Tivoli Storage Manager Part 3 CDP

1
Tivoli Storage ManagerPart 3 - CDP
Sriram Venkatesan IT SpecialistIBM Academic
Initiative
2

• TSM

3
Tape Reclamation

• Better utilizes tapes, thus, saving money
• Tape utilization constantly monitored
• User-defined reclamation threshold
• When free space reaches threshold
• Tape is mounted
• Valid data moved to another tape
• Original tape is returned to the scratch pool
• Can be scheduled to occur at specified times

100


4
Collocation

• Reduces tape mounts on restore, thus, expediting
  restores
• Stacks all data for a client , filesystem or
  group of clients onto the same tape or group of
  tapes

Hi Threshold
Disk Pool
A
C
A
B
B
B
Lo Threshold
C
A
B
Migration
Migration
Client A
A
Client A
Client B
Client C
Client B
B
Tape Pool
B
B
B
C
A
A
A
C
C
Client C
C
B
5

• TSM Express Video

6
CDP
7
Why is CDP for Files Needed?

• Most data protection solutions miss the most
  valuable data (what the user is working on now)
• Daily backup allows too much productivity loss
• WORSE, 70 of corporate data resides on
  desktops/laptops and fewer than 8 are
  backed-up at all
• End-users everywhere simply do not backup
  because no solution has been as transparent as
  air bags, until now

Continuous Data Protection is the next evolution
of backup
8
What is Tivoli CDP for Files?
A new backup paradigm using a unique1 hybrid
approach
Traditional Backup
Replication

• Versioning of files
• Point-in-time restore
• Central administration
• Archiving (vaulting)
• Retention
• Highly scalable

• To-Disk duplication
• Lightening fast
• Real-time for high-importance files
• Tolerant of transient networks
• Multiple targets

The new direction in Data Protection is
Real-Time, To-Disk, Native Format, Simplified
management
1 10 patents filed
9
IBM Tivoli Continuous Data Protection for Files

• IBM Tivoli Continuous Data Protection for Files
• Continuously protects important files without
  doing a thing.
• Protects all other files on a schedule.
• Tolerant of network spottiness.
• Up to three copies made.
• Local cache for highest performance, least
  impact, and greatest availability.
• Off-machine for protection against, well, life.
• Run any where anti-virus runs (file servers,
  laptops, workstations, etc.)

Replication (2)
File Server or USB or Removable
Replication (1)
Replication (3)
Primary Disk
Local CDP Cache
TSM Server
Transparent, always-on, airbag-like protection.
Whether connected or not.
10
Configuration

• High-Priority Files (Continuous Protection)
• Specify up to three target areas for high
  priority files.
• Target area(s) will capture every save of a file
  when it occurs.
• Local target area keeps versions of each file
  savedauto-managed to a specified pool size.
• Local area allows for un-tethered restore
• Option to specify a remote file server for
  off-machine protection still real-time, but
  tolerant of being disconnected.
• Option to specify a TSM server for off-machine
  protection still real-time, but tolerant of
  being disconnected.
• All other files (Scheduled Protection)
• All other files are protected on a periodic basis
  to the remote file server or TSM server.
• Change-journal eliminates ever having to scan the
  file system.
• Files versioned on remote file server allowing
  for point-in-time restore.
• Remote versions are automatically managed to a
  configurable pool size.

11
Configure CDP These files will be immediately
copied when they change sent to 3 different
target areas potentially. Storing locally, users
have local protection and restoration up to a
configured limit every save is a unique
version. Off-machine protects against primary
machine loss.
Recommended configuration is CDP Local only (not
to remote), and use Scheduled for remote no
reason to over-burden the network and the target
device.
12
Configure Scheduled Protection These files will
be protected on a schedule (typically all
files). Target is off-machine (file server,
external drive, external LUN, anything).
13
Status Screen Live update Recent activity
typically, files that have been copied to a
target area Summary of number of files protected
under CDP Summary of scheduled protection,
including accurate pending count.
14
Backup File Server Admin Visual summary of other
clients and their recent activity. Drill-down for
details. Any client not backed up within 24 hours
is shown in YELLOW. Any client that resulted in
failed transfers is shown in RED. Clicking on a
client displays full XML configuration of that
client (platform info, configuration database,
operational statistics, etc). Every backup
results in XML log files of the activity that can
be examined. All logs are retained (managed by
pool size) and can be viewed or analyzed by other
XML tools.
New Publish feature
15
Restore Screen Specify which area to restore from
(local or remote) If desired, specify a
point-in-time to restore-to. Typical select
and/or drill-down interface Restore single files
or trees. Typical backup products have to employ
a database to keep track of material CDP uses
the file system itself and dynamically builds the
restore views. File systems are actually pretty
powerful databases in and of themselves.
16
Restoring versions If a file has multiple
versions available, the count is
shown. Drilling-down into such a file will show
all the known versions users can select a
specific one to restore, or use point-in-time
specification. Again, this is accomplished
without a database clever on-disk file naming
conventions (patents pending) allow for tracking
material easily. All files are stored natively,
as files users can use their own tools to find
or index or search or restore, if desired.
17
New Advanced Features
18
Topics we should include

• Multiple Machines
• Changing the Reporting/Administration area
• Closing Applications During Scheduled Backup
• Scaling
• Purging and Removing Material
• Backing-up the backup file server
• Distributed Installs and Configuration updates
• Push/Pull install

• Using Tivoli CDP for Files on servers
• Sequencing Backup times
• Encryption and Security
• Delta Replication
• LAN-Free Backup
• SnapShots, Open Files, and Application Quiescing
• System directories and registry backup and
  recovery

19
Archive Retention

• Easy to Configure.
• Create folders defining a length of time for file
  retention.
• Drag-and-Drop, copy or save files to the
  appropriate VitalFile folder.
• Those files cannot be deleted or altered for the
  defined time period.

20
Advanced Settings Version control for remote
targets. Sub-file copy for files over a size
threshold. Throttlingin kbs per
replicator. Preemption. Skipping extra-big
files.
21
Advanced features (continued)

• Sub-file 2 techniques
• Bitmap for database-like files
• Delta-calculation for others (requires that user
  have the same files in the local CDP cache)
• Sub-file currently does not version on target
• Encryption
• WebDAV support coming can specify a web-server
  as a target, such as https//mysite.net (note
  this can use SSL)

22
Architecture
23
Code break-down
All C, plus HTML and jscript no Java, no C.
Kernel/User
Platform
Replaceable
Replication, audit/log, expunge, etc.
Rules processing
24
Main components

• Kernel
• Windows fp.sys Location WINDIR\System32\Driver
  s
• (Tipboot safe mode and delete or rename it to
  avoid it loading)
• (net stop/start filepath stop only if agent
  is stopped first and is still risky)
• User
• Windows filepathsrv.exe Location InstallDir
• Starts as a service, but exits right away if in
  non-file-server mode
• Starts as logonso that it runs as the Users
  credentials.
• To launch from a command prompt (or Startup
  Items) filepathsrv d
• Windows fpa.exe Location InstallDir
• Interactive program to set tracing and many other
  commands to the kernel.
• Same code as filepathsrv.exe (historical
  reasons for different name)
• Unix only has fpa.

25
File-based architecture

• No databases used
• File system and files used throughout as
  databases
• Several patents filed on some very unique usages,
  particularly the generations database
• Many filessee later slide/document for a
  complete listing of every file TCDP uses and
  purpose.

26
Use Case 1 Lone User

• Three different sub-use-cases
• Simplest No desire for off-machine protection
  simply enjoys the pool-based versioned
  replication to the local area of his most
  important files. This user installs-and-goes
  (doesnt change any defaults).
• USB/Firewire This user wants to periodically
  have his data automatically exported to a
  removable drive such as a USB or Firewire drive.
  Once in a while, hell hook-up this external
  drive, but typically it is not attached. This
  user will specify D\ (or whatever) as the CDP
  remote target. CDP will patiently wait for the
  existence of the drive and when connected it will
  catch up.
• ISP account (or broadband provider account) This
  user would like to exploit the storage space
  provided by an ISP. Hell configure the CDP
  remote target as \\pokgsa.ibm.com\stakutis (for
  example). This user may be more diligent in their
  configuration of the exclude list (due to
  longer distance network and more limited or
  costly target space).

CDP
27
Use Case 2 Homeowner SOHO
CDP
Broadband
CDP
CDP
CDP
CDP
Kitchen
Kid-1
Kid-2
Home Office
Basement old computer that is good enough as a
file server (add-on 80 200GB disk)

• Three different sub-use-cases
• Simplest Various household computers are
  configured to point to a re-deployed old computer
  as the remote target for CDP (inexpensively
  updated with a USB external hard disk), a cheap
  NAS from Staples/OfficeMax. Local real-time
  protection plus off-computer protection for both
  high-priority files and everything else.
• Vaulting As above, but this user also runs CDP
  on the target computer and activates the
  vaulting feature. He chooses to lock-down his
  home digital pictures directory and maybe any
  MP3s that get replicated to the box.
• ISP As above, plus configures the basement
  computer to also replicate some set of files
  (e.g. .doc) to his Internet Service provider
  which gets his critical files out of the house.

28
Use Case 3 Enterprise

• Three different sub-use-cases
• Simplest Customer deploys large-scale file
  server as the Backup File Server (possibly
  several). Customer is well-skilled with TSM and
  chooses to backup the BFS using TSM
• TSM Centric Users load both CDP and TSM Client.
  CDP does its normal high priority continuous
  protection to the local cache, and CDP is also
  configured for hourly backup into TSM of those
  same files (and patient if the network is not
  available)
• File Server Customer loads CDP on their
  important file servers (as opposed to the
  workstation endpoints). Users go about their
  normal operation. CDP makes real-time local cache
  copies of high-importance changed files and can
  also be configured to replicate to another target
  (remote disk, another file server,
  offsite-server, etc)

1,000s of desktop and laptop clients TSM shop
with TSM skills
CDP
CDP
CDP
CDP
CDP
CDP
CDP
CDP
BFS
CDP
CDP
CDP
CDP
29
Use Case 4 File Server

• Several sub-use-cases
• Simplest Users use their file server naturally
  (no changes). Add CDP to the file server.
  Activate real-time CDP to local LUN for
  high-valued files. Continue to backup the file
  server using existing tools and approaches (the
  main new value being CDP of high-valued files).
• Medium As above, but also have the file server
  replicate to a backup-file-server (BFS), most
  likely on a scheduled (hourly) basis and then
  back up the BFS using traditional tools. Likely
  the BFS will be a target of many file servers in
  the shop.

Any type of shop 10s-1,000 clients
CDP
File Server
BFS
30
Real-Time, All-the-Time Computer Backup, So You
Can Just Work ROY FURCHGOTT Published September
1, 2005 Even people who make regular backups of
their hard drives risk losing work done between
backup sessions. But what if you could save
everything all of the time, automatically? I.B.M.
is releasing a 35 program that keeps a
continuous record of every keystroke. "No one was
backing up at home, and companies thought backups
were taking too long," said Chris Stakutis, chief
technology officer for I.B.M.'s software
protection team, explaining the inspiration.
The new program, I.B.M. Tivoli Continuous Data
Protection for Files, will be available for
download at ibm.com on Sept. 16. The program,
which runs on Windows PC's, does not require you
to stop working while it backs up. It can copy
your latest updates to another computer or
network server over the Internet, or it can keep
track of them on your local hard drive. Changes
you make offline can be automatically copied to
the remote backup when you're back online. Saving
everything would fill a hard drive quickly, but
you can set the program to record only critical
work. About 1 percent of a 100-megabyte drive
should be enough, Mr. Stakutis said, "if you
aren't backing up all of your home movies."
31

• CDP Video

32
For EDUSAT Coordinators

• Please send the session wise attendance sheets
  to
• ibmvtu_at_gmail.com
• Students post your questions _at_
• http//forum.vtu.ac.in

33

• Thank You ?