Closed slagle closed 2 months ago
Build failed (check pipeline). Post recheck
(without leading slash)
to rerun all jobs. Make sure the failure cause has been resolved before
you rerun jobs.
https://review.rdoproject.org/zuul/buildset/6cef151bb28740809bdb956b2c0f7cf8
:heavy_check_mark: openstack-k8s-operators-content-provider SUCCESS in 42m 58s :x: podified-multinode-edpm-deployment-crc RETRY_LIMIT in 18m 18s :x: cifmw-crc-podified-edpm-baremetal FAILURE in 26m 29s :x: openstack-operator-tempest-multinode RETRY_LIMIT in 21m 48s
/test openstack-operator-build-deploy-kuttl
that seems to trigger another NodeSet reconcile which creates a loop.
Hmm.. Not sure why would service update/patch would trigger a nodeset reconciliation as nodeset controller neither owns or watches services? There is something strange going on for sure, if deleting the MutatingWebhookConfiguration stops the reconcile loop as the webhook configuration would be created again after deletion.
recheck
that seems to trigger another NodeSet reconcile which creates a loop.
Hmm.. Not sure why would service update/patch would trigger a nodeset reconciliation as nodeset controller neither owns or watches services? There is something strange going on for sure, if deleting the MutatingWebhookConfiguration stops the reconcile loop as the webhook configuration would be created again after deletion.
I deleted the webhook from the CSV, so it does not come back, unless you delete/recreate the Subscription.
Either way, I agree this is an unexplained behavior as I didn't think a webhook should trigger another reconcile loop, but it seems to based on the observations.
I deleted the webhook from the CSV, so it does not come back, unless you delete/recreate the Subscription.
Deleting the webhook with oc delete
would bring it back and it also seems to stop the reconcile loop in spite of the webhook re-appearing. Also, if you patch a service manually (not in the nodeset controller as we do), mutating webhook is called and the fields are defaulted, but it does not do the nodeset reconcile and hence the loop.
[APPROVALNOTIFIER] This PR is APPROVED
This pull-request has been approved by: bshephar, jpodivin, slagle
The full list of commands accepted by this bot can be found here.
The pull request process is described here
/cherrypick 18.0.0-proposed
@slagle: new pull request created: #964
I agree this is an unexplained behavior as I didn't think a webhook should trigger another reconcile loop, but it seems to based on the observations.
https://github.com/openstack-k8s-operators/openstack-operator/pull/968
Fixes a reconcile loop by not modifying existing DataPlaneServices in the CreateOrPatch call from EnsureService. If the webhook modifies the service instead, that seems to trigger another NodeSet reconcile which creates a loop.
Jira: OSPRH-8811 Signed-off-by: James Slagle jslagle@redhat.com