.status.hostIP}{"\n"}{end}' | \
awk 'BEGIN{FS="//"}{print $2}' | \
while read a b
do
echo ssh ${b} sudo du -sh /var/lib/docker/containers/${a}
done
Note that by default in k8s the docker json log driver is used so if the log is big for a container, that log will be in the container's directory. I'm not sure how to fix that . there is a config for it , but it looks like kubernetes is ont honouring it.
Also See
- kops - automated kubenetes cluster build.
- What I learned today Nov 2nd 2018
- Ingress Networking - 1
- https://kodekloud.com/p/practice-test-kubernetes-ckad-ingress-1
- Ingress Networking - 2
- https://kodekloud.com/p/practice-test-kubernetes-ckad-ingress-2-deploy-controller
