Where do single points of failure hide in a surveillance SAN?
Single points of failure hide anywhere one component sits alone in the path between camera streams and stored video. The usual locations are storage controllers, switches, host bus adapters, power supplies, network links, management software and the recording server itself. Disks are rarely the weakness, because RAID usually protects them. The surrounding infrastructure is where gaps appear.
Common hiding places include:
- A single storage controller: if the SAN has one controller, its failure takes every volume offline, regardless of how many drives are installed.
- A single switch or fabric: one network or Fibre Channel switch connecting servers to storage can interrupt all recording at once.
- Single cabling or network paths: one failed cable, transceiver or port can isolate a recording server from its volumes.
- One power feed or power supply: a lone supply or a shared circuit can switch off an otherwise resilient array.
- A single recording server: if one machine runs the CCTV software for a group of cameras, its failure stops recording even though the SAN is healthy.
- Shared management components: a single database, licence server or management node can create a dependency that is easy to overlook.
The most common mistake is assuming the SAN is resilient because the disk array is. A design is only as strong as its least duplicated element.
Which redundancy features remove SAN failure points?
Dual controllers, redundant power supplies, multipath connectivity, dual switches and hot spare drives together remove the main failure points in a surveillance SAN. Each feature protects one layer, so resilience comes from applying them all, with the recording servers and CCTV software configured to use the alternatives automatically.
The main features, layer by layer, are:
- Dual storage controllers: a second controller keeps volumes available if the first fails.
- Redundant, hot-swappable power supplies: separate supplies, ideally on separate feeds, remove power as a single point of failure.
- Multipath connectivity: two or more independent paths between server and storage let traffic continue if a cable or port fails.
- Dual switches or fabrics: duplicating the switching layer removes the network as a shared dependency.
- RAID with hot spares: parity-based RAID levels such as RAID 5, 6 or 60 tolerate drive failures, and a spare begins rebuilding without waiting for an engineer.
- Isolated operating system storage: mirroring the operating system on its own drives keeps a software fault or drive problem there from affecting the video arrays.
Secure Logiq builds this principle into its enterprise servers, which use dual or quad redundant, hot-swappable power supplies and separate hardware RAID 1 operating system drives. For multi-server designs, the system design service maps these layers against the project’s retention and uptime requirements.
How does failover keep cameras recording during a fault?
Failover keeps cameras recording by detecting a failed component and automatically switching its workload to a standby, so streams resume on the alternative path or server with minimal interruption. It works only if the standby is already configured, reachable and capable of carrying the full load.
There are two broad levels of failover in an IP CCTV camera system.
Storage path failover
Within the SAN, dual controllers and multipathing handle component failures. If a controller, switch or cable fails, the surviving path takes over, and the recording server continues writing to the same volumes. Depending on the design, this may be seamless or involve a brief pause while paths are renegotiated.
Recording server failover
At the application layer, the CCTV software or a virtualisation platform can restart or reassign camera streams to a standby server if the primary fails. Failover behaviour differs between VMS products, so the exact capabilities should be confirmed in the software vendor’s documentation. Secure Logiq’s Logiqal Core Architecture, described in its official announcement, includes automated failover and failback mechanisms for consolidated security platforms.
Failover can still lose a short window of video during the switch. Designers should define what gap is acceptable and, where necessary, use camera edge storage to bridge it.
What is the difference between RAID and SAN redundancy?
RAID protects data against drive failures within an array, while SAN redundancy protects the entire storage service, including controllers, power, switches and paths. RAID answers the question of what happens when a disk dies. SAN redundancy answers what happens when anything else in the chain dies.
The distinction matters because the two operate at different layers:
- Scope: RAID covers drives. SAN redundancy covers controllers, networks, power and connectivity.
- Failure types: RAID tolerates disk faults. SAN redundancy tolerates controller, switch, cable and power faults.
- Limits: RAID does not help if the controller or switch fails. Redundant paths do not help if too many drives fail at once.
Neither is a backup. RAID and redundant hardware keep systems running through component failures, but they do not protect against accidental deletion, corruption or a site-wide event. Archive and disaster recovery strategies need to be planned separately.
How can you test a SAN for hidden single points of failure?
Test a SAN by tracing the complete data path on paper first, then deliberately removing one component at a time in a controlled window while confirming that recording continues. Documentation shows where duplication should exist, and live failure testing proves whether it works.
A practical approach follows these steps:
- Map the path: draw every component from camera to switch, server, controller and disk, marking which are duplicated.
- Check power separation: confirm redundant supplies connect to independent feeds, not the same circuit or one shared power unit.
- Pull components under load: with cameras recording, disconnect one path, power supply or controller at a time and observe the result.
- Check alerts: verify that monitoring reports each fault, because a silent failure leaves the system running unprotected.
- Test recovery: restore the component and confirm failback and RAID rebuilds complete without disrupting recording.
- Validate performance during faults: a degraded array rebuilding under full camera load may struggle, so test with realistic stream counts.
Secure Logiq’s proprietary Logiqal Benchmark utility can simulate camera resolutions, frame rates and analytics workloads to stress-test server and storage throughput before physical deployment. Pairing that with fault testing shows whether a design copes when it is degraded, not only when it is healthy.
Does full redundancy mean a surveillance SAN never fails?
No. Full redundancy reduces the likelihood and impact of failure but cannot eliminate it. A redundant surveillance SAN can still fail through simultaneous faults, misconfiguration, software bugs, human error or site-level events that affect both halves of a duplicated design.
Several limits apply:
- Shared dependencies: duplicated hardware can still share a firmware version, a room or a power source, so one event hits both.
- Configuration errors: failover that was never tested may not work when needed.
- Capacity under fault: a surviving controller or path may not sustain the full recording load.
- Not a substitute for backup: redundancy does not recover deleted or corrupted footage.
The realistic goal is high availability with a clear understanding of residual risk, supported by monitoring, spares, maintenance and warranty cover. Secure Logiq’s enterprise products carry a five-year next business day onsite warranty, which addresses repair time, but good design and regular testing remain the main protection.
If you are planning or reviewing a resilient storage architecture, the Secure Logiq team can help you assess the full data path. Explore our system design service or contact us to discuss your surveillance storage requirements.
Frequently Asked Questions
Where should I start if I want to audit my existing surveillance SAN for single points of failure?
Start by drawing a simple diagram of the full path from camera to switch, recording server, SAN controller and disks, and mark every component that has no duplicate. Add power feeds, management software and licence servers to the diagram, since these are often overlooked. Prioritise gaps by the impact of failure (for example, a single controller affects every volume) and fix those first, then schedule controlled failure tests to confirm the duplicated parts actually work.
Can I add redundancy to an existing SAN without replacing everything?
Often, yes. Common upgrades include adding a second power supply and connecting it to a separate feed, enabling multipathing with an extra cable and port, adding a second switch, or assigning hot spare drives. A single-controller array usually cannot be upgraded to dual controllers, so that gap may require a hardware change. Check the manufacturer’s documentation for supported upgrades before purchasing parts.
How do I know if my failover is actually working if nothing has ever failed?
You can’t be sure until you test it. During an agreed maintenance window, with cameras recording under normal load, disconnect one path, power supply or controller at a time and confirm that recording continues and that alerts are raised. Repeat these tests periodically and after firmware or configuration changes, because a failover that worked at commissioning can silently stop working later.
What should I do about the short gap in video that can occur during failover?
First decide what gap is acceptable for each camera group, based on how critical the area is. For critical cameras, use edge storage on the camera (such as an SD card) so footage is recorded locally and can be back-filled once the recording server or storage path recovers. Also confirm in your VMS vendor’s documentation whether it supports automatic back-fill and how long the switchover typically takes.
What are the most common mistakes when designing a redundant surveillance SAN?
The most frequent mistakes are assuming RAID alone makes the system resilient, connecting redundant power supplies to the same circuit, and never testing failover. Others include sizing the surviving controller or path so it cannot carry the full recording load during a fault, and leaving shared dependencies such as a single database or licence server unprotected. A path-mapping exercise and realistic load testing catch most of these before deployment.
Do I still need backups or an archive if my SAN is fully redundant?
Yes. Redundancy keeps the system running through component failures, but it does not protect against accidental deletion, corruption, malicious activity or a site-wide event such as fire or flood. Plan a separate archive or disaster recovery strategy, such as replication to another location or exporting key footage to offline storage, based on your retention and legal requirements.
How should I monitor a redundant SAN so a failed component doesn't go unnoticed?
Configure alerts for controller faults, failed power supplies, path loss, degraded RAID arrays and drive failures, and send them to a monitored email address or management system rather than only a local dashboard. A silent failure is dangerous because the system keeps running with no remaining protection until the second fault occurs. Review alert tests regularly and keep spare drives and parts on hand so repairs happen quickly.
Related Articles
- How should SAN storage performance be tested before a CCTV system goes live?
- What is the role of a surveillance server in an enterprise VMS?
- How does an HD surveillance server work?
- When should you upgrade from an NVR to a surveillance server?
- Should you choose a surveillance server or an NVR for enterprise use?