Deep Battle Cards E–H
E — Development vs Test vs Production vs Sandbox
Definition: Development = first controlled place for changes; Test/QA/UAT = formal validation; Production = mission-critical live service; Sandbox = isolated experimentation/POC.
Purpose: protect Production while enabling repeatable change and safe experimentation.
Inputs: change/test objective, data sensitivity, Production risk.
When used: first patch → Development; UAT/integration/performance → Test; live service → Production; hypothesis/POC → Sandbox.
Common confusion: Sandbox is not formal Test and isolation does not remove data-security restrictions.
Scenario: exploratory model → Sandbox; regression/UAT before release → Test.
Exam trap: urgency or small change size does not justify direct Production testing.
Memory hook: Build → Prove → Run; Sandbox explores beside the path.
Source: pp. 182–184.
F — Retention vs Archive vs Purge
Definition: Retention = how long data must remain available; Archive = move less-active data while preserving retrieval; Purge = irreversible removal.
Purpose: manage lifecycle cost while honoring legal/business obligations.
Inputs: retention law/policy, usage, value, risk, storage cost.
When used: duration requirement → retention; rarely used but retrievable → archive; obligation expired + authorized no-need disposition → purge.
Common confusion: old/inactive does not mean purgeable.
Scenario: seven-year legal availability → retention/archive; period expired and disposition approved → purge.
Memory hook: Retention = how long; Archive = where kept; Purge = gone.
Source: pp. 186–188.
G — CDC vs Replication vs Mirroring vs Log Shipping
Definition: CDC identifies/transmits deltas; Replication maintains copies; Mirroring is relatively immediate secondary synchronization; Log shipping applies transaction logs at intervals.
Purpose: match movement/copy mechanism to change volume, latency, continuity, and cost.
When used: changed rows → CDC; same data multiple nodes → replication; near-current failover → mirroring; tolerable periodic lag → log shipping.
Common confusion: CDC can feed replication but is not synonymous with maintaining a replica.
Scenario: logs applied every 15 minutes → log shipping.
Exam trap: do not full-refresh a huge object when tiny deltas are all that must move unless requirements demand it.
Memory hook: CDC spots changes; Replication copies; Mirror stays close; Logs travel in batches.
Source: pp. 188–193.
H — Backup vs Recovery Test vs Resiliency vs Recovery
Definition: Backup creates protected copy; recovery test proves the copy/procedure can restore; resiliency tolerates failure while continuing; recovery restores after loss/failure.
Purpose: distinguish stored protection from proven recoverability and continuity-by-design.
Inputs: criticality, SLA, backup frequency, failure scenarios.
When used: “copy exists” → backup; “can it restore?” → recovery test; “continue despite failure” → resiliency/failover; “restore after failure” → recovery.
Common confusion: successful backup jobs are treated as proof of recovery capability.
Scenario: daily backup never restored = protection claim unproven.
Exam trap: same-site backup can share the failure mode of Production.
Memory hook: Back it up; Test it; Survive it; Recover it.
Source: pp. 194–197.