Docstoc

Documentation_ Backup and Archiving

Document Sample
Documentation_ Backup and Archiving Powered By Docstoc
					                           Backup
                         of computer files




Good Data Management Practice – Thursday May 8 2008 10.15-11.00
Gunnar Petersson, MEB-IT
 What is backup of computer files?

       The process of copying data so that this data can be restored
        after a data loss
       It’s simple but can also be very complicated depending on your
        environment (computer type, size of files, type of application, IT-
        Organization/Support etc)
       Backup vs. Archiving
        => Backup done everyday
        => Archiving is done a few times during a project
        => Backup purpose is to restore lost data
        => Archiving purpose is to preserve for future use
       But it’s VERY important that both are done.....and Why?

Gunnar Petersson MEB-IT                                              21 augusti 2011   2
 Why? – Some examples from ”my” world

  Spring 2006: 2 KI researchers found guilty of cheating –
   Documentation gone – And no backup and archiving
  Research projects never started because noone could find the
   old data – No backup and archiving
  Research projects spending weeks of work to recreate old data
  Doctorial students crying because they lost their only electronic
   copy of their thesis a week before submission
  Researchers loosing months of work when their harddisk
   crashes without backup
  Researchers spending thousands of Skr on trying to save
   crashed harddisks


Gunnar Petersson MEB-IT                                         21 augusti 2011   3
 Responsibilities

  Always, first and last: Yourself
  Don’t count on anybody else doing it for you. Always ask and
   confirm if it’s unclear what rules apply in the organization where
   you work and what’s demanded of you
  Have limited trust in technical systems. They can breakdown
   and people do mistakes




Gunnar Petersson MEB-IT                                         21 augusti 2011   4
 Backup – What are the threats?
  Threats:
   => Mechanical failure – Disk crash – Happens every
        month at MEB
   => Human mistakes – Unintentional delete/overwrite –
        Happens every week at MEB
   => Theft of computer (laptops) – Happens several times a
        year at MEB – Special consideration: Encryption
   => Malicious code as viruses, worms, trojans etc. – Has
        happened at MEB
   => Natural disasters (Fire, flooding, etc) – Has not happened
        at MEB (yet!)☺☺
  Backup protects your data and program files from getting lost or
   being destroyed = Will save you from disaster!

Gunnar Petersson MEB-IT                                       21 augusti 2011   5
 Backup – Different data file types

  Case 1:      Disaster Recovery – Restore of complete computer
                (Bare Metal Recovery, Automated System
                Recovery (ASR/XP))
  Case 2:      Independent files – Restore of individual files –
                Wordfiles, SAS-datasets etc
  Case 3:      Database files – Restore of complete databases –
                Attention for non-file-oriented databases like
                Oracle, MySQL, SQL Server = Special procedures
  You will have to use different tools and different procedures
   depending on what type of data files you are backing up




Gunnar Petersson MEB-IT                                     21 augusti 2011   6
 Backup – Different backup types
  First: Windows Archive Attribute shows if a file has been backed
   up or not. The Archive Attribute is reset by most backup
   programs but NOT all, e.g TSM. Dangerous running more than
   one backup program. The attribute is set by Windows when the
   file is modified in any way




Gunnar Petersson MEB-IT                                       21 augusti 2011   7
 Backup – Different backup types
 Full Backup
       Full backup = Complete backup of all files regardless of the archive
        attribute
       Pros: All files backed up. Easy to restore
       Cons: Takes lot of media space, can be slow depending on size

             Fr               Mo       Tu       We             Th   Fr


           Full               Full     Full     Full       Full     Full

                                                       Crash

                          Media needed to restore data

Gunnar Petersson MEB-IT                                                    21 augusti 2011   8
 Backup – Different backup types
 Differential Backup
  Differential backup = Backs up all files changed since last full
   backup – Archive attribute on file not reset – Cumulative
  Pros: Fast restore time
  Cons: Can take lot of media space depending on size
             Fr               Mo       Tu       We            Th    Fr


           Full                                                     Full

                                       Diff     Diff   Crash Diff
                             Diff


                          Media needed to restore data
Gunnar Petersson MEB-IT                                                    21 augusti 2011   9
 Backup – Different backup types
 Incremental Backup
  Incremental backup = Backup of files that have changed since
   last full, differential or incremental backup. Archive attribute
   reset
  Pros: Fast, needs smaller amount of media
  Cons: Restore can take long time.
            Fr                Mo       Tu       We            Th     Fr


           Full                                                      Full
                                                      Crash
                             Inc1     Inc2     Inc3           Inc4

                          Media needed to restore data
Gunnar Petersson MEB-IT                                                     21 augusti 2011   10
 Backup – Different backup types

  Adopt a Backup strategy based on these methods and the type
   of backup software and hardware that you have access to e.g A
   full backup every week and an incremental every night
  At MEB: TSM is incremental forever (Progressive) i.e One full
   backup one-time then only incremental. Information on backed
   up items are stored in an IBM DB2 database




Gunnar Petersson MEB-IT                                     21 augusti 2011   11
 Backup – Strategy Considerations
 Analyze your backup requirements and resources

       First: What data do I need to backup?
       Second: How much can I afford to loose?
       The answers here decide:
        1.     How much to invest in the backup system (technical solution)
        2.     Which combinations of full/incremental to use, time schedules
               and number of versions
       Third: Can I use my organizations backup facilities or can I buy from a
        Internet Service Providers (ISP), like Telia, or do I have to fix it myself?
       Fourth: Do my organization have a local policy for data protection?
       Fifth: If your organization doesn’t have neither facilities or policies,
        demand that they obtain it, it will payoff in the end. At KI the backup
        service also can be bought from ITC



Gunnar Petersson MEB-IT                                                       21 augusti 2011   12
 Backup – Strategy Considerations
 Some general advice

  Good folder structures facilitates backup
  Filenames are important
  Datasets are important but don’t forget other documents as
   programs and log books
  Document your backup process
  Make sure someone else know of your precautions. You can be
   indisposed in some way




Gunnar Petersson MEB-IT                                   21 augusti 2011   13
 Backup – Software Considerations

  If you use ISP or your organizations facilities; Inform you how it
   works. In the ISP case legal issues are important
  Different other approaches:
   => Use Windows built-in backup program
   => Look at some other easy-to-use backup program on the
         market
   => Make your own solution by simple copying to another
         media with or without help of zipping and encryption
         programs (e.g. Windows built-in zip, WinZip 9, Safeguard
         PrivateCrypto, Axcrypt, PGP)



Gunnar Petersson MEB-IT                                          21 augusti 2011   14
 Backup – Software Considerations
 Example of softwares

  Enterprise:
        IBM Tivoli Storage Management System (TSM)
        Symantec (Veritas) NetBackup
        EMC Legato Networker
  Small Server
        Symantec BackupExec
  Personal
        Microsoft Backup
        Symantec Norton Save & Restore
        Freeware/Shareware


  CDP – Continuous Data Protection
  HSM – Hierarchical Storage Management

Gunnar Petersson MEB-IT                              21 augusti 2011   15
 Backup – Media Considerations
  Diskettes (old 3.5”) are bad. Low capacity, low durability and
   sensitive. Newer types i.e Zip, Jazz are better
  USB sticks are easy to loose and have limited durability and
   capacity
  CD-ROMs and DVDs are better with higher capacity and higher
   stability. Might not be supported in backup program
  Magnetic tape (AIT,DAT,DLT,LTO) is not obsolete. Some of
   them is state-of-the-art. They have large capacity (LTO-3 900GB
   per tape) and high stability. But the drawpack is the price
  Disk (copy over network). Good capacity, stability and restore
   time. Removable disks another option
  Media should be stored at another location, fire-proof, access
   controlled and climate controlled
  Backups should be tested

Gunnar Petersson MEB-IT                                      21 augusti 2011   16
 Backup – Restore possibilities without backup
  A deleted file may still be in the Recycle Bin




Gunnar Petersson MEB-IT                             21 augusti 2011   17
 Backup – Restore possibilities without backup
  If you are connected to a Windows Server, Shadow Copy
   Service may be running which means you can restore older
   copies. It’s called Previous Versions in the Properties tab




Gunnar Petersson MEB-IT                                      21 augusti 2011   18
 Backup – Restore possibilities without backup

  You can use programs like Norton System Works or freeware to
   restore deleted files
  A crashed hard disk with important data can be sent to
   companies like ibas (nsec) but costs a lot of money (10000 Skr
   40GB). No guarantee that it works
  Precautions like a RAID-card and mirrored disks (Raid 1) or
   using some small NAS solutions like Netgear SC101/D-Link
   DNS323 for your disks can also save you. More advanced
   solutions are called like ReadyNAS/Multimedia Storage Server
  RAID 0, 1, 5




Gunnar Petersson MEB-IT                                     21 augusti 2011   19
 Backup – Some concluding remarks
       Recommendations:

           Backup everyday

           If you have little data and programs you can do full backups everytime. If not,
            only backup the files that have changed cf. incremental backup

           Name the backup folders/files with a timestamp

           Decide how many versions you need to store

           Beware that a mistake in a file also is backed up i.e. You may have no good
            copies if you don’t discover the mistake in time and delete older versions

           Backups should be done to at least 2 different medias since media sooner or
            later will fail

           Store the media at another location than the computer and preferably fire- and
            burglarproof. Consider encryption




Gunnar Petersson MEB-IT                                                                 21 augusti 2011   20
      Backup – How – The MEB case
       Central backup solution with IBM Tivoli Storage Manager and a
        FC-SAN based IBM DS4200 disk-backup-system (28TB, Raid
        5) + HP taperobot (8TB)

                                                                30 Servers
                                                       TSM      150 GB/day
                                                       Client




                          TSM Server
                           IBM p615


                                 DB
                                          IBM DS4200
                                          HP MSL5026




21   Gunnar Petersson MEB-IT                                        21 augusti 2011   21
 Backup – How – The MEB case
       Incremental forever = One full backup + continuing incremental

       Backup hardware stored in a fully protected Data Center:

          1.    Access control
          2.    Climate control
          3.    Redundant power
          4.    Automatic fire protection
          5.    Alarms connected to 7/24 emergency service centre




Gunnar Petersson MEB-IT                                             21 augusti 2011   22
 Backup – How – The MEB case
  Always 2 different total copies, Filepool + Copypool in IBM
   Taperobot at central ITC (another building, copied each night).
   Local taperobot used as fill-over protection and for special
   purpose backups

             Source MEB                               Target ITC
             TSM Server                              TSM Server

                                       Copies

                                           DB


       DB
                                                DB




                          IBM DS4200                       IBM 3584 Taperobot
                          HP MSL5026
                                                       MEB Safe

Gunnar Petersson MEB-IT                                             21 augusti 2011   23
 Backup – How – The MEB case
  A maximum of 3 versions of a file is stored for a maximum of 60
   days. A deleted file will exist for 90 days in the backup system
  Backup is automatically performed to disk once a day on P: and
   Z: (Fileserver Thor)
  Backup is also done 3 times a day at 07:00, 12:00, 15:00 by MS
   Volume Shadow Copy on P: and Z:. You can yourself restore
   files by using Windows Explorer, mark Z: -> Properties ->
   Previous versions -> View/Restore
  Backup on laptops are managed by Windows ”make available
   offline”. Mark Z: -> Mark wanted directory -> Right mouse ->
   Make available offline. Tip: Create a special folderstructure from
   the beginning. Syncronization may take some time when
   connecting/disconnecting the laptop
  E-mails are not backed up at MEB. IMAP-function should be
   used to store copies on central KI system
Gunnar Petersson MEB-IT                                         21 augusti 2011   24
              MEB Backup Schematic Example

07:00         Volume Shadow Copy is run on disk Z: (Thor)
09:00         Laptop is connected to MEB Network => Synchronization of changed files to
              Z: (Thor)
12:00         Volume Shadow Copy is run on disk Z: (Thor)
15:00         Volume Shadow Copy is run on disk Z: (Thor)
17:00         Laptop is disconnected from the MEB Network => Synchronization of
              changed files to Z: (Thor)
00:10         Changed files on Z: (Thor) are copied to the temporary BACKUPPOOL on
              backupserver Ragnarok
04:00         BACKUPPOOL on Ragnarok is emptied to the permanent FILEPOOL on the
              SAN disk system IBM DS4200
05:00         Changed files in FILEPOOL on the SAN disk system IBM DS4200 are copied to
              the ITC Tape Robot IBM 3584




  Gunnar Petersson MEB-IT                                                    21 augusti 2011   25
 The End


  Questions?

  Thank you!




Gunnar Petersson MEB-IT   21 augusti 2011   26

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:11
posted:8/21/2011
language:English
pages:26