NAME OpenMosix::HA -- High Availability (HA) layer for an openMosix cluster SYNOPSIS use OpenMosix::HA; my $ha = new OpenMosix::HA; # spawn all apps for resource group "foo", runlevel "run" $ha->tell("foo","run"); # spawn all apps for resource group "foo", runlevel "runmore" # (this stops everything started by runlevel "run") $ha->tell("foo","runmore"); # get status of all resource groups $ha->status(); DESCRIPTION This module provides basic "init" functionality, giving you a single inittab-like file to manage daemon startup and restart across a cluster of openMosix machines. This gives you a high-availability cluster with low hardware overhead. In contrast to traditional HA clusters, we use the openMosix cluster membership facility to provide heartbeat and to detect network partitions. All you need to do is build a relatively conventional openMosix cluster, install this module, and configure it to start and manage your HA processes. There is no need to use the heartbeat serial cables, spare ethernet cards, or high-end server machines which traditional HA requires. There is no need for chained SCSI buses -- you can share disks among many nodes via any number of other current technologies, including SAN, NAS, GFS, or Firewire (IEEE-1394). BACKGROUND Normally, a process-migration-based cluster computing technology (such as openMosix) is orthogonal to the intent of high availability. When openMosix nodes die, any processes migrated to those nodes will also die, regardless of where they were spawned. The higher the node count, the more frequently these failures are likely to occur. But if processes are started via OpenMosix::HA, any processes and resource groups which fail due to node failure can be configured to automatically restart on other nodes. OpenMosix::HA detects process failure, selects a new node out of all currently available, and deconflicts the selection so that two nodes don't restart the same process or resource group. In addition to the normal inittab format, the configuration file for OpenMosix::HA includes an extra "resource group" column -- this is what enables you to group processes, disk mounts, virtual IP addresses, and related resources into resource groups. Any given node only needs to be able to support a subset of all resource groups. OpenMosix::HA provides an extra "test" runmode (beyond init's normal 'wait', 'once', and 'respawn'), enabling the module to automatically test a given node for fitness before considering starting a given resource group there. There is no "head" or "supervisor" node in an OpenMosix::HA cluster -- there is no single point of failure. Each node makes its own observations and decisions about the start or restart of processes and resource groups. IO Fencing (also STOMITH or STONITH, the art of making sure that a partially-dead node doesn't continue to access shared resources) can be handled as it is in conventional HA clusters, by a combination of exclusive device logins when using Firewire, distributed locks when using GFS or other SAN, and brute-force methods such as X10 or network-controlled powerstrips. OpenMosix::HA provides a callback hook which can be used to trigger the latter. AVAILABILITY This module is based on my IS::Init module, which is already in production and available from CPAN. My wife and I had hoped to have a beta version of OpenMosix::HA available by the time of Moshe Bar's Feb 5 2003 openMosix talk at the Silicon Valley Linux Users Group. Then I unexpectedly became involved in data collection for Columbia's California transit -- SVLUG member Ian Kluft was one of the few witnesses. We decided it best to defer work on this module in favor of improving our understanding of where the shuttle's breakup actually began, relaying our results to Johnson Space Center and working with media to encourage others to do the same. These efforts by ourselves and others have been successful beyond what any of us expected -- NASA JSC emergency ops responded to us personally and as of this writing a search in California is already underway. But I don't have a Perl module for you yet. For a production version of OpenMosix::HA, check CPAN.org, TerraLuna.Org, or Infrastructures.Org in early March 2003, or contact me. Beta versions will become available as time permits before then. AUTHOR Steve Traugott CPAN ID: STEVEGT stevegt@TerraLuna.Org http://www.stevegt.com COPYRIGHT Copyright (c) 2003 Steve Traugott. All rights reserved. This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself. The full text of the license can be found in the LICENSE file included with this module. SEE ALSO IS::Init, openMosix.Org, qlusters.com