修复Restart Agent processes导致cpu100%卡住不动的问题，拆分Restart Agent processes 并优化Clone Agent Repository步骤

jcloud/playbooks/roles/agent/tasks/main.yml

@@ -6,6 +6,40 @@
     repo: '{{ agent_repository_url }}'
     dest: /home/jingrow/agent/repo
     remote: upstream
+    depth: 1
+    force: yes
+    clone: yes
+    update: no
+  register: clone_result
+  retries: 3
+  delay: 10
+  until: clone_result is success
+  ignore_errors: yes
+
+- name: Clean up failed clone
+  become: yes
+  become_user: jingrow
+  file:
+    path: /home/jingrow/agent/repo
+    state: absent
+  when: clone_result is failed
+  ignore_errors: yes
+
+- name: Verify clone success
+  become: yes
+  become_user: jingrow
+  stat:
+    path: /home/jingrow/agent/repo/.git
+  register: git_check
+  retries: 3
+  delay: 5
+  until: git_check.stat.exists
+  when: clone_result is success
+
+- name: Fail if clone not successful
+  fail:
+    msg: "Failed to clone agent repository after multiple attempts"
+  when: clone_result is failed or (clone_result is success and not git_check.stat.exists)
 
 - name: Install Agent
   become: yes

jcloud/playbooks/roles/docker/tasks/main.yml

@@ -81,7 +81,68 @@
 - name: Get Docker Info
   command: docker info
 
-- name: Restart Agent processes
+- name: Stop Agent processes
   supervisorctl:
     name: "agent:"
-    state: restarted
+    state: stopped
+  register: stop_result
+  retries: 3
+  delay: 5
+  until: stop_result is success
+  ignore_errors: yes
+
+- name: Wait for Agent processes to stop
+  shell: |
+    for i in $(seq 1 30); do
+      if ! supervisorctl status agent: | grep -q "RUNNING"; then
+        exit 0
+      fi
+      sleep 1
+    done
+    exit 1
+  register: wait_result
+  changed_when: false
+  when: stop_result is success
+
+- name: Start Agent processes in order
+  supervisorctl:
+    name: "{{ item }}"
+    state: started
+  loop:
+    - agent:redis
+    - agent:web
+    - agent:worker-0
+    - agent:worker-1
+  register: start_result
+  retries: 3
+  delay: 5
+  until: start_result is success
+  ignore_errors: yes
+
+- name: Wait for Agent processes to be ready
+  shell: |
+    for i in $(seq 1 30); do
+      if supervisorctl status agent: | grep -q "RUNNING" && ! supervisorctl status agent: | grep -q "STOPPED\|FATAL\|EXITED"; then
+        exit 0
+      fi
+      sleep 1
+    done
+    exit 1
+  register: wait_ready_result
+  changed_when: false
+  when: start_result is success
+
+- name: Verify Agent processes status
+  supervisorctl:
+    name: "agent:"
+    state: started
+  register: verify_result
+  retries: 3
+  delay: 5
+  until: verify_result is success
+  ignore_errors: yes
+
+- name: Fail if Agent processes not running
+  fail:
+    msg: "Agent processes failed to start properly"
+  when: verify_result is failed