Describe the bug:
cert-manager hang if an ingress with two hosts but twice the same secret is provided.
Example ingress:
spec:
tls:
- hosts:
- endpoint1.test.com
secretName: both-use-the-same
- hosts:
- endpoint2.test.com
secretName: both-use-the-same
This will block the cert-manager instance handling this ingress with an infinite loop of:
I0920 14:00:15.759615 1 controller.go:168] ingress-shim controller: syncing item 'example/example-ingress'
I0920 14:00:15.759640 1 sync.go:140] Certificate "both-use-the-same" for ingress "example-ingress" already exists
I0920 14:00:15.759664 1 sync.go:140] Certificate "both-use-the-same" for ingress "example-ingress" already exists
I0920 14:00:15.759678 1 sync.go:143] Certificate "both-use-the-same" for ingress "example-ingress" is up to date
I0920 14:00:15.953745 1 controller.go:182] ingress-shim controller: Finished processing work item "example/example-ingress"
Expected behaviour:
One of the following:
Steps to reproduce the bug:
Push an ingress with this spec for tls:
spec:
tls:
- hosts:
- endpoint1.test.com
secretName: both-use-the-same
- hosts:
- endpoint2.test.com
secretName: both-use-the-same
Anything else we need to know?:
I tried to find any issue with the same problem / documentation stating that this does not work but couldn't find anything.
Environment details::
v1.15.3quay.io/jetstack/cert-manager-controller:v0.5.2static manifests (via Kubespray addons)/kind bug
Hi there!
v0.5.2 is a very old version of cert-manager now - nearly a year. Would you be able to upgrade to the v0.10 and come back?
You can follow the docs here.
Please re-open/open a new issue once you've tested this with 0.10 - we've had some changes recently that prevent this sort of thing causing an infinite loop 馃槃