AgentSkillsCN

disaster-recovery

备份策略、灾难恢复计划和业务连续性。

SKILL.md
--- frontmatter
name: disaster-recovery
description: Backup strategies, disaster recovery planning, and business continuity.

Disaster Recovery

RTO vs RPO

MetricDefinitionExample
RTORecovery Time Objective4 hours
RPORecovery Point Objective1 hour

RTO: How long until service is restored? RPO: How much data can we lose?

DR Strategies

StrategyRTOCost
Backup & RestoreHours$
Pilot LightMinutes$$
Warm StandbyMinutes$$$
Multi-Site ActiveSeconds$$$$

Backup Strategy

Database

bash
# PostgreSQL
pg_dump -h host -U user dbname | gzip > backup.sql.gz

# MySQL
mysqldump -h host -u user -p dbname | gzip > backup.sql.gz

Automated Backups

hcl
# AWS RDS
resource "aws_db_instance" "main" {
  backup_retention_period = 7
  backup_window           = "03:00-04:00"

  # Cross-region replica for DR
  replicate_source_db = aws_db_instance.primary.arn
}

DR Runbook

Failover Steps

  1. Detect - Monitor alerts for primary failure
  2. Assess - Confirm failure, estimate recovery
  3. Decide - Failover if RTO exceeded
  4. Execute - Run failover procedure
  5. Verify - Test functionality
  6. Communicate - Update stakeholders

Failback Steps

  1. Verify primary is healthy
  2. Sync data from secondary
  3. Switch traffic back
  4. Monitor closely

Testing

Test TypeFrequency
Backup restoreMonthly
Failover drillQuarterly
Full DR testAnnually

Multi-Region

code
Primary (us-east-1)          Secondary (us-west-2)
┌─────────────────┐          ┌─────────────────┐
│    App + DB     │ ──sync── │   DB Replica    │
└─────────────────┘          └─────────────────┘
        │                            │
        └──────── Route 53 ──────────┘
                (failover)