发明名称 Consistent cluster operational data in a server cluster using a quorum of replicas
摘要 A method and system for increasing server cluster availability by requiring at a minimum only one node and a quorum replica set of replica members to form and operate a cluster. Replica members maintain cluster operational data. A cluster operates when one node possesses a majority of replica members, which ensures that any new or surviving cluster includes consistent cluster operational data via at least one replica member from the immediately prior cluster. Arbitration provides exclusive ownership by one node of the replica members, including at cluster formation, and when the owning node fails. Arbitration uses a fast mutual exclusion algorithm and a reservation mechanism to challenge for and defend the exclusive reservation of each member. A quorum replica set algorithm brings members online and offline with data consistency, including updating unreconciled replica members, and ensures consistent read and update operations.
申请公布号 US8850007(B2) 申请公布日期 2014.09.30
申请号 US201113156315 申请日期 2011.06.08
申请人 Microsoft Corporation 发明人 Massa Michael T.;Dion David A.;Das Rajsekhar;Doshi Rushabh A.;Lomet David B.;Nishanov Gor N.;Bernstein Philip A.;Gamache Rod N.;Jain Rohit;Shrivastava Sunita
分类号 G06F15/173;G06F11/08;G06F11/18;G06F11/14;G06F11/20;G06F17/30;G06F11/16 主分类号 G06F15/173
代理机构 代理人 Gabryjelski Henry;Drakos Kate;Minhas Micky
主权项 1. A method comprising: maintaining a cluster that includes cluster operational data and an independent replica that comprises a storage device that maintains the cluster operational data, that is not an element of any node of the cluster, that is not directly attached to the any node of the cluster, that is controlled by only one node of the cluster at a time, where the maintaining the cluster comprises the independent replica being exclusively controlled by the one node of the cluster based on the one node of the cluster obtaining or maintaining the exclusive control according to a node arbitration process that is configured for first reading a variable from the independent replica and, after a challenge period, second reading the variable from the independent replica, where the obtaining or the maintaining is based on the first reading and the second reading resulting in a same value, and where the independent replica is configured to be online in the cluster regardless of which of the any nodes are operational in the cluster.
地址 Redmond WA US