I am trying to deploy Mastodon to a GCP GKE Autopilot cluster. The Helm chart is failing due to container sysctl calls somewhere getting blocked because GKE does not allow them.
I don't know where to look to investigate where these calls are being made to see if I might avoid them from being made. Any ideas?
@kfekete That part is confirmed by GCP. I’m trying to find out why/where Mastodon is making them because they don’t seem like they should be necessary. The logs don’t point me anywhere unfortunately.
@Jeremiah Sounds like a perfect use-case for eBPF! You can drop execsnoop into the cluster and you could catch what process is trying to execute that. It's quite simple to do with our (solo.io) open-source project called BumbleBee, however I don't think we pushed an execsnoop example already to our repository. How urgent is this? :)
@mhamzahkhan Thanks for the reply! I am using GCP’s managed Redis and disabled it in the Helm chart. I am deploying Elasticsearch but it fails on the Mastodon containers. I also tried disabling the Elasticsearch container.
@Jeremiah Have you tried running the helm template with your values file? Then you can see through the rendered yaml to see where the sysctl container is being created?
Add comment