Add connection check before configuring tripleo-admin
This change adds a connection check to the "Run create admin" playbook which will ensure the target host is up and ready to recieve connection prior to attempting to execute the rest of the playbook. This is being done because there is a chance servers may not be up and available by the time this playbook is executed and we dont want to cause a bad user experience by failing early, forcing the deployer to rerun a deployment for no reason. Closes-Bug: #1873246 Change-Id: I2d2e3bdcaf0a31d8daf90656533ac0f55b70fa26 Signed-off-by: Kevin Carter <kecarter@redhat.com>
This commit is contained in:
parent
1c73d3ac64
commit
e50d06bbd0
|
@ -190,6 +190,24 @@
|
|||
hosts: localhost:tripleo_queues
|
||||
become: true
|
||||
any_errors_fatal: true
|
||||
gather_facts: false
|
||||
pre_tasks:
|
||||
# NOTE(cloudnull): The connection will allow for 40 minutes before failing. This time was chosen
|
||||
# because a server may take anywhere from 5 to 40 minutes to boot, and in large
|
||||
# deployments the number of "forks" may not accomodate all nodes running this
|
||||
# task in parallel. Because we don't know the all of the characteristics of the
|
||||
# machine being used, there's no way to compute the value appropriate for a
|
||||
# given node(s), so 40 minutes should accommodate most environments.
|
||||
- name: Wait for connection to become available
|
||||
wait_for_connection:
|
||||
sleep: 4
|
||||
timeout: 2400
|
||||
|
||||
- name: Gather facts with an active connection
|
||||
setup:
|
||||
gather_subset:
|
||||
- '!facter'
|
||||
- '!ohai'
|
||||
roles:
|
||||
- role: tripleo_create_admin
|
||||
tripleo_admin_user: tripleo-admin
|
||||
|
|
Loading…
Reference in New Issue