Re:Kubernaughty: A collection of documentation, how-tos, too...
by sk0g on Wednesday January 15, 2020 @07:50PM
(#22060595)
Not sure if this is in there, haven't had a chance to take a look yet, but does anyone know a good source (preferably with example) of how to get multiple images to deploy to the same node? From what I can tell I'm after the `nodeAffinity` field, but I found Kubernetes' documentation rather terse.
Re:Kubernaughty: A collection of documentation, how-tos, too...
by sk0g on Wednesday January 15, 2020 @08:51PM
(#22060944)
Do you mean manually? Should clarify, I use Google Kubernetes Engine that manages the deployments for me, and there's also CI/ CD happening in the background that updates the images every so often. Would need to find an automated way of doing this.
Re:Kubernaughty: A collection of documentation, how-tos, too...
by streetcat1 on Wednesday January 15, 2020 @09:46PM
(#22061254)
So there are two types of affinity: Node affinity and Pod affinity. Node affinity would tell the scheduler to prefer a Node. Pod Affinity would tell the scheduler to prefer other Pods location.
Re:Kubernaughty: A collection of documentation, how-tos, too...
by sk0g on Wednesday January 15, 2020 @09:55PM
(#22061292)
I've learned more from your two lines than I did reading through three different entries in the K8s wiki, all of which are rather long. Or rather, it's only clicked now.
Re:Kubernaughty: A collection of documentation, how-tos, too...
by jnoller on Wednesday January 15, 2020 @06:01PM
(#22059783)
I found my password to this site only to reply to this: I am, it's a work in progress, and theres so many different failure modes I'll just keep going.
But if you want several containers to always be deployed together, you want to put them in the same pod. You do this in the deployment yaml.
https://kubernetes.io/docs/concepts/workloads/pods/pod-overv... https://kubernetes.io/docs/tasks/run-application/run-statele...
Thanks!
I wonder if we can take the next step and create some sort of storage troubleshoot operator, that would implement the manual troubleshot algorithm.
Please keep up the great work!