Computer system and scale-up management method
It aims to make it possible to readily and rapidly scale up the server which executes one application. In a computer system which includes one or more compute server(s) which each has an application container which executes the one application and a management server which manages the compute server(s), the management server is configured to, in a case of increasing the number of the compute servers which each has the execution unit which executes the one application, specify a logic unit that a data unit that the execution unit of an existing compute server utilizes upon execution of an application is stored, and in a case where the execution unit of a newly added computer server executes the application, set the newly added compute server so as to refer to the specified logic unit.
1. A computer system comprising:
a plurality of compute server;
a plurality of storage servers;
a plurality of execution units that are each respectively executed on one of the plurality of compute servers, wherein each execution unit executes a particular application that requires data to execute; and
a management server which manages the plurality of compute servers and the plurality of storage servers to allocate the plurality of execution units among the plurality of compute servers, wherein the management server is configured to:
determine that an additional instance of a particular execution unit is required to execute the particular application, and
in response to determining that the additional instance of the particular execution unit is required:
determine a particular logic unit that stores the data required for the particular application, wherein the particular logic unit is stored on a particular storage server among the plurality of storage servers,
replicate the particular logic unit from the particular storage server to another storage server among the plurality of storage servers to form a replicated logic unit,
create the additional instance of the particular execution unit on another compute server amongst the plurality of compute servers, and
set the additional instance of the particular execution unit to refer to the replicated logic unit.
2. The computer system according to claim 1 , wherein
the another storage server among the plurality of storage servers is selected in accordance with a predetermined algorithm.
3. The computer system according to claim 2 , wherein
the predetermined algorithm is an algorithm includes a hash value which relates to information which contains a name of the particular logic unit.
4. The computer system according to claim 1 , wherein
the plurality of execution units are each a container which is configured on an OS of a respective compute server.
5. The computer system according to claim 1 , wherein
the management server is further configured to:
allocate a plurality of logic units to one block device and generates the logic unit such that a data volume of the logic unit becomes less than a predetermined volume.
6. The computer system according to claim 1 , wherein
the management server
outputs at least either load information or performance information on a resource of a respective server amongst the plurality of the compute servers that executes the particular execution unit.
7. The computer system according to claim 1 , wherein
the management server is further configured to:
output information which indicates performances of all of the plurality of the compute servers prior to creating the additional instance of the particular execution unit and after creating the additional instance of the particular execution unit.
8. A scale-up management method, the method comprising:
determining, by a management server, that an additional instance of a particular execution unit is required to execute a particular application; and
in response to determining that the additional instance of the particular execution unit is required:
determining, by the management server, a particular logic unit that stores data required for the particular application, wherein the particular logic unit is stored on a particular storage server among a plurality of storage servers,
replicating, by the management server, the particular logic unit from the particular storage server to another storage server among the plurality of storage servers to form a replicated logic unit,
creating, by the management server, the additional instance of the particular execution unit on another compute server amongst a plurality of compute servers, and
setting, by the management server, the additional instance of the particular execution unit to refer to the replicated logic unit.