CODEWITHSUNDEEP
javakubernetesinfinispanfault
Mauri
Mauri

Reputation: 41

Handle Infinispan POD Faults

I'm using the Java version of the Infinispan library.

I'm deploying Infinispan in two Kubernetes PODs and I have defined some replicated caches. Moreover, I have some microservices using such caches, connecting toward an Infinispan POD. Of course, the connection is made using the service name and the port. Indeed, my microservices are able to connect to one POD instance successfully at the end for using the remote caches.

Unluckily, when a POD had some fault and restarts, my microservice is not able to switch automatically to another replica POD of Infinispan, and issues begin to happen.

It's even worse when both POD restarts (I'm causing the restarts for robustness tests).

I was wondering if there's a way to notice when the POD a client is connected to becomes not available. For example, I use some Infinispan listeners to notice when certain operations are performed over the maps. Is there something similar for the service health? Maybe something keeping track of the distributed topology of the Infinispan cluster?

In case there's not, how do you handle these cases?

Upvotes: 0

Views: 142

Answers (0)

Related Questions