OrleansContrib / Orleans.Clustering.Kubernetes

Orleans Membership provider for Kubernetes
MIT License
188 stars 47 forks source link

Unable to update Silo Entry #57

Open matthewbrumpton-oddsbods opened 2 years ago

matthewbrumpton-oddsbods commented 2 years ago

We have a large number of KubeMembershipTable exceptions coming from our orleans application, the errors start after the application has been running for 5-9 hours.

Message: Unable to find silo entry 172.19.16.62-50000-374433332.

CategoryName: Orleans.Clustering.Kubernetes.KubeMembershipTable

Exception type: System.InvalidOperationException

Failed method: Orleans.Clustering.Kubernetes.KubeMembershipTable+d__11.MoveNext

Problem Id: System.InvalidOperationException at Orleans.Clustering.Kubernetes.KubeMembershipTable+d__11.MoveNext

Assembly: Orleans.Clustering.Kubernetes, Version=2.0.0.0, Culture=neutral, PublicKeyToken=null

Call Stack: System.InvalidOperationException: at Orleans.Clustering.Kubernetes.KubeMembershipTable+d__11.MoveNext (Orleans.Clustering.Kubernetes, Version=2.0.0.0, Culture=neutral, PublicKeyToken=null)

matthewbrumpton-oddsbods commented 2 years ago

Another observation is OrleansClusterVersion is deleted after a period of time and then being recreated after we restart the app.

OrleansClusterVersion after we restart the app:

apiVersion: v1 items: apiVersion: orleans.dot.net/v1 clusterId: clusterId clusterVersion: 4 kind: OrleansClusterVersion metadata: creationTimestamp: "2021-11-15T14:43:13Z" generation: 5 name: name namespace: namspace resourceVersion: "148545790" uid: d4a66cf5-18a9-4474-bf88-351d496ac61d kind: List metadata: resourceVersion: "" selfLink: ""

OrleansClusterVersion when deleted:

kubectl get clusterversions --namespace namespace -o yaml apiVersion: v1 items: [] kind: List metadata: resourceVersion: "" selfLink: ""