基于lt_probackup完成对lightdb数据库进行恢复

liangdong30629

已于 2022-07-13 20:04:31 修改

阅读量371

点赞数

文章标签：数据库 bash 开发语言

于 2022-07-13 02:45:00 首次发布

本文链接：https://blog.csdn.net/liangd407/article/details/125712391

版权

提示：文章写完后，目录可以自动生成，如何生成可参考右边的帮助文档

文章目录

前提
- 参考lt_probackup备份操作的链接
- 本地虚拟机环境准备
lt_probackup执行删除备份操作
lt_probackup执行恢复命令
执行时间点 (PITR) 恢复
在远程模式下使用lt_probackup执行备份(backup)与恢复(recovery)
附录
- 参考链接

前提

参考lt_probackup备份操作的链接

参考lt_probackup备份操作

本地虚拟机环境准备

操作系统：centos 7
平台：x86_64
数据库：已安装lightdb数据库
准备：对数据库做了全量备份(FULL)与增量备份(DELTA | PAGE)

lt_probackup执行删除备份操作

基本语法:

[lightdb@lightdb ~]$ lt_probackup delete --help

lt_probackup delete -B backup-path --instance=instance_name
                 [-i backup-id | --delete-expired | --merge-expired] [--delete-wal]
                 [-j num-threads] [--progress]
                 [--retention-redundancy=retention-redundancy]
                 [--retention-window=retention-window]
                 [--wal-depth=wal-depth]
                 [--no-validate] [--no-sync]

  -B, --backup-path=backup-path    location of the backup storage area # 备份目录
      --instance=instance_name     name of the instance # 备份实例名称
  -i, --backup-id=backup-id        backup to delete # 已经备份的标识
  -j, --threads=NUM                number of parallel threads
      --progress                   show progress
      --no-validate                disable validation during retention merge
      --no-sync                    do not sync merged files to disk

  Retention options:
      --delete-expired             delete backups expired according to current
                                   retention policy # 根据当前保留策略删除过期的备份
      --merge-expired              merge backups expired according to current
                                   retention policy # 根据当前保留策略删除过期备份
      --delete-wal                 remove redundant files in WAL archive # 基于WAL归档删除
      --retention-redundancy=retention-redundancy
                                   number of full backups to keep; 0 disables; (default: 0) # 要保留的全量备份数量
      --retention-window=retention-window
                                   number of days of recoverability; 0 disables; (default: 0) # 可以恢复的时间窗口
      --wal-depth=wal-depth        number of latest valid backups per timeline that must
                                   retain the ability to perform PITR; 0 disables; (default: 0)
      --dry-run                    perform a trial run without any changes # 只是做一个尝试不做出任何不可逆的行为
      --status=backup_status       delete all backups with specified status

  Logging options:
      --log-level-console=log-level-console
                                   level for console logging (default: info)
                                   available options: 'off', 'error', 'warning', 'info', 'log', 'verbose'
      --log-level-file=log-level-file
                                   level for file logging (default: off)
                                   available options: 'off', 'error', 'warning', 'info', 'log', 'verbose'
      --log-filename=log-filename
                                   filename for file logging (default: 'pg_probackup.log')
                                   support strftime format (example: pg_probackup-%Y-%m-%d_%H%M%S.log)
      --error-log-filename=error-log-filename
                                   filename for error logging (default: none)
      --log-directory=log-directory
                                   directory for file logging (default: BACKUP_PATH/log)
      --log-rotation-size=log-rotation-size
                                   rotate logfile if its size exceeds this value; 0 disables; (default: 0)
                                   available units: 'kB', 'MB', 'GB', 'TB' (default: kB)
      --log-rotation-age=log-rotation-age
                                   rotate logfile if its age exceeds this value; 0 disables; (default: 0)
                                   available units: 'ms', 's', 'min', 'h', 'd' (default: min)

删除不再需要的备份

# 查看备份信息
[lightdb@lightdb ~]$ lt_probackup show -B /home/lightdb/backup --instance instance_192.168.76.128_5432
========================================================================================================================================================
 Instance                      Version  ID      Recovery Time           Mode  WAL Mode  TLI  Time   Data    WAL  Zratio  Start LSN   Stop LSN    Status
========================================================================================================================================================
 instance_192.168.76.128_5432  13       RET37Y  2022-07-10 21:09:52+08  FULL  ARCHIVE   1/0   28s  361MB  512MB    1.00  5/120       5/2000A898  OK
 instance_192.168.76.128_5432  13       RET2UT  2022-07-10 21:01:59+08  FULL  ARCHIVE   1/0   27s  361MB  512MB    1.00  4/A0001690  4/C000A0C8  OK
 instance_192.168.76.128_5432  13       RET2R2  2022-07-10 20:59:43+08  FULL  ARCHIVE   1/0   26s  360MB  512MB    1.00  4/40000120  4/6000EE18  OK
 instance_192.168.76.128_5432  13       RET2MZ  2022-07-10 20:57:14+08  FULL  ARCHIVE   1/0   25s  360MB  512MB    1.00  3/E0000120  4/260C8     OK
 instance_192.168.76.128_5432  13       RET2M1  2022-07-10 20:56:42+08  FULL  ARCHIVE   1/0   25s  360MB  512MB    1.00  3/80000120  3/A000EEA8  OK
 instance_192.168.76.128_5432  13       RET2J1  2022-07-10 20:55:01+08  FULL  ARCHIVE   1/0   33s  360MB  512MB    1.00  3/20000068  3/40026F58  OK

# 做一次尝试看看是否可以删除某个backup id，这里以backup id RET37Y为例：
[lightdb@lightdb ~]$ lt_probackup delete -B /home/lightdb/backup --instance instance_192.168.76.128_5432 --dry-run -i RET37Y
INFO: Resident data size to free by delete of backup RET37Y : 361MB
# 可以删除backup id RET37Y，释放361MB磁盘空间
# 执行删除操作
[lightdb@lightdb ~]$ lt_probackup delete -B /home/lightdb/backup --instance instance_192.168.76.128_5432  -i RET37Y
INFO: Resident data size to free by delete of backup RET37Y : 361MB
INFO: Delete: RET37Y 2022-07-10 21:09:52+08

# 再次查看备份信息
[lightdb@lightdb ~]$ lt_probackup show -B /home/lightdb/backup --instance instance_192.168.76.128_5432
========================================================================================================================================================
 Instance                      Version  ID      Recovery Time           Mode  WAL Mode  TLI  Time   Data    WAL  Zratio  Start LSN   Stop LSN    Status
========================================================================================================================================================
 instance_192.168.76.128_5432  13       RET2UT  2022-07-10 21:01:59+08  FULL  ARCHIVE   1/0   27s  361MB  512MB    1.00  4/A0001690  4/C000A0C8  OK
 instance_192.168.76.128_5432  13       RET2R2  2022-07-10 20:59:43+08  FULL  ARCHIVE   1/0   26s  360MB  512MB    1.00  4/40000120  4/6000EE18  OK
 instance_192.168.76.128_5432  13       RET2MZ  2022-07-10 20:57:14+08  FULL  ARCHIVE   1/0   25s  360MB  512MB    1.00  3/E0000120  4/260C8     OK
 instance_192.168.76.128_5432  13       RET2M1  2022-07-10 20:56:42+08  FULL  ARCHIVE   1/0   25s  360MB  512MB    1.00  3/80000120  3/A000EEA8  OK
 instance_192.168.76.128_5432  13       RET2J1  2022-07-10 20:55:01+08  FULL  ARCHIVE   1/0   33s  360MB  512MB    1.00  3/20000068  3/40026F58  OK

# 在备份目录下分别查看backup id RET37Y对应的文件夹是否存在 
[lightdb@lightdb ~]$ ls /home/lightdb/backup/backups/instance_192.168.76.128_5432/RET37Y
ls: 无法访问/home/lightdb/backup/backups/instance_192.168.76.128_5432/RET37Y: 没有那个文件或目录
[lightdb@lightdb ~]$ ls /home/lightdb/backup/wal/instance_192.168.76.128_5432/RET37Y
ls: 无法访问/home/lightdb/backup/wal/instance_192.168.76.128_5432/RET37Y: 没有那个文件或目录
# RET37Y在备份数据目录和wal文件目录均已被删除

删除过时 WAL 文件，这些WAL文件不需要恢复任何剩余备份

# 查看备份目录下wal目录下实例的WAL文件(WAL文件所在目录为backup_dir/wal/instanceName)。 本地虚拟机环境，备份目录/home/lightdb/backup 备份实例为instance_192.168.76.128_5432
[lightdb@lightdb ~]$ ls /home/lightdb/backup/wal/instance_192.168.76.128_5432/ -rltha
总用量 15G
drwx------ 3 lightdb lightdb   42 7月  10 12:23 ..
-rw------- 1 lightdb lightdb 512M 7月  10 20:51 000000010000000200000007
-rw------- 1 lightdb lightdb 512M 7月  10 20:54 000000010000000300000000
-rw------- 1 lightdb lightdb 512M 7月  10 20:54 000000010000000300000001
-rw------- 1 lightdb lightdb 512M 7月  10 20:55 000000010000000300000002
-rw------- 1 lightdb lightdb  355 7月  10 20:55 000000010000000300000001.00000068.backup
-rw------- 1 lightdb lightdb 512M 7月  10 20:56 000000010000000300000003
-rw------- 1 lightdb lightdb 512M 7月  10 20:56 000000010000000300000004
-rw------- 1 lightdb lightdb 512M 7月  10 20:56 000000010000000300000005
-rw------- 1 lightdb lightdb  355 7月  10 20:56 000000010000000300000004.00000120.backup
-rw------- 1 lightdb lightdb 512M 7月  10 20:57 000000010000000300000006
-rw------- 1 lightdb lightdb 512M 7月  10 20:57 000000010000000300000007
-rw------- 1 lightdb lightdb  352 7月  10 20:57 000000010000000300000007.00000120.backup
-rw------- 1 lightdb lightdb 512M 7月  10 20:57 000000010000000400000000
-rw------- 1 lightdb lightdb 512M 7月  10 20:59 000000010000000400000001
-rw------- 1 lightdb lightdb 512M 7月  10 20:59 000000010000000400000002
-rw------- 1 lightdb lightdb 512M 7月  10 20:59 000000010000000400000003
-rw------- 1 lightdb lightdb  355 7月  10 20:59 000000010000000400000002.00000120.backup
-rw------- 1 lightdb lightdb 512M 7月  10 21:01 000000010000000400000004
-rw------- 1 lightdb lightdb 512M 7月  10 21:01 000000010000000400000005
-rw------- 1 lightdb lightdb  355 7月  10 21:02 000000010000000400000005.00001690.backup
-rw------- 1 lightdb lightdb 512M 7月  10 21:02 000000010000000400000006
-rw------- 1 lightdb lightdb 512M 7月  10 21:09 000000010000000400000007
-rw------- 1 lightdb lightdb 512M 7月  10 21:09 000000010000000500000000
-rw------- 1 lightdb lightdb 512M 7月  10 21:09 000000010000000500000001
-rw------- 1 lightdb lightdb  346 7月  10 21:10 000000010000000500000000.00000120.backup
-rw------- 1 lightdb lightdb 512M 7月  11 12:56 000000010000000500000002
-rw------- 1 lightdb lightdb 512M 7月  11 12:56 000000010000000500000003
-rw------- 1 lightdb lightdb  355 7月  11 12:56 000000010000000500000003.00000120.backup
-rw------- 1 lightdb lightdb 512M 7月  11 12:56 000000010000000500000004
-rw------- 1 lightdb lightdb 512M 7月  11 12:59 000000010000000500000005
-rw------- 1 lightdb lightdb 512M 7月  11 12:59 000000010000000500000006
-rw------- 1 lightdb lightdb 512M 7月  11 12:59 000000010000000500000007
-rw------- 1 lightdb lightdb  355 7月  11 12:59 000000010000000500000006.00000120.backup
-rw------- 1 lightdb lightdb 512M 7月  11 13:07 000000010000000600000000
-rw------- 1 lightdb lightdb 512M 7月  11 13:07 000000010000000600000001
-rw------- 1 lightdb lightdb 512M 7月  11 13:07 000000010000000600000002
-rw------- 1 lightdb lightdb  355 7月  11 13:07 000000010000000600000001.000000F8.backup
-rw------- 1 lightdb lightdb 512M 7月  11 13:10 000000010000000600000003
drwx------ 2 lightdb lightdb 4.0K 7月  11 13:11 .

# delete命令增加选项--dry-run进行尝试不做出真实删除行为
[lightdb@lightdb ~]$ lt_probackup delete -B /home/lightdb/backup --instance instance_192.168.76.128_5432 --delete-wal --dry-run
INFO: On timeline 1 WAL segments between 000000010000000200000007 and 000000010000000400000005 can be removed # 这里是时间线1上的WAL 段可以被删除(当前没有真正执行删除)
INFO: Logical WAL size to remove on timeline 1 : 7168MB
INFO: Resident WAL size to free on timeline 1 : 7168MB
INFO: There are no backups to merge by retention policy # 保留策略，没有备份需要被合并(merge)
INFO: There are no backups to delete by retention policy # 保留策略，没有备份需要被删除
INFO: There is no WAL to purge by retention policy # 保留策略，没有WAL文件需要被清除
# 执行真正的删除动作
[lightdb@lightdb ~]$ lt_probackup delete -B /home/lightdb/backup --instance instance_192.168.76.128_5432 --delete-wal
INFO: On timeline 1 WAL segments between 000000010000000200000007 and 000000010000000400000005 will be removed # 时间线1上的WAL 段将被删除
INFO: Logical WAL size to remove on timeline 1 : 7168MB
INFO: Resident WAL size to free on timeline 1 : 7168MB
INFO: There are no backups to merge by retention policy
INFO: There are no backups to delete by retention policy
# 查看备份目录下wal目录下实例的WAL文件，发现WAL文件段区间000000010000000200000007 and 000000010000000400000005已被删除
[lightdb@lightdb ~]$ ls /home/lightdb/backup/wal/instance_192.168.76.128_5432/ -rltha
总用量 7.6G
drwx------ 3 lightdb lightdb   42 7月  10 12:23 ..
-rw------- 1 lightdb lightdb 512M 7月  10 21:01 000000010000000400000005
-rw------- 1 lightdb lightdb  355 7月  10 21:02 000000010000000400000005.00001690.backup
-rw------- 1 lightdb lightdb 512M 7月  10 21:02 000000010000000400000006
-rw------- 1 lightdb lightdb 512M 7月  10 21:09 000000010000000400000007
-rw------- 1 lightdb lightdb 512M 7月  10 21:09 000000010000000500000000
-rw------- 1 lightdb lightdb 512M 7月  10 21:09 000000010000000500000001
-rw------- 1 lightdb lightdb  346 7月  10 21:10 000000010000000500000000.00000120.backup
-rw------- 1 lightdb lightdb 512M 7月  11 12:56 000000010000000500000002
-rw------- 1 lightdb lightdb 512M 7月  11 12:56 000000010000000500000003
-rw------- 1 lightdb lightdb  355 7月  11 12:56 000000010000000500000003.00000120.backup
-rw------- 1 lightdb lightdb 512M 7月  11 12:56 000000010000000500000004
-rw------- 1 lightdb lightdb 512M 7月  11 12:59 000000010000000500000005
-rw------- 1 lightdb lightdb 512M 7月  11 12:59 000000010000000500000006
-rw------- 1 lightdb lightdb 512M 7月  11 12:59 000000010000000500000007
-rw------- 1 lightdb lightdb  355 7月  11 12:59 000000010000000500000006.00000120.backup
-rw------- 1 lightdb lightdb 512M 7月  11 13:07 000000010000000600000000
-rw------- 1 lightdb lightdb 512M 7月  11 13:07 000000010000000600000001
-rw------- 1 lightdb lightdb 512M 7月  11 13:07 000000010000000600000002
-rw------- 1 lightdb lightdb  355 7月  11 13:07 000000010000000600000001.000000F8.backup
-rw------- 1 lightdb lightdb 512M 7月  11 13:10 000000010000000600000003
drwx------ 2 lightdb lightdb 4.0K 7月  11 17:34 .

删除根据当前保留策略过期的备份，使用参数–merge-expired

# 语法 lt_probackup delete -B backup_dir --instance instance_name --delete-expired
# 执行delete命令 添加选项 --dry-run 尝试执行一下不产生实际效果
[lightdb@lightdb ~]$ lt_probackup delete -B /home/lightdb/backup  --instance instance_192.168.76.128_5432 --delete-expired --dry-run
INFO: Evaluate backups by retention
WARNING: Failed to find parent FULL backup for REUB00 # backup id REUB00 的备份状态为ERROR 在备份过程中出现了意外而终止了。
INFO: Backup REUBKR, mode: DELTA, status: OK. Redundancy: 1/2, Time Window: 0d/14d. Active
INFO: Backup REUB6L, mode: PAGE, status: OK. Redundancy: 1/2, Time Window: 0d/14d. Active
INFO: Backup REUB0Z, mode: PAGE, status: OK. Redundancy: 1/2, Time Window: 0d/14d. Active
INFO: Backup REUB00, mode: PAGE, status: ERROR. Redundancy: 1/2, Time Window: 0d/14d. Active
INFO: Backup RET2UT, mode: FULL, status: OK. Redundancy: 1/2, Time Window: 1d/14d. Active
INFO: There are no backups to merge by retention policy
INFO: There are no backups to delete by retention polic y
INFO: There is no WAL to purge by retention policy

当至少一个满足保留策略的增量备份基于它们时，无法删除过期的备份。如果想尽量减少保持增量备份有效所需的备份数，使用参数–merge-expired

lt_probackup 会搜索满足保留策略的最旧的增量备份，并将此备份与已经过期的底层全量备份和增量备份合并(merge)，从而使其成为全量备份。合并完成后，剩余的过期备份将被删除。

# 基本语法：lt_probackup delete -B backup_dir --instance instance_name --delete-expired --merge-expired
[lightdb@lightdb ~]$ lt_probackup delete -B /home/lightdb/backup  --instance instance_192.168.76.128_5432 --delete-expired --merge-expired
INFO: Evaluate backups by retention
INFO: Backup REUBKR, mode: DELTA, status: OK. Redundancy: 1/2, Time Window: 0d/14d. Active
INFO: Backup REUB6L, mode: PAGE, status: OK. Redundancy: 1/2, Time Window: 0d/14d. Active
INFO: Backup REUB0Z, mode: PAGE, status: OK. Redundancy: 1/2, Time Window: 0d/14d. Active
INFO: Backup RET2UT, mode: FULL, status: OK. Redundancy: 1/2, Time Window: 1d/14d. Active
INFO: Retention merging finished
INFO: There are no backups to merge by retention policy
INFO: There are no backups to delete by retention policy
INFO: There is no WAL to purge by retention policy

根据备份状态(status)来删除备份可以忽略已经建立的保留策略

**注意：**根据状态(status)来删除备份的时候忽略已经建立的保留策略。

# 以删除状态为ERROR的备份为例，查询状态为ERROR的备份列表
[lightdb@lightdb ~]$ lt_probackup show -B /home/lightdb/backup  --instance instance_192.168.76.128_5432
==========================================================================================================================================================
 Instance                      Version  ID      Recovery Time           Mode   WAL Mode  TLI  Time    Data    WAL  Zratio  Start LSN   Stop LSN    Status
==========================================================================================================================================================
 instance_192.168.76.128_5432  13       REUBKR  2022-07-11 13:07:50+08  DELTA  ARCHIVE   1/1   18s  4054kB  512MB    1.00  6/200000F8  6/4000A850  OK
 instance_192.168.76.128_5432  13       REUB6L  2022-07-11 12:59:23+08  PAGE   ARCHIVE   1/1   21s   707kB  512MB    1.00  5/C0000120  5/E000B678  OK
 instance_192.168.76.128_5432  13       REUB0Z  2022-07-11 12:56:15+08  PAGE   ARCHIVE   1/1   38s    21MB  512MB    1.00  5/60000120  5/800358A8  OK
 instance_192.168.76.128_5432  ----     REUB00  ----                    PAGE   ARCHIVE   0/1     0       0      0    1.00  0/0         0/0         ERROR
 instance_192.168.76.128_5432  13       RET2UT  2022-07-10 21:01:59+08  FULL   ARCHIVE   1/0   27s   361MB  512MB    1.00  4/A0001690  4/C000A0C8  OK

# backup id REUB00在备份过程因为意外而终止，尝试删除该状态为ERROR的备份
# 查看backup id REUB00 在备份目录是否存在
[lightdb@lightdb ~]$ ls /home/lightdb/backup/backups/instance_192.168.76.128_5432/ -rltha
总用量 4.0K
drwx------ 3 lightdb lightdb  42 7月  10 12:23 ..
-rw-rw-r-- 1 lightdb lightdb 392 7月  10 14:26 pg_probackup.conf
drwx------ 3 lightdb lightdb  97 7月  10 22:06 RET2UT
drwx------ 3 lightdb lightdb  44 7月  11 12:55 REUB00
drwx------ 3 lightdb lightdb  97 7月  11 12:56 REUB0Z
drwx------ 3 lightdb lightdb  97 7月  11 12:59 REUB6L
drwx------ 7 lightdb lightdb 101 7月  11 13:07 .
drwx------ 3 lightdb lightdb  97 7月  11 13:07 REUBKR
# backup id REUB00在备份过程因为意外而终止，删除该状态为ERROR的备份
[lightdb@lightdb ~]$ lt_probackup delete -B /home/lightdb/backup  --instance instance_192.168.76.128_5432 --dry-run --status ERROR
INFO: Deleting all backups with status 'ERROR' in dry run mode
INFO: Backup REUB00 with status ERROR can be deleted
INFO: Resident data size to free by delete of 1 backups: 0

[lightdb@lightdb ~]$ lt_probackup delete -B /home/lightdb/backup  --instance instance_192.168.76.128_5432 --status ERROR
INFO: Deleting all backups with status 'ERROR'
INFO: Backup REUB00 with status ERROR will be deleted
INFO: Delete: REUB00 2022-07-11 12:55:12+08
INFO: Resident data size to free by delete of 1 backups: 0
INFO: Successfully deleted 1 backup from instance 'instance_192.168.76.128_5432'

备份状态的范围

Status— 备份状态。可能的值：

    OK— 备份完整且有效。

    DONE— 备份已完成，但未经验证。

    RUNNING— 备份正在进行中。

    MERGING— 正在合并备份。

    MERGED— 备份数据文件已成功合并，但其元数据正在更新中。只有完整备份才能具有此状态。

    DELETING— 正在删除备份文件。

    CORRUPT— 一些备份文件已损坏。

    ERROR— 备份因意外错误而中止。

    ORPHAN— 备份无效，因为其父备份之一已损坏或丢失。

查看当前备份实例的保留策略

[lightdb@lightdb ~]$ lt_probackup show-config  -B /home/lightdb/backup --instance instance_192.168.76.128_5432
# Backup instance information
pgdata = /home/lightdb/database/lightdb-x/13.3-22.2/data/defaultCluster
system-identifier = 7117222465357338420
xlog-seg-size = 536870912
# Connection parameters
pgdatabase = postgres
pghost = 192.168.76.128
pgport = 5432
pguser = lt_probackup
# Replica parameters
replica-timeout = 5min
# Archive parameters
archive-timeout = 5min
# Logging parameters
log-level-console = INFO
log-level-file = OFF
log-filename = pg_probackup.log
log-rotation-size = 0TB
log-rotation-age = 0d
# Retention parameters 保留参数
retention-redundancy = 2  # 可保持的全量副本的数量
retention-window = 14 # 可恢复的时间窗口
wal-depth = 0
# Compression parameters
compress-algorithm = none
compress-level = 2
# Remote access parameters
remote-proto = ssh

# 使用--dry-run参数做一个尝试不采取真正行为，比如要删除状态为OK的备份
[lightdb@lightdb ~]$ lt_probackup delete -B /home/lightdb/backup --instance instance_192.168.76.128_5432 --dry-run --status=OK
INFO: Deleting all backups with status 'OK' in dry run mode
INFO: Backup RET37Y with status OK can be deleted
INFO: Backup RET2UT with status OK can be deleted
INFO: Backup RET2R2 with status OK can be deleted
INFO: Backup RET2MZ with status OK can be deleted
INFO: Backup RET2M1 with status OK can be deleted
INFO: Backup RET2J1 with status OK can be deleted
INFO: Resident data size to free by delete of 6 backups: 2162MB

# 查看这儿写备份其实都还在
[lightdb@lightdb ~]$ ls /home/lightdb/backup/backups/instance_192.168.76.128_5432/ -rltha
总用量 4.0K
drwx------ 3 lightdb lightdb  42 7月  10 12:23 ..
-rw-rw-r-- 1 lightdb lightdb 392 7月  10 14:26 pg_probackup.conf
drwx------ 8 lightdb lightdb 115 7月  10 21:09 .
drwx------ 3 lightdb lightdb  97 7月  10 22:06 RET37Y
drwx------ 3 lightdb lightdb  97 7月  10 22:06 RET2UT
drwx------ 3 lightdb lightdb  97 7月  10 22:06 RET2R2
drwx------ 3 lightdb lightdb  97 7月  10 22:06 RET2MZ
drwx------ 3 lightdb lightdb  97 7月  10 22:06 RET2M1
drwx------ 3 lightdb lightdb  97 7月  10 22:06 RET2J1

lt_probackup执行恢复命令

在使用工具lt_probackup执行备份并校验备份有效之后，要从备份中恢复数据库实例需要执行恢复命令(restore)。
语法：

[lightdb@lightdb ~]$ lt_probackup restore --help

lt_probackup restore -B backup-path --instance=instance_name
                 [-D pgdata-path] [-i backup-id] [-j num-threads]
                 [--progress] [--force] [--no-sync]
                 [--no-validate] [--skip-block-validation]
                 [-T OLDDIR=NEWDIR]
                 [--external-mapping=OLDDIR=NEWDIR]
                 [--skip-external-dirs]
                 [-I | --incremental-mode=none|checksum|lsn]
                 [--db-include dbname | --db-exclude dbname]
                 [--recovery-target-time=time|--recovery-target-xid=xid
                  |--recovery-target-lsn=lsn [--recovery-target-inclusive=boolean]]
                 [--recovery-target-timeline=timeline]
                 [--recovery-target=immediate|latest]
                 [--recovery-target-name=target-name]
                 [--recovery-target-action=pause|promote|shutdown]
                 [--restore-command=cmdline]
                 [-R | --restore-as-replica]
                 [--primary-conninfo=primary_conninfo]
                 [-S | --primary-slot-name=slotname]
                 [--remote-proto] [--remote-host]
                 [--remote-port] [--remote-path] [--remote-user]
                 [--ssh-options]
                 [--archive-host=hostname] [--archive-port=port]
                 [--archive-user=username]

  -B, --backup-path=backup-path    location of the backup storage area # 备份目录
      --instance=instance_name     name of the instance # 备份实例
  -D, --pgdata=pgdata-path         location of the database storage area # 数据库实例的数据目录
  -i, --backup-id=backup-id        backup to restore # 待恢复的backup id
  -j, --threads=NUM                number of parallel threads
      --progress                   show progress
      --force                      ignore invalid status of the restored backup
      --no-sync                    do not sync restored files to disk
      --no-validate                disable backup validation during restore
      --skip-block-validation      set to validate only file-level checksum
  -T, --tablespace-mapping=OLDDIR=NEWDIR
                                   relocate the tablespace from directory OLDDIR to NEWDIR
      --external-mapping=OLDDIR=NEWDIR
                                   relocate the external directory from OLDDIR to NEWDIR
      --skip-external-dirs         do not restore all external directories

  Incremental restore options:  # 增量恢复选项
  -I, --incremental-mode=none|checksum|lsn
                                   reuse valid pages available in LTDATA if they have not changed
                                   (default: none)

  Partial restore options: # 部分恢复
      --db-include dbname          restore only specified databases # 只恢复指定的数据库
      --db-exclude dbname          do not restore specified databases # 不恢复指定的数据库

  Recovery options: # 恢复参数
      --recovery-target-time=time  time stamp up to which recovery will proceed
      --recovery-target-xid=xid    transaction ID up to which recovery will proceed
      --recovery-target-lsn=lsn    LSN of the write-ahead log location up to which recovery will proceed
      --recovery-target-inclusive=boolean
                                   whether we stop just after the recovery target
      --recovery-target-timeline=timeline
                                   recovering into a particular timeline
      --recovery-target=immediate|latest
                                   end recovery as soon as a consistent state is reached or as late as possible
      --recovery-target-name=target-name
                                   the named restore point to which recovery will proceed
      --recovery-target-action=pause|promote|shutdown
                                   action the server should take once the recovery target is reached
                                   (default: pause)
      --restore-command=cmdline    command to use as 'restore_command' in recovery.conf; 'none' disables

  Standby options:
  -R, --restore-as-replica         write a minimal recovery.conf in the output directory
                                   to ease setting up a standby server
      --primary-conninfo=primary_conninfo
                                   connection string to be used for establishing connection
                                   with the primary server
  -S, --primary-slot-name=slotname replication slot to be used for WAL streaming from the primary server

  Logging options:
      --log-level-console=log-level-console
                                   level for console logging (default: info)
                                   available options: 'off', 'error', 'warning', 'info', 'log', 'verbose'
      --log-level-file=log-level-file
                                   level for file logging (default: off)
                                   available options: 'off', 'error', 'warning', 'info', 'log', 'verbose'
      --log-filename=log-filename
                                   filename for file logging (default: 'pg_probackup.log')
                                   support strftime format (example: pg_probackup-%Y-%m-%d_%H%M%S.log)
      --error-log-filename=error-log-filename
                                   filename for error logging (default: none)
      --log-directory=log-directory
                                   directory for file logging (default: BACKUP_PATH/log)
      --log-rotation-size=log-rotation-size
                                   rotate logfile if its size exceeds this value; 0 disables; (default: 0)
                                   available units: 'kB', 'MB', 'GB', 'TB' (default: kB)
      --log-rotation-age=log-rotation-age
                                   rotate logfile if its age exceeds this value; 0 disables; (default: 0)
                                   available units: 'ms', 's', 'min', 'h', 'd' (default: min)

  Remote options: # 远程参数
      --remote-proto=protocol      remote protocol to use
                                   available options: 'ssh', 'none' (default: ssh)
      --remote-host=destination    remote host address or hostname
      --remote-port=port           remote host port (default: 22)
      --remote-path=path           path to directory with lt_probackup binary on remote host
                                   (default: current binary path)
      --remote-user=username       user name for ssh connection (default: current user)
      --ssh-options=ssh_options    additional ssh options (default: none)
                                   (example: --ssh-options='-c cipher_spec -F configfile')

  Remote WAL archive options:
      --archive-host=destination   address or hostname for ssh connection to archive host
      --archive-port=port          port for ssh connection to archive host (default: 22)
      --archive-user=username      user name for ssh connection to archive host (default: LightDB user)

命令完成后restore，启动数据库服务。
基本使用：lt_probackup restore -B backup_dir --instance instance_name -i backup_id
backup_dir是存储所有备份文件和元信息的备份目录。
instance_name是要恢复的集群的备份实例。
backup_id 指定要恢复的备份的标识。如果省略此选项，lt_probackup将使用可用于指定实例的最新有效备份。如果指定要恢复的增量备份，lt_probackup 会自动恢复底层全量备份，然后按顺序应用所有必要的增量。

# 本地备份恢复，备份目录与数据库实例在同一个机器之上,使用可用于指定实例的最新有效备份(若该备份为增量备份，优先全量备份，再依次应用必要的增量备份)
[lightdb@lightdb ~]$ lt_probackup restore -B /home/lightdb/backup --instance=instance_192.168.76.128_5432
ERROR: Restore destination is not empty: "/home/lightdb/database/lightdb-x/13.3-22.2/data/defaultCluster"# 待恢复的数据库实例的数据目录($LTDATA)不空则不允许进行恢复(restore)。

# 查看备份列表
[lightdb@lightdb ~]$ lt_probackup show -B /home/lightdb/backup --instance=instance_192.168.76.128_5432
==========================================================================================================================================================
 Instance                      Version  ID      Recovery Time           Mode   WAL Mode  TLI  Time    Data    WAL  Zratio  Start LSN   Stop LSN    Status
==========================================================================================================================================================
 instance_192.168.76.128_5432  13       REUBKR  2022-07-11 13:07:50+08  DELTA  ARCHIVE   1/1   18s  4054kB  512MB    1.00  6/200000F8  6/4000A850  OK
 instance_192.168.76.128_5432  13       REUB6L  2022-07-11 12:59:23+08  PAGE   ARCHIVE   1/1   21s   707kB  512MB    1.00  5/C0000120  5/E000B678  OK
 instance_192.168.76.128_5432  13       REUB0Z  2022-07-11 12:56:15+08  PAGE   ARCHIVE   1/1   38s    21MB  512MB    1.00  5/60000120  5/800358A8  OK
 instance_192.168.76.128_5432  13       RET2UT  2022-07-10 21:01:59+08  FULL   ARCHIVE   1/0   27s   361MB  512MB    1.00  4/A0001690  4/C000A0C8  OK

# 清空数据库实例的数据目录($LTDATA)之后再次进行恢复
[lightdb@lightdb data]$ lt_probackup restore -B /home/lightdb/backup --instance=instance_192.168.76.128_5432
INFO: Validating parents for backup REUBKR # 最新有效备份REUBKR为增量备份，需要应用全量备份和按顺序应用必要增量
INFO: Validating backup RET2UT # 验证备份是否有效
INFO: Backup RET2UT data files are valid
INFO: Validating backup REUB0Z
INFO: Backup REUB0Z data files are valid
INFO: Validating backup REUB6L
INFO: Backup REUB6L data files are valid
INFO: Validating backup REUBKR
INFO: Backup REUBKR data files are valid
INFO: Backup REUBKR WAL segments are valid
INFO: Backup REUBKR is valid.
INFO: Restoring the database from backup at 2022-07-11 13:07:39+08 # 从备份恢复在时间点"2022-07-11 13:07:39+08"的数据库数据
INFO: Start restoring backup files. LTDATA size: 361MB
INFO: Backup files are restored. Transfered bytes: 361MB, time elapsed: 1s
INFO: Restore incremental ratio (less is better): 100% (361MB/361MB)
INFO: Syncing restored files to disk
INFO: Restored backup files are synced, time elapsed: 1s
INFO: Restore of backup REUBKR completed. # 备份标识REUBKR的恢复已完成

# 重启数据库实例(指定数据目录)
[lightdb@lightdb data]$ lt_ctl start -D $LTDATA
2022-07-11 22:27:03.297937T  @  postmaster  00000[2022-07-11 22:27:03 CST] 0 [123172] LOG:  LightDB autoprewarm: prewarm dbnum=0
waiting for server to start....2022-07-11 22:27:03.302378T  @  postmaster  00000[2022-07-11 22:27:03 CST] 0 [123172] LOG:  ltaudit extension initialized
....2022-07-11 22:27:07.938711T  @  postmaster  00000[2022-07-11 22:27:03 CST] 0 [123172] LOG:  redirecting log output to logging collector process
2022-07-11 22:27:07.938711T  @  postmaster  00000[2022-07-11 22:27:03 CST] 0 [123172] HINT:  Future log output will appear in directory "log".
.................... done
server started

增量恢复

使用“增量恢复选项”仅替换现有Postgres Pro数据目录中的无效和更改页面，可以显着提高从备份恢复的速度。

语法：lt_probackup restore -B backup_dir --instance instance_name -D data_dir -I incremental_mode
-D data_dir # 表示lightdb数据库的数据目录
-I incremental_mode # 增量恢复选项

增量恢复选项(incremental_mode) 介绍

CHECKSUM — 读取数据目录中的所有数据文件，验证每个页面中的 header 和 checksum，只替换无效页面以及那些 checksum 和 LSN 与备份中相应页面不匹配的页面。这是最简单的增量模式。建议默认使用。

LSN — 读取pg_control数据目录中的数据以获得重做 LSN 和重做 TLI，这允许确定历史中的一个点 (shiftpoint)，其中数据目录状态从目标备份链历史转移。如果 shiftpoint 不在备份链历史的范围内，则恢复被中止。如果 shiftpoint 在备份链历史的范围内，则读取数据目录中的所有数据文件，验证每个页面中的 header 和校验和，仅替换无效页面和 LSN 大于 shiftpoint 的页面。与 CHECKSUM 相比，此模式提供了更大的加速，但需要满足两个条件。一、data_checksums必须在数据目录中启用参数（以避免由于提示位而损坏）。此条件将在增量恢复开始时检查，如果禁用校验和，操作将中止。其次，pg_control文件必须与数据目录的状态同步。在恢复开始时无法检查此条件，因此用户有责任确保pg_control包含有效信息。因此，不建议在任何情况下使用 LSN 模式，其中 pg_control 不可信或被篡改：pg_resetxlog执行后，从备份恢复后未运行恢复等。

NONE — 没有任何增量优化的常规恢复。

注意：无论选择哪种增量模式，lt_probackup 都会检查给定目标目录中的 postmaster 是否未运行并且system-identifier与备份中的相同。

[lightdb@lightdb ~]$ lt_probackup restore -B /home/lightdb/backup --instance=instance_192.168.76.128_5432 -I checksum
INFO: Running incremental restore into nonempty directory: "/home/lightdb/database/lightdb-x/13.3-22.2/data/defaultCluster"
WARNING: Postmaster with pid 1580 is running in destination directory "/home/lightdb/database/lightdb-x/13.3-22.2/data/defaultCluster"
INFO: Trying to read lt_control file in destination direstory
ERROR: Incremental restore is not allowed # 增量恢复，需要停止当前数据库实例(pid 1580)

# 停止运行的数据库实例
[lightdb@lightdb defaultCluster]$ lt_ctl stop
waiting for server to shut down.......... done
server stopped

# 查看备份列表
[lightdb@lightdb defaultCluster]$ lt_probackup show -B /home/lightdb/backup --instance=instance_192.168.76.128_5432
==========================================================================================================================================================
 Instance                      Version  ID      Recovery Time           Mode   WAL Mode  TLI  Time    Data    WAL  Zratio  Start LSN   Stop LSN    Status
==========================================================================================================================================================
 instance_192.168.76.128_5432  13       REUBKR  2022-07-11 13:07:50+08  DELTA  ARCHIVE   1/1   18s  4054kB  512MB    1.00  6/200000F8  6/4000A850  OK
 instance_192.168.76.128_5432  13       REUB6L  2022-07-11 12:59:23+08  PAGE   ARCHIVE   1/1   21s   707kB  512MB    1.00  5/C0000120  5/E000B678  OK
 instance_192.168.76.128_5432  13       REUB0Z  2022-07-11 12:56:15+08  PAGE   ARCHIVE   1/1   38s    21MB  512MB    1.00  5/60000120  5/800358A8  OK
 instance_192.168.76.128_5432  13       RET2UT  2022-07-10 21:01:59+08  FULL   ARCHIVE   1/0   27s   361MB  512MB    1.00  4/A0001690  4/C000A0C8  OK

# 再次进行增量恢复，增量恢复选项使用checksum
[lightdb@lightdb defaultCluster]$ lt_probackup restore -B /home/lightdb/backup --instance=instance_192.168.76.128_5432 -I checksum
INFO: Running incremental restore into nonempty directory: "/home/lightdb/database/lightdb-x/13.3-22.2/data/defaultCluster"
INFO: Trying to read lt_control file in destination direstory
INFO: Validating parents for backup REUBKR
INFO: Validating backup RET2UT
INFO: Backup RET2UT data files are valid
INFO: Validating backup REUB0Z
INFO: Backup REUB0Z data files are valid
INFO: Validating backup REUB6L
INFO: Backup REUB6L data files are valid
INFO: Validating backup REUBKR
INFO: Backup REUBKR data files are valid
INFO: Backup REUBKR WAL segments are valid
INFO: Backup REUBKR is valid.
INFO: Restoring the database from backup at 2022-07-11 13:07:39+08 # 恢复在时间点2022-07-11 13:07:39+08 的备份
INFO: Extracting the content of destination directory for incremental restore
INFO: Destination directory content extracted, time elapsed: 0
INFO: Removing redundant files in destination directory
INFO: Redundant files are removed, time elapsed: 0
INFO: Start restoring backup files. LTDATA size: 361MB
INFO: Backup files are restored. Transfered bytes: 6435kB, time elapsed: 1s
INFO: Restore incremental ratio (less is better): 2% (6435kB/361MB) # 恢复增量百分比 2% 越小越好
INFO: Syncing restored files to disk
INFO: Restored backup files are synced, time elapsed: 0
INFO: Restore of backup REUBKR completed.

# 再次启动lightdb数据库实例
[lightdb@lightdb defaultCluster]$ lt_ctl start -D $LTDATA
2022-07-12 10:37:45.197079T  @  postmaster  00000[2022-07-12 10:37:45 CST] 0 [24478] LOG:  LightDB autoprewarm: prewarm dbnum=0
2022-07-12 10:37:45.199313T  @  postmaster  00000[2022-07-12 10:37:45 CST] 0 [24478] LOG:  ltaudit extension initialized
waiting for server to start........2022-07-12 10:37:50.013407T  @  postmaster  00000[2022-07-12 10:37:45 CST] 0 [24478] LOG:  redirecting log output to logging collector process
2022-07-12 10:37:50.013407T  @  postmaster  00000[2022-07-12 10:37:45 CST] 0 [24478] HINT:  Future log output will appear in directory "log".
.......... done
server started

部分恢复

如果在进行备份之前启用了部分还原，则可以使用部分还原选项(partial restore options)和还原命令(restore)仅还原部分数据库。

设置部分还原

语法：GRANT SELECT ON TABLE pg_catalog.pg_database TO backup;
为备份用户backup赋予数据库表pg_catalog.pg_database的只读权限

#为备份用户lt_probackup设置部分还原
[lightdb@lightdb ~]$ ltsql
ltsql (13.3-22.2)
Type "help" for help.

lightdb@postgres=# \du
                                     List of roles
  Role name   |                         Attributes                         | Member of
--------------+------------------------------------------------------------+-----------
 em           | Superuser                                                  | {}
 lightdb      | Superuser, Create role, Create DB, Replication, Bypass RLS | {}
 lt_probackup |                                                            | {}

lightdb@postgres=# GRANT SELECT ON TABLE pg_catalog.pg_database TO lt_probackup;
GRANT
lightdb@postgres=# \q
# 为数据库的备份用户lt_probackup赋予部分还原的权限。

部分恢复选项说明

  --db-include dbname          恢复指定的数据库
  --db-exclude dbname          排除指定数据库

# 恢复指定数据库em
[lightdb@lightdb ~]$ lt_probackup restore -B /home/lightdb/backup --instance=instance_192.168.76.128_5432 --db-include em
ERROR: Restore destination is not empty: "/home/lightdb/database/lightdb-x/13.3-22.2/data/defaultCluster"
# 需要清空数据库目录之后才能进行恢复指定数据库

# 停止正在运行的数据库实例
[lightdb@lightdb ~]$ lt_ctl stop
waiting for server to shut down......... done
server stopped

# 清空数据目录 /home/lightdb/database/lightdb-x/13.3-22.2/data/defaultCluster

# 恢复指定数据库em 使用--db-include选项
[lightdb@lightdb data]$ lt_probackup restore -B /home/lightdb/backup --instance=instance_192.168.76.128_5432 --db-include em
INFO: Validating parents for backup REUBKR
INFO: Validating backup RET2UT
INFO: Backup RET2UT data files are valid
INFO: Validating backup REUB0Z
INFO: Backup REUB0Z data files are valid
INFO: Validating backup REUB6L
INFO: Backup REUB6L data files are valid
INFO: Validating backup REUBKR
INFO: Backup REUBKR data files are valid
INFO: Backup REUBKR WAL segments are valid
INFO: Backup REUBKR is valid.
INFO: Restoring the database from backup at 2022-07-11 13:07:39+08
INFO: Start restoring backup files. LTDATA size: 361MB
INFO: Backup files are restored. Transfered bytes: 320MB, time elapsed: 1s
INFO: Restore incremental ratio (less is better): 89% (320MB/361MB)
INFO: Syncing restored files to disk
INFO: Restored backup files are synced, time elapsed: 0
INFO: Restore of backup REUBKR completed.

# 重启数据库实例，并进行连接
[lightdb@lightdb ~]$ lt_ctl start -D $LTDATA
2022-07-12 11:15:03.492031T  @  postmaster  00000[2022-07-12 11:15:03 CST] 0 [39627] LOG:  LightDB autoprewarm: prewarm dbnum=0
waiting for server to start....2022-07-12 11:15:03.496560T  @  postmaster  00000[2022-07-12 11:15:03 CST] 0 [39627] LOG:  ltaudit extension initialized
...2022-07-12 11:15:06.848920T  @  postmaster  00000[2022-07-12 11:15:03 CST] 0 [39627] LOG:  redirecting log output to logging collector process
2022-07-12 11:15:06.848920T  @  postmaster  00000[2022-07-12 11:15:03 CST] 0 [39627] HINT:  Future log output will appear in directory "log".
.... done
server started
[lightdb@lightdb ~]$ ltsql
ltsql: error: FATAL:  "base/13592" is not a valid data directory
DETAIL:  File "base/13592/LT_VERSION" does not contain valid data.
HINT:  You might need to initdb. # 提示需要执行initdb命令

# 要从还原中排除一个或多个数据库，使用--db-exclude选项，部分恢复排除数据库em
[lightdb@lightdb ~]$ lt_probackup restore -B /home/lightdb/backup --instance=instance_192.168.76.128_5432 --db-exclude=em
ERROR: Restore destination is not empty: "/home/lightdb/database/lightdb-x/13.3-22.2/data/defaultCluster"
[lightdb@lightdb ~]$ rm -rf /home/lightdb/database/lightdb-x/13.3-22.2/data/defaultCluster # 只是本地虚拟机这么暴力删除 需要提前将该数据目录进行备份。

# 再次进行部分恢复，排除数据库em
[lightdb@lightdb ~]$ lt_probackup restore -B /home/lightdb/backup --instance=instance_192.168.76.128_5432 --db-exclude=em
INFO: Validating parents for backup REUBKR
INFO: Validating backup RET2UT
INFO: Backup RET2UT data files are valid
INFO: Validating backup REUB0Z
INFO: Backup REUB0Z data files are valid
INFO: Validating backup REUB6L
INFO: Backup REUB6L data files are valid
INFO: Validating backup REUBKR
INFO: Backup REUBKR data files are valid
INFO: Backup REUBKR WAL segments are valid
INFO: Backup REUBKR is valid.
INFO: Restoring the database from backup at 2022-07-11 13:07:39+08
INFO: Start restoring backup files. LTDATA size: 361MB
INFO: Backup files are restored. Transfered bytes: 59MB, time elapsed: 0
INFO: Restore incremental ratio (less is better): 16% (59MB/361MB)
INFO: Syncing restored files to disk
INFO: Restored backup files are synced, time elapsed: 0
INFO: Restore of backup REUBKR completed.

# 启动数据库实例
[lightdb@lightdb ~]$ lt_ctl start -D $LTDATA
2022-07-12 11:19:13.184998T  @  postmaster  00000[2022-07-12 11:19:13 CST] 0 [41826] LOG:  LightDB autoprewarm: prewarm dbnum=0
waiting for server to start....2022-07-12 11:19:13.189054T  @  postmaster  00000[2022-07-12 11:19:13 CST] 0 [41826] LOG:  ltaudit extension initialized
....2022-07-12 11:19:17.402657T  @  postmaster  00000[2022-07-12 11:19:13 CST] 0 [41826] LOG:  redirecting log output to logging collector process
2022-07-12 11:19:17.402657T  @  postmaster  00000[2022-07-12 11:19:13 CST] 0 [41826] HINT:  Future log output will appear in directory "log".
...... done
server started

# 验证部分恢复之后，数据库em是否已经不存在
[lightdb@lightdb ~]$ ltsql em
ltsql: error: FATAL:  "base/28848" is not a valid data directory
DETAIL:  File "base/28848/LT_VERSION" does not contain valid data.
HINT:  You might need to initdb.

执行时间点 (PITR) 恢复

准备

如果在进行备份(backup命令)之前启用了连续 WAL 归档，则可以使用restore命令的恢复目标选项(recovery target options)将集群恢复到任意时间点（恢复目标）的状态。

restore命令的恢复目标选项(recovery target options)说明

如果配置了连续 WAL 归档，可以将这些选项之一与restore或validate命令一起使用来指定数据库实例必须恢复或验证的时刻。

–recovery-target=immediate|latest
定义何时停止恢复：

	该immediate值在达到指定备份的一致状态后停止恢复，或者如果省略-i/--backup-id选项，则停止最新的可用备份。这是 STREAM 备份的默认行为。

	该latest值将继续恢复，直到应用了归档中所有可用的 WAL 段。这是 ARCHIVE 备份的默认行为。

–recovery-target-timeline=timeline
指定用于恢复的特定时间线。默认情况下，使用指定备份的时间线。

–recovery-target-lsn=lsn
指定恢复将继续到的预写日志位置的 LSN。只能在恢复主版本 10 或更高版本的数据库实例时使用。

–recovery-target-name=recovery_target_name
指定将数据库实例恢复到的命名保存点。

–recovery-target-time=time
指定恢复到的时间戳。如果未指定时区偏移量，则使用本地时区。例子：–recovery-target-time=“2020-01-01 00:00:00+03”

–recovery-target-xid=xid
指定恢复到的事务 ID。

–recovery-target-inclusive=boolean
指定是在指定的恢复目标 ( true) 之后停止，还是在恢复目标 ( false) 之前停止。此选项只能与 --recovery-target-name, --recovery-target-time, --recovery-target-lsn or --recovery-target-xid一起使用。默认值取决于recovery_target_inclusive参数。

–recovery-target-action=pause|promote|shutdown
指定达到恢复目标时服务器应采取的 recovery_target_action。默认：pause

实践基于时间点恢复

可以同时使用 STREAM 和 ARCHIVE 备份进行时间点恢复，只要 WAL 存档至少从进行备份开始的时候可用。如果省略-i/选项， lt_probackup 会自动选择最接近指定恢复目标的备份并启动恢复过程，否则lt_probackup将尝试将指定备份(--backup-id ,-I)恢复到指定恢复目标。

# 查看备份目录列表
[lightdb@lightdb ~]$ lt_probackup show -B /home/lightdb/backup --instance=instance_192.168.76.128_5432
==========================================================================================================================================================
 Instance                      Version  ID      Recovery Time           Mode   WAL Mode  TLI  Time    Data    WAL  Zratio  Start LSN   Stop LSN    Status
==========================================================================================================================================================
 instance_192.168.76.128_5432  13       REUBKR  2022-07-11 13:07:50+08  DELTA  ARCHIVE   1/1   18s  4054kB  512MB    1.00  6/200000F8  6/4000A850  OK
 instance_192.168.76.128_5432  13       REUB6L  2022-07-11 12:59:23+08  PAGE   ARCHIVE   1/1   21s   707kB  512MB    1.00  5/C0000120  5/E000B678  OK
 instance_192.168.76.128_5432  13       REUB0Z  2022-07-11 12:56:15+08  PAGE   ARCHIVE   1/1   38s    21MB  512MB    1.00  5/60000120  5/800358A8  OK
 instance_192.168.76.128_5432  13       RET2UT  2022-07-10 21:01:59+08  FULL   ARCHIVE   1/0   27s   361MB  512MB    1.00  4/A0001690  4/C000A0C8  OK

# 在准确时间恢复数据库状态，使用恢复目标选项--recovery-target-time以时间戳格式指定内容。
# 语法 lt_probackup restore -B backup_dir --instance instance_name --recovery-target-time="2017-05-18 14:18:11+03"
# 恢复到时间点 2022-07-11 13:00:50+08
[lightdb@lightdb ~]$ lt_probackup restore -B /home/lightdb/backup --instance=instance_192.168.76.128_5432 --recovery-target-time="2022-07-11 13:00:50+08"
ERROR: Restore destination is not empty: "/home/lightdb/database/lightdb-x/13.3-22.2/data/defaultCluster"
# 清空数据目录，本次只是学习分享，已提前备份该数据库数据目录，清空该目录
[lightdb@lightdb ~]$ rm -rf /home/lightdb/database/lightdb-x/13.3-22.2/data/defaultCluster
[lightdb@lightdb ~]$ lt_probackup restore -B /home/lightdb/backup --instance=instance_192.168.76.128_5432 --recovery-target-time="2022-07-11 13:00:50+08"
INFO: Validating parents for backup REUB6L
INFO: Validating backup RET2UT
INFO: Backup RET2UT data files are valid
INFO: Validating backup REUB0Z
INFO: Backup REUB0Z data files are valid
INFO: Validating backup REUB6L
INFO: Backup REUB6L data files are valid
INFO: Backup validation completed successfully on time 2022-07-11 13:00:50+08, xid 770785 and LSN 6/306EB0 # 恢复之前验证备份在待恢复的时间点"2022-07-11 13:00:50+08"是否有效
INFO: Backup REUB6L is valid.
INFO: Restoring the database from backup at 2022-07-11 12:59:09+08
INFO: Start restoring backup files. LTDATA size: 360MB
INFO: Backup files are restored. Transfered bytes: 360MB, time elapsed: 0
INFO: Restore incremental ratio (less is better): 100% (360MB/360MB)
INFO: Syncing restored files to disk
INFO: Restored backup files are synced, time elapsed: 2s
INFO: Restore of backup REUB6L completed.

# 要将集群状态恢复到特定 LSN，使用恢复目标选项--recovery-target-lsn
# 查看备份列表
[lightdb@lightdb ~]$ lt_probackup show -B /home/lightdb/backup --instance=instance_192.168.76.128_5432
==========================================================================================================================================================
 Instance                      Version  ID      Recovery Time           Mode   WAL Mode  TLI  Time    Data    WAL  Zratio  Start LSN   Stop LSN    Status
==========================================================================================================================================================
 instance_192.168.76.128_5432  13       REUBKR  2022-07-11 13:07:50+08  DELTA  ARCHIVE   1/1   18s  4054kB  512MB    1.00  6/200000F8  6/4000A850  OK
 instance_192.168.76.128_5432  13       REUB6L  2022-07-11 12:59:23+08  PAGE   ARCHIVE   1/1   21s   707kB  512MB    1.00  5/C0000120  5/E000B678  OK
 instance_192.168.76.128_5432  13       REUB0Z  2022-07-11 12:56:15+08  PAGE   ARCHIVE   1/1   38s    21MB  512MB    1.00  5/60000120  5/800358A8  OK
 instance_192.168.76.128_5432  13       RET2UT  2022-07-10 21:01:59+08  FULL   ARCHIVE   1/0   27s   361MB  512MB    1.00  4/A0001690  4/C000A0C8  OK

# 选项lsn 5/800358A8 对应的bakckup id为 REUB0Z
[lightdb@lightdb ~]$ lt_probackup restore -B /home/lightdb/backup --instance=instance_192.168.76.128_5432 --recovery-target-lsn=5/800358A8
ERROR: Restore destination is not empty: "/home/lightdb/database/lightdb-x/13.3-22.2/data/defaultCluster"
[lightdb@lightdb ~]$ rm -rf /home/lightdb/database/lightdb-x/13.3-22.2/data/defaultCluster # 仅仅学习分享使用，直接删除了数据目录，测试和生产环境不要这么做
[lightdb@lightdb ~]$ lt_probackup restore -B /home/lightdb/backup --instance=instance_192.168.76.128_5432 --recovery-target-lsn=5/800358A8
INFO: Validating parents for backup REUB0Z # 增量备份，先验证其父备份
INFO: Validating backup RET2UT #  全量备份 
INFO: Backup RET2UT data files are valid
INFO: Validating backup REUB0Z
INFO: Backup REUB0Z data files are valid
INFO: Backup validation completed successfully on time 2022-07-11 12:56:15+08, xid 769271 and LSN 5/800358A8
INFO: Backup REUB0Z is valid.
INFO: Restoring the database from backup at 2022-07-11 12:55:47+08
INFO: Start restoring backup files. LTDATA size: 360MB
INFO: Backup files are restored. Transfered bytes: 360MB, time elapsed: 0
INFO: Restore incremental ratio (less is better): 100% (360MB/360MB)
INFO: Syncing restored files to disk
INFO: Restored backup files are synced, time elapsed: 2s
INFO: Restore of backup REUB0Z completed.


# 将数据库实例恢复到特定的事务 ID，使用恢复目标选项--recovery-target-xid选项，xid 769271 为例
[lightdb@lightdb ~]$ lt_probackup restore -B /home/lightdb/backup --instance=instance_192.168.76.128_5432 --recovery-target-xid=769271
ERROR: Restore destination is not empty: "/home/lightdb/database/lightdb-x/13.3-22.2/data/defaultCluster" # 仅仅为例验证工功能，这里直接暴力删除了数据库数据目录 
[lightdb@lightdb ~]$ rm -rf /home/lightdb/database/lightdb-x/13.3-22.2/data/defaultCluster
[lightdb@lightdb ~]$ lt_probackup restore -B /home/lightdb/backup --instance=instance_192.168.76.128_5432 --recovery-target-xid=769271
INFO: Validating parents for backup REUB0Z
INFO: Validating backup RET2UT
INFO: Backup RET2UT data files are valid
INFO: Validating backup REUB0Z
INFO: Backup REUB0Z data files are valid
INFO: Backup validation completed successfully on time 2022-07-11 12:56:15+08, xid 769271 and LSN 5/800358A8
INFO: Backup REUB0Z is valid.
INFO: Restoring the database from backup at 2022-07-11 12:55:47+08
INFO: Start restoring backup files. LTDATA size: 360MB
INFO: Backup files are restored. Transfered bytes: 360MB, time elapsed: 0
INFO: Restore incremental ratio (less is better): 100% (360MB/360MB)
INFO: Syncing restored files to disk
INFO: Restored backup files are synced, time elapsed: 1s
INFO: Restore of backup REUB0Z completed.


# 查看备份列表
[lightdb@lightdb ~]$ lt_probackup show -B /home/lightdb/backup --instance=instance_192.168.76.128_5432
==========================================================================================================================================================
 Instance                      Version  ID      Recovery Time           Mode   WAL Mode  TLI  Time    Data    WAL  Zratio  Start LSN   Stop LSN    Status
==========================================================================================================================================================
 instance_192.168.76.128_5432  13       REUBKR  2022-07-11 13:07:50+08  DELTA  ARCHIVE   1/1   18s  4054kB  512MB    1.00  6/200000F8  6/4000A850  OK
 instance_192.168.76.128_5432  13       REUB6L  2022-07-11 12:59:23+08  PAGE   ARCHIVE   1/1   21s   707kB  512MB    1.00  5/C0000120  5/E000B678  OK
 instance_192.168.76.128_5432  13       REUB0Z  2022-07-11 12:56:15+08  PAGE   ARCHIVE   1/1   38s    21MB  512MB    1.00  5/60000120  5/800358A8  OK
 instance_192.168.76.128_5432  13       RET2UT  2022-07-10 21:01:59+08  FULL   ARCHIVE   1/0   27s   361MB  512MB    1.00  4/A0001690  4/C000A0C8  OK
 
# --recovery-target=latest 直到应用了归档中所有可用的 WAL 段。这是 ARCHIVE 备份的默认行为。
[lightdb@lightdb ~]$ lt_probackup restore -B /home/lightdb/backup --instance=instance_192.168.76.128_5432 --recovery-target=latest
INFO: Validating parents for backup REUBKR
INFO: Validating backup RET2UT
INFO: Backup RET2UT data files are valid
INFO: Validating backup REUB0Z
INFO: Backup REUB0Z data files are valid
INFO: Validating backup REUB6L
INFO: Backup REUB6L data files are valid
INFO: Validating backup REUBKR
INFO: Backup REUBKR data files are valid
INFO: Backup REUBKR WAL segments are valid
INFO: Backup REUBKR is valid.
INFO: Restoring the database from backup at 2022-07-11 13:07:39+08
INFO: Start restoring backup files. LTDATA size: 361MB
INFO: Backup files are restored. Transfered bytes: 361MB, time elapsed: 0
INFO: Restore incremental ratio (less is better): 100% (361MB/361MB)
INFO: Syncing restored files to disk
INFO: Restored backup files are synced, time elapsed: 2s
INFO: Restore of backup REUBKR completed.

在远程模式下使用lt_probackup执行备份(backup)与恢复(recovery)

准备

在本地准备两个虚拟机(192.168.76.129 和 192.168.76.128 )，操作系统centos7 ，平台x86_64，一台做备份服务器(192.168.76.129)，一台做为待备份的数据库实例(192.168.76.128)。
lt_probackup支持通过ssh远程地执行备份(backup)和恢复(restore)动作。
配置远程模式(备份服务器与数据库实例服务器ssh免密互通)。
两台虚拟机上均安装了lightdb数据库实例，其中备份服务器上不启动数据库实例，数据库实例的安装目录($LTHOME)下的bin目录下面有工具lt_probackup。

典型的工作流程

在备份服务器（192.168.76.129）上，按照安装和设置部分中的说明(1、初始化备份目录，2、在备份目录添加备份实例，3、配置lt_probackup)配置lt_probackup。对于add-instance(添加实例)和set-config(为lt_probackup添加配置) 命令，确保包含指向具有lightdb数据库实例的数据库服务器的远程选项(remote options)。
如果要在 PAGE 模式执行远程备份，或者依赖归档WAL 传输模式，或者使用 PITR(基于时间点恢复)，需要配置从数据库实例主机(192.168.76.128)到备份主机(192.168.76.129)的连续 WAL 归档。对于archive-push(归档推送)和archive-get(归档获取)命令，必须指定具有备份目录的备份主机的远程选项(remote options)。
在备份主机(192.168.76.128)上使用**远程选项(remote options)运行备份(backup)或还原(recovery)**命令。lt_probackup通过 SSH 连接到远程系统(192.168.76.129)，并在本地创建备份或分别在远程系统上还原以前创建的备份。
额外补充**远程选项(remote options)**的主要内容：
–remote-proto=proto 指定用于远程操作的协议。目前仅支持 SSH 协议。可能的值包括：
ssh 通过 SSH 启用远程模式。默认值。
none 显式禁用远程模式。
如果指定了此选项，则可以省略-remote-host选项
–remote-host=destination 指定要连接到的远程主机 IP 地址或主机名。
–remote-port=port 指定要连接到的远程主机端口。默认：22
–remote-user=username 指定用于 SSH 连接的远程主机用户。如果省略此选项，则使用启动 SSH 连接的当前用户。
–remote-path=path 指定远程主机上lt_probackup安装目录。
–ssh-options=ssh_options SSH 命令行选项。例如，以下选项可用于为由 lt_probackup 打开的 SSH 连接设置保持活动状态：
–ssh-options=“-o ServerAliveCountMax=5 -o ServerAliveInterval=60”，有关ssh命令行选项可以参考链接：ssh命令行链接

配置远程模式(ssh免密互通)

配置ssh免密互通基本操作：
[backup@backup_host] ssh-copy-id postgres@db_host
[postgres@db_host] ssh-copy-id backup@backup_host
backup_host 备份服务器
backup 备份服务器上的操作系统用户
db_host 数据库实例所在主机
postgres 数据库实例所在主机上的操作系统用户

 # 分别在本地虚拟机192.168.76.129和192.168.76.128执行如下操作完成ssh免密互通
 #ssh-keygen -t rsa -P "" -f ~/.ssh/id_rsa
 #ssh-copy-id lightdb@192.168.76.129 # 128免密登录129配置
 # 测试128环境的lightdb用户可以ssh免密登录129
 [lightdb@lightdb_128 scripts]$ ssh 'lightdb@192.168.76.129'
Last login: Wed Jul 13 13:17:18 2022 from 192.168.76.3
　　／＼／＼
　 (_人｜人_)
　　 ／‥＼
　 ミ(_Ｙ_)ミ
　　 ＞　＜
　　(／　＼)
　 _(　　　)_
　(＿＞―＜＿)

没有bug
[lightdb@lightdb_backup_129 ~]$ 
 # 测试129环境的lightdb用户可以ssh免密登录128
 [lightdb@lightdb_backup_129 ~]$ ssh 'lightdb@192.168.76.128'
Last login: Wed Jul 13 09:37:59 2022 from 192.168.76.3

 　　　＿＿＿
彡　 _／ ⌒ ⌒
　　/)) (●) (●)ヽ
　 ｜∩　(_人_)　 |
　 / ノ､＿ヽノ＿ノ￣)
　/ /　　　　　 /ﾌ_/
　L＿／＼　　　 ＼(
没有bug
[lightdb@lightdb_128 ~]$

使用ssh在远程模式下运行lt_probackup工作流程如下：
本地备份机器：192.168.76.129
远程机器：192.168.76.128

在远程模式下只能启动命令：添加实例(add-instance)、备份(backup)、恢复(restore)、追加(catchup)、归档推送(archive-push)和归档获取(archjve-get)。
在远程模式下在本地备份机器(192.168.76.129)和远程机器(192.168.76.128)上安装lt_probackup并且lt_probackup的版本必须相同。
远程模式下启动时，本地备份机器(192.168.76.129)上的lt_probackup进程通过 SSH 连接到远程机器(192.168.76.128)，并在远程机器(192.168.76.128)上启动一个或多个代理进程，这些进程称为远程代理。远程代理的数量来自配置项 -j/–threads
本地备份服务器(192.168.76.129)上的lt_probackup进程使用远程代理访问远程文件并在本地备份机器(192.168.76.129)和远程机器(192.168.76.128)之间传输数据。
远程机器(192.168.76.128)上远程代理尝试最小化网络流量和主机之间(192.168.76.128与192.168.76.129)的往返次数。
主进程通常在备份服务器(192.168.76.129)启动并连接到数据库实例所在主机(192.168.76.128)，但在发生归档推送(archive-push)和归档获取(archive-get)命令的情况下，主进程在数据库实例所在服务器(192.168.76.128)启动并连接到备份服务器(192.168.76.129)。
数据传输完成后，远程代理将终止，SSH 连接将关闭。
如果远程代理遇到错误情况，则所有代理都将终止，并且主lt_probackup进程将报告错误详细信息，该进程将退出并显示错误。
压缩总是在数据库实例所在机器上完成，而解压缩总是在备份服务器上完成。