Automated virtual interface failover in a mass storage cluster
First Claim
Patent Images
1. A method of reconfiguring virtual interfaces on a cluster formed by a plurality of servers in a mass data storage system;
- wherein each server includes a port for interfacing to a computer network for communicating data service requests, each virtual interface is formed by an operative port which hosts an internet protocol (IP) address used in communicating the data service requests on the computer network, the servers of the cluster are interconnected in server failover pairs, the cluster includes multiple server failover pairs, each server is connected to at least one data storage unit by which to perform the data service requests communicated on operative ports attached to the server, and each server of each server failover pair is interconnected to the data storage unit of each other server of the server failover pair to enable either server of the server failover pair to perform data service requests with respect to the data storage unit connected to either server of each server failover pair;
the method comprising;
establishing two groups of operative servers of the cluster, each of the two groups including a different one of the two servers from each server failover pair;
communicating status messages among all of the servers of each group which indicate the operative status of each server within the corresponding group and each port of each server within the corresponding group;
establishing a list for each group of servers describing operative ports of each server of the group;
recognizing a change in operative status of any port of any server of each group from the status messages; and
transferring the IP address from one port of the one server with the change in operative status to another operative port of one server within the same group in response to recognizing the change in operative status, whereby the server to which the IP address is transferred is not the other server of the server failover pair which includes the server having the port with the change in operative status.
1 Assignment
0 Petitions
Accused Products
Abstract
Virtual interfaces on a cluster of servers in a mass data storage system are automatically reconfigured by transferring an internet protocol (IP) address hosting responsibility from one operative port to another operative port within the cluster, in response to a change in operative status of any port within the cluster.
56 Citations
22 Claims
-
1. A method of reconfiguring virtual interfaces on a cluster formed by a plurality of servers in a mass data storage system;
- wherein each server includes a port for interfacing to a computer network for communicating data service requests, each virtual interface is formed by an operative port which hosts an internet protocol (IP) address used in communicating the data service requests on the computer network, the servers of the cluster are interconnected in server failover pairs, the cluster includes multiple server failover pairs, each server is connected to at least one data storage unit by which to perform the data service requests communicated on operative ports attached to the server, and each server of each server failover pair is interconnected to the data storage unit of each other server of the server failover pair to enable either server of the server failover pair to perform data service requests with respect to the data storage unit connected to either server of each server failover pair;
the method comprising;establishing two groups of operative servers of the cluster, each of the two groups including a different one of the two servers from each server failover pair; communicating status messages among all of the servers of each group which indicate the operative status of each server within the corresponding group and each port of each server within the corresponding group; establishing a list for each group of servers describing operative ports of each server of the group; recognizing a change in operative status of any port of any server of each group from the status messages; and transferring the IP address from one port of the one server with the change in operative status to another operative port of one server within the same group in response to recognizing the change in operative status, whereby the server to which the IP address is transferred is not the other server of the server failover pair which includes the server having the port with the change in operative status. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13)
- wherein each server includes a port for interfacing to a computer network for communicating data service requests, each virtual interface is formed by an operative port which hosts an internet protocol (IP) address used in communicating the data service requests on the computer network, the servers of the cluster are interconnected in server failover pairs, the cluster includes multiple server failover pairs, each server is connected to at least one data storage unit by which to perform the data service requests communicated on operative ports attached to the server, and each server of each server failover pair is interconnected to the data storage unit of each other server of the server failover pair to enable either server of the server failover pair to perform data service requests with respect to the data storage unit connected to either server of each server failover pair;
-
14. A method of reconfiguring virtual interfaces on a cluster formed by a plurality of servers in a mass data storage system, wherein each server includes a port for interfacing to a computer network for communicating data service requests, each virtual interface is formed by an operative port which hosts an internet protocol (IP) address used in communicating the data service requests on the computer network, and each server performs the data service requests, the method comprising:
-
establishing pairs of servers from the plurality of servers to operate as server failover pairs; establishing a first group of servers within the cluster by recognizing one server from each server failover pair which has at least one operative port; establishing a second group of servers within the cluster by recognizing the other server within each server failover pair which has at least one operative port; recognizing a change in operative status of any port of any server of either the first or second group; and transferring the IP address from one port of one server of one of the first or second groups to another operative port of one server of the one group in response to the change of operative status, without transferring the IP address from the one server to the other server within the same failover pair.
-
-
15. A mass data storage system for communicating data service requests with client computers over a computer network, comprising:
-
a plurality of servers organized in a cluster, each server including at least one port for interfacing with the computer network by which to communicate the data service requests between the server and the client computers, at least one port of each server hosting an internet protocol (IP) address used in communicating the data service requests on the computer network, each server also including at least one data storage unit by which to perform the data service requests communicated on each operative port attached to the server, the servers are interconnected in server failover pairs, the cluster includes multiple server failover pairs, and each server of each server failover pair is interconnected to the data storage unit of each other server of the server failover pair to enable either server of the server failover pair to perform data service requests with respect to the data storage unit connected to either server of each server failover pair; each server further including at least one processor and a memory unit which includes operating instructions for performing the data service requests; the memory unit of each server further including a configuration rules table containing information defining the operative status of servers in the cluster and information defining the operative status of each port of each server in the cluster; the servers of the cluster communicating status messages among themselves to maintain the operative server status information and operative port status information current in the configuration rules table; the operating instructions in the memory unit causing the servers to reconfigure virtual interfaces on the servers by transferring hosting of IP addresses from one operative port of the servers to another operative port of the servers, each IP address used in communicating the data service requests on the computer network; and
wherein the execution of the operating instructions to cause reconfiguration of the virtual interfaces operatively;establishes first and second groups of operative servers of the cluster from the server status information in the configuration rules table in response to the status messages, each group including no more than one server from each server failover pair, establishes a list of operative ports of each server of the first and second groups from the port status information in the configuration rules table in response to status messages, recognizes a change in operative status of any port of any server of the first and second groups from the server status and port status information in the configuration rules table, and transfers the IP address from one port of one server of one of the first and second groups to another operative port of one server of the one group in response to a recognized change in status information, without transferring the IP address from the one server to the other server within the same failover pair. - View Dependent Claims (16, 17, 18, 19, 20)
-
-
21. A mass data storage system connected to a computer network, comprising:
-
a plurality of servers each including at least one port for interfacing with the computer network, at least one port of each server is assigned an Internet Protocol (IP) address, pairs of servers in the plurality of servers being interconnected in server failover pairs; and
wherein;each server from each server failover pair is assigned to a separate one of two groups; and upon the failure of the port of any one of the plurality of servers, the IP address assigned to the failed port is transferred to an operative port on another server within the same group as the server having the failed port, whereby the transfer of the IP address from the server with the failed port is not to the other server of the same server failover pair. - View Dependent Claims (22)
-
Specification