![[Altibase] 13 backup and recovery](https://image.slidesharecdn.com/altibase13backupandrecovery-160126072259/85/Altibase-13-backup-and-recovery-1-320.jpg)
![Backup Considerations
When performing offline backup, there are more things to check other than
the default directory in the properties file
Data files in other directories might be missed
Check data file-related performance views for data files in other directories.
Online backup and checkpointing cannot be performed simultaneously
If checkpointing is underway when the online backup procedure starts, the
backup procedure waits for checkpointing to finish
If online backup is underway when checkpointing is attempted, an error
occurs
[ERR-110A4 : Backup is in progress. Please wait until the current backup process
completes.]
Do one of the followings to prepare for recovery when modifying a
tablespace:
Back up the tablespace, the dictionary tablespace, and log anchor files
Perform online or offline database backup
If any replication objects exist, replication information is also backed up.
If the database is restored on a system with a different IP address, the
replication error will occur.](https://image.slidesharecdn.com/altibase13backupandrecovery-160126072259/85/Altibase-13-backup-and-recovery-13-320.jpg)
![Restart Recovery
Automatic recovery in the event of an abnormal shutdown, such as a
power outage or a system crash.
………
TRANSITION TO PHASE : META
[SM] Recovery Phase - 1 : Preparing Database
: Dynamic Memory Version => Parallel Loading
[SM] Recovery Phase - 2 : Loading Database
[SM] Recovery Phase - 3 : Starting Recovery
Initializing Active Transaction List
Redo
Refine Disk Table..
Undo
[SM] Refine Memory Table :
........................................................................................................................................ [SUCCESS]](https://image.slidesharecdn.com/altibase13backupandrecovery-160126072259/85/Altibase-13-backup-and-recovery-14-320.jpg)
[Altibase] 13 backup and recovery
- 2. Checkpointing Saves information about the current status of, and changes to, the database to data files Management tasks after checkpointing has completed Change the current recovery start point to the end point of checkpointing Delete unneeded redo log files that have been applied to data files Regular checkpointing reduces the amount of time required to recover the database in the event of a fault. ALTIBASE HDB Checkpointing features Fuzzy & Ping-pong checkpointing Transaction performance is not reduced during checkpointing Two sets of data files are maintained Only applies to memory tablespaces
- 3. Performing Checkpointing Automatic execution (Default) Set auto-execution using the CHECKPOINT_ENABLED property Checkpointing interval: - CHECKPOINT_INTERVAL_IN_SEC Number of redo log files- CHECKPOINT_INTERVAL_IN_LOG Manual execution User command - Only the SYS user can perform checkpointing manually. Reasons why redo log files might not be deleted after checkpointing is completed: They might contain information about transactions that are still underway In Archivelog mode, they have not yet been archived If using replication, they have not yet been converted to XLogs and sent for replication iSQL> ALTER SYSTEM CHECKPOINT;
- 4. Database mode Redo log files are managed in one of two different ways after checkpointing: Archivelog mode – Archived to a specified path and then deleted Noarchivelog mode – Deleted without being archived Set when the CREATE DATABASE statement is executed Can be subsequently changed in the CONTROL startup phase (restart required) Comparison Database Mode Pros Cons Archivelog mode Media recovery possible If archive log files are not lost, recovery up to the present time is possible. Disk space for saving archive log files must be provided. DBA is burdened with additional archive log file management tasks Noarchivelog mode No disk space for saving archive log files is needed. DBA need not perform archive log file management tasks Media recovery is impossible. Recovery is only possible up to the point at which the most recent offline backup was conducted.
- 5. Checking the current database mode More detailed information can be checked by querying the v$archive performance view Changing the database mode after the database is created Start up ALTIBASE HDB in the CONTROL phase Change the database mode Can only be changed by the SYS user in SYSDBA mode iSQL> connect sys/manager as sysdba iSQL(sysdba)> shutdown immediate; iSQL(sysdba)> startup control; iSQL(sysdba)> ALTER DATABASE ARCHIVELOG; iSQL(sysdba)> startup; iSQL(sysdba)> SELECT archive_mode, archive_dest FROM v$archive; ARCHIVE_MODE ARCHIVE_DEST -------------------------------------------------------------------- 0 /edu/alti1/altibase_home/arch_logs
- 6. Types of database backup and objects that are backed up Online backup-related commands can only be executed by the “SYS” user, in SYSDBA mode Cannot be executed from multiple sessions, as there can be only one SYSDBA mode session at any time. Backup Type Target Object Recovery Point DB Mode Online Database Backup The entire database (data files, log anchor files) except redo log files Present time point Must be Archivelog mode Online Tablespace Backup The data files for a particular tablespaceOnline Tablespace backup by “file system copy” Online log anchor backup log anchor files Offline backup The entire database (data files, log anchor files, redo log files) Time of last backup Noarchivelog or Archivelog mode Backup using aexport and iloader Schema creation script, data in a particular table
- 7. Online Database Backup Back up the entire database at particular time points while ALTIBASE HDB is running When the command is executed, all data files and redo loganchor files are automatically backed up to the specified directory Only one of two (#0 and #1) checkpoint image files is backed up Temporary tablespace data files are not backed up, as they are only used for processing queries iSQL(sysdba)> ALTER DATABASE BACKUP DATABASE TO 'backup_dir ';
- 8. Online Tablespace Backup Back up a particular tablespace while ALTIBASE HDB is running The data files related to the tablespace are automatically backed up to the specified path when the command is executed. Only one of two (#0 and #1) memory tablespace-related data files is backed up iSQL(sysdba)> ALTER DATABASE BACKUP TABLESPACE tablespace_name TO 'backup_dir ';
- 9. Online Tablespace Backup by DBA Backup particular tablespace while DB is operating Copy tablespace datafile using OS copy command after the backup using backup command Log switching is essential to forcibly archive the related redo log file Interoperable with 3rd party backup solution as the multiple tablespaces can be backed up simultaneously 1. iSQL(sysdba)> ALTER TABLESPACE tablespace_name BEGIN BACKUP; 2. Copy tablespace datafile using OS copy command 3. iSQL(sysdba)> ALTER TABLESPACE tablespace_name END BACKUP; 4. iSQL(sysdba)> ALTER SYSTEM SWITCH LOGFILE;
- 10. Online log anchor file backup Back up log anchor files while ALTIBASE HDB is running All three log anchor files are automatically backed up to the specified path when the query is executed When altering the tablespace (ALTER DATABASE), the modified tablespace and dictionary tablespace are also backed up, in case they are needed for recovery iSQL(sysdba)> ALTER DATABASE BACKUP LOGANCHOR TO 'backup_dir '; iSQL(sysdba)> ALTER DATABASE BACKUP 2 TABLESPACE tablespace_name TO 'backup_dir '; iSQL(sysdba)> ALTER DATABASE BACKUP 2 TABLESPACE SYS_TBS_MEM_DIC TO 'backup_dir ';
- 11. Offline database backup Back up the entire database after ALTIBASE HDB is shut down. Back up data files, redo log files and log anchor files using an OS copy command. 1. iSQL(sysdba)> shutdown immediate; 2. Backup data files, redo log files and loganchor files by using OS copy command. 3. iSQL(sysdba)> startup;
- 12. Example of offline database backup altibase.properties settings Backup procedure This example assumes that the tablespace data files are located in the default directory MEM_DB_DIR = ?/mem_dbs DEFAULT_DISK_DB_DIR = ?/dbs LOG_DIR = ?/logs LOGANCHOR_DIR = ?/logs iSQL(sysdba)> shutdown immediate; shell> cp –r $ALTIBASE_HOME/mem_dbs/* /backup shell> cp –r $ALTIBASE_HOME/dbs/* /backup shell> cp –r $AL TIBASE_HOME/logs /backup iSQL(sysdba)> startup;
- 13. Backup Considerations When performing offline backup, there are more things to check other than the default directory in the properties file Data files in other directories might be missed Check data file-related performance views for data files in other directories. Online backup and checkpointing cannot be performed simultaneously If checkpointing is underway when the online backup procedure starts, the backup procedure waits for checkpointing to finish If online backup is underway when checkpointing is attempted, an error occurs [ERR-110A4 : Backup is in progress. Please wait until the current backup process completes.] Do one of the followings to prepare for recovery when modifying a tablespace: Back up the tablespace, the dictionary tablespace, and log anchor files Perform online or offline database backup If any replication objects exist, replication information is also backed up. If the database is restored on a system with a different IP address, the replication error will occur.
- 14. Restart Recovery Automatic recovery in the event of an abnormal shutdown, such as a power outage or a system crash. ……… TRANSITION TO PHASE : META [SM] Recovery Phase - 1 : Preparing Database : Dynamic Memory Version => Parallel Loading [SM] Recovery Phase - 2 : Loading Database [SM] Recovery Phase - 3 : Starting Recovery Initializing Active Transaction List Redo Refine Disk Table.. Undo [SM] Refine Memory Table : ........................................................................................................................................ [SUCCESS]
- 15. Media recovery Performed when data files are missing or corrupted. The database should be manually restored using: archive log files backup data files log anchor files Recovery commands must be executed in the CONTROL phase. (Online recovery is not possible.) Is possible only if ALTIBASE HDB is operating in Archivelog mode. Both complete and partial recovery are possible
- 16. Complete recovery Execute the command shown below to restore the database up to the present: Considerations The archive log files AND the redo log files generated since the most recent complete database backup must not be corrupted. The most recently updated log anchor files are needed. iSQL(sysdba)> ALTER DATABASE RECOVER DATABASE;
- 17. Partial recovery Restore the database to a specific point in time by executing the command shown below: If redo log files or archive log files are missing or corrupted: To restore the database to a specified point in time: Redo log files must be reset after performing partial recovery. This step is mandatory in order to avoid restart recovery when the system is restarted. Consideration In order to restore the database to a specific point in time, the log anchor files up to that time are needed. iSQL(sysdba)> ALTER DATABASE db_name META RESETLOGS; iSQL(sysdba)> ALTER DATABASE RECOVER DATABASE UNTIL CANCEL; iSQL(sysdba)> ALTER DATABASE RECOVER DATABASE UNTIL TIME ‘2010-09-10:17:55:00’;
- 18. Recovery example in case of missing data file (1) – recovery using backup data files Use the redo log files whose LSN (Log Sequence Number) is between the LSN of the backup data files and the current LSN of the database. Log Anchor User1.dbf information Creation LSN(20:012284) Checkpoint LSN(102:172168) Archive Logs Directory Log File 32 Log File 100… Log File 101 Log File 102 Online Redo Logs directory User1.dbf (backup file) Recovery command executed User1.dbf LSN 102:172168 Replay Redo logs LSN 32:345698User1.dbf Media Recovery Process Log File 0 …
- 19. Recovery example in case of missing data file (2) – recovery by creating empty data files Perform recovery using archive log files and active redo log files after creating data files. Use the redo log files whose LSN (Log Sequence Number) is between the LSN of the backup data files and the current LSN of the database. Log Anchor User1.dbf information Creation LSN(20:012284) Checkpoint LSN(102:172168) Archive Logs Directory Log File 20 Log File 100… Log File 101 Log File 102 Online Logs Directory Create missing User1.dbf Recovery command executed User1.dbf LSN 102:172168 Replay Redo logs User1.dbf Media Recovery Process Log File 0 …
- 20. Create an empty data file for media recovery Create one in the CONTROL phase if no backup data files exist. Create an empty memory tablespace data file: Create an empty disk tablespace data file: Consideration The location of the data file must be same as the directory recorded in the log anchor file. iSQL(sysdba)> ALTER DATABASE CREATE DATAFILE 'user_disk_data001.dbf'; iSQL(sysdba)> ALTER DATABASE CREATE CHECKPOINT IMAGE 'USER_MEM_DATA-1-0';
- 21. Considerations during media recovery Some data files must be created by the user. Memory tablespace data files must be copied to form a pair. Empty temporary tablespace data files must be created in the CONTROL phase. Considerations after media recovery Back up the database after performing partial recovery. If the database is not backed up at this time, and recovery is again subsequently performed, it will be impossible to recover any changes made to the database after the redo logs were initialized. shell> cp USER_MEM_DATA-1-0 USER_MEM_DATA-0-0; iSQL(sysdba)> ALTER DATABASE CREATE DATAFILE 'temp001.dbf';