Default timeout for service creation in k8s bootstrap on openstack too short and not configurable

Bug #1920835 reported by Michael Skalka
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Canonical Juju
Fix Released
High
Yang Kelvin Liu

Bug Description

Attempting to bootstrap a kubernetes controller on top of Ussuri Focal Openstack using an Octavia loadbalancer the Juju 2.8.10 candidate. Also verified the same behavior on 2.9/beta. The LB comes online but only after the 3 minute default timeout for the service creation, causing the bootstrap to fail.

"bootstrap-timeout" in this case was set to 1800 seconds, which is more than enough time for the amphora to come online, but seems to have no effect on the services timeout. This should be a configurable timeout or should respect the overall "bootstrap-timeout" value.

Logs:

2021-03-22-22:49:21 root DEBUG ['juju', 'bootstrap', '--config', 'bootstrap-timeout=1800', '--model-default', '/home/ubuntu/project/generated/model_defaults.yaml', 'kubernetes_cloud', 'foundations-k8s']
2021-03-22-22:49:21 root DEBUG [localhost]: juju bootstrap --config bootstrap-timeout=1800 --model-default /home/ubuntu/project/generated/model_defaults.yaml kubernetes_cloud foundations-k8s
2021-03-22 22:49:23.005716 Creating Juju controller "foundations-k8s" on kubernetes_cloud
2021-03-22 22:49:23.005872 Bootstrap to Kubernetes cluster identified as openstack
2021-03-22 22:49:25.132280 Fetching Juju Dashboard 0.3.0
2021-03-22 22:49:25.153648 Creating k8s resources for controller "controller-foundations-k8s"
2021-03-22 22:52:23.732358 ERROR failed to bootstrap model: creating controller stack: creating service for controller: max duration exceeded: controller service address not provisioned

Despite this about twenty seconds later:

+--------------------------------------+--------------------------------------------------------------------------------------------------------+----------------------------------+---------------+---------------------+------------------+----------+
| id | name | project_id | vip_address | provisioning_status | operating_status | provider |
+--------------------------------------+--------------------------------------------------------------------------------------------------------+----------------------------------+---------------+---------------------+------------------+----------+
| cc35b8fe-fe6a-44d5-a5ff-5277b3f43571 | kube_service_kubernetes-pbpdwh0lgfgtsv2cu7fvvzwn7tk1fjfq_controller-foundations-k8s_controller-service | 678e60a33ad14f34811b287825ac4016 | 10.244.32.127 | ACTIVE | OFFLINE | amphora |
+--------------------------------------+--------------------------------------------------------------------------------------------------------+----------------------------------+---------------+---------------------+------------------+----------+

Ian Booth (wallyworld)
Changed in juju:
milestone: none → 2.8.11
status: New → Triaged
importance: Undecided → High
Revision history for this message
Michael Skalka (mskalka) wrote :

Ian,

Can you confirm there is no workaround or configurable on the controller to alleviate this? We intend to expand the release gate for Juju to include a test which covers this scenario in the future as it's critical to the operation of both Kubeflow and OSM on top of Openstack.

Michael Skalka (mskalka)
Changed in juju:
status: Triaged → New
Changed in juju:
status: New → Triaged
assignee: nobody → Yang Kelvin Liu (kelvin.liu)
Changed in juju:
status: Triaged → In Progress
Revision history for this message
Yang Kelvin Liu (kelvin.liu) wrote :

https://github.com/juju/juju/pull/12817 will be landed to 2.8 to fix this issue

Ian Booth (wallyworld)
Changed in juju:
status: In Progress → Fix Committed
Changed in juju:
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.