class documentation

class NodeServiceClient(client.Client): (source)

Constructor: NodeServiceClient(channel)

View In Hierarchy

Undocumented

Method set_unhealthy SetUnhealthy marks the node underlying the Compute VM as unhealthy, which has the following effect:
Constant __PB2_DESCRIPTOR__ Undocumented
Class Variable __service_name__ Undocumented

Inherited from Client:

Method __init__ Undocumented
Method request Undocumented
Instance Variable _channel Undocumented
def set_unhealthy(self, request: NodeSetUnhealthyRequest, metadata: abc.Iterable[builtins.tuple[builtins.str, builtins.str]] | None = None, timeout: builtins.float | None = None, credentials: grpc.CallCredentials | None = None, compression: grpc.Compression | None = None, retries: builtins.int | None = 3, per_retry_timeout: builtins.float | None = None) -> request_1.Request[NodeSetUnhealthyRequest, NodeSetUnhealthyResponse]: (source)

SetUnhealthy marks the node underlying the Compute VM as unhealthy, which has the following effect: 1. Scheduler makes the best effort not to assign new VMs to the unhealthy node, but in case of no capacity, the VM can be assigned to an unhealthy node. 2. The existing VMs continue to work on the node, but after stop/start via Compute API they most probably will be assigned to a different node. If the node was already marked unhealthy, the consecutive calls to SetUnhealthy will return grpc code AlreadyExists. To use this rpc one needs to obtain `compute.node.setUnhealthy` permission for the VM's parent container. The permission is granted to the TSA inside the VM. .

__PB2_DESCRIPTOR__ = (source)

Undocumented

Value
descriptor.DescriptorWrap[descriptor_1.ServiceDescriptor]('.nebius.compute.v1.No
deService',
                                                          node_service_pb2.DESCRIPTOR
,
                                                          descriptor_1.ServiceDescriptor
)
__service_name__: str = (source)