DB2备份+日志恢复并前滚数据库
任何数据库都是这样的，我们会做数据库的全量备份，增量备份，并记录日志。如果数据库出现意外宕机，则可以用全备+增量备份+日志来进行数据库恢复。
本文主要展示如何使用备份+日志来恢复DB2的数据库，为了更清晰一点儿，使用两台机器151和152；在152机器上利用备份和日志恢复到151的数据库的状态。
151机器上数据操作步骤
1.创建测试库
#1.创建测试数据库
db2 create database yynewlog
#2.启用归档日志
db2 update db cfg for yynewlog using logarchmeth1 DISK:/data/db2data
#3.启用增量备份
db2 update db cfg for yynewlog using TRACKMOD YES
#3.解除backup pending状态
db2 backup database yynewlog
#4.创建测试表
db2 connect to yynewlog
db2 "create table student (id int not null,
name varchar(20),
sex smallint,
phone varchar(12),
address varchar(30)
)"
#5.填充数据：id=1,id=2
db2 "insert into student values(1,'charles ',0,'19089076756','BEIJING ')"
db2 "insert into student values(2,'zhouzhou ',0,'1076758886','BEIJING ')"2.全量备份
使用在线备份数据库，with include logs保留日志，这个选项好像是默认的，但是为了语句的可读理解，最好还是加上。
[db2inst1@db22 ~]$ db2 backup db yynewlog online to /tmp include logs
Backup successful. The timestamp for this backup image is : 20171030175924
/tmp目录下有了名为YYNEWLOG.0.db2inst1.DBPART000.20171030175924.001的全量备份，全量备份里面是id=1,id=2的记录。
3.增量备份
#新插入id=3,id=4的数据。
db2 "insert into student values(3,'xiaohong',0,'67583886','TAIYUAN')"
db2 "insert into student values(4,'lianliang',0,'13667583886','SHIJIAZHAUNG')"
#增量备份
[db2inst1@db22 ~]$ db2 backup db yynewlog ONLINE INCREMENTAL to /tmp include logs
Backup successful. The timestamp for this backup image is : 20171030180235
/tmp目录下有了名为YYNEWLOG.0.db2inst1.DBPART000.20171030180235.001的备份文件。
全量备份+增量备份里面是id=1，id=2,id=3,id=4的数据。
语法
>--+------------------------+----------------------------------->
'-INCREMENTAL--+-------+-'
'-DELTA-' - INCREMENTAL
Specifies a cumulative (incremental) backup image. An incremental backup image is a copy of all database data that has changed since the most recent successful, full backup operation.
指定一个累积的(增量)备份。一个INCREMENTAL备份镜像是一个 从 距离最近的成功备份的全量备份 之后 所有改变的数据的副本。
- DELTA
Specifies a noncumulative (delta) backup image. A delta backup image is a copy of all database data that has changed since the most recent successful backup operation of any type.
指定一个非 累积的(delta:差值)。一个delta备份镜像是一个 自动最近的一次成功的任何类型的备份操作之后 所有改变的数据库数据的副本。
- DELTA
4.增量delta备份
db2 "insert into student values(5,'shuang',0,'15067583886','SHENZHEN')"
db2 "insert into student values(6,'licy',0,'15683886675','XIAMEN')"
[db2inst1@db22 ~]$ db2 backup db yynewlog ONLINE INCREMENTAL DELTA to /tmp include logs
Backup successful. The timestamp for this backup image is : 201710301819035.活动日志
db2 "insert into student values(7,'shuang',0,'15067583886','SHENZHEN')"
db2 "insert into student values(8,'licy',0,'15683886675','XIAMEN')"
此时数据如下：
[db2inst1@db22 ~]$ db2 "select * from student"
ID NAME SEX PHONE ADDRESS
----------- -------------------- ------ ------------ ------------------------------
1 charles 0 19089076756 BEIJING
2 zhouzhou 0 1076758886 BEIJING
3 xiaohong 0 67583886 TAIYUAN
4 lianliang 0 13667583886 SHIJIAZHAUNG
5 shuang 0 15067583886 SHENZHEN
6 licy 0 15683886675 XIAMEN
7 shuang 0 15067583886 SHENZHEN
8 licy 0 15683886675 XIAMEN
8 record(s) selected.
152机器上的恢复步骤
1.拷贝文件到目标机器
scp /tmp/YYNEWLOG.* db2inst1@192.168.225.152:/tmp
将拷贝YYNEWLOG.0.db2inst1.DBPART000.20171030175924.001 ,
YYNEWLOG.0.db2inst1.DBPART000.20171030180235.001 ,
YYNEWLOG.0.db2inst1.DBPART000.20171030181903.001 这三个文件。
2.备份恢复步骤：
三个时间戳
20171030175924:全备
20171030180235：增量1
20171030181903：差异增量2
#1.指定通过增量备份恢复的形式最后要恢复到 20171030181903这个是时间戳。
[db2inst1@db22 ~]$ db2 restore db YYNEWLOG INCREMENTAL from /tmp taken at 20171030181903
DB20000I The RESTORE DATABASE command completed successfully.
#2.第一步恢复全备
[db2inst1@db22 ~]$ db2 restore db YYNEWLOG INCREMENTAL from /tmp taken at 20171030175924 logtarget /data/db2data/logs
DB20000I The RESTORE DATABASE command completed successfully.
#3.第二步恢复备份1：
[db2inst1@db22 ~]$ db2 restore db YYNEWLOG INCREMENTAL from /tmp taken at 20171030180235 logtarget /data/db2data/logs
SQL2580W Warning! Restoring logs to a path which contains existing log files.
Attempting to overwrite an existing log file during restore will cause the
restore operation to fail.
Do you want to continue ? (y/n) y
DB20000I The RESTORE DATABASE command completed successfully.
#4.第三步恢复差异增量备份2：
[db2inst1@db22 ~]$ db2 restore db YYNEWLOG INCREMENTAL from /tmp taken at 20171030181903 logtarget /data/db2data/logs
SQL2580W Warning! Restoring logs to a path which contains existing log files.
Attempting to overwrite an existing log file during restore will cause the
restore operation to fail.
Do you want to continue ? (y/n) y
DB20000I The RESTORE DATABASE command completed successfully.此时，logtarget下有前滚所需的日志：
[root@db22 logs]# ls NODE0000/LOGSTREAM0000/
S0000007.LOG S0000010.LOG S0000013.LOG3.确定所需的活动日志：
3.1 151机器上的活动日志：
[db2inst1@db22 ~]$ db2 get db cfg for yynewlog | grep -i 'First active log'
First active log file = S0000015.LOG3.1 152机器上的活动日志：
[db2inst1@db22 ~]$ db2 get db cfg for yynewlog | grep -i 'First active log'
First active log file = S0000013.LOG3.3 确定所需的日志
所以恢复的话是需要S0000014.LOG和S0000015.LOG的，151上的第一个活动日志是S0000015.LOG，故S0000014.LOG已经归档了，要把该日志拷贝到归档日志的目录中（恢复时的overflow log path
中，在此我们指定overflow log path为/data/db2data/logs下，刚刚restore时的logtarget）：
3.3.1 拷贝归档日志
#先找一下归档日志路径
[root@db22 tmp]# find / -name S0000014.LOG
/db/log/db2inst1/TESTMOVE/NODE0000/LOGSTREAM0000/C0000000/S0000014.LOG
/data/bak/db2inst1/SAMPLE/NODE0000/LOGSTREAM0000/C0000001/S0000014.LOG
/data/db2data/db2inst1/SCHOOL/NODE0000/LOGSTREAM0000/C0000000/S0000014.LOG
/data/db2data/db2inst1/YYNEWLOG/NODE0000/LOGSTREAM0000/C0000000/S0000014.LOG
/home/db2inst1/db2inst1/NODE0000/SQL00003/LOGSTREAM0000/S0000014.LOG
/home/db2inst1/db2inst1/NODE0000/SQL00002/LOGSTREAM0000/S0000014.LOG
#拷贝归档日志：
scp /data/db2data/db2inst1/YYNEWLOG/NODE0000/LOGSTREAM0000/C0000000/S0000014.LOG db2inst1@192.168.225.152:/data/db2data/logs/NODE0000/LOGSTREAM0000
#overflow log path 下有该日志了
[root@db22 logs]# ls NODE0000/LOGSTREAM0000/
S0000007.LOG S0000010.LOG S0000013.LOG S0000014.LOG
3.3.2 拷贝归档日志
#152机器上的活动日志目录
[db2inst1@db22 ~]$ db2 get db cfg for yynewlog | grep -i 'log'
Changed path to log files (NEWLOGPATH) =
Path to log files = /home/db2inst1/db2inst1/NODE0000/SQL00003/LOGSTREAM0000/
Overflow log path (OVERFLOWLOGPATH) =
Mirror log path (MIRRORLOGPATH) =
所以152机器上的活动日志目录是：/home/db2inst1/db2inst1/NODE0000/SQL00003/LOGSTREAM0000/
拷贝活动日志
[root@db22 tmp]# scp -r /home/db2inst1/db2inst1/NODE0000/SQL00005/LOGSTREAM0000 [email protected]:/home/db2inst1/db2inst1/NODE0000/[email protected]'s password:
S0000024.LOG 100% 4104KB 4.0MB/s 00:00
S0000020.LOG 100% 4104KB 4.0MB/s 00:00
S0000022.LOG 100% 4104KB 4.0MB/s 00:01
S0000015.LOG 100% 4104KB 4.0MB/s 00:00
S0000017.LOG 100% 4104KB 4.0MB/s 00:00
S0000026.LOG 100% 4104KB 4.0MB/s 00:01
S0000023.LOG 100% 4104KB 4.0MB/s 00:00
S0000027.LOG 100% 4104KB 4.0MB/s 00:01
S0000021.LOG 100% 4104KB 4.0MB/s 00:00
S0000019.LOG 100% 4104KB 4.0MB/s 00:00
S0000016.LOG 100% 4104KB 4.0MB/s 00:01
S0000025.LOG 100% 4104KB 4.0MB/s 00:01
S0000018.LOG 100% 4104KB 4.0MB/s 00:01
4.前滚恢复
[db2inst1@db22 ~]$ db2 "rollforward db yynewlog to end of logs overflow log path (/data/db2data/logs)"
Rollforward Status
Input database alias = yynewlog
Number of members have returned status = 1
Member ID = 0
Rollforward status = DB working
Next log file to be read = S0000016.LOG
Log files processed = S0000013.LOG - S0000014.LOG
Last committed transaction = 2017-10-31-08.43.15.000000 Local
DB20000I The ROLLFORWARD command completed successfully.
[db2inst1@db22 ~]$ db2 "rollforward db yynewlog complete"
Rollforward Status
Input database alias = yynewlog
Number of members have returned status = 1
Member ID = 0
Rollforward status = not pending
Next log file to be read =
Log files processed = S0000013.LOG - S0000014.LOG
Last committed transaction = 2017-10-31-08.43.15.000000 Local
DB20000I The ROLLFORWARD command completed successfully.
5.校验
[db2inst1@db22 ~]$ db2 "connect to yynewlog"
Database Connection Information
Database server = DB2/LINUXX8664 11.1.1.1
SQL authorization ID = DB2INST1
Local database alias = YYNEWLOG
[db2inst1@db22 ~]$ db2 "select * from student"
ID NAME SEX PHONE ADDRESS
----------- -------------------- ------ ------------ ------------------------------
1 charles 0 19089076756 BEIJING
2 zhouzhou 0 1076758886 BEIJING
3 xiaohong 0 67583886 TAIYUAN
4 lianliang 0 13667583886 SHIJIAZHAUNG
5 shuang 0 15067583886 SHENZHEN
6 licy 0 15683886675 XIAMEN
7 shuang 0 15067583886 SHENZHEN
8 licy 0 15683886675 XIAMEN
6.成功
7.疑问解析
7.1 Log files processed = S0000013.LOG - S0000014.LOG，没有15.LOG；是否并不需要活动日志？
前滚恢复返回：
Next log file to be read = S0000016.LOG
Log files processed = S0000013.LOG - S0000014.LOG不拷贝活动日志的情况下来试试恢复会如何
db2 restore db YYNEWLOG INCREMENTAL from /tmp taken at 20171030181903
db2 restore db YYNEWLOG INCREMENTAL from /tmp taken at 20171030175924
db2 restore db YYNEWLOG INCREMENTAL from /tmp taken at 20171030180235
db2 restore db YYNEWLOG INCREMENTAL from /tmp taken at 20171030181903 logtarget /data/db2data/logs
#拷贝14.LOG日
[root@db22 logs]# ls NODE0000/LOGSTREAM0000/
S0000013.LOG S0000014.LOG
[root@db22 logs]# ls /home/db2inst1/db2inst1/NODE0000/SQL00003/LOGSTREAM0000/
[root@db22 logs]#
#前滚
[db2inst1@db22 ~]$ db2 "rollforward db yynewlog to end of logs overflow log path (/data/db2data/logs)"
Rollforward Status
Input database alias = yynewlog
Number of members have returned status = 1
Member ID = 0
Rollforward status = DB working
Next log file to be read = S0000015.LOG
Log files processed = S0000013.LOG - S0000013.LOG
Last committed transaction = 2017-10-30-10.19.11.000000 UTC
DB20000I The ROLLFORWARD command completed successfully.
[db2inst1@db22 ~]$ db2 "rollforward db yynewlog complete"
SQL1273N An operation reading the logs on database "YYNEWLOG" cannot continue
because of a missing log file "S0000014.LOG" on database partition "0" and log
stream "0".
由上面的实验可以看到我们溢出目录中是有14.LOG文件的，但是却报找不到日志文件的错误
将活动日志拷贝过来之后：
[db2inst1@db22 ~]$ db2 "rollforward db yynewlog complete"
SQL1273N An operation reading the logs on database "YYNEWLOG" cannot continue
because of a missing log file "S0000014.LOG" on database partition "0" and log
stream "0".
[db2inst1@db22 ~]$ db2 "rollforward db yynewlog to end of logs overflow log path (/data/db2data/logs)"
Rollforward Status
Input database alias = yynewlog
Number of members have returned status = 1
Member ID = 0
Rollforward status = DB working
Next log file to be read = S0000016.LOG
Log files processed = S0000013.LOG - S0000014.LOG
Last committed transaction = 2017-10-31-02.03.51.000000 UTC
DB20000I The ROLLFORWARD command completed successfully.
[db2inst1@db22 ~]$ db2 "rollforward db yynewlog complete"
Rollforward Status
Input database alias = yynewlog
Number of members have returned status = 1
Member ID = 0
Rollforward status = not pending
Next log file to be read =
Log files processed = S0000013.LOG - S0000014.LOG
Last committed transaction = 2017-10-31-02.03.51.000000 UTC
DB20000I The ROLLFORWARD command completed successfully.
将活动日志拷贝过来之后再rollforward就可以成功了。
7.2 我只想恢复id =7 的数据，不想要id=8的数据怎么办？重点怎么确定时间戳？
db2 rollforward 命令时读取日志的顺序
数据库备份恢复的一点总结