dataprotection-microsoft-kubernetes-agent-upgrade-crds job killed with OOM

Johannes Göhr 0 Reputation points
2023-12-17T10:36:45.73+00:00

Installing the azure backup extension on a new cluster work well, but we try to install it on an existing cluster.

The job dataprotection-microsoft-kubernetes-agent-upgrade-crds fails every time with a OOM. Our analysis shows, that the job fails after applying the CRDs (they all exists after the first run). It seems that kubectl requires more memory, when there are more CRDs available e.g. azure service operator v2 crds.

grafik

With more memory the job will complete and also the installation can complete successfully, when we delete the job right after the first failure. Is there any configuration we can do to increase the memory limit of the job?

Azure Backup
Azure Backup
An Azure backup service that provides built-in management at scale.
1,490 questions
Azure Kubernetes Service
Azure Kubernetes Service
An Azure service that provides serverless Kubernetes, an integrated continuous integration and continuous delivery experience, and enterprise-grade security and governance.
2,447 questions
{count} votes

1 answer

Sort by: Most helpful
  1. Johannes Göhr 0 Reputation points
    2023-12-18T18:00:12.8266667+00:00

    I found a solution, there is a parameter upgradeJobResources.limits.memory for the job. But I don't know if it's possible with az. I'm using terraform.

    The result looks good.

    grafik


Your answer

Answers can be marked as Accepted Answers by the question author, which helps users to know the answer solved the author's problem.