[Kubespray] make requirement check not to fail immediately (#4567)

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

contrib/kubespray/environment-check.yml

@@ -14,6 +14,13 @@
     - name: "SSH test from dev-box to all machine of infra and worker"
       ping:
 
+- hosts: all,localhost
+  gather_facts: false
+  tasks:
+    - name: "set up unmet_requirements"
+      set_fact:
+        unmet_requirements: []
+
 - hosts: localhost
   become: true
   become_user: root
@@ -35,3 +42,10 @@
   gather_facts: true
   roles:
     - { role: requirement/worker }
+
+- hosts: all,localhost
+  tasks:
+    - name: "display unmet requirements"
+      fail:
+        msg: "The following requirements are not met: {{ unmet_requirements | join('. ') }}"
+      when: unmet_requirements|length > 0

contrib/kubespray/quick-start-kubespray.sh

@@ -48,7 +48,17 @@ echo "Ping Test"
 
 ansible all -i ${HOME}/pai-deploy/cluster-cfg/hosts.yml -m ping || exit $?
 
-/bin/bash requirement.sh -m ${MASTER_LIST} -w ${WORKER_LIST} -c ${CLUSTER_CONFIG} || exit $?
+/bin/bash requirement.sh -m ${MASTER_LIST} -w ${WORKER_LIST} -c ${CLUSTER_CONFIG}
+ret_code_check=$?
+if [ $ret_code_check -ne 0 ]; then
+  echo ""
+  echo "Please press ENTER to stop the script, check the log, and modify the cluster setting to meet the requirements."
+  echo "If you are very sure about the configuration, and still want to continue, you can type in \"continue\" to force the script to proceed."
+  read user_input
+  if [ "${user_input}"x != "continue"x ]; then
+    exit $ret_code_check
+  fi
+fi
 
 /bin/bash preinstall.sh -c ${CLUSTER_CONFIG} || exit $?
 

contrib/kubespray/requirement.sh

@@ -34,10 +34,9 @@ if [ $ret_code_check -eq 0 ]
 then
   echo "Pass: Cluster meets the requirements"
 else
-  echo "Faild: Please check the output, and modify the cluster setting to meet the requirement"
+  echo "Failed: There are unmet requirements in your cluster, the installation will be very likely to fail."
   rm -rf ${HOME}/pai-pre-check/
   exit $ret_code_check
 fi
 
 rm -rf ${HOME}/pai-pre-check/
-

contrib/kubespray/roles/kubectl/tasks/main.yml

@@ -52,7 +52,7 @@
   when: ping_cloud_google.rc == 0
 
 - name: Update source with mirrors.aliyun.com
-  include_tasks: cloud.google.yml
+  include_tasks: mirrors.aliyun.yml
   when: ping_cloud_google.rc != 0
 
 - name: run the equivalent of "apt-get update" as a separate step

contrib/kubespray/roles/requirement/devbox/tasks/ubuntu.yml

@@ -8,8 +8,8 @@
   environment: {}
 
 - name: "Dev-box 1.2 Check if docker is installed on your dev-box or not."
-  fail:
-    msg: "Unable to find docker in your dev-box machine"
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['Unable to find docker in your dev-box machine'] }}"
   changed_when: false
   check_mode: false
   environment: {}
@@ -29,32 +29,15 @@
       dev_box_docker_version_replace: "{{ dev_box_docker_version.stdout | replace('\"','') }}"
 
 - name: "Dev-box 2.3 Check docker version in dev-box"
-  fail:
-    msg: "The docker version in your dev-box is too low, please update it. And make sure its higher then 1.10"
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['The docker version in your dev-box is too low, please update it. And make sure its higher then 1.10'] }}"
   changed_when: false
   check_mode: false
   environment: {}
   when:
     - dev_box_docker_version_replace is version('1.10.0', '<=')
 
-- name: "Dev-box 3.1 Check whether the vm can raw.githubusercontent.com"
-  raw: curl https://raw.githubusercontent.com/microsoft/pai/master/README.md
-  register: devbox_curl_githubusercontent
-  failed_when: false
-  changed_when: false
-  check_mode: false
-  environment: {}
-
-- name: "Dev-box 3.2 Check whether the vm can access to raw.githubusercontent.com"
-  fail:
-    msg: "Unable to access raw.githubusercontent.com"
-  changed_when: false
-  check_mode: false
-  environment: {}
-  when:
-    - devbox_curl_githubusercontent.rc != 0
-
-- name: "Dev-box 4.1 Check whether the vm can access to docker.io"
+- name: "Dev-box 3.1 Check whether the vm can access to docker.io"
   raw: nslookup index.docker.io
   register: devbox_nslookup_docker_io
   failed_when: false
@@ -62,9 +45,9 @@
   check_mode: false
   environment: {}
 
-- name: "Dev-box 4.2 Check whether the vm can access to docker.io"
-  fail:
-    msg: "Unable to access docker.io"
+- name: "Dev-box 3.2 Check whether the vm can access to docker.io"
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['Unable to access docker.io'] }}"
   changed_when: false
   check_mode: false
   environment: {}

contrib/kubespray/roles/requirement/infra/tasks/resource.yml

@@ -1,12 +1,12 @@
 ---
 - name: Check memory resource requirement for openpai service
-  fail:
-    msg: "OpenPAI's infra node should have 40 Gi free memory for service "
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + [\"OpenPAI's infra node should have 40 Gi free memory for service\"] }}"
   when:
     - ansible_memory_mb["nocache"]["free"] < 40000
 
 - name: Check cpu resource requirement for openpai service
-  fail:
-    msg: "OpenPAI's infra node should have 1 CPU vcore for service "
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + [\"OpenPAI's infra node should have 1 CPU vcore for service\"] }}"
   when:
     - ansible_processor_vcpus < 1

contrib/kubespray/roles/requirement/infra/tasks/ubuntu-docker.yml

@@ -8,8 +8,8 @@
   environment: {}
 
 - name: "Fail, if docker is not installed on your master machine."
-  fail:
-    msg: "Unable to find docker in your master machine"
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['Unable to find docker in your master machine'] }}"
   changed_when: false
   check_mode: false
   environment: {}

contrib/kubespray/roles/requirement/infra/tasks/ubuntu.yml

@@ -8,8 +8,8 @@
   environment: {}
 
 - name: "Infra 1.2 Check whether the vm can access to docker.io"
-  fail:
-    msg: "Unable to access docker.io"
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['Unable to access docker.io'] }}"
   changed_when: false
   check_mode: false
   environment: {}
@@ -33,8 +33,8 @@
   environment: {}
 
 - name: "Infra 2.3 Check whether the NTP is installed and enabled."
-  fail:
-    msg: "NTP is not enabled on your infra machines."
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['NTP is not enabled on your infra machines'] }}"
   changed_when: false
   check_mode: false
   environment: {}
@@ -42,8 +42,8 @@
     - infra_ntpq.rc != 0 or infra_nslookup_docker_io.rc != 0
 
 - name: "Infra 3.1 Ensure dev-box is not an infra machines"
-  fail:
-    msg: "Dev-box should be a separated machine from the cluster."
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['Dev-box should be a separated machine from the cluster'] }}"
   when:
     - ansible_control_host_address == ansible_default_ipv4.address
 
@@ -56,8 +56,8 @@
   environment: {}
 
 - name: "Infra 4.2 Check whether the vm can access to gcr.io"
-  fail:
-    msg: "Unable to access gcr.io"
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['Unable to access gcr.io'] }}"
   changed_when: false
   check_mode: false
   environment: {}
@@ -73,8 +73,8 @@
   environment: {}
 
 - name: "Infra 5.2 Check whether the vm can access to quay.io"
-  fail:
-    msg: "Unable to access quay.io"
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['Unable to access quay.io'] }}"
   changed_when: false
   check_mode: false
   environment: {}

contrib/kubespray/roles/requirement/worker/tasks/resource.yml

@@ -1,6 +1,6 @@
 ---
 - name: Check memory resource requirement for openpai service
-  fail:
-    msg: "OpenPAI's worker node should have 16 Gi free memory for service "
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + [\"OpenPAI's worker node should have 16 Gi free memory for service\"] }}"
   when:
     - ansible_memory_mb["nocache"]["free"] < 16000

contrib/kubespray/roles/requirement/worker/tasks/ubuntu-docker.yml

@@ -8,8 +8,8 @@
   environment: {}
 
 - name: "Failed, if docker is not installed on your worker machine."
-  fail:
-    msg: "Unable to find docker in your master machine"
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['Unable to find docker in your worker machine'] }}"
   changed_when: false
   check_mode: false
   environment: {}
@@ -29,8 +29,8 @@
     default_runtime_processed: "{{ default_runtime.stdout_lines[0] | replace('\"','') }}"
 
 - name: "Check the default runtime is set correctly"
-  fail:
-    msg: "The default runtime is not set correctly"
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['The default runtime is not set correctly'] }}"
   changed_when: false
   check_mode: false
   environment: {}

contrib/kubespray/roles/requirement/worker/tasks/ubuntu-nvidia.yml

@@ -8,8 +8,8 @@
   environment: {}
 
 - name: "Check NVIDIA GPU exits or not"
-  fail:
-    msg: "NVIDIA GPU card is not detected in your worker machines."
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['NVIDIA GPU card is not detected in your worker machines'] }}"
   changed_when: false
   check_mode: false
   environment: {}
@@ -25,8 +25,8 @@
   environment: {}
 
 - name: "Check NVIDIA GPU drivers is installed or not "
-  fail:
-    msg: "NVIDIA GPU drivers is not detected in your worker machines."
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['NVIDIA GPU drivers is not detected in your worker machines'] }}"
   changed_when: false
   check_mode: false
   environment: {}
@@ -42,8 +42,8 @@
   environment: {}
 
 - name: "Check NVIDIA container runtime is installed or not"
-  fail:
-    msg: "NVIDIA container runtime is not detected in your worker machines."
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['NVIDIA container runtime is not detected in your worker machines'] }}"
   changed_when: false
   check_mode: false
   environment: {}

contrib/kubespray/roles/requirement/worker/tasks/ubuntu.yml

@@ -8,8 +8,8 @@
   environment: {}
 
 - name: "Worker 1.2 Check whether the vm can access to docker.io"
-  fail:
-    msg: "Unable to access docker.io"
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['Unable to access docker.io'] }}"
   changed_when: false
   check_mode: false
   environment: {}
@@ -17,8 +17,8 @@
     - worker_nslookup_docker_io.rc != 0
 
 - name: "Worker 2.1 Ensure dev-box is not an worker machines"
-  fail:
-    msg: "Dev-box should be a separated machine from the cluster."
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['Dev-box should be a separated machine from the cluster'] }}"
   when:
     - ansible_control_host_address == ansible_default_ipv4.address
 
@@ -31,8 +31,8 @@
   environment: {}
 
 - name: "Worker 3.2 Check whether the vm can access to gcr.io"
-  fail:
-    msg: "Unable to access gcr.io"
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['Unable to access gcr.io'] }}"
   changed_when: false
   check_mode: false
   environment: {}
@@ -48,8 +48,8 @@
   environment: {}
 
 - name: "Worker 4.2 Check whether the vm can access to quay.io"
-  fail:
-    msg: "Unable to access quay.io"
+  set_fact:
+    unmet_requirements: "{{ unmet_requirements + ['Unable to access quay.io'] }}"
   changed_when: false
   check_mode: false
   environment: {}