 | ||
| About disaster recovery for Red Hat Enterprise Virtualization virtual machines | ||
|---|---|---|
| Prev | Disaster recovery for virtual machines in the Red Hat Enterprise Virtualization environment | Next |
| ||
| About disaster recovery for Red Hat Enterprise Virtualization virtual machines | ||
|---|---|---|
| Prev | Disaster recovery for virtual machines in the Red Hat Enterprise Virtualization environment | Next |
Red Hat Enterprise Virtualization (RHEV) virtual machines can be configured for disaster recovery (DR) by replicating their boot disks using replication methods such as Volume Replicator (VVR), File Replicator (VFR), Hitachi TrueCopy or EMC SRDF. The network configuration for the virtual machines in the primary site may not be effective in the secondary site if the two sites are in different IP subnets. Hence you must make some additional configuration changes to the KVMGuest resource managing the virtual machine.
Supported technologies for replicating virtual machines include:
Disaster recovery use cases for virtual machines work in the following way:
The replication agent takes care of the replication direction. After a disaster event at the primary site, VCS tries to online the replication service group at the secondary site (according to the ClusterFailoverPolicy). The replication resource reverses the replication direction. Reversing the replication direction makes sure that the old secondary LUNs become the new primary LUNs and also are Read-Write enabled on the RHEL-H hosts at the secondary site. This helps RHEV-M activate the Fibre Channel (FC) Storage Domain on the secondary site RHEL-H hosts.
Before the virtual machine (VM) service group can be brought online, the Storage Pool Manager (SPM) in the datacenter needs to failover to the secondary site. This is achieved by the pre-online trigger script configured on the VM service group. This trigger script checks whether the SPM is still active in the primary RHEV cluster. If so, it deactivates all the RHEL-H hosts in the primary RHEV cluster. Additionally, if the SPM host in the primary RHEV cluster is in the NON_RESPONSIVE state, the trigger fences out the host to enable SPM failover. The trigger script then waits for the SPM to failover to the secondary RHEV cluster. When the SPM successfully fails over to the secondary RHEV cluster, the pre-online trigger script reactivates all the RHEL-H hosts in the primary RHEV cluster, which were deactivated earlier and proceeds to online the VM service group in the secondary site
