Return to TODO CVS log

Up to [Development] / failsafe / FailSafe

Initial Checkin

                                Linux FailSafe 1.0
                        http://oss.sgi.com/projects/failsafe/
                        ====================================


Short term tasks (6-9 months)
-----------------------------

Scalability
    Increase scalability to 8 nodes.
    Testing with large number of resources and resource groups.

Infrastructure enhancements
    Support for volume managers (such as LVM)
    Support for Network mirroring (such as DRBD)
    Testing Shared SCSI solutions
    Testing Storage area network solutions
    Supporting for reset devices using STONITH API

API
    Develop application (FailSafe resource type) API
    Develop cluster API (Group communication services layer interface)

Integration with other products
    Integration with IPVS to provide load balancing solution
    Integration with monitoring tools such as "mon"

Usability enhancements (GUI and command line interfaces)
    Provide interim status of long running tasks (such as creating
    cluster, stopping HA services in the cluster)
    Improve error messages and logging mechanisms
    Generate cluster_mgr CLI scripts for a configuration database
    Generate cluster_mgr CLI scripts for a set of GUI tasks
    Executing cluster_mgr CLI scripts from the GUI
    Develop GUI plugin model for resource types
    Develop GUI plugin model for reset mechanisms

FailSafe applications
    Oracle, DB2 database agents

Documentation
    Improve FailSafe administration manual
    Improve FailSafe programmers manual
    Develop FailSafe developers manual (document failsafe internals)

Long term roadmap (more than 6-9 months)
---------------------------------------

Scalability
    Scalability beyond 8 nodes. Improve protocol scalability.

Infrastructure enhancements
    Configuration database based on MySQL or some other database
    Common heartbeat service that can be used by configuration
    database daemon (cdbd) and cluster membership (ha_cmsd)
    Better I/O fencing mechanism (GRITS API from Oracle)
    Developing 64bit version of FailSafe for 64bit platforms

Integration with other products
    Integration with DLM using cluster API
    Integration with cluster filesystems (GFS, CXFS)
    Supporting Oracle parallel server using cluster API
    Testing FailSafe on non-i386 platforms (alpha, ultraSparc, IA64)

Security
    Encryption and authentication of cluster messages

Disaster recovery
    Geographic clustering and failover.