Abstract:
A machine is disclosed for transferring data between two storage locations in a high-performance computing system including a enterprise storage system, a distributed parallel storage system connected to an application processor, and a storage bridge having a plurality of parallel data transfer nodes connecting the enterprise storage system to the distributed parallel storage system. A computer- implemented process directed by software executing on the machine performs scheduling a data transfer to minimize interference with another data transfer; balancing a load amongst the plurality of parallel data transfer nodes by dividing the data transfer into plural segments and distributing the plural segments amongst the plurality of parallel data transfer nodes; and transferring the plural segments substantially in parallel with respect to time using the plurality of parallel data transfer nodes.
Abstract:
A rackmountable storage enclosure includes a chassis. The chassis includes one or more power supplies and a plurality of drawers, each extendable through a front surface of the chassis. Each drawer of the plurality of drawers provides mounting for one or more storage devices. All storage devices in any drawer of the plurality of drawers are inserted or removed through a common side of the drawer. Any storage device may be inserted or removed from any drawer of the plurality of drawers even if the side surfaces of the chassis are each parallel to and in contact with a wall, each of the walls extending forward at least to a fully extended length of any drawer. Electrical failures in any one drawer are prevented from affecting any other drawer of the plurality of drawers. The one or more power supplies provides DC power to the plurality of drawers.
Abstract:
Since the whole storage device is blocked according to the conventional data saving method when failure occurs to the storage device in a storage subsystem, so that when failure occurs to two storage devices at the same time within a same RAID group, double failure is caused and data loss occurs. In order to solve the problem, the present invention divides a storage device into storage areas of predetermined units, constructs RAID groups from two or more storage areas, and when failure occurs to the storage area, selects a data migration destination storage area from either the RAID group in which failure has occurred or the RAID group other than the RAID group in which failure has occurred, migrates the data stored in the storage area where failure has occurred to the selected data migration destination storage area, and blocks only the storage area where failure has occurred.
Abstract:
Each of SAS expanders ("expander(s)" hereinafter) has a switch device for switching whether to bypass the expander in each communication path or not. Of the plurality of switch devices, an actual connection destination of a switch device bypassing the expander is a switch device in a upper-level and/or a lower-level of the switch device. Of the plurality of switch devices, an actual connection destination of a switch device that does not bypass an expander is the expander.
Abstract:
Systems are disclosed in which a first receiver circuit of an integrated circuit die, the die having conductive die pads for electrical signal transfer between conductive pathways inside the die and conductive pathways outside the die, the first receiver circuit having an input connected to a first die pad and an output connected to a first logic circuit of the die and a second receiver circuit of the integrated circuit die, the second receiver circuit having an input connected to a second die pad and an output connected to a second logic circuit of the integrated circuit die, where both receiver circuits are configured to receive the same signals.
Abstract:
Provided are a method, system and program for selecting a path comprising ports on primary and secondary clusters to use to transmit data at a primary volume to a secondary volume. A request is received to copy data from a primary storage location to a secondary storage location. A determination is made from a plurality of primary clusters of an owner primary cluster for the primary storage location, wherein the primary clusters are configured to access the primary storage location. A determination is made as to whether there is at least one port on the owner primary cluster providing an available path to the secondary storage location. One port on the owner primary cluster is selected to use to copy the data to the secondary storage location in response to determining that there is at least one port on the owner primary cluster available to transmit to the secondary storage location.
Abstract:
A network storage appliance is disclosed. The appliance includes a single chassis that encloses a plurality of servers and a plurality of storage controllers coupled together via a chassis backplane. The storage controllers control the transfer of data between the plurality of servers and a plurality of storage devices coupled to the storage controllers. The servers and storage controllers include a plurality of field replaceable unit (FRUs) hot-pluggable into the backplane such that any one of the FRUs may fail without loss of availability to the storage devices' data. In various embodiments, the chassis fits in a 19" wide rack; is 1U high; the servers are standard PCs configured to execute off-the-shelf server applications and to facilitate porting of popular operating systems with little modification; the servers include disk-on-chip memory rather than a hard drive; local buses (e.g., PCIX) on the backplane interface the various FRUs.
Abstract:
Systems and methods for performing error recovery in a system utilizing redundant communication links. In one embodiment, a host device (11) is coupled to a sequential device (15) such as a tape drive by a pair of Fibre Channel links (12, 13). The host is configured to associate an identifier with each command. Upon detecting a failure of the transmission of the command via the first link, the host re-transmits the command via the second link. When the link failure is detected, the status of the last command transmitted before the failure is requested. After receiving the status of the command, a request that recovery start (possibly at a point indicated by the status information) is transmitted to the sequential device. If the recovery request is accepted, performance of the command is completed.
Abstract:
Failover processing in storage server system utilizes policies for managing fault tolerance (FT) and high availability (HA) configurations. The approach encapsulates the knowledge of failover recovery between components within a storage server and between storage server systems. This knowledge includes information about what components are participating in a Failover Set, how they are configured for failover, what is the Fail-Stop policy, and what are the steps to perform when "failing-over" a component.