Archiwizacja danych - przyklady

1
Archiwizacja danych - przyklady

• Hierarchical Storage Management
• Hierarchiczne zarzadzanie zasobami

2
Ograniczenia Backupu

• Backup nie zabezpiecza ciaglosci bezpieczenstwa
  danych
• Najszybciej zuzywajace sie nosniki
• Tasmy nie sa interaktywnym nosnikiem
• Informacja rozlozona jest na wielu nosnikach
  (dlugi dostep)

3
Co to jest HSM?

• Hierarchical Storage ManagementSystem
  przechowywania danych, który automatycznie
  migruje je pomiedzy drozszymi i tanszymi mediami.
  Koszt jednostkowy szybkich urzadzen pamieci
  masowej, jak np. dysków twardych jest duzo
  wiekszy od wolniejszych jak napedy MO lub napedy
  tasmowe. Rozwiazanie azeby wszystkie dane
  znajdowaly sie na szybkich i drogich urzadzeniach
  nie jest do zaakceptowania przez wiele firm. HSM
  przechowuje dane na wolnych urzadzeniach i
  stamtad kopiuje je do szybkiej pamieci dyskowej,
  które pracuje jako cache.
• Webopaedia

4
Konfiguracja bez HSM
5
Konfiguracja z HSM
6
HSM w sieci
IJ
X
Serwer
CD
7
Zasada pracy HSM
Dysk Logiczny (np. X)
SecondaryStorage
PrimaryStorage
Cache
Dane
HDDRAID
Biblioteka- MO- CD- DVD
8
Rozwiazanie HSM
SecondaryStorage
PrimaryStorage
Cache
Biblioteka MO
HDDRAID
Dane
SecondaryStorage
Cache
Biblioteka CD
HDDRAID
9
Dzialanie cache
Serwer
Cache
CD/MO
70 700kB/s (1naped)
Do 20MB/s
SCSI
SCSI
siec
Cache pozwala uzytkownikom szybciej przeslac dane
Aktywne pliki sa dostepne z wydajnoscia RAID
10
Systemy zarzadzania danymi (podsumowanie)

• Tworzenie zapasowych kopii bezpieczenstwa
  (backup)
• strategie
• pelna
• przyrostowa
• róznicowa
• schematy rotacji
• ojciec-syn
• dziadek-ojciec-syn
• wieze Hanoi
• Przyklady systemów
• HP OpenView OmniBack II
• Legato Networker
• CA ARCserveIT
• Veritas Software

• Archiwizacja danych (archive)
• Skladowanie hierarchiczne (HSM)
• dysk - cache dyskowy - biblioteka
• Przyklady systemów
• UniTree Central File Manager
• HP OmniStore
• Legato Networker
• Veritas Software

11
Przykladowe cechy UniTree CFM

• Opracowany w Lawrence Livermore National Lab.
• Modele pracy
• podstawowy jedna instancja zarzadza pamiecia
  cache, robotem i systemem plików, zasoby
  alokowane na jednej/kilku maszynach
• rozszerzony wirtualny system plików - uzytkownik
  nie posiada dostepu do plików innego uzytkownika.
• Rozproszony system zasobów tasmowych - zasoby
  lokalne i zdalne
• Przykladowe zastosowania Centra Informatyczne -
  ACK Cyfronet, TASK, PCSS, WCSS

12
Przyklad zestawu oprogramowania

• VERITAS Software

13
Przeglad zastosowan

• Veritas Solutions
• Backing up Windows 2000
• Backup Management
• Storage Area Networking (SAN)
• Disaster Recovery

14
VERITAS Software

• The industrys largest independent application
  storage management software vendor
• 1 company in 5.5 Billion storage management
  software market
• 48 Billion market capitalization
• Analysts projections for 1999
• Revenue gt600M Growth gt50
• Actual 700M Growth 80
• 2000 projection 1.2B
• Global company with 3,000 employees
• A premier strategic partner to most major vendors
• Customer-focused with enterprise-class, worldwide
  service and support

15
(No Transcript)
16
Wspólpraca
17
Hewlett PackardServing HP Customers Enterprise
Storage Needs

• VERITAS File System included in HP-UX (jfs)
• VERITAS Volume Manager to be embeddedin HP-UX
  11.X
• Available today as standalone product
• HP resells VERITAS NetBackup, Storage Migrator
  for UNIX, and Storage Edition for Oracle
• VERITAS Cluster Server for HP available
• VERITAS Foundation Suite for HP en route

18
MicrosoftProviding Enterprise Storage for
Windows NT Server

• VERITAS selected for four applets expected to
  ship in Windows 2000
• NTBackup (Backup Exec also in NT4, NT 3.5,
  Win98)
• Remote Storage Service HSM applet
• VERITAS Volume Manager (Logical Disk Manager)
• WinInstall LE component
• Disaster Recovery component
• VERITAS enterprise solutions for Windows NT/2000
• Backup Exec, NetBackup, Replication Exec,
  ClusterX
• Backup Exec used internally by Microsoft ITG
  worldwide
• First to deliver Windows 2000 backup
• Only fully certified Windows 2000 solution
• VERITAS VM GUI selected by Microsoft
• Developer of Microsoft Tape Format

19
VERITAS Solutions
20
Data Availability Index (DAI)
21
VERITAS Technologies
Professional Services
Education
ApplicationSolutions
Support
File Volume Mgmt
Clustering Replication
Backup
SAN Management
22
VERITAS Foundation Technologies
Professional Services
Education
ApplicationSolutions
Support
Clustering Replication
Backup
SAN Management
23
VERITAS Volume Manager

• High-performance, high-availability tool for
  maximizing availability of on-line storage.
• On-line data movement and relayout
• RAID configurations 0, 1, 01, 10, 5
• Low-impact Oracle mirror resilvering
• Dynamic multipathing forperformance and
  availability
• Simple Java interface
• Broad device support

24
VERITAS File System

• High-performance, journaling file system for UNIX
  environments
• Fast crash recovery
• Easy manageability
• Online grow, shrink, move, etc.
• Native throughput approximates raw
• Robust error isolation
• No calls to panic

25
VERITAS Backup Technologies
Professional Services
Education
ApplicationSolutions
Support
File Volume Mgmt
Clustering Replication
SAN Management
26
VERITAS NetBackup

• One solution for heterogeneous environments and
  all major databases
• Oracle, Sybase, Exchange, SQL Server, SAP, DB2
• UNIX, Windows NT/9x, NetApp, VMS, NetWare, Linux,
  etc.
• Centralized administration
• Simple, low-cost management for large enterprises
• Record-shattering performance
• Scalable to terabyte-size systems, thousands of
  nodes
• Advanced functionality
• Bundled, resold, or OEMed by Sun, Bull, HP, NEC,
  Red Hat, sgi

Centralized Master Server
Distributed Media Servers
Broad Client Support
27
VERITAS

• Industry-leading solution for Windows 2000,
  Windows NT, and Novell NetWare
• Used by Microsoft ITG and Novell internally
• OEMed or bundled by Microsoft, Dell, Compaq,
  Quantum/ATL, NEC
• Ease of use, great disaster recovery, and
  powerful functionality with a low entry price
• Supports Microsoft MTF, COM, RSM standards
• Agents/clients for Exchange, Databases, Notes,
  UNIX (including Linux), SAP R/3, Windows 9x, Mac
• Integrated open file backup, virus scanning
• ExecView for single-console management

28
Global Data Manager Enterprise Backup Management

• Management of geographically dispersed backup
  master servers

NetBackup for NT
Backup Exec NT
NetBackup for UNIX
Backup Exec NetWare
29
Protection of all Storage DomainsDesktop to Data
Center
30
BLIBBlock Level Incremental Backup

• Breakthrough technology that redefines database
  backup
• FAST - 10-100X faster than File Level backup
• HIGH CAPACITY - Protects the largest
  multi-terabyte databases.
• EFFICIENT - Backs up changed blocks only, min.
  disk I/O server overhead.
• SAFE - Non-disruptive. Tested and endorsed by
  Oracle.
• EASY - Simple as file backup, only faster.
  Integrated with Oracle8 recovery manager.
• INEXPENSIVE - Fraction the cost of alternatives.
  Use any disks and less of them.

New Data
File Level Backup
31
VERITAS Storage Migrator (Hierarchical Storage
Management)

• Automatic migration of stale data to
  secondarymedia, preserving access
• Cuts out-of-space incidents, management effort
• Cuts backup times
• Transparent recall uponreference
• Supports Solaris, NT,HP, SGI, Auspex
• Embedded in EnigmaPARS Migrator
• Integrated withVERITAS backup solutions

Optical disk, tape,or vault
32
VERITAS Clustering and Replication Technologies
Professional Services
Education
ApplicationSolutions
Support
File Volume Mgmt
Backup
SAN Management
33
Clustering for High Availability
CLUSTERING

• Heartbeats and application agents monitor
  services for problems
• Service and network addresses can be restarted in
  place or failed over to alternate server
• Users see little or no disruption to application

34
VERITAS Cluster Server

• Application service oriented failover solution
• Cost-effective scalability up to 32 nodes
• Manage planned and unplanned downtime
• Easy to install, configure and modify
• Simple Java interface
• SAN-ready today
• High availability from the industry leader
• For HP-UX or Solaris

35
VERITAS ClusterX

• Single-point cluster management for Windows NT
  environments
• Versions for
• Microsoft Cluster Service (MSCS)
• Windows Load Balancing Service (WLBS)
• Architected to support other clustering
  technologies
• OEMed by Data General, Dell, Unisys

36
VERITAS Storage Replicators
An automated way to keep the changing data states
of distributed nodes continuously
synchronized (NT UNIX)
SYNCHRONOUS
ASYNCHRONOUS
37
VERITAS SAN Management Technologies
Professional Services
Education
ApplicationSolutions
Support
File Volume Mgmt
Clustering Replication
Backup
38
VERITAS SAN Exploitation

• VERITAS shipping SAN-aware products today
• Further SAN exploitation en route
• VERITAS built and demonstrated the worlds
  largest SAN (NY, September 98)
• Partnerships with key SAN vendors

39
Some Key VERITAS SAN Partnerships
40
VERITAS Application Solutions
Professional Services
Education
Support
File Volume Mgmt
Clustering Replication
Backup
SAN Management
41
VERITAS Editions

• Application-optimized solutions
• For Oracle, Sybase, SAP, and File Servers
• File System and Volume Manager-based performance,
  availability, and simplicity of management
• Special application-oriented performance
  accelerators
• Exploitation of large memory sizes
• Block-Level Incremental Backups (Oracle)
• The ultimate in low-utilization, online backups
• High Availability options

42
The VERITAS Information Availability Index

VAI
99.9999
99.999
99.99
99.9
99.5
99.0
98.0
95.0
Potential Savings ()
43
Backing up Windows NT/2000
44
Windows 2000 Support

• Windows 2000 Support
• Automatic protection of all Windows 2000
  components both local and remote (via Agent
  Accelerator)
• Encrypted Files
• Sparse Files
• Active directory
• Special handling of
• Mount Points and Links
• Remote Storage Services data
• Disk Quota data
• Integration with Removable Storage Manager (RSM)

45
Mount Points and Links

• Backup
• Backup mount points which direct a file or
  directory to a local disk or directory path
• Allows you to backup data that is linked to local
  files, but stored on another device without
  having to actually select the device.
• Remote Storage Services (RSS)
• Automatic protection of RSS database when system
  directory is selected for backup
• Automatically protect RSS reparse points without
  generating a recall of data from secondary media
• Option to back up data migrated from primary
  storage to secondary storage without recalling
  data to primary storage
• Data is backed up directly to backup media
• Provides complete backup of your entire system
  for disaster recovery purposes

46
Disk Quota

• Disk Quota
• Tracks and controls disk usage on a per user, per
  volume basis
• Disk Quota data is automatically backed up when
  the root directory of the volume is selected for
  backup
• Values can be restored to limits that were set
  before backup
• Removable Storage Manager
• Formerly known as NT Media Services (NTMS)
• Being provided by HighGround Systems
• Allows different applications to share devices
  supported by RMS
• Enhanced ADAMM
• User can use ADAMM to manage devices and media
  without having to use RSM
• User can have ADAMM work in conjunction with RSM
  to share devices and media with other
  applications

47
MS Cluster Support

• Active/Passive server support
• Server A is Active, Sever B is Passive
• Server A fails, Server B
• Automatically detects Backup Exec failure from A
• Automatically starts up its own Backup Exec
  Services
• Automatically restarts the job that fails
• Automatically takes over the rest of the jobs
  that A had

A
B
48
MS Cluster Support

• Supported in 3 different configurations
• Backup Device on both servers A and B
• Backup Device shared between servers A and B via
  dedicated SCSI bus
• Backup Device shared between servers A and B via
  the Shared Storage Option

49
Backup Server Management

• Network Storage Executive

50
NSE 8.0 Key Features

• Data-centric orientation (not backup server
  centric)
• Centralized administration of backup, restore and
  device operations
• An intuitive user interface based on the
  Microsoft Management Console (MMC)
• Policy-based management offers a more scalable
  way to manage and automate backup
• Backup server pooling permits load balancing
  across Backup Exec servers with greater fault
  tolerance
• Microsoft SQL Server 7.0 database integration
  provides a central database that stores all job,
  device, media and configuration information
• SSO and GDM integration

51
Policy-based Job Management

• Backup policies define how when data will be
  backed up
• Task list identifies the different types of jobs
  within a policy

• Each task defines the details of a specific job
• Backup method
• Device media
• Schedule
• Advanced settings

52
Backup Server Pooling

• Group backup servers using drive pools

• Target policies and jobs to drive pools

• Load balancing
• Jobs dispatch to first available backup server
• Simplifies capacity planning
• Fault tolerance
• Offline backup server will not prevent job
  execution

53
Centralized History Logging

• Results from all jobs are rolled up to the Master
  database
• Job history details are available for viewing or
  reporting

• Problem files (skipped or in use) are also stored
  in the database

• History pruning controls depth of history and
  size of the database

54
Alerting Notification

• Flexible variety of notification methods
• Email
• Exchange
• Lotus Notes
• SMTP
• SNMP
• Alpha-numeric pager
• Event Log
• Circuit Breaker stops alert overload

55
Consolidated Reporting

• Over 20 new reports using Crystal Reports 7.0
• All data stored in an ODBC-compliant database
• Create custom reports if desired
• Automate operations based on database states

• Spot trends
• Predict growth
• Optimize performance
• Manage media pools
• Track service levels

56
Offsite Media Tracking

• Track media location onsite and offsite

• Reports direct operators to move proper media
• Media discovered online will be moved
  automatically
• Restore jobs will inform the user where the media
  is currently located
• Create multiple offsite vaults to store media

57
Centralized Restore

• Centralized catalog repository enables fast
  restore from any backup server
• Not necessary to restore data using the original
  backup server
• Dedicate restore tasks to a specific machine

• This capability is impossible without a
  centralized catalog repository
• Catalog partializing and pruning control catalog
  growth

58
NSE Benefits

• Reduce Total Cost of Ownership by consolidating
  the administration of backup restore tasks
• Easily scales with you as your network grows
• Pool storage resources for more efficient
  resource utilization greater fault tolerance
• Manage by exception with NSEs rich variety of
  reporting and alerting capabilities
• Integrate with other Windows management tools by
  taking advantage of NSEs Microsoft
  component-based design (MMC SQL)
• Trust the leaders in NT storage management for
  the highest level of NT compatibility

59
Storage area networkingSAN
60
Issues

• Backups impact business productivity
• Backups take too long
• LAN bandwidth insufficient
• Data grows exponentially (50-100 per year)
• Certain applications have little or no backup
  window and require greater availability
• Widely dispersed backup systems are too costly
  for IT to maintain and manage effectively
• Each backup system requires care feeding
• Difficult to automate

61
Traditional Approaches to Backup

• Backup to locally attached tape devices
• High performance
• Does not require network bandwidth
• Backup over the network
• Consolidated device hardware
• Centralized device and media management

62
Backup to Locally Attached Devices
Each Server has its own backup device
LAN
Server
Tape Device
63
Problems with this Approach

• Hardware and media required for every server on
  the network
• Cost of a library on every server is prohibitive
• Considerable manual effort required to maintain
  many individual systems

64
Backup Over Network
Data Flow
Tape Library
65
Problems with this Approach

• Significant network bandwidth is required
• Network performance is seriously degraded
• Network users are severely affected
• Does not meet the needs of 7 x 24 operations
• Modern tape drive performance far exceeds network
  performance
• The network becomes the bottleneck

66
The Ultimate Solution

• Achieve speeds of local backup
• Dont utilize network bandwidth
• Consolidate device hardware
• Centralize device and media management

67
The Storage Area Network (SAN)

• Replaces the conventional SCSI bus
• Connects storage devices to a separate network on
  the back end of the server
• Storage devices may be simultaneously connected
  to many servers at once
• Usually implemented using Fibre Channel
  technology
• Current SAN topologies use a Fibre
  Channel-Arbitrated Loop (FC-AL)

68
Shared Storage Option Architecture
Data Flow
Storage devices
Tape Libraries
69
Backup Exec Shared Storage Option

• Allows multiple Backup Exec servers to share
  backup devices directly attached across a
  storage area network or SAN
• Consolidates storage to fewer centralized tape
  devices with greater capacity backup
• Performance comparable to local speeds
• Backup traffic removed from the main network
• Support for Windows NT and Novell NetWare
• Currently NT NetWare servers must reside on
  separate loops

70
Disaster Recovery
71
What Happens IF .......
72
The Recovery Problem

• Without automated solution one has to
• Collect and install system configuration files
• Manually re-partition and format H/D
• Manually re-install profiles and drivers
• Manually re-install Backup program
• Identify last Backup Tape
• Re-catalog the backup tape
• The above is all prone to Human Error

73
How IDR works ?

• Three Phases of Implementation
• Initial preparation of Disaster Disks
• On Installation of IDR creates four disks
• Three are modified Windows NT setup disks
• The fourth is the IDR disk that contains
• Restore Engine for Backup Exec
• And server specific configuration Information
• Automatic update of IDR information via regular
  backup
• IDR is integrated into Backup Exec all
  information about full / Incremental backup is
  stored and any new server configuration changes
  is recorded.
• This ensures that current information will be
  available on a Recovery
• Recover after a Disaster

74
Recovery phase

• First replace faulty HD or Server
• Boot with IDR disks and NT CD
• Disk partition sizing configuration and profile
  information is all restored with IDR
• All applications and data is restored
  automatically off the most current backup tape
• Can restore remote servers