- Failed pods associated with a job with `parallelism = 1` are recreated by the job controller honoring exponential backoff delay again. However, for jobs with `parallelism > 1`, pods might be created without exponential backoff delay. ([#115020](https://github.com/kubernetes/kubernetes/pull/115020), [@nikhita](https://github.com/nikhita)) [SIG Apps]

      scope, OutputsFromTFOutputs(parent_inputs, num_loop_vars), cond_fn,
      body_fn, params->name, &loop_outputs);

  // Update name_map with newly-created ops.
  // TODO(skyewm): right now BuildWhileLoop() may alter the graph if it returns
  // a bad status. Once we fix this, we may want to return early instead of
  // executing the following code.

                              &Deallocator, &deallocator_called);

  TF_Tensor* o = TF_TensorMaybeMove(t);
  ASSERT_TRUE(o == nullptr);  // It is unsafe to move memory TF might not own.
  TF_DeleteTensor(t);
  EXPECT_TRUE(deallocator_called);
}

TEST(CAPI, LibraryLoadFunctions) {
  // TODO(b/73318067): Fix linking for the GPU test generated by the

remaining limbs, reducing modulo m each time. for i >= 0 { out.shiftIn(x.limbs[i], m) i-- } return out } // ExpandFor ensures x has the right size to work with operations modulo m. // // The announced size of x must be smaller than or equal to that of m. func (x *Nat) ExpandFor(m *Modulus) *Nat { return x.expand(len(m.nat.limbs)) } // resetFor ensures out has the right size to work with operations modulo m. // // out is zeroed and may start at any size. func (out *Nat) resetFor(m *Modulus) *Nat { return...

					Object:     objInfo,
					UserAgent:  "Internal: [Replication]",
					Host:       globalLocalNodeName,
				})
			}
			// object with same VersionID already exists, replication kicked off by
			// PutObject might have completed
			if objInfo.TargetReplicationStatus(tgt.ARN) == replication.Pending ||
				objInfo.TargetReplicationStatus(tgt.ARN) == replication.Failed ||
				ri.OpType == replication.ExistingObjectReplicationType {

remaining limbs, reducing modulo m each time. for i >= 0 { out.shiftIn(x.limbs[i], m) i-- } return out } // ExpandFor ensures x has the right size to work with operations modulo m. // // The announced size of x must be smaller than or equal to that of m. func (x *Nat) ExpandFor(m *Modulus) *Nat { return x.expand(len(m.nat.limbs)) } // resetFor ensures out has the right size to work with operations modulo m. // // out is zeroed and may start at any size. func (out *Nat) resetFor(m *Modulus) *Nat { return...

    * action required: Cloud providers other than GCE might want to update their deployments or sample yaml files to reuse the role created via add-on.

### Other notable changes

	In Namespaces that previously did not have the "DefaultDeny" annotation, you should delete any existing NetworkPolicy objects. These had no effect in the v1beta1 API, but with v1 semantics they might cause some traffic to be unintentionally blocked.


### Storage

- Failed pods associated with a job with `parallelism = 1` are recreated by the job controller honoring exponential backoff delay again. However, for jobs with `parallelism > 1`, pods might be created without exponential backoff delay. ([#115021](https://github.com/kubernetes/kubernetes/pull/115021), [@nikhita](https://github.com/nikhita)) [SIG Apps]

  * Note that because any taints you might have created with Kubernetes 1.5 can only affect the scheduling of new pods (the `NoExecute` taint effect is introduced in 1.6), neither the master upgrade nor your running `kubectl taint` to re-create the taints will affect pods that are already running....