OOMKilled during install/upgrade ServiceCatalog in cluster with Helm v3
jasiu001 opened this issue · comments
What happened:
When the service-catalog release is upgraded by helm in version 3, there is a rapid increase in the memory requirement on controller-manager Pod (which may cause OOMKilled state if the set memory limit is insufficient).
The more releases are on the cluster, the more memory consumption is increased. This is because Helm releases are kept in Secrets resources and controller-manager listens to those resources.
What you expected to happen:
I propose to make a setting in which the controller-manager ignores specific secrets after giving the appropriate labels.
How to reproduce it (as minimally and precisely as possible):
Create a cluster with helm v3 releases (more of them) and make a service-catalog install/upgrade with a default memory limit.
Environment:
- service-catalog version: v0.3.0
- helm version: v3.2.1