This repository has been archived by the owner on Nov 1, 2022. It is now read-only.
Make auto-release commits and events more accurate #1036
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
So that we can (shortly) use the same definition of containers whether we get them the cluster or from a manifest, move the Container type to the `resource`. This also updates the type so that images are parsed when constructed, i.e., in one place.
This commit adds an interface `resource.Workload`, and implements it for the "pod controller" resources. This makes it possible to calculate based on the images used in the resources, rather than those used in the cluster (without implementing exactly that, yet).
We will want to be able to call CollectAvailableImages with values from the cluster (as in `daemon.ListImages(...)`), but also with values from the manifests (when calculating what images can be updated).
Part of the problem being addressed here is that releases decide what to do based on the images used in the cluster, and that does not always correspond to changes in the repo. This commit removes the possibility for automated releases, by - skipping any containers that _already_ use the supposedly new image - building the commit message by looking only at the images that were successfully changed.
tamarakaufler
approved these changes
Apr 4, 2018
| Name string | ||
| Image string | ||
| } | ||
|
|
||
| // Sometimes we care if we can't find the containers for a service, | ||
| // sometimes we just want the information we can get. |
This comment was marked as abuse.
This comment was marked as abuse.
Sorry, something went wrong.
This comment was marked as abuse.
This comment was marked as abuse.
Sorry, something went wrong.
| } | ||
|
|
||
| func (s Controller) ContainersOrNil() []Container { | ||
| func (s Controller) ContainersOrNil() []resource.Container { | ||
| return s.Containers.Containers |
This comment was marked as abuse.
This comment was marked as abuse.
Sorry, something went wrong.
This comment was marked as abuse.
This comment was marked as abuse.
Sorry, something went wrong.
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
As described in #1026, it's possible to calculate a set of image updates that do not end up being carried out, usually because the cluster doesn't quite reflect what's in the git repo.
As a conservative fix that doesn't change behaviour, this PR makes the reporting of auto-releases more accurate by
There's also a bit of tidying and making way for deeper changes -- future work can, for example,
remove the separate preparation and application steps of auto-releases. At the minute, when it's time to do an "automation run", a list of updates is compiled by comparing the images running in the cluster with those available in the image registry. It's applied after going through the job queue; but since it has to look at the git repo in both steps anyway, it may as well just do everything when the job runs.
remove the Calculate and ApplyChanges steps, or at least stop the latter from just splatting bytes into the given file, since that will prevent updates to multidoc files and List resources from working properly.