Level
Practitioner
Reading time
300 min
Entries
36 lessons
You get
Certificate (70% to pass)
/ Courses

Backups & Disaster Recovery

For anyone whose data matters. Most backup disasters are discovered at the worst possible moment. Learn to design a backup and recovery plan that holds up when a disk dies, a server is hacked, or someone deletes the wrong thing.

42 quiz questions

Read the first lesson

The whole of lesson one, as written. The rest of the course reads the same way.

A scheduled job that prints SUCCESS tells you a process ran. It does not tell you the data is readable, complete, or that you still own the tooling to put it back. A copy earns the name backup only once you have restored from it and checked the result. Until then you have bytes and a hope.

The gap is not theoretical. In 2017 GitLab lost production data during an incident and discovered that five separate backup and replication methods had all failed silently: dumps produced empty files, cloud sync was misconfigured, and the snapshots they counted on were never taken. They recovered only because an engineer happened to have a six-hour-old copy sitting on a staging server. Every one of those five methods had been reporting healthy.

Treat the restore, not the job, as the deliverable. A backup you have never restored is an untested assumption parked next to your money. The job log is the least trustworthy signal in the whole chain, because it goes green whether or not the bytes are any good.

CopyBytes a job wrote somewhere. Unproven until read back.
BackupA copy you have restored and verified end to end.
Test restoreReading the copy into a usable system and confirming the data is correct.
Why it matters

If you have never restored a given backup set, plan as though you do not have it. Book the first test restore this week, not after the next outage.

Lesson 1 of 36.

Syllabus

  1. What a backup actually is

    6 lessons · module test

    1. 01A backup is a copy you have restored; everything else is hope
    2. 02Backup vs snapshot vs replication vs archive
    3. 03The seven ways backups fail silently
    4. 04What you protect against: hardware, human error, ransomware, provider loss
    5. 05Coverage gaps: what is quietly not in your backup
    6. 06Crash-consistent vs application-consistent backups
  2. 3-2-1 and where copies live

    6 lessons · module test

    1. 07The 3-2-1 rule, stated plainly
    2. 08The second medium: one disk type is not enough
    3. 09The off-site copy: geography and provider risk
    4. 10Off-line and immutable copies vs ransomware (3-2-1-1-0)
    5. 11Cloud object storage, versioning, and object lock
    6. 12Encryption at rest and who holds the key
  3. RPO and RTO

    6 lessons · module test

    1. 13RPO: how much data you can afford to lose
    2. 14RTO: how long you can afford to be down
    3. 15Reading RPO and RTO off a backup schedule
    4. 16Tiering systems by criticality
    5. 17The cost curve: why RPO of zero is rarely worth it
    6. 18Frequency vs retention: two different dials
  4. Schedules, retention and verification

    6 lessons · module test

    1. 19Full, incremental, and differential
    2. 20Grandfather-Father-Son retention
    3. 21Building a cron schedule for backups
    4. 22Monitoring: alert on failure AND on no backup in X hours
    5. 23Verifying integrity: checksums and automated test restores
    6. 24Off-host: never keep the only copy on the source
  5. Restore drills

    6 lessons · module test

    1. 25The restore runbook: exact steps, written down
    2. 26Single-file vs full-system restore
    3. 27Rebuilding a server from bare metal
    4. 28Database point-in-time recovery
    5. 29The drill cadence: quarterly and after big changes
    6. 30Measuring the real RTO during a drill
  6. The DR plan and the incident

    6 lessons · module test

    1. 31A DR plan is more than backups
    2. 32Dependencies and order of recovery
    3. 33Credentials, contacts, and escrow
    4. 34Communicating during an outage
    5. 35Full rebuild from nothing: the walkthrough
    6. 36The post-incident review
Enroll
$79