IBM Global Services - Federal announces the High Availability option for HPSS. HA-HPSS* provides hardware redundancy for HPSS core servers, complementing the ability to configure redundancy in the data movers. The result is a mass storage system that remains on-line, in spite of failures in key hardware components.

High Availability Features:	Components:
Minimize downtime Automatically recover from: Total failure of primary server computer Power failure Disk failure Disk adapter failure Network adapter failure and more...	Standard HPSS software HACMP for AIX clustering software from IBM Redhat Cluster Suite for Linux clustering software from Redhat Redundant hardware components

• Two HPSS core server machines: one primary and one standby.
• Primary computer services requests using "service interface."
• Metadata is mirrored across busses.
• "Keep alive" messages in AIX are exchanged between primary and standby over Ethernet and TTY cable (to prevent isolation).
• "Keep alive" messages in Linux are exchanged between primary and standby over Ethernet and shared disk.
• Network adapter failure causes a standby adapter to assume the IP and hardware address of the failed adapter.
• Failures of disk, disk adapter, and cables are resolved by AIX mirroring over dual busses.
• Primary computer failure results in the standby computer assuming the IP and hardware addresses of the service interface, taking control of the metadata disks, and restarting HPSS.

*HA-HPSS is available from IBM by special bid arrangements.

< Home