Use role instead of playbooks - 99-logs.yml#3032
Conversation
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
fmount
left a comment
There was a problem hiding this comment.
This seems more converging the entire ci-framework logging topic into the existing must-gather role.
I see this as a wrong move because we risk to confuse the actions we execute in the role that shouldn't break the principle of managing a single thing (which is running openstack-must-gather).
I'm not sure why removing a playbook and unpacking it into this role results in a simplification, but I see this as an attempt to rename must-gather into a logging role, while we should move the common code into a logging role.
|
Build failed (check pipeline). Post https://softwarefactory-project.io/zuul/t/rdoproject.org/buildset/91ae268010a94a25ada04e2d938a09ca ✔️ openstack-k8s-operators-content-provider SUCCESS in 1h 37m 33s |
If we are watching on that from this PR - no, that "simplification" is just an illusion, but if we check another playbooks what we were executing - then yes, it is simplification. Import playbook does not give transparent workflow, where moving that to roles:
About this PR: So if you think that openstack-must-gather is better for collecting logs, than let's remove the 99-logs play and use it directly. Is it ok @fmount ? |
@danpawlik ack, my point is simply related to the fact that I see tasks moved to the must-gather role that are not related to |
|
@fmount At this point my suggestion is:
Alternative way:
is it ok @fmount ? |
The alternative way sounds good. You should include must-gather where you need, not the other way around. |
2a9af59 to
b800a34
Compare
|
Build failed (check pipeline). Post https://softwarefactory-project.io/zuul/t/rdoproject.org/buildset/d9f0fc0886eb4e5c9abddbedf23dc974 ✔️ openstack-k8s-operators-content-provider SUCCESS in 36m 49s |
b800a34 to
fdae026
Compare
|
Build failed (check pipeline). Post https://softwarefactory-project.io/zuul/t/rdoproject.org/buildset/2059a3510eb943a49089abdeb5ebec37 ✔️ openstack-k8s-operators-content-provider SUCCESS in 36m 14s |
fdae026 to
f9a59bd
Compare
f9a59bd to
d4b8e4f
Compare
|
Build failed (check pipeline). Post https://softwarefactory-project.io/zuul/t/rdoproject.org/buildset/28d7c2413989454c8cedd6f6eb9980ba ❌ openstack-k8s-operators-content-provider FAILURE in 4m 04s |
d4b8e4f to
3362125
Compare
|
Build failed (check pipeline). Post https://softwarefactory-project.io/zuul/t/rdoproject.org/buildset/1064f149b47f4ecea216a2a2a8181d62 ❌ openstack-k8s-operators-content-provider FAILURE in 4m 03s |
3362125 to
6f96679
Compare
|
IMO, We can't just replace the 99-logs.yml playbook with os_must_gather role (as suggested in the comments) as we are calling roles inside the 99-logs.yml so created new playbook in the ci_setup role name - run_logs.yml to replace 99-logs.yml . As it's specific to our cifmw related task so better to add that in the ci_setup role instead of calling that in the other roles like os_must_gather |
|
Build failed (check pipeline). Post https://softwarefactory-project.io/zuul/t/rdoproject.org/buildset/532a9d55d21c494e8dc0f1549a6fffb3 ✔️ openstack-k8s-operators-content-provider SUCCESS in 1h 53m 13s |
|
recheck POST_FAILURE |
|
recheck |
6f96679 to
b302564
Compare
|
Build failed (check pipeline). Post https://softwarefactory-project.io/zuul/t/rdoproject.org/buildset/3f8233ac2ecd45848132f5bc3d601308 ✔️ openstack-k8s-operators-content-provider SUCCESS in 1h 42m 27s |
|
This PR is stale because it has been for over 15 days with no activity. |
It is continuation of simplification job execution [1]. [1]: #2929
9b277a7 to
08d6230
Compare
|
please check @fmount if its ok |
|
I think this patch needs a follow up to run the must-gather role on the existing jobs :/ |
After merging[1], jobs have stopped running must-gather to collect service logs, e.g [2,3]. This seems to be caused by the change in [4]. Where before the 99-logs.yml playbook was run from an 'ansible-playbook' invocation, without any inventory and thus the check for 'zuul_log_collection' was never triggered because the variable was never set. Now, when importing the cifmw_setup role, the cifmw variables are loaded and that variable which is in most jobs set to True prevents the log collection, and most critically, must-gather collection. This change removes the condition to preserve the previous behaviour. [1] openstack-k8s-operators#3032 [2] https://softwarefactory-project.io/zuul/t/rdoproject.org/build/63673310b4f94883a9aff2db719a2095 [3] https://softwarefactory-project.io/zuul/t/rdoproject.org/build/03e87102fbf04dc7b482ecdba7128085 [4] openstack-k8s-operators@cc87f54#diff-691dae244b4fb4bfc1270fb5d3b708c6be6073b0e33d997db15850edfa3fd976L23
It is continuation of simplification job execution 1.