David Gageot

Rolling updates with Swarm plugin stop after the Leader is recreated

3

I have a Swarm with 3 managers and 3 workers. All managers run the swarm flavor plugin with two committed groups: `workers` and `managers`. All nodes, but the leader, were...

Mitigate the risk of creating a LOT of instances or deleting instances that we don't manage?

1

If there is a bug in an Infrakit plugin or a misconfiguration in the templates, Infrakit could create too many instances or it could delete instances it doesn't manage or...

Let's say you have a group of 1 instance. You commit the group with a change that ups the number of instances to 100. Quickly after that, you commit the...

A nice feature for docker swarm would be to create a Swarm without telling the number of managers and workers but only the number of nodes. The two kinds of...

Signed-off-by: David Gageot

--pull can only pull public images

2

The code that runs `docker pull` currently doesn't load the credentials from `~/.docker/config.json`. So only public images can be pulled.

Documentation is missing details about volumes

2

/kind doc This [page](https://github.com/elafros/build/blob/master/builds.md#volumes) describes what kind of volumes can be made available to "Builds". Three types are listed. Secrets, `emptyDir`, host Paths. It would be useful to show a...

No sample shows me how to build a non trivial Go application

4

/kind doc ## Expected Behavior As a developer who's going to use Build CRD to build a Go application, I'd like to have a sample that shows me how to...

Debugging a build description is painful

8

This is more of a feature request. When I work on a build description, I find myself doing those steps over and over again: 1. kubectl delete -f build.yaml --ignore-not-found=true...

Strange logs when the git repo doesn't exist

4

/kind bug ## Expected Behavior If I provide a non-existent git repository as a source. I expect to get a clear error message when the build fails. ## Actual Behavior...