Container deployment method and apparatus
A container deployment method and apparatus are disclosed. The method includes receiving a first container creation request that carries a required resource capacity and a required image identifier, where the required image identifier is an identifier of an image corresponding to an application that needs to be run. From a cluster of working nodes, at least two first working nodes whose unused resource capacity is greater than the required resource capacity are selected. A target working node is determined from the at least two working nodes according to images stored on the at least two first working nodes and the required image identifier. A second container creation request is sent to the target working node carrying the required resource capacity and the required image identifier.
1. A container deployment method, comprising:
receiving a first container creation request that carries a required resource capacity and a required image identifier, wherein the required image identifier is an identifier of an image corresponding to an application that needs to be run;
selecting, from a cluster of working nodes, at least two first working nodes, wherein an unused resource capacity of each of the at least two first working nodes is greater than the required resource capacity;
determining, from the at least two first working nodes, a target working node according to images stored on the at least two first working nodes and the required image identifier; and
sending, to the target working node, a second container creation request that carries the required resource capacity and the required image identifier,
wherein the determining, from the at least two first working nodes, the target working node according to images stored on the at least two first working nodes and the required image identifier comprises:
obtaining an image identifier corresponding to each image stored on each of the at least two first working nodes;
determining, according to the image identifier corresponding to each image, that none of the at least two first working nodes stores an image corresponding to the required image identifier;
obtaining, according to the required image identifier and from a registry repository, image layer identifiers corresponding to image layers comprised in the image corresponding to the required image identifier;
obtaining image layer identifiers corresponding to image layers stored on each of the at least two first working nodes;
determining that there is at least one third working node in the image layer identifiers corresponding to the stored image layers, wherein image layer identifiers of each of the at least one third working node are the same as one or more image layer identifiers in the image layer identifiers corresponding to the image layers comprised in the image corresponding to the required image identifier; and
selecting, from the determined at least one third working node, the target working node.
2. The method according to claim 1 , wherein the selecting, from the determined at least one third working node, the target working node comprises:
if a quantity of the determined at least one third working node is one, determining the determined at least one third working node as the target working node; or
if a quantity of the determined at least one third working node is greater than one, selecting, from the determined at least one third working node, a working node with a largest quantity of the one or more image layer identifiers as the target working node.
3. The method of claim 1 wherein the required resource capacity includes at least one of a required CPU capacity or a required memory capacity.
4. The method of claim 1 wherein the required resource capacity includes a required CPU capacity and a required memory capacity.
5. The method of claim 1 wherein the selecting, from a cluster of working nodes, at least two first working nodes, comprises comparing the unused resource capacity of each of the at least two working nodes with the required resource capacity.
6. The method of claim 5 wherein the selecting, from a cluster of working nodes, at least two first working nodes, further comprises determining, based on the comparing, that the unused resource capacity of each of the at least two working nodes can meet the required resource capacity.
7. The method of claim 1 including determining the unused resource capacity of each of the at least two working nodes, wherein determining the unused resource capacity includes subtracting a used resource capacity of a respective one of the working nodes from a total resource capacity of the respective one of the working nodes.
8. The method of claim 1 wherein the sending, to the target working node, a second container creation request triggers the target working node to create a container according to the required resource capacity and the required image identifier.
9. The method of claim 8 wherein creation of the container according to the required resource capacity and the required image identifier allows the application to be run on the target working node.