Migrate replication controllers to generic etcd #5746
Conversation
|
Still need to do a little bit of cleanup and write a few more uts, but for the most part I'm convinced this is OK. |
|
@lavalamp is out until Wednesday (I believe) and I feel bad explicitly giving work to @smarterclayton (as he's in a different org), so temporarily assigning to @davidopp; feel free to redistribute. |
|
I'll review - I'm in that part of the codebase right now anyway. Don't few bad assigning things to me, but it's somewhat like assigning to Tim or Brian in terms of our ability to review :)
|
|
I am happy to review as well. I tend to have more time than Clayton and also know this area well :-) Sent from my iPhone
|
|
Clayton or Derek, if you want to review this please assign it to yourself, otherwise I will review i tthis week. |
|
|
||
| Helper: h, | ||
| } | ||
| store.CreateStrategy = controller.Strategy |
There was a problem hiding this comment.
Move these to struct init fields
There was a problem hiding this comment.
I'm going to add a subresource for /status, but I guess I'll just do that in the next cl.
There was a problem hiding this comment.
Ok, that's fine then. Leave it.
----- Original Message -----
KeyFunc: func(ctx api.Context, name string) (string, error) {return etcdgeneric.NamespaceKeyFunc(ctx, controllerPrefix, name)},// Retrieve the name field of a replication controllerObjectNameFunc: func(obj runtime.Object) (string, error) {return obj.(*api.ReplicationController).Name, nil},// Used to match objects based on labels/fields for list and watchPredicateFunc: func(label labels.Selector, field fields.Selector)return controller.MatchController(label, field)},EndpointName: "replicationControllers",
Helper: h,- }
- store.CreateStrategy = controller.Strategy
I'm going to add a subresource for /status, but I guess I'll just do that in
the next cl.
Reply to this email directly or view it on GitHub:
https://github.com/GoogleCloudPlatform/kubernetes/pull/5746/files#r26963687
There was a problem hiding this comment.
Changed anyway, things might happen that prevent that cl from going in so it's cleaner this way.
|
Two minor comments |
bprashanth
force-pushed
the
rc_genericetcd_interface
branch
from
344fae6 to
eda8716
Compare
bprashanth
force-pushed
the
rc_genericetcd_interface
branch
from
eda8716 to
ef66016
Compare
|
PTAL |
|
LGTM |
smarterclayton
added
the
lgtm
|
Haven't run e2e post rebase yet |
|
e2e done, this should be good to go |
Migrate replication controllers to generic etcd
The following will silently happen as a byproduct of the migration:
Listwon't list try to match labels in controller/rest (because I deleted most of the file).PodLister.Also dropped controller tick to 5s, because if a pod dies it almost always does so within this time