AI - Red Hat Cluster Suite
Empowering scalable, load balancing, and highly-available clusters with advanced HA virtual machine management for OpenNebula through RHCS integration.
- Name
- Red Hat Cluster Suite -https://github.com/CERIT-SC/opennebula-rhcs
- Last Audited At
About Red Hat Cluster Suite
Red Hat Cluster Suite is a set of components for building scalable, load balancing, and highly-available clusters developed by Red Hat. The suite is part of the High Availability Add-On in Red Hat Enterprise Linux, but it is also freely available in RHEL-like systems such as CentOS, SL, and Oracle Linux.
Red Hat Cluster Suite provides advanced features for managing truly HA virtual machines with OpenNebula (ON) through its integration with ON using RHCS. The supported features include hypervisor management via libvirt for KVM, virtual machine actions like deploy, cancel, power off, reboot, reset, live migration, and node/host name binding for failover domains to ensure virtual machines run on suitable hosts.
Multi-node configurations offer several advantages such as having each node's status visible to ON, following ON's deployment scheduler configuration, and supporting true HA virtual machine management through RHCS. However, one disadvantage is that virtual machines on failed nodes cannot be managed until the nodes are up again.
To set up a multi-node cluster using Red Hat Cluster Suite with OpenNebula, certain requirements must be met: latest RHEL 6.x (or CentOS, SL, Oracle Linux), a cluster-wide shared filesystem mounted on /opt/opennebula, a working RHCS cluster, and the hypervisor configured for OpenNebula, among other things. Cluster configuration is managed using the ccs command with ricci daemon configured and running on all nodes.
The scripts in this context add support for managing virtual machines via ON using RHCS. Features like deploying, canceling, powering off, rebooting, resetting, live migration, and node/host name binding are supported for KVM hypervisors. However, some features such as suspend, resume, stop, offline migration, disk attach/detach are not currently supported.
For multi-node clusters to work correctly, RHCS cluster node names must be identical with OpenNebula hosts to ensure proper virtual machine binding and migration. The supported combination of drivers for OpenNebula hosts is an information driver: kvm, a virtualization driver: kvm_rhcs, and a network driver: dummy.