Open jnatten opened 3 months ago
After some investigation i figured out the problem goes away if i add a rule to our security group that allows all tcp traffic from control-plane to worker nodes.
Not sure why it is needed or why it worked without previously, but this seems to solve the issue consistently for us for now. Is there a specific port that is needed?
@davinchia can you check if this issue?
@jnatten strange. Does your cluster have special security rules set up? We run Airbyte Cloud on EKS and have never seen this issue.
Not sure if they are special, but the previous security group setup were something like this:
Worker node -> Cluster: 443 Worker node -> Worker node 53,1025 - 65535 Cluster -> worker node: 443,4443,6443,8443,9443,10250 Worker node -> outside world: all open
Think all of it is from the terraform eks module, but i could be wrong on that.
After allowing all ports from cluster -> worker nodes it started working. Not sure if we need all or just some, but i don't think its an issue for us to keep them open.
Happened to me after trying to upgrade a cluster. Had to helm uninstall and re-install and then it worked fine.
Helm Chart Version
0.445.3
What step the error happened?
During the Sync
Relevant information
When running a sync-job. Creating a new destination or anything that spawns a new pod the frontend complains about unknown error (HTTP 504) and The provided log appears.
I have a similar test-cluster with the exact same configuration that works just fine. And I have attempted to install a completely fresh airbyte install in a new namespace.
Running on AWS EKS if it matters.
Any suggestions on how to fix it or how i should continue debugging would be greatly appreciated!
Relevant log output