HA deployment of Solace on Google Kubernetes Engine (GKE) cluster
Hi,
I'm trying to perform an HA deployment of Solace (with Primary, Backup and Monitor nodes) in a cluster.
ref: https://github.com/SolaceProducts/pubsubplus-gke-quickstart
I want to assign a smaller CPU size of node for the monitor node, and regular CPU size of node for primary and backup nodes.
Supposed that I have already created two set of node pools in the cluster (one is 4 vCPU 15GB memory, another one is 1vCPU 3.75GB memory).
Does anyone have experience on how to separately/manually assign the monitor node in the node pool with the smaller size?
Thanks
Comments
-
The current Kubernetes quickstart indeed doesn't support specifying different pod size for the monitor node. There may be hacks to do this in your environment but the results would be brittle. The way we plan to implement this will be using kubernetes mutating admission webhooks.
0 -
@Balazs said:
The current Kubernetes quickstart indeed doesn't support specifying different pod size for the monitor node. There may be hacks to do this in your environment but the results would be brittle. The way we plan to implement this will be using kubernetes mutating admission webhooks.Hi, thanks for your advice. Could you have more explanation on how to apply kubernetes mutating admission webhooks on my case?
0