Retry the git clone in role fetcher
I see failures in my local CI 10-20% of the time on cloning one or two roles, so the test fails to bootstrap. I'm not sure if there is some rate limiting in place on the Openstack git repos, but it seems to fail quite often on a very reliable link. The failure being seen is: Failed to connect to git.openstack.org port 443: No route to host This patch adds retries to the git clone task used to fetch roles to hopefully avoid this failure. Change-Id: I1ea74e3ac4b2d0da88f3155d593eaceb0a205d46
This commit is contained in:
parent
2e1c6a4dce
commit
c8606c443c
@ -39,7 +39,13 @@
|
||||
when:
|
||||
- item.scm == "git" or item.scm is undefined
|
||||
with_items: "{{ roles }}"
|
||||
register: git_clone
|
||||
until: git_clone | success
|
||||
retries: "{{ git_clone_retries }}"
|
||||
delay: "{{ git_clone_retry_delay }}"
|
||||
vars:
|
||||
roles: "{{ lookup('file', role_file) | from_yaml }}"
|
||||
role_file: '../ansible-role-requirements.yml'
|
||||
role_path_default: '/etc/ansible/roles'
|
||||
git_clone_retries: 2
|
||||
git_clone_retry_delay: 5
|
||||
|
Loading…
x
Reference in New Issue
Block a user