[BUG] ALBs for Ingress might be recreated while updating corresponding IngressClass's spec.parameters #2731
Labels
kind/bug
Categorizes issue or PR as related to a bug.
priority/critical-urgent
Highest priority. Must be actively worked on as someone's top priority right now.
Describe the bug
Users using the default provided "alb" IngressClass and are currently with controller installed by HelmChart with version 1.3.1-1.3.3 could be impacted when upgrade to version 1.4.0-1.4.2.
Since when upgrading the chart, there is a change that will configure the default IngressClass
alb
to reference a newly added IngressClassParams namedalb
, but thealb
IngressClassParams is created after thealb
IngressClass modification.Current mitigation is set
--set ingressClassParams.create=false
if you were upgrading from HelmChart with version 1.3.1-1.3.3If the IngressGroup's
spec.parameters
was updated, the ALBs for Ingresses might be recreated due to a race condition.When updating the IngressGroup's
spec.parameters
referencing a "non-exists" IngressClassParams, the controller will treat the Ingress using that IngressGroup was invalid, thus remove the rules/ALB for that Ingress.Steps to reproduce
Create ALBs using a IngressClass, update the IngressClass to reference an un-exists IngressClassParams.
Expected outcome
The AWSLoadBalancerController should stop reconcile and ask users to correct configuration errors if a previously managed Ingresses's IngressClass was changed to be in an invalid state, instead of just treat the Ingress as invalid.
This is necessary since even if the IngressClassParam is created before the IngressClass modification(but within a tiny period), the Controller might receive the IngressClassParam creation event after the IngressClass modification.
Environment
Additional Context:
The text was updated successfully, but these errors were encountered: