Chapter 8 · Reading 4 of 5
Finding the dependency both sites share
Security+ SY0-801 · Objectives 3.1, 3.4
The college has paid for a second appointment server. During an exercise the primary server stops, the standby comes online, and students still cannot book appointments. The technician can show that both machines behaved according to their configuration. The advisor can show that the service failed. To reconcile those accounts, draw the dependencies between the machines and the people who need them.
Ask what can fail together
A failure domain is a set of components vulnerable to a common failure. Two servers in different racks can share a power feed. Two regions can rely on one identity provider. Two independent storage copies can both accept the same mistaken deletion. For each layer, imagine losing one dependency and follow the effects outward. State the failure you are trying to survive before treating redundancy as proof of resilience.
Distinguish capacity from independence. A standby might have its own power and network but too little capacity to serve registration traffic. An application might support a read-only mode that lets advisors view existing bookings while new requests queue. A restricted service can be valuable if the business owner knows which work remains possible and how to reconcile it later. Let the advisors define what “useful during an outage” actually means.
Read a diagram from the bottom up
Take the student’s successful booking and work backward. The confirmation depends on an application decision. That decision depends on an available record store and a valid identity. Identity validation may depend on name resolution, current time and a reachable signing-key service. You may discover that parts labeled independent in two purchasing agreements rely on one internally managed service. Ask the owner how they would detect a fault in that service from outside its own monitoring system.
Assign each shared dependency an owner and one observation that would distinguish its failure from a site outage. If sign-in fails at both sites while authorized existing users can still view appointments, that difference narrows the investigation. Preserve the distinction in status updates so people who still have working access can use the limited service appropriately.
Compare two failover arrangements
An active-passive arrangement keeps one environment ready to take over. Its value depends on current configuration, accessible keys and a tested transition. An active-active arrangement can serve traffic from more than one location during routine operation. It brings its own questions about consistency, routing and simultaneous updates. In either arrangement, identify who can declare failover, where the evidence will be collected and how the team avoids two systems accepting conflicting changes.
Try a short tabletop. The primary application is healthy, the identity service is unavailable and existing sessions still work. What would a monitor that checks only the landing page report? What would a newly arriving student experience? Monitor a representative sign-in and booking alongside infrastructure health so the operations team sees the service from both perspectives.
When the dependency returns, determine what must happen to requests queued during the incident. Processing every queued request twice could create duplicate appointments. Discarding the queue could lose legitimate work. Preserve a stable transaction identifier, reconcile outcomes and ask an advisor to verify a sample. The time and work involved in that reconciliation belong in the recovery conversation.
Inspect the three records, then use the evidence to answer the question.
1. Primary site
Site A serves bookings through identity service I1.
2. Standby site
Site B has a separate application and database replica.
3. Authentication
Both sites require I1 for a new user sign-in.
Apply the reading
Which service transaction would reveal a shared identity failure?
Compare your reasoning
A fresh sign-in followed by a booking exercises a dependency that a public landing-page check can miss.