RecoverPoint for VMs and vSphere Replication on the Same Virtual Machine
Article ID: 453034
Updated On:
Products
Issue/Introduction
Running Dell RecoverPoint for VMs (RP4VMs) and VMware vSphere Replication against the same virtual machine at the same time is a design that should be treated with extreme caution. Both products provide host-level VM replication, but they use different replication architectures, maintain independent replication state, and have separate recovery workflows.
The key concern is not simply that the VM is being "replicated twice." The concern is that two independent data-protection mechanisms are operating on the same VM and its virtual disks, potentially introducing I/O, snapshot, resource, networking, and recovery interactions.
Resolution
1. The Basic Architecture
At a high level, the configuration would look like this:
The two products do not form a single replication chain.
RecoverPoint does not normally replicate the vSphere Replication copy, and vSphere Replication does not normally replicate the RecoverPoint copy. Each system independently protects the production VM.
Dell documents that the RP4VM splitter intercepts writes for protected VMs and sends a copy of the write toward the assigned vRPA and designated storage.
Broadcom describes vSphere Replication as host-based replication. After initial synchronization, changes to the protected VM are tracked and transferred as lightweight delta synchronizations according to the configured RPO.
2. Why the Configuration Is Potentially Problematic
The same production VM becomes the source for two independent protection systems:
This means both systems must:
- Track changes to the VM,
- Consume storage and compute resources,
- Transfer replication data,
- Maintain their own replication state,
- Manage their own recovery points, and
- Perform their own recovery procedures.
The resulting replicas are not necessarily identical in terms of recovery point or application state.
For example:
Production:
12:00:00 Transaction A
12:00:05 Transaction B
12:00:10 Transaction C
12:00:20 Site failure
The RP4VM recovery point selected by an administrator might contain:
A + B + C
while the VR replica could represent an earlier point such as:
A + B
depending on replication progress and RPO.
This means "two replicas" does not necessarily mean "two identical recovery copies."
3. I/O Filtering and the ESXi Storage Stack
RecoverPoint for VMs uses a hypervisor-level write splitter/filter. Dell's current RP4VM documentation describes the splitter as proprietary software installed on ESXi hosts involved in RP4VM replication; it intercepts writes from protected VMs and sends a copy toward the vRPA (Virtual RecoverPoint Appliance).
For RP4VM 6.x, Dell also documents that the splitter uses VMware VAIO/I/O-filter mechanisms.
vSphere Replication, meanwhile, is a host-based replication technology that tracks VM changes and performs regular delta synchronizations.
Therefore, the simplified conceptual model is:
Multiple replication technologies operating on the same VM introduce additional I/O-processing and interoperability dependencies, and therefore must be explicitly supported and validated rather than assumed to be interchangeable.Therefore, the interoperability of both these products hasn't been tested by VMware by Broadcom, so it's not recommended to use them together.
4. Storage and I/O Overhead
Replicating the same VM twice naturally creates additional workload.
For every change generated by the VM, the infrastructure must support:
This can increase:
- ESXi host processing requirements
- Storage I/O
- Replication network traffic
- WAN bandwidth consumption
- Target-side storage consumption
- Replication appliance workload
Broadcom specifically identifies network performance, storage latency, compute resources, aggressive RPO settings, and replication bandwidth as factors that can affect vSphere Replication performance and RPO compliance.
Consequently, running two replication mechanisms against the same high-change-rate VM can increase the probability of resource contention.
Dual replication may increases the number of components involved in the I/O and data-protection path and can increase resource consumption. In an environment that is already close to its CPU, storage-I/O, or network limits, this additional workload may contribute to latency or replication instability leading to a massive I/O latency or a PSOD.
5. VM Freeze and Application Consistency
Because both products have overlapping workflows for obtaining consistent recovery points, and whether guest quiescing occurs depends on the configuration, VMware Tools state, application-consistency requirements, and the specific operation being performed.
If multiple protection or backup tasks attempt to manipulate the VM at the same time, one operation can interfere with another or fail because another protection software is performing the same task or overlapping with the ongoing tasks which may lead to simultaneously freezing the guest operating system and creating locks on VMDK files.
6. RP4VM and vSphere Replication Have Different Strengths
These products are not simply two versions of the same technology.
RecoverPoint for VMs
RecoverPoint for VMs is designed around continuous VM/data protection, consistency groups, journals and recovery to previous points in time. Dell describes RP4VM as supporting local and remote replication and recovery to any point in time.
A typical use case is:
This is particularly useful when the business requirement is: "I need to recover the VM to a specific point immediately before data corruption or a ransomware event."
vSphere Replication
vSphere Replication is VMware's host-based replication technology and integrates with VMware recovery orchestration products.
Broadcom documents that vSphere Replication can use a 5-minute RPO under supported conditions, while newer Enhanced vSphere Replication configurations can support a 1-minute RPO with the required VMware Live Recovery licensing.
The important point is that the exact RPO capability depends on:
- vSphere Replication version
- Enhanced vs. legacy replication
- Licensing
- Source/target storage
- Available bandwidth
- VM workload
7. When to Choose RecoverPoint for VMs
RecoverPoint for VMs is a strong fit when the requirement is:
- Continuous protection,
- Granular point-in-time recovery,
- Journal-based recovery,
- Recovery before a specific corruption event,
- Application consistency across multiple VMs,
- Recovery from ransomware or logical corruption.
For example:
14:00 Normal
14:05 Normal
14:10 Ransomware begins
14:11 Files encrypted
14:12 Attack detected
With journal-based recovery, the objective may be to select a recovery point such as:
14:09:30
rather than simply recovering to the latest available replica.
Dell explicitly describes RP4VM as providing recovery to any point in time.
8. When to Choose vSphere Replication
vSphere Replication is generally attractive when the requirement is:
- VMware-native replication
- Straightforward VM-level protection
- Integration with Site Recovery Manager
- RPO-based replication
9. What About Using Both for the Same VM?
Currently, there is no testing performed by VMware by Broadcom and Dell on using both these products simultaneously to protect the same or to be used in ESXi. Only choose to use one protection product at a time or get in touch with VMware by Broadcom PSO team to understand the benefits of using Cyber Recovery software. VMware Live Cyber Recovery is a cloud service that helps businesses recover quickly and safely from ransomware attacks and cyber threats and this product can be used together with an array of VMware products.
Final Recommendation
Do not configure RecoverPoint for VMs and vSphere Replication to independently protect the same VM merely as a way to obtain "double protection."
The two technologies have different replication engines, different recovery models and different management planes. RP4VM introduces its own write-splitting and vRPA architecture, while vSphere Replication independently tracks and transfers VM changes.
Running both can increase:
I/O processing + storage workload + network traffic + operational complexity + recovery ambiguity.
Choose RecoverPoint for VMs when granular, journal-based, point-in-time recovery is the primary requirement.
Choose vSphere Replication / VMware's current recovery stack when VMware-native RPO-based replication and recovery orchestration are the primary requirements.
For a mission-critical VM, the most important thing is not having two replication products. It is having one clearly supported replication architecture, one authoritative recovery process, and a tested DR runbook.
Feedback
