Hard to think of anything worse than a data center going dark, right? Business doesn’t wait, and your services shouldn’t either. That’s where Azure Site Recovery (ASR) steps in. It’s not magic, but it’s close enough when your infrastructure’s on fire and you need a Plan B that actually works.
Why Disaster Recovery Isn’t Just a Buzzword
Sure, some folks treat outages like natural disasters—unavoidable, unpredictable, just pray and hope. But no, Microsoft gives you tools to keep things running: Hyper-V Replica and Azure Site Recovery.
Hyper-V Replica lets you keep copies of critical VMs in a second site, and that’s great… until you realize most companies don’t have the luxury of a second site just lying around. That’s where ASR comes in: instead of a physical second site, it uses Azure itself as the fallback.
TL;DR: What You’ll Need
To get rolling with ASR:
- A working connection between your on-prem (or primary Azure region) and Azure.
- A list of VMs you want to protect — with replication settings nailed down.
- A clear recovery plan: boot order, scripts, timing, manual steps… the whole thing.
ASR: The Two Flavors of Protection
1. Site Recovery Service
This is the main event. It keeps your apps running during planned or unplanned outages by replicating workloads (VMs, physical machines, etc.) to another location — either another Azure region or the Azure cloud from on-prem. When stuff breaks, you fail over and keep working. When things are back to normal, fail back and carry on.
2. Azure Backup
This isn’t a replacement for ASR, but more of a sidekick. It protects your data (file-level, system state, etc.), not entire machines or workloads. You want both: backup for your data, ASR for your running apps and infra.
What ASR Can Replicate (as of 2024)
- Azure VMs across regions
- VMware and Hyper-V VMs on-prem → Azure
- Physical servers (Linux and Windows) → Azure
- VMs under System Center VMM → secondary site or Azure
That covers pretty much all the usual suspects.
Main Perks (Yes, There Are a Few)
- Runs directly in Azure: No extra software needed.
- No production impact testing: You can test failovers without touching live workloads.
- RBAC-ready: Built-in roles (Contributor, Operator, Reader) and support for custom permissions.
- Supports app-level consistency: Even for multi-VM setups.
- Handles SQL Always On, Exchange DAG, Oracle Data Guard — these still work in failover scenarios if you configure them right.
- Custom recovery plans: Automate stuff, or throw in manual steps where needed.
- RPOs down to 30 seconds (Hyper-V only, using near-synchronous replication).
- Can replicate every 30s, 5min, or hourly depending on app criticality.
Quick Note on Networking
Failing over isn’t just spinning up VMs — your network has to come with them. ASR helps here too:
- Reserve static IPs in Azure
- Automate DNS updates
- Integrate with Azure Traffic Manager or load balancers
- Use isolated virtual networks for testing
A Word on Testing
ASR lets you simulate a disaster without causing one. That’s big. You spin up test copies of your replicated VMs in an isolated Azure network, click through your recovery plan, and see what actually happens.
In the test, ASR makes a clone like MyVM-test, drops it into your test network, and you can see how apps behave, if dependencies load, and if your recovery steps make sense. Best part? Replication keeps running the whole time.
Real-World Scenario Hints
Even if your setup is solid and everything looks clean on paper, recovery often needs a human touch. ASR supports pausing a recovery plan at certain stages so an admin can manually step in, do what’s needed, and resume.
Think DNS changes, service checks, credential tweaks — all the fun stuff that only shows up in real life.
Costs (aka: Will It Bankrupt You?)
Nope. ASR is pay-as-you-go:
- You’re billed per protected instance while replication is active.
- During failover, you’re billed for actual Azure compute.
- No need to run hot standby hardware or maintain a second data center.
Snapshot and Replication Details
- Snapshots can be stored for up to 2 hours.
- Recommended snapshot frequency: every hour.
- Replication starts with a full snapshot, then sends deltas.
- Failback is supported — you can reverse the replication back on-prem once your infra recovers.
Final Thoughts
ASR won’t fix all your problems, but it’ll give you a solid shot at keeping your apps online when stuff hits the fan. Setup takes a bit of prep, but once it’s live, it’s a surprisingly low-maintenance safety net.
Want to test it? You should. Just please, please don’t wait until the building’s already on fire.
And yeah — take backups anyway. Always.
