November 10, 2017
10 VM Replication Best Practices
Virtual machine replication (VM replication) is often confused with VM backup, but these VM data protection methods pursue different objectives and should not substitute each other (see our blog post VM Backup vs. VM Replication). A VM replica becomes a real lifesaver in case of disaster allowing you to perform VM recovery within seconds. Therefore, in recent years, VM replication has gained momentum in protecting virtual infrastructures from failures or downtimes of any origin. However, you may still experience difficulties and headaches while running VM replication software: app failures and errors, lost data, inadequately high consumption of resources, to say nothing of common troubles you may suffer from at the deployment, management, or maintenance stage. Fortunately, such complications can be avoided, and we strongly recommend that you follow the VM replication best practices listed below as part of your efficient VM data protection strategy.
1. Define Your Mission-Critical Data Areas and Establish Their Priorities.
Decide which of your virtual machine and application data require the highest protection and the fastest recovery in case of downtime or disaster. You do not need to replicate everything, and not everything needs to be recovered within the same time frame. For example, your online store maintains continuous operation and allows conducting direct order and purchase transactions with customers. In case your web server fails, the website, inventory, and CRM will be extremely critical to be recovered in the first place and as soon as possible to minimize the website downtime effect for customers, while, surely, back-office apps can wait.
2. Outline Your Data Protection Plan
Failing to plan is planning to fail. Plan your data protection activities with regard to business continuity demands, mission-critical data areas, priorities, individual data protection techniques, VM backup/replication job frequency and duration, required resources (data storage capacity, network bandwidth, VM backup/replication windows, etc.).
Make sure that your Data Protection Plan also covers personnel responsibilities, dedicated hardware and software components, and disaster recovery scenarios.
3. Use a Proper VM Backup and Replication Ratio
Although VM backup and VM replication seem to do the same thing, they have different objectives and cannot substitute each other. VM backup, whatever the type, is designed for long-term safekeeping of VM data, while VM replication provides for fast VM recovery, or in other words, VM high availability.
Considering VM replication to be a constituent element of your Data Protection Plan, find an efficient combination of regular VM backups and VM replication to fit your individual business needs and budget. Every combination will have its price, so thoroughly consider the opportunities to keep your expenses low.
4. Establish Measurable Criteria for Your VM Data Replication Sequence
Establishing measurable criteria may be of help when deciding which VM replication solution to choose. With regard to high-availability VM replication, you mainly need to consider two aspects: speed and fault tolerance (i.e., your system’s ability to maintain a good level of performance in case one or more components have failed).
Ask yourself, how fast you want your VM data recovery to happen. Although a VM replica is a full copy of the original VM, you may still need some time to access it, e.g. by remounting and bringing back a DBMS.
Consider, what fault tolerance degree your system should keep up with to make your VM data available in case a disaster strikes. While deciding how often you are going to employ VM replication for fault tolerance, also take into account your hardware and network capacity.
Depending on your business continuity policy, establish an optimal proximity of the recovery point to the point of failure, that is, the Recovery Point Objective (RPO), which is a tradeoff between minimizing your data loss and cost of additional resources. For some businesses, it can be 24 hours, 7 days, or even 1 month, while others might need it to be merely 10 seconds.
5. Prepare Your Failover and Failback Roadmap
To keep services and applications running with minimum disruption in case a production VM goes down, determine your failover and failback procedures in advance. Work out production VM failover and failback action scenarios to be followed under different adverse circumstances.
Depending on the established data protection system, failover and failback procedures may be either manual or automated for accidental and planned shutdowns.
6. Enable Verification of VM Replicas
Regular verification of your VM replicas to make sure that they are alright is a good practice. This will save you from many unpleasant surprises. It can happen that your VM replica turns out to be damaged or corrupted. Whatever the reason of the defect, always verify your VM replicas, or at least once in a while, if you do it manually.
When using 3rd party VM replication solutions, it is recommended to enable the feature of automatic verification of VM replicas, if any. Thus, all your VM replicas will be automatically verified for integrity. The VM replication software performs verification by powering VM replicas on with networking frozen and making screenshots of the booted OS. Then the product sends you screenshots of recovered VMs as a proof of your VM replica consistency.
7. Create Application-aware Replicas
Modern data protection solutions provide you with a special application-aware replication mode allowing you to create application-aware VM replicas for applications, such as Microsoft Active Directory, Microsoft SQL Server, Microsoft SharePoint, Microsoft Exchange, or Oracle, by relying on VSS writers.
“Application-aware” or, in other words, application-consistent VM replication mode means that in the process of replication the in-memory data and pending I/O operations are flushed to disk before a snapshot is made. Consequently, your VM replicas are application consistent.
8. Build Automation around Your Recovery Tools
Building automation around recovery tools brings down unnecessary delays, if you have to restore VMs as soon as possible in case of disaster. Make sure automatic scripts are properly maintained. On the other hand, modern advanced data protection solutions offer integrated automation features.
9. Determine Your Retention and Rotation Policy
This policy establishes the frequency of creating VM replicas and their retention time. VM replication data changes are continuously added to the replication changes database. Without its periodical purging, it would grow dramatically, until it consumed all the available disk space. This period between purges is called a Replication Purge Delay. The default replication purge delay is set in your data protection software and varies from one to more days.
So, when establishing your retention and rotation policy, consider your replication purge delay. Changes older than the purge delay are cleared off from the replication changes database. Make sure that the replication changes database is backed up more often than the replication purge delay period; otherwise, some changes may be lost.
10. Choose the Right VM Replication Solution
If you need to regularly perform VM replication, you should select the right solution, which will fit your specific needs and budget.
When comparing VM replication products currently available on the market, you should consider their specific features through the perspective of both functionality and performance. For instance, some products use synchronous VM replication, while others run semi-synchronous or asynchronous VM replication. While synchronous VM replication seems to be a better choice for failover of transactional applications, full synchronization used in a slow network may critically bring down performance. In contrast, the asynchronous VM replication has almost no impact on performance in this situation. On the other hand, if the primary server fails, asynchronous VM replication can result in the loss of data committed.
Another point to take into consideration is the speed of your VM replication and restore operation, since it might require high-capacity hardware.
While remote replication of virtual machines to offsite location or a cloud keeps your data safe even if a disaster strikes your whole datacenter, make sure your infrastructure network has an appropriate bandwidth.
Also, when deciding on which VM replication solution to use, pay attention to a set of VM replication-related features this product offers. Modern VM replication products, such as NAKIVO Backup & Replication, offer you many useful functions to make your VMware replication, Hyper-V replication, and even AWS EC2 instance replication painless and straightforward.
VM Replication Best Practices Summary
Until not so long ago, VM replication was considered to be a costly and sluggish data protection method in a traditional environment. However, the abovementioned drawbacks are now easy to avoid, if you follow the time-proven VM replication best practices. Besides that, the growing virtualization market now offers new highly innovative software tools to handle your VM replication processes. Apart from the features that are usually on demand, the best solutions offer unique improvements in terms of data processing speed, workload optimization and balancing, integration with clouds, etc.
Being a modern VM replication and backup solution for VMware, Hyper-V, and AWS EC2 environments, NAKIVO Backup & Replication is designed to go ahead of your expectations. Request Live Demo or Download Free Trial of the product below to expand your vision to the most advanced backup and replication technologies in the world of virtualization.