A database outage with different connection outcomes
A fictional on-premises Citrix site has two zones. Each zone has Controllers and VDAs, and the design includes Local Host Cache. During an authorized test, connectivity to the site database is interrupted. Existing sessions continue to run, and some new launches to eligible local resources succeed. A requested launch requiring a resource in another zone does not succeed during the outage. The project sponsor concludes that Local Host Cache is broken because its name sounds like a complete copy of the normal control plane. The operations team has recorded the relevant component events and test timestamps but has not compared the requested launch with the documented outage restrictions. No one has authorization to make broad production configuration changes. The architect needs to distinguish an expected design boundary from evidence of a separate fault.
1. How should the architect interpret the different outcomes?
Read the worked answer
Existing sessions and new brokering have different dependencies. LHC supports specified cached brokering behaviour and does not guarantee arbitrary cross-zone launches during outage mode. Compare the request, prerequisites and recorded events with the documented version-specific restrictions before declaring the mechanism broken.
2. What evidence and follow-up are appropriate?
Read the worked answer
Check the selected/elected local broker, cache readiness, eligible resources, available capacity and zone of the requested VDA, using correlated events. Document supported and unsupported outage cases, test remaining requirements in a bounded environment and retain database recovery in the runbook.