Nan Yu
Nan Yu
I talked with @sdowell briefly on reviewing this PR. We both agreed that it would be beneficial for the reviewers if this PR can be broken down into multiple smaller...
I confirmed that the `insecure` flag is the root cause of the OCI failure. After adding the flag, the updated `oci-sync` container is able to sync the image from `image:...
> Removing the finalizer on the RepoSync when its parent Namespace is being deleted DOES solve the case where deletion ordering is unimportant. That is the intent of this fix....
> In the case of namespace garbage collection, temporary deletion errors are ok, because it will just keep retrying, giving the RepoSync reconciler a chance to do it's work. In...
/unhold #1194 is related, but a different issue. It shouldn't block this.
Memory usage is not solely determined by the object count. For instance, 1k objects of the same Kubernetes Kind may consume less memory than 100 objects with various kinds. Additionally,...
I accidentally pushed from the wrong fork branch. I'll push a new patch after addressing these comments.
replaced by https://github.com/GoogleContainerTools/kpt-config-sync/pull/1121
I think using the remote default branch makes sense if neither flag is specified, but it behaves differently from v3. v3 forces a default which may not be the same...
> That said, how important is this? It's more about avoiding breaking changes to customers. Config Sync uses git-sync. With this change, the upgrade process won't be seamless. They have...