Transcript Backup & Restore

Backup & Restore Load & Unload
15 June 2011
© 2011 IBM Corporation
Backup & Restore
 What does DBA stands for ?
• Default and Backup for Anything.
 Why a backup ?
• Disaster Recovery.
• High Availability Data Replication (HDR) setup.
• Cloning (Testing, development, QA, etc).
2
© 2011 IBM Corporation
Backup & Restore
 Why a database backup? – Disaster recovery
• When a disaster strikes, you need a way to recover the data. The
recovery process requires procedures and processes in place ahead of
time that allow for data recovery when needed. A backup.
• What might cause you a need to recover?
•
•
•
•
•
•
Unintended deletion of a database object (rows, columns, tables).
Unintended deletion of a server object (databases, chunks, dbspace).
Data corruption or incorrect data created.
Hardware failure (such as when a disk that contains chunk files fails).
Database server failure.
Natural disaster.
• What to backup ?
• Depending on the recovery plans.
3
© 2011 IBM Corporation
Backup & Restore – Disaster Recovery
 Goals and Objectives:
•
•
•
•
•
4
How much data loss, if any, is acceptable?
How long can your business function without the data?
How long can your production system be down during a restore?
How much transaction time can be lost?
How much budget is available for a recovery plan?
© 2011 IBM Corporation
Backup & Restore – Disaster Recovery
 Type –Time
• Type:
• Knowing the type or usage of data can help you determine how quickly it
needs to be recovered.
• Why was this data backed up to begin with?
• Time:
• How much transaction time can be lost?
• Can we re-apply the transactions / load data via other means?
5
© 2011 IBM Corporation
Backup & Restore – Disaster Recovery
 Quantity – Distribution:
• Quantity
• What quantity of data has been lost? (e.g., we lost 25% of our customer profiles).
• Distribution
• What is the extent of the loss? (e.g., we lost the entire Midwest regional sales figures, but
the East/West coast and international figures are intact).
6
© 2011 IBM Corporation
Backup & Restore
 Disaster Recovery – Right, Fast, or Cheap
• There are three ways to implement recovery:
right, fast, or cheap.
• Choose the two highest on your priority list. These choices should
directly relate to your data and time-loss recovery goals.
• Most people will choose the right option as their first priority. The
difficult step is to weigh the advantages and disadvantages of speed
and cost.
7
© 2011 IBM Corporation
Backup & Restore
 Recovery strategies and scheduling backups:
• Recovery strategy requires planning with an understanding of the
business critical data, and comprehension of the backup tools available
and their capabilities.
• The first step in planning your recovery is to outline recovery goals
based on the understanding of the business data.
• Define what is a successful recovery for your system in terms of
tolerance for data loss and acceptable time loss. Use key questions
presented earlier.
• Have a recovery checklist in place.
8
© 2011 IBM Corporation
Backup & Restore
 Recovery planning cycle:
9
© 2011 IBM Corporation
Backup & Restore
 Why a database backup – HDR setting:
How it works?
Database Logging (DDL, DML)
Logical log transmission
The need for a backup / restore to set it up
Physical
Logical
HDR as backup solution in itself
Primary
Secondary
10
© 2011 IBM Corporation
Backup & Restore
 Why a database backup – Cloning:
•
•
•
•
11
Testing
Development
QA
etc
© 2011 IBM Corporation
Backup & Restore
 Informix concepts:
•
•
•
•
•
12
Database Server or Instance
Dbspace
Database
Tablespace
Logical Logs
© 2011 IBM Corporation
Backup & Restore
onconf
ig
Emergency
Boot file
ONINIT
VPs
rootdbs
reserve pages
online
log
sysmaster
sysutils
Physical Log
Logical Logs
/informix/rootchunk1
/informix/rootchunk2
sbspace
dbs1
dbs2
Backup Media
/informix/sbck1 /informix/s1ck1/informix/s2ck1
/informix/sbck2 /informix/s1ck2/informix/s2ck2
13
© 2011 IBM Corporation
Backup & Restore
 Informix Backup
A backup occurs when all or some server database objects are copied to
storage media
Physical
Dbspaces
backup
Or
Logical
Logs
14
Logical
backup
© 2011 IBM Corporation
Backup & Restore
 Physical and Logical Backup:
• Physical
• Focuses on current contents of storage spaces (dbspaces, blobspaces,
smart blobspaces).
• Foundation of all backup and restore operations.
• Works on page level.
• Backup only used pages.
• Logical
• Focuses on transactions stored in logical log files.
• Backup can be continuous (as log files fill) or manual.
• How we can restore as much data as possible, even data updated between
backups.
15
© 2011 IBM Corporation
Backup & Restore
 Physical Backup
• Performed while server is On-Line or Quiescent:
• To reduce impact on production.
• Off-Line backup is possible, but is not supported because data consistency
can’t be guaranteed.
• Incremental Backup
• To save time and space on backup media.
• Backup only the pages that have changed since the last backup.
• Three levels:
• 0 - backup all used pages.
• 1 - backup all changes since the last level 0 backup.
• 2 - backup all changes since the last level 1 backup.
16
© 2011 IBM Corporation
Backup & Restore
 Logical Backup
Log 1
Log 2
……
Log ‘n’
A logical-log backup is the process
of copying the contents of a logical
log file to secondary storage
media. This allows the logical log
to be reused.
In order to successfully complete a logical restore, the logical-log backup
must have been created with the same backup utility used to create the
physical backup.
17
© 2011 IBM Corporation
Backup & Restore
 Logical Backup
• Log file cannot be re-used before it is backed up.
• No log backup:
• Set LTAPEDEV to /dev/null or NUL
• Continuous log backup.
• Manual log backup.
• Log salvage:
• Backup logical log files when server is off-line.
• Can only backup undamaged files still on disk.
18
© 2011 IBM Corporation
Backup & Restore
 Physical & Logical restore
19
© 2011 IBM Corporation
Backup & Restore
 Physical Restore
• Cold:
• Server is off-line.
• Must restore all critical media.
• Salvage logical log files still on disk, physical restore, logical restore.
• Warm:
•
•
•
•
20
Server is On-Line or Quiescent.
Restore of critical media not allowed.
Small impact on production in most cases.
Physical restore of storage spaces, followed by logical log backup, then
logical log restore.
© 2011 IBM Corporation
Backup & Restore
 Physical Restore (cont’d)
• Mixed
• Get back to production as soon as possible.
• First step: Cold restore all critical and application significant spaces to bring
server On-Line.
• Second step: At a later time complete the restore with a warm restore of the
remaining spaces.
• Overall restore takes longer because there are 2 logical restores, one for
each step.
21
© 2011 IBM Corporation
Backup & Restore
 Restore Options
• Point in Time / Point in Log:
• Restore to a specific time or to the end of a specific log.
• All changes to data after the point in time are lost.
• MUST restore all storage spaces to the same time.
• Imported
• Restore on one server a backup taken from another.
• Intended for testing or to start up replication.
• Creates a new server instance:
• You cannot merge data from two servers into one.
• All storage spaces:
• This is not the way to get rid of old data.
• Restartable and Suspended.
• Redirected:
• Allows chunk pathnames to be changed while doing the restore.
22
© 2011 IBM Corporation
Backup & Restore
 Logical Restore
• Logical restore has 2 parts:
• Copy the backed up log file to disk.
• Read the data in the log file and replay the transactions.
• Logical restore is performed only on the spaces just physically restored:
• Can’t do logical restore by itself, must do physical restore first.
23
© 2011 IBM Corporation
Backup & Restore
 Select Tools:
• Now that recovery goals have been set, it is time to identify the IBM Informix
backup and recovery tools. In addition, other useful features and utilities that
provide added data redundancy and high availability will be discussed.
24
© 2011 IBM Corporation
Backup & Restore Utilities - ontape
 Backup and restore utility for Informix that provides:
•
•
•
•
25
Backup at the server level.
Support for incremental backup.
Manual or continuous logical-log backup.
Restore at the system or dbspace level.
© 2011 IBM Corporation
Backup & Restore Utilities - ontape
 Backup & Restore
• Advantages
• Reliability - Data can be restore up to point of failure because of logical-log
backups.
• Ability to restore at the system or dbspace level .
• (restore not backup at dbspace level.
• Support for remote tape devices.
• Support for external backup and restore.
• Simple configuration.
• Easy to use command syntax.
• Generally used for small database servers.
26
© 2011 IBM Corporation
Backup & Restore Utilities - ontape
 Backup & Restore
• Disadvantages:
• Backup can only be done at the system level.
• Each dbspace is backed up serially to a single device. For large systems,
this can mean very slow backups.
• Requires user input when device media must be changed.
• No media management.
27
© 2011 IBM Corporation
Backup & Restore Utilities - onbar
 Backup & Restore Utility for Informix that provides:
•
•
•
•
•
•
•
28
Parallel backup and restore.
System level and dbspace level backup and restore.
Support for incremental backups.
Manual or automatic backup of logical logs.
Point-in-time recovery for an instance.
An open interface for communication with storage manager (XBSA).
Third-party storage managers may provide support for sophisticated
tape stacker and jukebox devices.
© 2011 IBM Corporation
Backup & Restore Utilities - onbar
 Backup & Restore Utility for Informix that provides:
• Advantages
• Parallel backup and restore provides speed.
• Reliability-Data can be restore up to point of failure because of logical-log
backups.
• Support for multiple devices.
• Ability to restore at the server or dbspace level. This means the server can
be made available to users once the critical dbspaces, i.e. root, logs, etc.,
have been restored:
• Non-critical dbspaces can then be restored in order of importance.
29
© 2011 IBM Corporation
Backup & Restore Utilities - onbar
 onbar
• Disadvantages:
• Requires configuration and use of the IBM Informix Storage Manager or an
additional cost third-party storage manager.
30
© 2011 IBM Corporation
Backup & Restore - onbar
 onbar enhancements over ontape:
•
•
•
•
High-speed parallel backup and restore.
Unattended operations.
Automatic backup of logical logs via the ALARMPROGRAM parameter.
Support for sophisticated backup storage devices through a storage
manager.
• Backup of selected dbspaces.
• Restartable restore.
31
© 2011 IBM Corporation
External Backup & Restore
 Methods of backing up data at the O/S level
– UNIX commands such as:
• cpio
• tar
• od
–
–
–
–
–
32
File system snapshots.
3rd mirror break off.
Storage manager “raw” backup.
External backup/restore feature of ON-Bar allows logical recovery.
Can be taken from a MACH-11 RSS of logged objects.
© 2011 IBM Corporation
External Backup & Restore
 External Backup & Restore
• Advantages:
• Availability, speed-Data can be backed up and restored more quickly,
making the data available sooner.
• Multiple copies of the same data can be kept.
• Ability to copy raw physical devices.
• Logical restore available if combined with ON-Bar External Restore.
33
© 2011 IBM Corporation
External Backup & Restore
 External Backup & Restore
• Disadvantages:
• No logical recovery available when using external backup utilities unless you
use the ON-Bar External Backup/Restore feature.
• Greater costs associated with third mirror break off solutions.
• No media tracking for external backups is done by ON-Bar.
34
© 2011 IBM Corporation
Backup & Restore – High Performance Loader
 Load / Unload utilities – High Performance Loader (HPL):
•
•
•
•
35
Enables rapid loading/unloading of large amounts of data in parallel.
Performs data conversion.
Supports a variety of devices.
Stores load/unload operations as reproducible jobs.
© 2011 IBM Corporation
Backup & Restore – High Performance Loader
 Load / Unload utilities – High Performance Loader (HPL)
• Advantages:
• Enables rapid unloading/loading of large quantities of data. I/O to the
devices and the server is performed in parallel.
• Supports a variety of devices.
• Supports multiple formats for load/unload files.
• Data can be loaded from multiple sources.
• Recovery at the table and row level.
• Reusability of load/unload definition. Load/unload jobs are defined and
stored in a database for reuse.
• Could be used as a backup utility for a database with a few static tables.
36
© 2011 IBM Corporation
Backup & Restore – High Performance Loader
 Load / Unload utilities – High Performance Loader (HPL)
• Disadvantages:
• No transaction recovery available. Recovery only occurs up to the time data
was last unloaded.
• DBA must assure relational integrity of the database during the load and
unload process.
• Tool is more complex than other load/unload utilities.
• Does not append to output devices.
37
© 2011 IBM Corporation
Backup & Restore – dbimport/dbexport
 Load / Unload utilities – dbimport / dbexport
 dbimport:
• Creates the database.
• Imports ASCII data into the specified database.
• Loads from disk or tape.
 dbexport:
• Unloads data and schema for an entire database.
• Data stored in ASCII files.
• Unloads to disk or tape.
38
© 2011 IBM Corporation
Backup & Restore – dbimport/dbexport
 Load / Unload utilities – dbimport / dbexport
• Advantages:
•
•
•
•
39
Ability to capture data and schema.
ASCII files provide portability.
Recovery at the database level.
Inexpensive and easy way to backup small databases.
© 2011 IBM Corporation
Backup & Restore – dbimport/dbexport
 Load / Unload utilities – dbimport / dbexport
• Disadvantages:
• No transaction recovery available. Recovery only occurs up to the time the
last export was taken.
• Large databases will consume tape/disk resources; processing will be slow.
• Requires an exclusive lock on the database when processing.
40
© 2011 IBM Corporation
Backup & Restore – SQL load/unload
 Load / Unload utilities – SQL load / unload:
• SQL statements for loading and unloading data from ASCII files.
• Adds rows to existing tables.
• Must be run through DB-Access or ISQL.
41
© 2011 IBM Corporation
Backup & Restore – SQL load/unload
 Load / Unload utilities – SQL load / unload
• Advantages
• Data is added to existing tables:
• dbimport and onload create new tables.
•
•
•
•
42
Specific rows can be unloaded.
Data unloaded is based upon an SQL SELECT statement.
ASCII files provide portability.
Recovery at the table and row level.
© 2011 IBM Corporation
Backup & Restore – SQL load/unload
 Load / Unload utilities – SQL load / unload
• Disadvantages:
•
•
•
•
•
43
Dependent upon DB-Access or ISQL.
Does not load from tape.
No transaction recovery available.
Recovery only occurs up to the time the last export was taken.
Large amounts of data will be loaded and unloaded slowly.
© 2011 IBM Corporation
Backup & Restore – onload/onunload
 Load / Unload utilities – onload / onunload
 onload
• Creates database or tables in specified dbspaces.
• Loads data created by onunload only.
• Page size must be compatible.
 onunload
• Transfers data in binary pages.
• Default write to tape unless –t option used to an existing file.
44
© 2011 IBM Corporation
Backup & Restore – onload/onunload
 Load / Unload utilities – onload / onunload
• Advantages
• Efficient data transfer. Data is read and written in binary pages, providing
speed.
• Recovery at the database and table level.
45
© 2011 IBM Corporation
Backup & Restore – onload/onunload
 Load / Unload utilities – onload / onunload
• Disadvantages
•
•
•
•
Cannot be run for databases across a network.
Cannot be used for file I/O.
Loads data created by onunload only.
Binary data and page size limit portability. Page size varies from machine to
machine (check your configuration file, onconfig, for the page size for your
server system). You cannot use the utilities to transfer data from one server
system to another with a different page size. Most server systems have
either 4K or 2K pages.
• No transaction recovery available. Recovery only occurs up to the time data
was last unloaded.
• You must read the file that onunload creates with the onload utility of the
same version of your database server. You cannot use onunload and onload
to move data from one version to another.
• onunload does not preserve access privileges, synonyms, views,
constraints, triggers, or default values that were associated with the original
tables.
46
© 2011 IBM Corporation
Backup & Restore - dbload
 Load / Unload utilities – dbload
• Loads data from multiple ASCII files on disk into one or more existing
tables.
• ASCII files can be created by the UNLOAD utility or by another means
(c program, text editor or any other unload utility file that unloads data
to ASCII).
• Dbload offers more flexibility than LOAD.
47
© 2011 IBM Corporation
Backup & Restore - dbload
 Load / Unload utilities – dbload
• Advantages
• Data can be loaded from multiple sources.
• Data is added to existing tables:
• dbimport and onload create new tables.
•
•
•
•
•
•
48
ASCII files provide portability.
Recovery at the table and row level.
Provides more flexibility than LOAD.
Specify a starting point in the load file.
Add transaction logic (commit after every x rows).
Limit the number of bad rows read, at which time dbload terminates.
© 2011 IBM Corporation
Backup & Restore - dbload
 Load / Unload utilities – dbload
• Disadvantages:
•
•
•
•
•
49
Doesn’t load from tape.
No transaction recovery available.
Recovery only occurs up to the time the last export was taken.
Large amounts of data will be loaded slowly.
DBA must assure relational integrity of the database during the load process.
© 2011 IBM Corporation
Backup & Restore - Summary
 Summary
– Informix provides different Backup & Restore options
•
•
•
•
ontape
onbar
External Backup & Restore
Load & Unload utilities:
–
–
–
–
–
–
50
Dbexport / dbimport
Dbschema
HPL
SQL load / unload
Onload / onunload
Dbload
© 2011 IBM Corporation
Backup & Restore - Resources
 Resources
– The Online Informix Information Center:
http://publib.boulder.ibm.com/infocenter/idshelp/v117/index.jsp
• One-stop shop for Informix product documentation.
• Supports book marking favorite topics, narrowing the scope to refine
searches, printing subsets of topics.
– IBM Informix DeveloperWorks Technical Articles:
• http://www.ibm.com/developerworks/db2/products/informix/index.html
• Premium technical resource site for DBAs and developers.
• Features explained with examples/sample code.
• Contributions from IBM experts as well as customers.
51
© 2011 IBM Corporation