Merge pull request #1167 from ksatchit/v1.0.x-ramp-time-support

[Cherry pick for patch release 1.0.1]

.gitlab-ci.yml

@@ -20,8 +20,33 @@ Test Litmus Playbook Syntax:
   script:
     - sudo docker run litmuschaos/ansible-runner:ci ./syntax-check 
 
+baseline-image:
+
+  when: always
+  stage: baseline
+  tags:
+    - test
+  only:
+    refs:
+      - /^(v[0-9][.][0-9][.]x|master)?$/
+  script:
+     - pwd
+     - export BRANCH=${CI_COMMIT_REF_NAME}
+     - echo $BRANCH
+     - export COMMIT=${CI_COMMIT_SHORT_SHA}
+     - echo $COMMIT
+     - git clone https://github.com/litmuschaos/litmus-e2e.git
+     - cd litmus-e2e
+     - git checkout ${BASELINE_BRANCH}
+     - cd baseline
+     - ansible-playbook commit-writer.yml --extra-vars "branch=$BRANCH repo=$CI_PROJECT_NAME commit=$COMMIT"
+     - git status
+     - git add baseline
+     - git status
+     - git commit -m "updated $CI_PROJECT_NAME commit:$COMMIT"
+     - git push http://${YOUR_USERNAME}:${PERSONAL_ACCESS_TOKEN}@github.com/litmuschaos/litmus-e2e.git --all
+
 Push Litmus Image:
   stage: litmusImagePush
   script:
     - REPONAME="litmuschaos" IMGNAME="ansible-runner" IMGTAG="ci" ./hack/push
-

chaoslib/litmus/container_kill/containerd_chaos/containerd.j2

@@ -1,8 +1,11 @@
-apiVersion: extensions/apps/v1
+apiVersion: apps/v1
 kind: DaemonSet
 metadata:
   name: containerd-chaos
 spec:
+  selector:
+    matchLabels:
+      app: crictl
   template:
     metadata:
       labels:

chaoslib/litmus/container_kill/containerd_chaos/crictl-chaos.yml

@@ -1,20 +1,20 @@
 ---
-- name: Patch the chaoslib image
-  template:
-    src:  /chaoslib/litmus/container_kill/containerd_chaos/containerd.j2
-    dest: /chaoslib/litmus/container_kill/containerd_chaos/containerd-chaos-ds.yml
-  vars:
-    containerd_image: "{{ lib_image }}"
-
 - block:
 
+    - name: Patch the chaoslib image
+      template:
+        src:  /chaoslib/litmus/container_kill/containerd_chaos/containerd.j2
+        dest: /chaoslib/litmus/container_kill/containerd_chaos/containerd-chaos-ds.yml
+      vars:
+        containerd_image: "{{ lib_image }}"
+
     - name: Setup containerd chaos infrastructure.
       shell: >
         kubectl apply -f /chaoslib/litmus/container_kill/containerd_chaos/containerd-chaos-ds.yml
         -n {{ namespace }}
       args:
         executable: /bin/bash
-      register: result
+      register: crictl_ds_result
 
     - name: Confirm that the containerd-chaos ds is running on all nodes.
       shell: >
@@ -131,27 +131,45 @@
       delay: 3
       retries: 30
 
-  when: action == "killapp"
-
-- block:
-
     - name: Delete the crictl-chaos daemonset
       shell: >
-        kubectl delete -f /chaoslib/litmus/container_kill/containerd_chaos/containerd-chaos-ds.yml
-        -n {{ namespace }}
+        kubectl delete -f /chaoslib/litmus/container_kill/containerd_chaos/containerd-chaos-ds.yml -n {{ namespace }} 
       args:
         executable: /bin/bash
-      register: result
-
+
     - name: Confirm that the containerd-chaos pod is deleted successfully
       shell: >
-        kubectl get pod -l app=crictl
-        --no-headers -n {{ namespace }}
+        kubectl get pods -l app=crictl --no-headers -n {{ namespace }}
       args:
         executable: /bin/bash
       register: result
-      until: "result.stdout == ''"
-      delay: 3
-      retries: 50
+      until: "'No resources found' in result.stderr"
+      delay: 5
+      retries: 60
+
 
-  when: action == "delete-containerd"
+  rescue: 
+
+    - block: 
+
+        - name: Delete the crictl-chaos daemonset
+          shell: >
+            kubectl delete -f /chaoslib/litmus/container_kill/containerd_chaos/containerd-chaos-ds.yml -n {{ namespace }}
+          args:
+            executable: /bin/bash
+          when: crictl_ds_result.rc == 0
+
+        - name: Confirm that the containerd-chaos pod is deleted successfully
+          shell: >
+            kubectl get pods -l app=crictl --no-headers -n {{ namespace }}
+          args:
+            executable: /bin/bash
+          register: result
+          until: "'No resources found' in result.stderr"
+          delay: 5
+          retries: 60
+      when: " crictl_ds_result is defined"
+
+    - fail:
+        msg: "crictl-chaos lib failed"
+      when: true

chaoslib/litmus/cordon_drain_node.yaml

@@ -1,5 +1,9 @@
 - block:
 
+  - name: Wait for the specified ramp time before injecting chaos
+    wait_for: timeout="{{ ramp_time }}"
+    when: "ramp_time is defined and ramp_time != ''"
+
   - name: Drain the application node
     shell: >
       kubectl drain {{ app_node }}

chaoslib/litmus/cpu_hog/app_cpu_stress.j2

@@ -32,6 +32,10 @@ spec:
             ## default: 1
           - name: CORES
             value: "{{ c_cores }}"
+{% if ramp_time is defined and ramp_time != '' %}
+          - name: RAMP_TIME
+            value: "{{ ramp_time }}"
+{% endif %}
         volumeMounts:
           - name: dockersocket
             mountPath: /var/run/docker.sock

chaoslib/litmus/cpu_hog/pod_cpu_hog.yml

@@ -87,6 +87,20 @@
         executable: /bin/bash
       register: cpu_app_deploy_result
 
+    - name: Calculate the total wait time for cpu stress job
+      ## considering duration in milliseconds from spec
+      ## with a grace period of 2 min before cpu stress job terminate
+      set_fact:
+        job_wait_time: "{{ ((c_duration|int)/1000 + 120)| int }}"
+      when: "ramp_time is not defined or ramp_time == ''"
+
+    - name: Calculate the total wait time for cpu stress job
+      ## considering duration in milliseconds from spec
+      ## with a grace period of 2 min before cpu stress job terminate
+      set_fact:
+        job_wait_time: "{{ ((c_duration|int)/1000 + (ramp_time|int) + 120)| int }}"
+      when: "ramp_time is defined and ramp_time != ''"
+
     - name: Wait until the cpu stress job is completed
       shell: >
         kubectl get pods -l job-name=app-cpu-stress-{{ run_id }} --no-headers -n {{ app_ns }}
@@ -96,9 +110,7 @@
       register: result
       until: "result.stdout == 'Succeeded'"
       delay: 1
-      ## considering duration in milliseconds from spec
-      ## with a grace period of 1 min before cpu stress job terminate
-      retries: "{{ ((c_duration|int)/1000 + 120)| int }}"
+      retries: "{{ job_wait_time }}"
 
     - name: Tear down cpu chaos infrastructure
       shell: >

chaoslib/litmus/disk_fill/disk_fill_by_litmus.yml

@@ -84,6 +84,10 @@
         app_ns: "{{ a_ns }}"  
         delay: 1
         retries: 60
+
+    - name: Wait for the specified ramp time before injecting chaos
+      wait_for: timeout="{{ ramp_time }}"
+      when: "ramp_time is defined and ramp_time != ''"
 
     - name: Record the disk-fill pod on app node
       shell: >

chaoslib/litmus/platform/aws/disk_loss.yml

@@ -1,3 +1,7 @@
+- name: Wait for the specified ramp time before injecting chaos
+  wait_for: timeout="{{ ramp_time }}"
+  when: "ramp_time is defined and ramp_time != ''"
+
 - name: Detaching the disk
   ec2_vol:
     id: "{{ disk_name }}"

chaoslib/litmus/platform/gke/disk_loss.yml

@@ -1,3 +1,7 @@
+- name: Wait for the specified ramp time before injecting chaos
+  wait_for: timeout="{{ ramp_time }}"
+  when: "ramp_time is defined and ramp_time != ''"
+
 - name: Detaching the disk
   shell: gcloud compute instances detach-disk {{ node_name }} --disk {{ disk_name }} --zone {{ zone_name }}
 
@@ -11,4 +15,4 @@
 - block:
   - name: If disk is not attached, it will attach manually
     shell: gcloud compute instances attach-disk {{ node_name }} --device-name {{ disk_name }} --disk {{ disk_name }} --zone {{ zone_name }}
-  when: "inuse == false"
+  when: "inuse == false"

chaoslib/litmus/platform/gke/pod_cpu_consumption.yml

@@ -45,6 +45,10 @@
     - set_fact:
         pod_ds_app: "{{ po_ds_app | json_query('resources[0].metadata.name') }}"
 
+    - name: Wait for the specified ramp time before injecting chaos
+      wait_for: timeout="{{ ramp_time }}"
+      when: "ramp_time is defined and ramp_time != ''"
+
     ## GETTING THE NODE CPU CAPACITY
     - name: Getting the node cpu capacity
       k8s_facts:

chaoslib/litmus/pod_failure_by_litmus.yml

@@ -11,9 +11,13 @@
 - name: Set min chaos count to 1 if interval > duration
   set_fact: 
     c_iterations: 1
-  when: "c_iterations == '0'" 
+  when: "c_iterations == '0'"
+
+- name: Wait for the specified ramp time before injecting chaos
+  wait_for: timeout="{{ ramp_time }}"
+  when: "ramp_time is defined and ramp_time != ''" 
 
 - name: Kill random pod
   include: kill_random_pod.yml
   with_sequence: start=1 end={{ c_iterations }}
-
+

chaoslib/powerfulseal/pod_failure_by_powerfulseal.yml

@@ -1,4 +1,8 @@
 - block:
+
+    - name: Wait for the specified ramp time before injecting chaos
+      wait_for: timeout="{{ ramp_time }}"
+      when: "ramp_time is defined and ramp_time != ''"
 
     - name: Generate the powerfulseal deployment spec from template
       template:
@@ -24,7 +28,7 @@
       retries: 90
 
     - name: Wait for chaos duration
-      wait_for: timeout={{ c_duration }}
+      wait_for: timeout="{{ c_duration }}"
 
     - name: Tear down the powerfulseal deployment
       shell: 

chaoslib/pumba/network_chaos/network_chaos.yml

@@ -55,6 +55,13 @@
             c_container: "{{ container.stdout }}"
 
       when: c_container is not defined or c_container == ''
+
+    - name: Wait for the specified ramp time before injecting chaos
+      ## the ramp_time is spent before launching the pumba job instead 
+      ## of in the pumba job, as with the new scratch-based pumba images
+      ## there is no shell available by default
+      wait_for: timeout="{{ ramp_time }}"
+      when: "ramp_time is defined and ramp_time != ''"
 
     - name: Patch the chaoslib image
       template: