F5Networks / k8s-bigip-ctlr

Repository for F5 Container Ingress Services for Kubernetes & OpenShift.
Apache License 2.0
364 stars 195 forks source link

RFE: avoid multi-cluster CIS split brain situations when using redundant CIS instances #3415

Open alonsocamaro opened 6 months ago

alonsocamaro commented 6 months ago

Title

RFE: avoid CIS split brain situations when using redundant CIS instances

Description

Created as per customer request

At present, there is no mechanism that avoids a Secondary CIS to push the configuration to BIG-IP when it believes that the Primary CIS is down but it is not down. This situation is called in the industry split brain and introduces many problems. This situation should be avoided by design.

Actual Problem

Two CIS instances submitting configurations simultaneously:

Solution Proposed

To avoid this situation I don't think there is a better solution than a logic in the BIG-IP to be the one that decides from which CIS it wants to receive the configuration push.

This RFE affects current work https://github.com/F5Networks/k8s-bigip-ctlr/issues/3354