leaves = leaves[1:]
	}
}

// forceCompact will force compact the cache of the top entry.
// If the number of children is more than limit*100, it will compact self.
// When above the limit a cleanup will also be performed to remove any possible abandoned entries.
func (d *dataUsageCache) forceCompact(limit int) {
	if d == nil || len(d.Cache) <= limit {
		return
	}
	top := hashPath(d.Info.Name).Key()

        "cpu": "k8",
        "copt": "-mno-sse4.2",
    },
    visibility = ["//visibility:public"],
)

# This condition takes precedence over :linux_x86_64
# TODO(b/290533709): Remove this with PJRT build rule cleanup.
config_setting(
    name = "linux_x86_64_with_weightwatcher",
    constraint_values = if_google(
        ["//third_party/bazel_platforms/os:linux"],
        [],
    ),

* Update cluster-proportional-autoscaler, etcd-empty-dir-cleanup, fluentd-gcp, and kube-addon-manager addons with refreshed base images containing fixes for CVE-2015-8271, CVE-2016-7543, CVE-2016-9841, CVE-2016-9843, CVE-2017-1000366, CVE-2017-2616, and CVE-2017-7507. ([#48011](https://github.com/kubernetes/ku...

* Fixes issue where subpath volume content was deleted during orphaned pod cleanup for Local volumes that are directories (and not mount points) on the root filesystem. ([#72291](https://github.com/kubernetes/kubernetes/pull/72291), [@msau42](https://github.com/msau42))

## Version 3.14.1

_2019-04-10_

 *  Fix: Don't crash when an interceptor retries when there are no more routes. This was an
    edge-case regression introduced with the events cleanup in 3.14.0.

 *  Fix: Provide actionable advice when the exchange is non-null. Prior to 3.14, OkHttp would
    silently leak connections when an interceptor retries without closing the response body. With

    * [metatada-proxy addon] Bump prometheus-to-sd v0.5.0 to pick up security fixes.

  // by sending a graceful termination signal to the containers in the pod. After that 30 seconds,
  // the Kubelet will send a hard termination signal (SIGKILL) to the container and after cleanup,
  // remove the pod from the API. In the presence of network partitions, this object may still
  // exist after this timestamp, until an administrator or automated process can determine the
  // resource is fully terminated.

* locking fix in vSphere cleanup function ([#68388](https://github.com/kubernetes/kubernetes/pull/68388), [@wgliang](https://github.com/wgliang))
    * The vSphere cloud provider has a cleanup function with a loop,
    * inside of which a write lock was intended to prevent races against

	}

	pr, pw := xioutil.WaitPipe()
	go func() {
		pw.CloseWithError(er.getObjectWithFileInfo(ctx, bucket, object, off, length, pw, fi, metaArr, onlineDisks))
	}()

	// Cleanup function to cause the go routine above to exit, in
	// case of incomplete read.
	pipeCloser := func() {
		pr.CloseWithError(nil)
	}

	if !unlockOnDefer {
		return fn(pr, h, pipeCloser, nsUnlocker)
	}

	if z, ok := objLayer.(*erasureServerPools); ok {
		formattedDisks = z.serverPools[0].GetDisks(0)()
	}

	// Success.
	return objLayer, formattedDisks, nil
}

// removeRoots - Cleans up initialized directories during tests.
func removeRoots(roots []string) {
	for _, root := range roots {
		os.RemoveAll(root)
	}
}

// creates a bucket for the tests and returns the bucket name.