Show page

Differences

This shows you the differences between two versions of the page.

--- devops:laboratoare:05 [2021/05/24 17:37]
127.0.0.1 external edit
+++ devops:laboratoare:05 [2021/07/28 16:54] (current)
bogdan.croitoru
@@ Line 1: / Line 1: @@
-===== Laboratorul 05. =====
+===== Laboratorul 05. Kubernetes =====
+==== 5.1. Pods ====
+<note tip>Nota: pe parcursul laboratoarelor de mai jos, va rog sa inlocuiti **"id"** cu numarul aferent fiecarui student.</note>
+  * Adaugam credentialele pentru clusterul de kubernetes
+<note>mkdir ~/.kube
+cp config ~/.kube
+kubectl get nodes</note>
+  * Observam contextul actual folosit
+  * Verificam conexiunea si observam contextul folosit
+<note>kubectl config view</note>
+  * Creare namespace student.
+  * Fiecare student isi va crea propriul namespace, folosind id-ul student aferent.
+<note>kubectl create ns student[id]
+kubectl get ns</note>
+  * Configuram kubectl sa foloseasca ca namespace implicit namespace-ul propriu.
+<note>kubectl config set-context --current --namespace=student[id]
+kubectl config view</note>
+  * Editam fisierul "pod01.yaml" cu id-ul aferent fiecarui student, apoi creem un pod conform definitiei din fisier. Pod-ul se va crea in aproximativ 10 secunde. Observati output-ul celor doua comenzi get. Folositi id-ul aferent.
+<note>kubectl apply -f pod01.yaml
+kubectl get pods
+kubectl get pods webserver-student-id -owide
+kubectl get pods webserver-student-id -oyaml
+</note>
+  * Intram in pod-ul nou creat si observam.
+<note>kubectl exec -it webserver-student-id -- bash</note>
+<note>root@webserver-student-id:/# cat /etc/os-release
+PRETTY_NAME="Debian GNU/Linux 10 (buster)"
+NAME="Debian GNU/Linux"
+VERSION_ID="10"
+VERSION="10 (buster)"
+VERSION_CODENAME=buster
+ID=debian
+HOME_URL="https://www.debian.org/"
+SUPPORT_URL="https://www.debian.org/support"
+BUG_REPORT_URL="https://bugs.debian.org/"
+root@webserver-student-id:/# exit
+kubectl logs webserver-student-id</note>
+  * Stergem pod-ul creat anterior.
+<note>kubectl delete pod webserver-student-id
+kubectl get po -o wide # Should output "No resources found."</note>
+  * Cream un pod ce contine doua containere
+<note>kubectl apply -f pod02.yaml
+kubectl get po -o wide
+kubectl describe pod webserver</note>
+  * Rulati comenzile urmatoare una cate una si observati ca va aflati in containere diferite ale aceluiasi pod.
+<note>kubectl exec -it webserver -c webwatcher -- /bin/bash
+root@webserver:/# cat /etc/os-release
+kubectl exec -it webserver -c webserver -- /bin/bash
+root@webserver:/# cat /etc/os-release
+</note>
+  * Stergem ceea ce am creat folosind comanda de mai jos
+<note>kubectl delete -f pod02.yaml</note>
+  * Cream pod-ul 3 folosind fisierul pod03.yaml. Avem dooua containere care impart acelasi volum.
+<note>kubectl apply -f pod03.yaml</note>
+  * Observati in definitia pod-ului cum este montat volumul html in cele doua containere
+  * Mai departe verificam faptul ca aceleasi date sunt prezente in ambele containere
+<note>kubectl exec webserver -c c1 -- /bin/cat /usr/share/nginx/html/index.html
+kubectl exec webserver -c c2 -- /bin/cat /html/index.html
+</note>
+  * Stergem pod-ul creat anterior
+<note>kubectl delete pod webserver</note>
+  * Pentru partea a 2-a (din 3) a laboratorului, rulati comanda:
+<note>cd ../Laborator2</note>
+==== 5.2. Deployments, stateless/stateful ====
+=== Deployments - de obicei folosit de aplicatii stateless ===
+  * Creeaza primul tau deployment folosind commanda urmatoare:
+<note>kubectl apply -f deploy-01.yaml</note>
+  * Observam detaliile deployment-ului
+<note>kubectl get deployments
+kubectl get pods
+kubectl get replicaset</note>
+  * Scalam deployment-ul in sus, crescand numarul de replici
+<note>kubectl scale deployments/nginx-deployment --replicas=4
+kubectl get pods -o wide # vom vedea 4 pod-uri</note>
+  * Scalam deployment-ul in jos. Reducem numarul de pod-uri
+<note>kubectl scale deployments/nginx-deployment --replicas=2
+kubectl get pods -o wide</note>
+  * Schimbam imaginea folosita in deployment utilizand comanda urmatoare:
+<note>kubectl set image  deployments/nginx-deployment nginx=nginx:1.9.1</note>
+  * Facem revert la deployment-ul anterior cu imaginea de docker anterioara
+<note>kubectl rollout undo deployments/nginx-deployment
+kubectl rollout status deployments/nginx-deployment
+kubectl describe deploy nginx-deployment
+</note>
+  * Stergem deployment-ul
+<note>kubectl delete deployment nginx-deployment</note>
+=== Statefulset - aplicatii stateful ===
+  * Pe fiecare masina student exista instalat si configurat nfs server cu export-urile:
+<note>/exports/data-0001
+/exports/data-0002</note>
+  * Verificam ca functioneaza mount-urile:
+<note>sudo showmount -e localhost</note>
+  * Fiecare student va folosi propria adresa ip, mai departe in cadrul acestui laborator.
+  * Adresa ip poate fi obtinuta folosind comanda de mai jos
+<note>ip -f inet addr show ens3</note>
+  * Editati fisierele nfs-0001.yaml si nfs-0002.yaml, adaugand adresa ip proprie in campul spec.nfs.server si actualizati campul metadata.name cu id-ul de student aferent.
+  * Cream volume persistente astfel:
+<note>kubectl apply -f nfs-0001.yaml
+kubectl apply -f nfs-0002.yaml</note>
+  * Observam volumele create
+<note>kubectl get pv | grep studentid
+kubectl describe pv <pv_name></note>
+  * Cream un statefulset folosind comanda de mai jos:
+<note>kubectl apply -f statefulset-01.yaml</note>
+  * Observam statefulset-ul creat
+<note>kubectl get statefulset
+kubectl describe statefulset
+kubectl get pods</note>
+  * Verificam ca datele de pe disk sunt disponibile in container
+  * Scriem date in volumele de pe disk
+echo 'Hello World'  /exports/data-0001/index.html
+echo "Hello $USER" > /exports/data-0002/index.html
+  * Verificam in pod-uri. Comenzile urmatoare ar trebui sa afiseze continutul fisierului index.html de pe disk
+<note>kubectl exec -it web-0 -- cat /usr/share/nginx/html/index.html
+kubectl exec -it web-1 -- cat /usr/share/nginx/html/index.html</note>
+  * Cream un serviciu care sa acceseze statefulset-ul
+<note>kubectl create -f service.yaml</note>
+  * Observam cum serviciul se conecteaza la backend (campul Endpoints:)
+<note>kubectl describe svc nginx</note>
+  * Curatam mediul
+<note>kubectl delete -f statefulset-01.yaml</note>
+  * Trecem la urmatorea parte din laborator
+<note>cd ../Laborator3</note>
+==== 5.3. Daemonsets ====
+  * Daemonset-urile sunt resurse stateless, folosite pentru a putea rula, spre exemplu, un proces pe toate nodurile din cluster.
+  * Daemonset-urile se asigura ca anumite nod-uri Kubernetes sau chiar toate ruleaza o copie a unui pod.
+  * Editam continutul fisierului daemonset.yaml, actualizand "student[id]" cu id-ul corespunzator
+<note>vi daemonset.yaml</note>
+  * Aplicam fisierul "daemonset.yaml"
+<note>kubectl apply -f daemonset.yaml</note>
+  * Verificam ca pod-urile au pornit. Vom vedea ca s-au creat 3 pod-uri, cate unul pe fiecare worker node.
+<note>kubectl get pods -owide
+kubectl describe daemonset daemonset-studentid</note>
+  * Editam fisierul "daemonset.yaml", decomentam liniile 16,17 si 18 .
+<note>vi daemonset.yaml</note>
+  * Aplicam din nou fisierul "daemonset.yaml"
+<note>kubectl apply -f daemonset.yaml</note>
+  * Unordered List ItemVerificam starea daemonset-ului si a pod-urilor. Observam ca s-a creat un pod si pe nodul master.
+<note>kubectl get daemonset
+kubectl get pods -owide</note>
+  * Incercati sa stergeti un pod creat de daemonset.
+<note>kubectl delete pod daemonset-student[id]-xxxxx</note>
+  * Curatam mediul
+<note>kubectl delete daemonset daemonset-student[id]</note>
+sau
+<note>kubectl delete -f daemonset.yaml</note>

Cursuri

TBD

Laboratoare

TBD

Resurse

TBD

devops/laboratoare/05.txt · Last modified: 2021/07/28 16:54 by bogdan.croitoru

Show page Old revisions

Media Manager Back to top