Last change
on this file since 14960 was
14763,
checked in by uli, 7 years ago
|
Fix reboot playbook.
There were two problems with the former version of
this playbook which should be fixed now:
- we do not restart virtual nodes (as their hosts
are restarted already)
- we act asyncronously when restarting machines
The latter led to predictable failures when rebooting
Ubuntu 16.04 machines. With 14.04, however, the
async: 0 setting worked fine.
|
File size:
888 bytes
|
Line | |
---|
1 | --- |
---|
2 | # This ansible-playbook reboots productive |
---|
3 | # servers |
---|
4 | # |
---|
5 | # What it does: |
---|
6 | # stop all services in need of manual stop |
---|
7 | # then reboot |
---|
8 | # then start all services in need of manual action. |
---|
9 | # |
---|
10 | # It will normally be run with ``ask-pass`` |
---|
11 | # |
---|
12 | # Sample: |
---|
13 | # |
---|
14 | # $ ansible-playbook -i hosts -K -l h5.waeup.org playbook-reboot.yml |
---|
15 | # |
---|
16 | |
---|
17 | - include: playbook-prep-reboot.yml |
---|
18 | |
---|
19 | - hosts: nonvirtual |
---|
20 | become: yes |
---|
21 | become_user: root |
---|
22 | tasks: |
---|
23 | - name: Restart machine |
---|
24 | shell: sleep 2 && /sbin/shutdown -r now "Ansible triggered reboot" |
---|
25 | async: 1 |
---|
26 | poll: 0 |
---|
27 | ignore_errors: true |
---|
28 | |
---|
29 | - name: Wait for server to come back |
---|
30 | become: no |
---|
31 | local_action: wait_for |
---|
32 | host={{ inventory_hostname }} |
---|
33 | port=22 |
---|
34 | state=started |
---|
35 | delay=30 |
---|
36 | timeout=300 |
---|
37 | connect_timeout=5 |
---|
38 | |
---|
39 | - include: playbook-post-reboot.yml |
---|
Note: See
TracBrowser for help on using the repository browser.