Hi All,
AWX can be deployed on K8s which can provide a measure of high availability by increasing the pod replica count above 1 and spreading pods across multiple K8s nodes. Is there a higher order way of creating AWX redundancy with more than one deployment, with each in different clusters being configurable to be redundant for each other?
Rod
I have come across https://github.com/ansible/awx/blob/22.5.0/docs/clustering.md which describes clustering. This document refers to communication between AWX control nodes in a cluster. Is this deployable via AWX Operator? If so, how can it be done?
The linked setup is the default deployment of AWX. As described in the linked site, scaling up is accomplished by increasing the amount of pods within AWX´s replicaset. The pods from that replicaset belonging to one AWX deployment form a cluster automatically and will therefore manage the same queue of automation jobs.
Hi Tian,
Thanks for the response. My misunderstanding, I thought that high availability concept that it described was to permit AWX instances running on different clusters to be able to be in sync. Is there any way to may this happen?
Best regards
Rod
As far as I know, clustering AWX instances beyond (openshift) clusters is not supported.
Our attempt to make AWX as HA as possible ist to keep each AWX instance related pod on different nodes so on the event of node drainage, never more than one pod shall restart at a time