Systems and methods for enhancing performance of a clustered source code management system

摘要

A clustered source code management system is described. The system comprises a plurality of cluster nodes, a shared file server storing repository data, and a load balancer. Each of the plurality of cluster nodes is configured to receive an incoming request sent from a client computer, the incoming request being a source code management request to access repository data stored on the shared file server and process the incoming request to determine if a type of the incoming request is a cacheable request type. If the incoming request is of a cacheable request type, the cluster node is configured to determine if a valid cached response to the incoming request is available on storage media locally accessible to the cluster node; and, if so, respond to the incoming request by communicating the valid cached response to the client computer.

主权项

1. A computer implemented method for operating a cluster node, the cluster node being one of a plurality of cluster nodes of a clustered source code management system, the method comprising:
receiving an incoming request from a client computer, the incoming request being a source code management protocol request to access repository data stored on a shared file server of the source code management system, wherein the repository data on the shared file server is accessible to the cluster node via an external source code management server process; processing the incoming request to determine if a type of the incoming request is a cacheable request type; based on determining that the type of the incoming request is a cacheable request type, determining if a valid cached response to the incoming request is available on storage media locally accessible to the cluster node, and wherein:
based on determining that a valid cached response to the incoming request is available, the method further comprises responding to the incoming request by communicating the valid cached response to the client computer; andbased on determining that a valid cached response to the incoming request is not available, the method further comprises:
initiating an external source code management server process in order to access repository data stored on the shared file server;passing the incoming request to the external source code management server process;intercepting a source code management server process response to the incoming request from the external source code management server process;storing the source code management server process response on the storage media locally accessible to the cluster node; andresponding to the incoming request by communicating the source code management server process response to the client computer;wherein storing the source code management server process response on the storage media locally accessible to the cluster node comprises generating and storing a new cache entry in a cache entry data structure, the new cache entry comprising:
a key generated in respect of the incoming request, the key being based on data included in the incoming request; andthe source code management server process response.