Backup & Restore

1. Backup & Restore • The purpose of backup is to protect data from loss. • The purpose of restore is to recover data that is temporarily unavailable due to some unexpected event.

2. Backup To backup or not to back up, that is the question. • Backup is not free. • No backup is risky.

3. Proper Backup Procedure • Choose your application • Scheduling • Implementation • Inventory (content and media) • Verify • Automate • Secure

4. Factors • Determine which data is critical. • Determine frequency and types of backups to be used. • Full • Differential • Incremental

5. Factors • Determine which data is static and which is dynamic. • Some OS installations are changed infrequently; few backups required • E-commerce may require continuous backups. • Understand the changing state of your client’s data to determine an appropriate backup sched. • Organize with partitions

6. Factors • Determine the appropriate media storage for your backups: • CD • DVD • Tape • Disk • Solid State?

7. Factors • Partitioning of disk space is used to manage backups

8. Choose your Backup App • Mac OS X : • Time Machine • Linux/Unix : • tar (tape archive), cpio, dump • MS Windows : • MS Windows XP & 7 includes Backup & Restore capability • Many commercial apps are available

9. Enterprise Level Backup Apps • Paragon Backup & Recovery includes customer support • Backup4All Professional • GRBackPro7

10. Schedule • Determine the categories of data so you can schedule the backups accordingly

11. Schedule • Partitions are often used to manage backups • Examples: • The OS has its own partition and may require infrequent backups if changes are quarterly • User data may require nightly backups • Users must know what partitions have backup and the frequency (SLA).

12. Schedule • Full Archival Backup • image backup implies copying the unused space. • Differential Backup – what has changed since the last backup • Incremental Backup – what has changed since the last backup of anytype

13. Full Archival Backup • Mirror – every last bit on the disk is duplicated. • Many full backups ignore empty space.

14. Full Archival Backup (Pros) Pros • Provides a complete copy of data • Easy to manage: • Done less frequently than other types of backups due to cost and resource requirements: Monthly, Quarterly, semi-annually, annually.

15. Full Archival Backup (Cons) Cons • Usually requires more media space than either differential or incremental. • Takes a long time to recover the full backup to a new disk.

16. Full Archival Backup • Consider making multiple backup copies • Full backup media should be stored offsite to protect data from disasters • Fire, Flood, Earthquake, Terrorist attack, Sabotage, Hacker attacks

17. Full Archival Backups • The trend is to reduce the cycle of full backups. This is because of liability. Files that are not backed up cannot be subpoenaed. • Statute may require destruction of some records.

18. Differential Backup • Copy files changed since the last full backup. • Differential backups grow with time. They can eventually grow larger than the last full backup. • Scheduled less frequently than a full backups: Weekly, monthly.

19. Differential Backup (Pros) • Redundancy • Usually takes up less time and space than a full backup. • If the differential backup grows to the size of the last full backup, then schedule a new full backup.

20. Differential Backup (Cons) • Redundancy – potentially many unneeded copies of the same data. • Subsequent differentials take longer and use more media space.

21. Incremental Backups • A backup of what has changed since the last previous backup of any type. • Frequency of incremental backups depends on the client needs. • Weekly, daily, hourly, continuously.

22. Incremental Backups • Pros • Keeps a revision history of actively changing files • Fastest backup type • Uses the least amount of media to complete a single backup • Cons • Much more difficult to manage

23. Schedule Example • Full backup twice per year • Differential each first Saturday morning of each month that is not scheduled for a full backup • Incremental each Saturday morning that is not scheduled for a Full or Differential

24. Other Schedule Considerations • Consider completing a backup in conjunction with and before any major system changes are scheduled.

25. The Actual Backup • Assignment of responsibilities • Written in the SA’s job description • Acceptance of accountability • A signed form indicating that the backup was complete, verified and secured

26. The Actual Backup • Hardware • Choose the media type • Centralize the backup to reduce redundancy of hardware • Not everyone archives the local “C-Drive” on general purpose workstations. (SLA and user awareness)

27. Backup Inventory • Inventory the backup media • Tapes and other writable media use barcodes or hand-written labels • Inventory the content of the media • The backup should have a table of contents included in the backup

28. Backup Inventory • The media label information: • Date • System identifier • Partition name(s) • Backup category: full, differential, incremental

29. Verify • The only time you know the quality of your backup media is when you are doing a restore. • This is the worse time to discover you have problems. • Restore a small subset of random files from the backup. Verify their integrity through differences or checksums.

30. Verify: Firedrills • When new equipment arrives, test your backup procedure on the new equipment to verify it works correctly

31. Backup Automation • Automation reduces human errors. • Many pre-packaged apps include automatic scheduling • Linux/Unix backup scripts can be submitted using the cron utility. Logs can be kept in /var/log, and e-mail can be sent to the admin.

32. Secure The Backup • Offsite storage • Encryption: to encrypt or not to encrypt, … • Will the encryption key always be available? • Statute or contract (SLA) may require encryption

33. Secure: Off-Site Storage Off site storage has risks all its own • Data can be lost/destroyed in transit • How important is it to have a backup of the backup? • Some backups can be kept on site

34. Secure: Example Strategy Where the backup is stored will impact service response to restore requests: • Consider keeping incremental backups on site. • Differential and full backups could be stored offsite.

35. Data Compression • Risks – if the media is damaged, recovery may be difficult or impossible. • Lossy • some data tolerates degradation (loss of information) • No-loss • Some data should not be compressed. Know your data!

36. Secure Backups • Contract to store your data in a “secret” offsite location. (Secret implies a need to know basis)

37. Backup Considerations • Backups slow down service. This should be included in the SLA • Files should be write-locked during backup. • Avoid doing backups during peak service hours. Schedule during early AM hours on the weekend and holidays.

38. Restore • Common reasons for restores • Accidental file deletion • Disk failure • Disaster recovery • Fire, flood, earthquake, hacker attack, sabotage, terrorist attack, etc.

39. Accidental File Deletion • If backups are once per day, lost work is limited to one day for a given file. (RAID does not help) • As storage technology gets cheaper by the Gbyte, it becomes easier to implement more sophisticated storage procedures that are more timely. (HDD backup). • A user wants the restoration to be immediate. The quicker the turnaround, the happier your customer.

40. Disk Failure • A disk failure causes two problems • Loss of data • Loss of service • Critical systems should implement RAID so that disk failures do not cause a loss of service. • Restoring an entire disk is slow. Service is hampered until the last bit is recovered. • Consider using hot spares and hot swap

41. Disk Restore from Tape • Restoring from tape can interrupt service. • Restoring from tape slows the restore process by a factor of about 5-10 times compared to a simple disk to disk copy.

42. Tape Backup • Large amounts of data historically favored tape media for backup: • Tapes are portable and fairly durable.

43. Tapes • Tape historically has been the preferred backup media for very large data storage environments. • Tape has a useful life span. • Tape can be very robust for storage • Easy to transport • Some tape formats are more reliable than others.

44. Tape Inventory • Backup tapes must be • Properly labeled • Properly stored • Proper inventory is needed to do restores in a timely fashion. • Hand written labels are ok • Bar codes and printed labels are better.

45. Tape Inventory • Inventory is not limited to the physical tape itself. • The contents of tapes must be inventoried. • The number of accesses must be logged because tapes ware out. • Tape equipment is not free. You don’t want to purchase any more hardware than is needed.

46. Tape • Rotate media • Incremental backup stored on site can be reused.

47. Tape Standards • 8mm • DLT • DAT (4mm) • QIC • http://www.pctechguide.com/15tape.htm • Imation.com

48. Tape Technology • Tape technology expands in leaps. • Tape hardware purchases are not made on a constant basis (like disk storage). • Tape technology is purchased in leaps. Three year intervals are more practical.

49. Firedrills • An occasional test of a full partition restore is not unreasonable. This would be done if • A change is made in the backup software release. • A change of vendor for the backup software product. • When a new server with new unused disk arrives.

50. Centralization • Without centralization, a tape drive is needed for each server location. • Equipment can be interchanged more easily when centralized.