6.5.0 (2023-03-10)

[manusa]

Bugs

• Fix Race condition triggered in java-generator for very big schemas #4723: (java-generator) Fix a race in the use of JavaParser hitting large CRDs
• Fix json-schema draft-05 (404). Update to draft-07 #4784: json-schema draft-05 is a (404). Updating to draft-07 (http://json-schema.org/draft-07/schema#)
• Fix Unhandled NullPointerException in loadRelated #4791: handle the NullPointerException in Thread.currentThread().getContextClassLoader()
• Fix Add a convenient way to make sure a SharedIndexInformer is done watching #4794: improving the semantics of manually calling informer stop
• Fix Restore ability to make OkHttp's Dispatcher use daemon threads #4797: OkHttpClientFactory.additionalConfig can be used to override the default OkHttp Dispatcher
• Fix LeaderElection: LockException when trying to release leadership status #4798: fix leader election release on cancel
• Fix Windows-specific .kube/config file can't be deserialized (autoConfigured Client can't be instantiated) #4803: Windows .kube/config file with spaces in command can be deserialized
• Fix [java-generator] Fix the download directory creation steps #4815: (java-generator) create target download directory if it doesn't exist
• Fix Java generation from CustomResourceDefinition yaml: descriptions containing "*/" lead to compile error #4818: (java-generator) Escape */ in generated JavaDocs
• Fix Java generation from CustomResourceDefinition yaml: field names containing dots, whitespace or quotes lead to compile error #4823: (java-generator) handle special characters in field names
• Fix NO_PROXY rule with CIDR containing a bit suffix <10 appears to be ignored in the fabric client #4832: NO_PROXY can match cidr with bit suffix <10
• Fix pod read unable to distinguish between empty result and missing file #4846: allowed for pod read / copy operations to distinguish when the target doesn't exist
• Fix WebSocket send should account for buffer reuse #4851: adding buffer cloning to ensure buffers cannot be modified after sending
• Fix JDK client hanging on PodIT #4885: addresses a potential hang in the jdk client with exec stream reading
• Fix Initial resourceVersion if initialState is used #4888: narrowing where the 0 initial list resourceVersion is used for informers - in particular if a limit is set or initialState is used, then we should not use 0. Additionally for the informOnCondition / wait methods we'll also not use 0 - it's not expected that the user should test any state prior to the latest.
• Fix vertx may not complete read exec stream #4891: address vertx not completely reading exec streams
• Fix fix: BuildConfigs.instantiateBinary().fromFile() does not time out #4899: BuildConfigs.instantiateBinary().fromFile() does not time out
• Fix vertx httpresponse reference request headers #4908: using the response headers in the vertx response
• Fix Issues with upload / failures with PodIT #4910: addressing inconsistent behavior with pod exec operations
• Fix Failures with PodIT.copyFile #4923: addressing inconsistent behavior with pod exec operations
• Fix fix #4924: handles non-okhttp clients with invalid watch status #4928: allows non-okhttp clients to handle invalid status
• Fix Better mock server concurrency handling for put / patch processing #4931: using coarse grain locking for all mock server operations
• Fix fix: typo in HttpClient.Factory scoring system logic #4947: typo in HttpClient.Factory scoring system logic
• Fix JDK issue with buildconfig #4954: disabling expect continue support by default in the jdk client

Improvements

• Fix Set a timeout client side in informer watches #4675: adding a fully client side timeout for informer watches
• Fix DeletionTimestamp and Finalizer support in Mock server #3805: DeletionTimestamp and Finalizer support in Mock server.
• Fix LeaderElection: Lease resource left behind #4638: adding a way to set the full object meta on a leadership election lock, this can be used to set owner references
• Fix improve: generate crds in parallel and optimize code #4644: generate CRDs in parallel and optimize code
• Fix Remove SupportTestingClient interface #4659: added a generic support(apiversion, kind) method in addition to the class based check
• Fix Private configuration classes cause trouble with Java native (reflection) #4724: Private configuration classes cause trouble with Java native (reflection)
• Fix Mock server doesn't handle optimistic concurrency control #4739: honor optimistic concurrency control semantics in the mock server for PUT and PATCH requests.
• Fix Possible to improve the usage of snakeyaml due to CVE-2022-1471 #4747: migrate to SnakeYAML Engine
• Fix Move exponential backoff handling to standard client layer #4788: moved retry logic into the standard client so that it applies to all requests, including websockets
• Fix Don't print warning message when service account token property is unset #4795: don't print warning message when service account token property is unset
• Fix Question: How to setup the client to access cluster wide (custom) resources #4800: (java-generator) Reflect the scope field when implementing the Namespaced interface
• Fix Pod ready wait missing from log operations #4853: adding a wait on the pod for log operations
• Fix fix: disable Vertx DNS resolver when creating Vertx instance #4848: Vert.x async DNS resolver is disabled
• Fix DEFAULT_REQUEST_RETRY_BACKOFFLIMIT value #4863: default HttpClient retry logic to 100ms interval
• Fix DEFAULT_REQUEST_RETRY_BACKOFFLIMIT value #4863: default HttpClient retry logic to 10 attempts
• Fix Improve performance of the java-generator #4865: (java-generator) performance improvements
• Fix Unable to create a pod using example YAML InputStream with client >=6.4 #4873: Update all samples in kubernetes-examples/ module to use up to date code
• Fix URLFromIngressImpl doesn't consider Ingress in networking.k8s.io apiGroup #4906: URLFromIngressImpl considers Ingress in networking.k8s.io apiGroup while resolving Ingress

Dependency Upgrade

• Fix Upgrade Fabric8 Kubernetes Model to Kubernetes v1.26 #4655: Upgrade Fabric8 Kubernetes Model to Kubernetes v1.26.0
• Fix The attribute of ‘responseHeaderModifier’ is not in class named io.fabric8.kubernetes.api.model.gatewayapi.v1beta1.HTTPRouteFilter #4804: Update Kubernetes Model GatewayApi to v0.6.0
• Fix The attribute of ‘responseHeaderModifier’ is not in class named io.fabric8.kubernetes.api.model.gatewayapi.v1beta1.HTTPRouteFilter #4804: Update CertManager Model to v1.11.0

New Features

• Fix Add pod ephemeral containers operations #4758: added support for pod ephemeral container operations
• Fix Add kustomize api #4929: added support for kustomize api v1beta1

Note: Breaking changes

• Fix Refine what is possible via withRequestConfig #4708: The signature of the Interceptor methods changed to pass the full HttpRequest, rather than just the headers, and explicitly pass request tags - in particular the RequestConfig. To simplify authentication concerns the following fields have been removed from RequestConfig: username, password, oauthToken, and oauthTokenProvider. Not all HttpClient implementation support setting the connectionTimeout at a request level, thus it was removed from the RequestConfig as well.
• Fix Remove SupportTestingClient interface #4659: The SupportTestingClient interface has been deprecated. Please use one of the supports methods or getApiGroup to determine what is available on the api server.
• Fix config.refresh() erases token specified when building initial config #4802: to ease developer burden, and potentially bad behavior with the vertx client, the callbacks for Watcher and ResourceEventHandler will no longer be made by an HttpClient thread, but rather from the thread pool associated with the KubernetesClient. Please ensure that if you are customizing the Executor supplied to the client that it has sufficient threads to handle these callbacks.
• Fix fix #4788 moving retry logic into the standard layer #4825: removed or deprecated/moved methods that are unrelated to the rolling timeout from ImageEditReplacePatchable. Deprecated rollout methods for timeout and edit - future versions will not support
• Fix Config / RequestConfig rollingTimeout is unused #4826: removed RequestConfig upload connection and rolling timeouts. Both were no longer used with no plans to re-introduce their usage.
• Fix Revisit resourceVersion and locking behavior #4861: several breaking changes related to resourceVersion handling and the replace operation:
  • replace is deprecated, you should use update instead. If you set the resourceVersion to null it will not be optimistically locked.
  • createOrReplace is deprecated, you should use server side apply instead.
  • edit uses now optimistic locking by default. To disable locking you should change your methods to follow this pattern: .edit(pod -> new PodBuilder(pod).editMetadata().withResourceVersion(null)//...
  • JSON patch methods using an item for the diff generation such as edit or patch will no longer omit the resourceVersion in the patch. If you want the patch to be unlocked, then set the resourceVersion to null on the item to be patched.
  • internal logic to mimic an apply that modify an item prior to a JSON patch is deprecated - you should instead build the item to be patched off of base version, such as with the edit method.

Note: Deprecations

• Fix [java-generator] Remove unused generation options #4875: Deprecated all the java-generator options identified for deletion in a following release

Full Changelog: v6.4.1...v6.5.0

---

This discussion was created from the release 6.5.0 (2023-03-10).

[maheshreddy77]

Regarding "createOrReplace is deprecated, you should use server side apply instead"
Can you please help share working sample method to use "server side apply instead" using this client

[rohanKanojia]

It should be available on resource(). Something like this:

client.network().v1().ingresses().inNamespace("default").resource(ing).serverSideApply();