transfer_replica()
This operation transfers a specific number of replicas between resource groups.
Request Syntax
transfer_replica(
source_group: str,
target_group: str,
collection_name: str,
num_replicas: int,
using: str = "default",
timeout: float | None,
)
PARAMETERS:
source_group (str) -
[REQUIRED]
The name of the source resource group from which the query nodes are moved.
Setting this to a resource group that does not exist results in a MilvusException.
target_group (str) -
[REQUIRED]
The name of the source resource group to which the query nodes are moved.
Setting this to a resource group that does not exist results in a MilvusException.
num_replicas (int) -
[REQUIRED]
The number of replicas to move between the source and target resource groups.
what is a replica?
With replicas, Milvus can load the same segments on multiple query nodes. If one query node has failed or is busy with a current search request when another arrives, the system can send new requests to an idle query node that has a replication of the same segment.
Replicas are organized as replica groups. Each replica group contains shard replicas. Each shard replica has a streaming replica and a historical replica that correspond to the growing and sealed segments in the shard.
Shards can be regarded as DML channels for distributed data write operations among multiple nodes to make the most of the parallel computing potential out of a Milvus cluster.
using (str) -
The alias of the employed connection.
The default value is default, indicating that this operation employs the default connection.
timeout (float | None)
The timeout duration for this operation. Setting this to None indicates that this operation timeouts when any response arrives or any error occurs.
RETURN TYPE:
NoneType
RETURNS:
None.
EXCEPTIONS:
MilvusException
This exception will be raised when any error occurs during this operation.
EXAMPLE:
from pymilvus import (
connections,
Collection,
CollectionSchema,
FieldSchema,
DataType,
utility,
)
# Connect to localhost:19530
connections.connect()
# Create a collection
collection = Collection(
name="test_collection",
schema=CollectionSchema([
FieldSchema("id", DataType.INT64, is_primary=True),
FieldSchema("vector", DataType.FLOAT_VECTOR, dim=5)
])
)
# Get the currently loaded replicas
collection.get_replicas()
# Create a new resource group
utility.create_resource_group(
name="rg_01",
using="default"
)
# Transfer replicas between resource groups
utility.transfer_node(
source_group="__default_resource_group",
target_group="rg_01",
num_nodes=1
)
Related operations
The following operations are related to transfer_replica():