SageIE: Backups and Disaster Recovery

1
Sage-IEBackups and Disaster Recovery

• Andrew Barnes
• http//www.ramsesit.co.uk
• andrewbarnes_at_ramsesit.co.uk

2
Covered in this session

• Backups
• Disaster Recovery
• As SysAdmins, the focus may appear to be on the
  data-recovery component gt
• Examples
• Questions

3
Why Backup?

• The obvious answer for System Administrators
• To protect data against a myriad of unforseen
  circumstances
• The not-so-obvious answer for System
  Administrators
• To protect business processes and procedures
  against the loss of critical staff or other
  non-technical disasters

4
What to back up?

• Depends upon your organisational needs
• Doesnt just cover the actual data
• System Information/Configuration
• System Information
• Disk Information
• Network Configuration
• Application Information/Configuration
• Business Backups
• Business Processes
• Technical Procedures
• Staff Details

5
When to back up?

• A number of factors
• What the data consists of
• How often your data changes
• What level of redunancy you require
• How much data you have that requires backing up
• How much data you can afford to be without
• How long can you afford to be without the data
• Only for a short period of time, or until data
  can be re-entered (if at all)
• Depends on backup medium utilised
• Can become a Need versus Cost question

6
How to back up?

• Purely Hardware
• Disk Mirroring
• On local disks
• On off-site disks (EMC, HDS, etc)
• Synchronous versus Asynchronous Mirroring
  Performance implications
• RAID
• Software with Hardware
• Backup application
• Unix/Windows/Macintosh applications tar, gzip,
  dump, dd, winzip, StuffIt, others
• Commerical Applications TSM, Legato, Backup
  Exec, others
• Backup Levels
• The best solution? A combination of both

7
RAID 5 on Floppy Disks?

• One would argue this is not an option but you
  gotta admit it looks kinda cool gt

8
Verifying Integrity of Backups

• Two of the most crucial components of backups
  consist of
• Making sure you have backed up the correct/enough
  data
• Making sure you can recover specific data as and
  when required
• Defining the backup retention period is crucial
  espectially if there are legal requirements (aka
  Traffic Data Retention!)
• There is an important point here Data Recovery
  must be tested on a regular basis
• An email/message stating that a backup has
  completed successfully is not sufficient
• If backing up your data is half the battle,
  recovering it makes up the other half
• This is further complicated the longer your
  backups are required

9
Storage of Backup Media

• Another consideration for organisations is
  whether off-site backups are required
• Offsite backups can take different forms
• The home of a trusted employee
• Another office
• Third Party Organisation
• Care must be taken to carefully evaluate the need
  of off-site storage versus the cost, versus the
  security associated with each of the options
• In particular, the cost of data retrieval

10
Security of Backups

• There are three main ways to get access to your
  data
• Remote access to a system
• Physical (local) access to a system
• Access to the backup system used
• Your companys data is only as secure as its
  backups

11
Why so much time on backups?

• The answer here should be obvious
• The ability to retrieve important corporate
  information is crucial to business continuity
• I have seen too many companies, and individuals,
  trivialise the issue of failed backups
• It is important to note that data recovery is
  only one part of of Backups and Disaster
  Recovery
• Ala the next section of this presentation

12
Disaster Recovery

• What constitutes Disaster Recover?
• One persons distaster might be another persons
  inconvenience
• Are you faced with a Disaster, or with an
  application/service outage
• This depends upon your terms of service SLA
• There are two aspects two Disaster Recovery
• Data Recovery
• Business Recover

13
Business Recovery

• Business Continuity Plan(ing) (BCP)
• Business Continuity Management (BCM)
• Both the above acronyms describe a critical
  component of Disaster Recover
• Planning
• Management
• BCP protects an organisation from loss of
• Site
• Critical Employees
• Organisation/Operational Procedures

14
Components of BCP 1

• Identify Key Personnel
• In the event of a disaster, what is the minimum
  number of staff required to either
• Continue service Business As Usual (BAU)
• Restore service
• Ensure BAU where possible, and restore those that
  have been lost
• Contact Details of these personnel
• Alternative Site
• Alternative Site in the event of a particular
  data centre
• Hot or Cold Site
• Alternative Equipment
• Depends on site used (hot/cold)
• Use of On-Call equipment is often useful

15
Components of BCP 2

• Identify Applications required to support
  services
• Firewall GUIs
• Application Management Utilities
• You may be required to (re)build a new management
  workstation
• Identify Access Mechanisms
• Physical Access requirements
• Network Access requirements
• Dont firewall yourself off from your network
• Single/Dual Point of Access

16
Components of BCP 3

• Critical Login Information
• Disaster Recovery Pack
• Stationary
• Installation Media
• Other
• Dont bite off more than you can chew consider
  the use of a Service Continuity Plan
• A more modular approach to Business Continuity
  Planning and Management
• Also useful for identifying priorities in the
  event of problems

17
System/Data Recovery

• As per previous slides
• Remember - backing up data (and the ability to
  recover it), is all for nothing if you dont know
  the system configuration
• System Information/Configuration
• Network Configuration
• Application Information/Configuration
• There is no point in having the above information
  stored on a system that is itself lost
• Hard copy
• rsync of System Information Site to an
  alternative location

18
Data Centres

• Again, the number of data centres depends on your
  companies requirements
• Speed of recovery
• Cost to implement and maintain
• One, Two or even more?
• Hot or Cold data centres
• The more data centres required, the more complex
  the network (and system) infrastructure
• Especially where both sites are live at any one
  time
• The more access points to your network, the
  greater the risk of system/network compromise
• Did I mention cost?

19
Planning for a Disaster

• You need to collect
• Business D/R Information
• System/Network D/R Information
• Make sure that all staff know
• That you even have a D/R plans
• Where to find your D/R plan
• How to activate the D/R plan
• Under what circumstances a D/R plan should be
  activated
• Where possible, identify a recovery priority
  (incorporating dependancies)

20
Planning to Recover

• Documentation
• BCP (Business and Data)
• Every time you change a component, you should
  update your BRP
• TEST your recovery solution, on a regular basis
• If you factor in BCP as part of any new service,
  you will help ensure the continuity of your
  business in the event of any number of potential
  disasters

21
Examples

• Downtime Table
• Useful when setting SLAs
• Tape Backup Levels
• Data Storage
• Iron Mountain
• Disaster Recovery
• The Wall Street Journal September 11th 2001
• Lehman Brothers September 11th 2001

22
Downtime Per Year (count of 9s)
23
Tape Backup Levels
24
Iron Mountain

• Paper Records Stored approx 207 million cubic
  feet
• Computer Media Stored gt 80 million pieces
• Microfilm Stored gt 50 million pieces
• Number of Retrievals and Re-files each year gt
  72 million
• There is a key point here The cost of off-site
  storage with a 3rd party organisation could end
  up costing more than the actual storage itself

25
The Wall Street Journal 9/11

• Had extensive BRP
• Based on experiences of the 1990 power blackouts
  in NY
• Primary site located across the road from the
  Twin Towers
• BRP Plan activated by senior management
• Elected to abandon the primary site
• Full edition of paper released the following day
• Complete with current stories, including
  eye-witness reports on the collapse of the Twin
  Towers

26
Lehman Brothers 9/11

• Primary Sites in both Tower 1, as well as 1, 2, 3
  WTC (across the road from the Towers)
• Moved to BRP Site
• Performed Cash Management functions the same day
• Fixed-Income Securities trading on next day
• 400 traders online when NYSE re-opened on the
  17th September
• Testimony to the excellent BRP solution planned
  by the company

27
The End

• Questions?

28
References

• http//www.uni-mainz.de/neuffer/scsi/what_is_raid
  .html
• http//ohlssonvox.8k.com/fdd_raid.htm
• http//www.ironmountain.com
• http//www.sysadminmag.com
• http//www.tldp.org/LDP/sag/x2615.html