Closed eumel8 closed 3 years ago
@eumel8 have you have chance to test latest 1.22 or master? whether the problem still occurs due to #1252 ?
@jichenjc unfortunatelly the same error
E0906 08:12:18.182017 1 controller.go:307] error processing service ingress-nginx/otc-lb (will retry): failed to delete load balancer: error deleting obsolete pool 3316574c-6bf6-4973-acb8-724af0f7db1d for listener 3f0552a3-d171-4603-9b89-6f2403f70b6f: Expected HTTP response code [202 204] when accessing [DELETE https://vpc.eu-de.otc.t-systems.com/v2.0/lbaas/pools/3316574c-6bf6-4973-acb8-724af0f7db1d], but got 409 instead
{"NeutronError": {"detail": "", "message": "member 91a6b67e-81da-4839-ba1b-511724a3cc00 is using this pool", "type": "EntityInUse"}}
I0906 08:12:18.182079 1 event.go:291] "Event occurred" object="ingress-nginx/otc-lb" kind="Service" apiVersion="v1" type="Warning" reason="SyncLoadBalancerFailed" message="Error syncing load balancer: failed to delete load balancer: error deleting obsolete pool 3316574c-6bf6-4973-acb8-724af0f7db1d for listener 3f0552a3-d171-4603-9b89-6f2403f70b6f: Expected HTTP response code [202 204] when accessing [DELETE https://vpc.eu-de.otc.t-systems.com/v2.0/lbaas/pools/3316574c-6bf6-4973-acb8-724af0f7db1d], but got 409 instead\n{\"NeutronError\": {\"detail\": \"\", \"message\": \"member 91a6b67e-81da-4839-ba1b-511724a3cc00 is using this pool\", \"type\": \"EntityInUse\"}}"
ok, I searched the code in CPO and didn't find it but found in CAPI https://github.com/kubernetes-sigs/cluster-api-provider-openstack/blob/master/controllers/openstackcluster_controller.go#L148 so I assume you are using CAPI + CPO ?
and @lingxiankong do we have any LB test in e2e? if so curious why we didn't find the error maybe due to different env? as #1252 is an optimization so maybe we can revert that and in the mean time dig into what's wrong in @eumel8 's env then provide another fix..
@eumel8 are you using Octavia or Neutron-LBaaS?
and @lingxiankong do we have any LB test in e2e? if so curious why we didn't find the error maybe due to different env? as #1252 is an optimization so maybe we can revert that and in the mean time dig into what's wrong in @eumel8 's env then provide another fix..
Let's grab more info before deciding.
@lingxiankong cloud.conf runs with use-octavia = false
. This would have the only option to run Neutron-LBaaS
Unfortunately, we stopped maintaining Neutron-LBaaS since release 1.17 (or 1.18?)
Is this a BUG REPORT or FEATURE REQUEST?:
/kind bug
What happened:
Since 1.20 release we have problems delete loadbalancer service:
What you expected to happen:
How to reproduce it:
This error occured between release-1.19 and release-1.20. After comparing the commits I spotted https://github.com/kubernetes/cloud-provider-openstack/pull/1252 After reverting this commit loadbalancer will deleted as expected very fast. Not sure how can this more tuned.
Anything else we need to know?:
@sbueringer FYI
Environment: