Implement incremental update for StatefulSet #307
Conversation
lblackstone
force-pushed
the
lblackstone/statefulset-awaiter
branch
from
07905e0
to
19743da
Compare
lblackstone
changed the title
| obj, err := decodeUnstructured(fmt.Sprintf(`{ | ||
| "kind": "StatefulSet", | ||
| "apiVersion": "apps/v1beta1", | ||
| "metadata": { |
There was a problem hiding this comment.
I think we're intermingling spaces and tabs here. I don't think it particularly matters which we pick, but we're using spaces in the deployment examples.
| return obj | ||
| } | ||
|
|
||
| func statefulsetUpdate(namespace, name, targetService string) *unstructured.Unstructured { |
There was a problem hiding this comment.
Is this supposed to be the user-provided inputs that update StatefulSet, or the StatefulSet after the new user inputs are provided? I believe it's the first, in which case is it intentional that .status is missing?
There was a problem hiding this comment.
Yeah, was supposed to be the user-provided StatefulSet
| return obj | ||
| } | ||
|
|
||
| func statefulsetAdded(namespace, name, targetService string) *unstructured.Unstructured { |
There was a problem hiding this comment.
The difference between each of these is kind of subtle -- what do you think about putting a little comment above each to point out the major differences?
pkg/await/apps_statefulset.go
Outdated
| replicasReady bool | ||
| currentGeneration int64 | ||
|
|
||
| statefulsetErrors map[string]string |
There was a problem hiding this comment.
Hmm, doesn't seem like this is actually populated?
There was a problem hiding this comment.
Ha, looks like you're right. Will fix.
| // | ||
| // The success conditions are somewhat complex: | ||
| // | ||
| // 1. `.status.replicas`, `.status.currentReplicas` and `.status.readyReplicas` match the |
There was a problem hiding this comment.
Ah, looks like we also have to check .status.updatedReplicas if it's an update rather than a create?
There was a problem hiding this comment.
I'm using it for the status messages, but it's not actually part of the success condition.
Here's a table that's illustrative (I'll add to the doc as well):
Current replicas Ready replicas Updated replicas Notes
3 3 --
<Update image>
2 3 -- observedGeneration/updateRevision changes
2 2 --
2 2 1
1 3 1
1 2 1
1 2 2
-- 3 2
-- 2 2
-- 2 3
3 3 3 currentRevision updated
3 3 --
lblackstone
force-pushed
the
lblackstone/statefulset-awaiter
branch
from
75b71be
to
1aac165
Compare
|
@hausdorff I think I've addressed all your feedback. |
|
@lblackstone @hausdorff Ready to merge? |
|
@joeduffy This was delayed by kubecon, I told @lblackstone that I'd pick up the remaining work and make the tweaks necessary. |
|
Per my earlier comment, I'm merging this because (a) looks reasonable, and (b) I understand this to be smoke-tested fairly extensively. |
Fixes #300