Remove unnecessary retry/delay task parameters
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack-Ansible |
Won't Fix
|
Wishlist
|
Unassigned |
Bug Description
In many, many tasks we make use of delays and retries. This effectively implements mandatory delays in the execution of tasks, slowing the deployment down.
Sometimes the retries are important to do, and something they may not be.
For example - it is really necessary to enact the delay/retry loop for apt, yum and pip installs?
I suspect that our use of of this mechanism is a hangover from Ansible 1.9.x and before where task execution wasn't reliable and our deployment architecture was different (Juno, Icehouse - no repo server).
I think that we could very likely cut execution time down considerably by removing the unnecessary use of the delay/retry/until task options where they are unnecessary.
summary: |
- Audit and evaluate the appropriate use of retries and delays + Remove unnecessary retry/delay task parameters |
Changed in openstack-ansible: | |
assignee: | nobody → Praveen N (praveenn) |
Changed in openstack-ansible: | |
status: | New → Confirmed |
importance: | Undecided → Wishlist |
Changed in openstack-ansible: | |
assignee: | Praveen N (praveenn) → nobody |
This is not necessary any more, especially given the recent surge in apt cache failures.