达梦数据守护（单主备）集群实践

最新推荐文章于 2023-08-07 23:40:20 发布

ora_ckpt_Lu

最新推荐文章于 2023-08-07 23:40:20 发布

阅读量464

点赞数

本文链接：https://blog.csdn.net/weixin_42145355/article/details/109722631

版权

数据库集群故障切换备库重建自动切换联机备份

关键词由CSDN通过智能技术生成

1.0自杀实验

故障自动切换模式下，必须配置确认监视器，且确认监视器最多只能配置一个。

#关闭主库服务器
[root@Centos7-STD root]# reboot

查看监视器

#捕捉到主库异常
[monitor]         2020-11-13 15:22:52: Received message timeout from(GRP1_PR)
                  WTIME                WSTATUS        INST_OK   INAME            ISTATUS     IMODE     RSTAT    N_OPEN   FLSN            CLSN            
                  2020-11-13 02:22:43  ERROR          OK        GRP1_PR          OPEN        PRIMARY   VALID    7        45393           45393

开始自动切换主备库角色

[monitor]         2020-11-13 15:22:52: Check primary instance error in group(GRP1), start to auto takeover

[monitor]         2020-11-13 15:22:52: Notify group(GRP1)'s active dmwatcher to set MID
[monitor]         2020-11-13 15:22:52: Notify group(GRP1)'s active dmwatcher to set MID success
[monitor]         2020-11-13 15:22:52: Start to takeover use instance GRP1_SD
[monitor]         2020-11-13 15:22:52: Notify dmwatcher(GRP1_SD) switch to TAKEOVER status
[monitor]         2020-11-13 15:22:52: Dmwatcher process GRP1_SD status switching [OPEN-->TAKEOVER] 
[monitor]         2020-11-13 15:22:52: Switch dmwatcher GRP1_SD to TAKEOVER status success
[monitor]         2020-11-13 15:22:52: Instance GRP1_SD start to execute sql SP_SET_GLOBAL_DW_STATUS(0, 7)
[monitor]         2020-11-13 15:22:52: Instance GRP1_SD execute sql SP_SET_GLOBAL_DW_STATUS(0, 7) success
[monitor]         2020-11-13 15:22:52: Instance GRP1_SD start to execute sql SP_APPLY_KEEP_PKG()
[monitor]         2020-11-13 15:22:52: Instance GRP1_SD execute sql SP_APPLY_KEEP_PKG() success
[monitor]         2020-11-13 15:22:52: Instance GRP1_SD start to execute sql ALTER DATABASE MOUNT
[monitor]         2020-11-13 15:22:54: Instance GRP1_SD execute sql ALTER DATABASE MOUNT success
[monitor]         2020-11-13 15:22:54: Instance GRP1_SD start to execute sql ALTER DATABASE PRIMARY
[monitor]         2020-11-13 15:22:54: Instance GRP1_SD execute sql ALTER DATABASE PRIMARY success
[monitor]         2020-11-13 15:22:57: Notify instance GRP1_SD to change all arch status to be invalid
[monitor]         2020-11-13 15:22:57: Succeed to change all instances arch status to be invalid
[monitor]         2020-11-13 15:22:57: Instance GRP1_SD start to execute sql ALTER DATABASE OPEN FORCE
[monitor]         2020-11-13 15:22:59: Instance GRP1_SD execute sql ALTER DATABASE OPEN FORCE success
[monitor]         2020-11-13 15:22:59: Instance GRP1_SD start to execute sql SP_SET_GLOBAL_DW_STATUS(7, 0)
[monitor]         2020-11-13 15:22:59: Instance GRP1_SD execute sql SP_SET_GLOBAL_DW_STATUS(7, 0) success
[monitor]         2020-11-13 15:22:59: Notify dmwatcher(GRP1_SD) switch to OPEN status
[monitor]         2020-11-13 15:22:59: Dmwatcher process GRP1_SD status switching [TAKEOVER-->OPEN] 
[monitor]         2020-11-13 15:22:59: Switch dmwatcher GRP1_SD to OPEN status success
[monitor]         2020-11-13 15:22:59: Notify group(GRP1)'s dmwatcher to do clear
[monitor]         2020-11-13 15:22:59: Clean request of dmwatcher processer GRP1_SD success
[monitor]         2020-11-13 15:22:59: Success to takeover use instance GRP1_SD

[monitor]         2020-11-13 15:22:59: Group(GRP1) use instance GRP1_SD auto takeover success

主库节点重启成功后再次以备库角色加入集群

[monitor]         2020-11-13 15:23:26: Dmwatcher process GRP1_PR status switching [NONE-->STARTUP] 
[monitor]         2020-11-13 15:23:45: Dmwatcher process GRP1_PR status switching [STARTUP-->UNIFY EP] 
                  WTIME                WSTATUS        INST_OK   INAME            ISTATUS     IMODE     RSTAT    N_OPEN   FLSN            CLSN            
                  2020-11-13 02:23:47  UNIFY EP       OK        GRP1_PR          MOUNT       PRIMARY   VALID    7        45624           45624           

[monitor]         2020-11-13 15:23:45: Dmwatcher process GRP1_PR status switching [UNIFY EP-->STARTUP] 
                  WTIME                WSTATUS        INST_OK   INAME            ISTATUS     IMODE     RSTAT    N_OPEN   FLSN            CLSN            
                  2020-11-13 02:23:48  STARTUP        OK        GRP1_PR          MOUNT       STANDBY   INVALID  7        45624           45624           

[monitor]         2020-11-13 15:23:45: Dmwatcher process GRP1_PR status switching [STARTUP-->UNIFY EP] 
                  WTIME                WSTATUS        INST_OK   INAME            ISTATUS     IMODE     RSTAT    N_OPEN   FLSN            CLSN            
                  2020-11-13 02:23:48  UNIFY EP       OK        GRP1_PR          MOUNT       STANDBY   INVALID  7        45624           45624           

[monitor]         2020-11-13 15:23:45: Dmwatcher process GRP1_PR status switching [UNIFY EP-->STARTUP] 
                  WTIME                WSTATUS        INST_OK   INAME            ISTATUS     IMODE     RSTAT    N_OPEN   FLSN            CLSN            
                  2020-11-13 02:23:48  STARTUP        OK        GRP1_PR          OPEN        STANDBY   INVALID  7        45624           45624           

[monitor]         2020-11-13 15:23:46: Dmwatcher process GRP1_PR status switching [STARTUP-->OPEN] 
                  WTIME                WSTATUS        INST_OK   INAME            ISTATUS     IMODE     RSTAT    N_OPEN   FLSN            CLSN            
                  2020-11-13 02:23:48  OPEN           OK        GRP1_PR          OPEN        STANDBY   INVALID  7        45624           45624           

[monitor]         2020-11-13 15:23:48: Dmwatcher process GRP1_SD status switching [OPEN-->RECOVERY] 
                  WTIME                WSTATUS        INST_OK   INAME            ISTATUS     IMODE     RSTAT    N_OPEN   FLSN            CLSN            
                  2020-11-13 15:23:48  RECOVERY       OK        GRP1_SD          OPEN        PRIMARY   VALID    8        46880           46880           

[monitor]         2020-11-13 15:23:52: Dmwatcher process GRP1_SD status switching [RECOVERY-->OPEN] 
                  WTIME                WSTATUS        INST_OK   INAME            ISTATUS     IMODE     RSTAT    N_OPEN   FLSN            CLSN            
                  2020-11-13 15:23:52  OPEN           OK        GRP1_SD          OPEN        PRIMARY   VALID    8        46880           46880

查看结果

GROUP            OGUID       MON_CONFIRM     MODE            MPP_FLAG  
GRP1             453332      TRUE            AUTO            FALSE     


<<DATABASE GLOBAL INFO:>>
IP              MAL_DW_PORT  WTIME                WTYPE     WCTLSTAT  WSTATUS        INAME            INST_OK   N_EP  N_OK  ISTATUS     IMODE     DSC_STATUS     RTYPE     RSTAT    
10.10.3.104     52142        2020-11-13 15:26:39  GLOBAL    VALID     OPEN           GRP1_SD          OK        1     1     OPEN        PRIMARY   DSC_OPEN       REALTIME  VALID    

EP INFO:
INST_PORT  INST_OK   INAME            ISTATUS     IMODE     DSC_SEQNO  DSC_CTL_NODE RTYPE     RSTAT    FSEQ            FLSN            CSEQ            CLSN            DW_STAT_FLAG          
32142      OK        GRP1_SD          OPEN        PRIMARY   0          0            REALTIME  VALID    3942            46880           3942            46880           NONE                  

<<DATABASE GLOBAL INFO:>>
IP              MAL_DW_PORT  WTIME                WTYPE     WCTLSTAT  WSTATUS        INAME            INST_OK   N_EP  N_OK  ISTATUS     IMODE     DSC_STATUS     RTYPE     RSTAT    
10.10.3.103     52141        2020-11-13 02:26:41  GLOBAL    VALID     OPEN           GRP1_PR          OK        1     1     OPEN        STANDBY   DSC_OPEN       REALTIME  VALID    

EP INFO:
INST_PORT  INST_OK   INAME            ISTATUS     IMODE     DSC_SEQNO  DSC_CTL_NODE RTYPE     RSTAT    FSEQ            FLSN            CSEQ            CLSN            DW_STAT_FLAG          
32141      OK        GRP1_PR          OPEN        STANDBY   0          0            REALTIME  VALID    3941            46880           3941            46880           NONE                  

DATABASE(GRP1_PR) APPLY INFO FROM (GRP1_SD):
DSC_SEQNO[0], (ASEQ, SSEQ, KSEQ)[3942, 3942, 3942], (ALSN, SLSN, KLSN)[46880, 46880, 46880], N_TSK[0], TSK_MEM_USE[0]

1.1手工切换

手动切换刚才的重启节点为主库

#先登陆

login
username:
password:
[monitor]         2020-11-13 15:27:22: Login dmmonitor success!

#手动switchover

switchover grp1.grp1_pr

#过程

[monitor]         2020-11-13 15:27:50: Start to switchover instance GRP1_PR
[monitor]         2020-11-13 15:27:50: Notify dmwatcher(GRP1_SD) switch to SWITCHOVER status
[monitor]         2020-11-13 15:27:50: Dmwatcher process GRP1_SD status switching [OPEN-->SWITCHOVER] 
[monitor]         2020-11-13 15:27:50: Switch dmwatcher GRP1_SD to SWITCHOVER status success
[monitor]         2020-11-13 15:27:50: Notify dmwatcher(GRP1_PR) switch to SWITCHOVER status
[monitor]         2020-11-13 15:27:50: Dmwatcher process GRP1_PR status switching [OPEN-->SWITCHOVER] 
[monitor]         2020-11-13 15:27:50: Switch dmwatcher GRP1_PR to SWITCHOVER status success
[monitor]         2020-11-13 15:27:50: Instance GRP1_SD start to execute sql SP_SET_GLOBAL_DW_STATUS(0, 6)
[monitor]         2020-11-13 15:27:50: Instance GRP1_SD execute sql SP_SET_GLOBAL_DW_STATUS(0, 6) success
[monitor]         2020-11-13 15:27:50: Instance GRP1_PR start to execute sql SP_SET_GLOBAL_DW_STATUS(0, 6)
[monitor]         2020-11-13 15:27:50: Instance GRP1_PR execute sql SP_SET_GLOBAL_DW_STATUS(0, 6) success
[monitor]         2020-11-13 15:27:50: Instance GRP1_SD start to execute sql ALTER DATABASE MOUNT
[monitor]         2020-11-13 15:27:52: Instance GRP1_SD execute sql ALTER DATABASE MOUNT success
[monitor]         2020-11-13 15:27:52: Instance GRP1_PR start to execute sql SP_APPLY_KEEP_PKG()
[monitor]         2020-11-13 15:27:52: Instance GRP1_PR execute sql SP_APPLY_KEEP_PKG() success
[monitor]         2020-11-13 15:27:52: Instance GRP1_PR start to execute sql ALTER DATABASE MOUNT
[monitor]         2020-11-13 15:27:54: Instance GRP1_PR execute sql ALTER DATABASE MOUNT success
[monitor]         2020-11-13 15:27:54: Instance GRP1_SD start to execute sql ALTER DATABASE STANDBY
[monitor]         2020-11-13 15:27:54: Instance GRP1_SD execute sql ALTER DATABASE STANDBY success
[monitor]         2020-11-13 15:27:54: Instance GRP1_PR start to execute sql ALTER DATABASE PRIMARY
[monitor]         2020-11-13 15:27:54: Instance GRP1_PR execute sql ALTER DATABASE PRIMARY success
[monitor]         2020-11-13 15:27:54: Notify instance GRP1_PR to change all arch status to be invalid
[monitor]         2020-11-13 15:27:54: Succeed to change all instances arch status to be invalid
[monitor]         2020-11-13 15:27:54: Instance GRP1_SD start to execute sql ALTER DATABASE OPEN FORCE
[monitor]         2020-11-13 15:27:54: Instance GRP1_SD execute sql ALTER DATABASE OPEN FORCE success
[monitor]         2020-11-13 15:27:54: Instance GRP1_PR start to execute sql ALTER DATABASE OPEN FORCE
[monitor]         2020-11-13 15:27:56: Instance GRP1_PR execute sql ALTER DATABASE OPEN FORCE success
[monitor]         2020-11-13 15:27:56: Instance GRP1_SD start to execute sql SP_SET_GLOBAL_DW_STATUS(6, 0)
[monitor]         2020-11-13 15:27:56: Instance GRP1_SD execute sql SP_SET_GLOBAL_DW_STATUS(6, 0) success
[monitor]         2020-11-13 15:27:56: Instance GRP1_PR start to execute sql SP_SET_GLOBAL_DW_STATUS(6, 0)
[monitor]         2020-11-13 15:27:56: Instance GRP1_PR execute sql SP_SET_GLOBAL_DW_STATUS(6, 0) success
[monitor]         2020-11-13 15:27:56: Notify dmwatcher(GRP1_SD) switch to OPEN status
[monitor]         2020-11-13 15:27:56: Dmwatcher process GRP1_SD status switching [SWITCHOVER-->OPEN] 
[monitor]         2020-11-13 15:27:56: Switch dmwatcher GRP1_SD to OPEN status success
[monitor]         2020-11-13 15:27:56: Notify dmwatcher(GRP1_PR) switch to OPEN status
[monitor]         2020-11-13 15:27:56: Dmwatcher process GRP1_PR status switching [SWITCHOVER-->OPEN] 
[monitor]         2020-11-13 15:27:57: Switch dmwatcher GRP1_PR to OPEN status success
[monitor]         2020-11-13 15:27:57: Notify group(GRP1)'s dmwatcher to do clear
[monitor]         2020-11-13 15:27:57: Clean request of dmwatcher processer GRP1_PR success
2020-11-13 15:27:57 
#================================================================================#
GROUP            OGUID       MON_CONFIRM     MODE            MPP_FLAG  
GRP1             453332      TRUE            AUTO            FALSE     


<<DATABASE GLOBAL INFO:>>
IP              MAL_DW_PORT  WTIME                WTYPE     WCTLSTAT  WSTATUS        INAME            INST_OK   N_EP  N_OK  ISTATUS     IMODE     DSC_STATUS     RTYPE     RSTAT    
10.10.3.103     52141        2020-11-13 02:27:59  GLOBAL    VALID     OPEN           GRP1_PR          OK        1     1     OPEN        PRIMARY   DSC_OPEN       REALTIME  VALID    

EP INFO:
INST_PORT  INST_OK   INAME            ISTATUS     IMODE     DSC_SEQNO  DSC_CTL_NODE RTYPE     RSTAT    FSEQ            FLSN            CSEQ            CLSN            DW_STAT_FLAG          
32141      OK        GRP1_PR          OPEN        PRIMARY   0          0            REALTIME  VALID    3943            46880           3943            48238           NONE                  

<<DATABASE GLOBAL INFO:>>
IP              MAL_DW_PORT  WTIME                WTYPE     WCTLSTAT  WSTATUS        INAME            INST_OK   N_EP  N_OK  ISTATUS     IMODE     DSC_STATUS     RTYPE     RSTAT    
10.10.3.104     52142        2020-11-13 15:27:57  GLOBAL    VALID     OPEN           GRP1_SD          OK        1     1     OPEN        STANDBY   DSC_OPEN       REALTIME  INVALID  

EP INFO:
INST_PORT  INST_OK   INAME            ISTATUS     IMODE     DSC_SEQNO  DSC_CTL_NODE RTYPE     RSTAT    FSEQ            FLSN            CSEQ            CLSN            DW_STAT_FLAG          
32142      OK        GRP1_SD          OPEN        STANDBY   0          0            REALTIME  INVALID  3943            46880           3943            46880           NONE                  

DATABASE(GRP1_SD) APPLY INFO FROM (GRP1_PR):
DSC_SEQNO[0], (ASEQ, SSEQ, KSEQ)[3943, 3943, 3943], (ALSN, SLSN, KLSN)[46880, 46880, 46880], N_TSK[0], TSK_MEM_USE[0]

结果查看

GROUP            OGUID       MON_CONFIRM     MODE            MPP_FLAG  
GRP1             453332      TRUE            AUTO            FALSE     


<<DATABASE GLOBAL INFO:>>
IP              MAL_DW_PORT  WTIME                WTYPE     WCTLSTAT  WSTATUS        INAME            INST_OK   N_EP  N_OK  ISTATUS     IMODE     DSC_STATUS     RTYPE     RSTAT    
10.10.3.103     52141        2020-11-13 02:28:40  GLOBAL    VALID     OPEN           GRP1_PR          OK        1     1     OPEN        PRIMARY   DSC_OPEN       REALTIME  VALID    

EP INFO:
INST_PORT  INST_OK   INAME            ISTATUS     IMODE     DSC_SEQNO  DSC_CTL_NODE RTYPE     RSTAT    FSEQ            FLSN            CSEQ            CLSN            DW_STAT_FLAG          
32141      OK        GRP1_PR          OPEN        PRIMARY   0          0            REALTIME  VALID    3944            48238           3944            48238           NONE                  

<<DATABASE GLOBAL INFO:>>
IP              MAL_DW_PORT  WTIME                WTYPE     WCTLSTAT  WSTATUS        INAME            INST_OK   N_EP  N_OK  ISTATUS     IMODE     DSC_STATUS     RTYPE     RSTAT    
10.10.3.104     52142        2020-11-13 15:28:37  GLOBAL    VALID     OPEN           GRP1_SD          OK        1     1     OPEN        STANDBY   DSC_OPEN       REALTIME  VALID    

EP INFO:
INST_PORT  INST_OK   INAME            ISTATUS     IMODE     DSC_SEQNO  DSC_CTL_NODE RTYPE     RSTAT    FSEQ            FLSN            CSEQ            CLSN            DW_STAT_FLAG          
32142      OK        GRP1_SD          OPEN        STANDBY   0          0            REALTIME  VALID    3943            48238           3943            48238           NONE                  

DATABASE(GRP1_SD) APPLY INFO FROM (GRP1_PR):
DSC_SEQNO[0], (ASEQ, SSEQ, KSEQ)[3944, 3944, 3944], (ALSN, SLSN, KLSN)[48238, 48238, 48238], N_TSK[0], TSK_MEM_USE[0]

1.3备库重建

现实环境中，我们会遇到服务器硬件故障，或者误操作导致集群高可用环境失败。
由于本例中我们使用的是自动切换，当主库发生dang机后，备库会自动接管升级为主库，这样我们只需要单方面考虑备库重建的问题了。

模拟误删除环境

#直接删除备库数据目录
[root@Centos7-STD log]# rm -rf /data/dmdbs/data/sd/sd/

#这里删除后发现dmserver还在运行
[root@Centos7-STD log]# ps -ef|grep dmserver
dmdba     1258     1  0 15:32 ?        00:00:01 /data/dmdbs/bin/dmserver /data/dmdbs/data/sd/sd/dm.ini mount

这里由于我们的数据库还太小，删除文件后，相关的文件句柄依旧存在缓存中

[dmdba@Centos7-STD 1258]$ cd /proc/1258/fd
lrwx------ 1 dmdba dinstall 64 Nov 13 15:49 10 -> /data/dmdbs/data/sd/sd/sd01.log (deleted)
lrwx------ 1 dmdba dinstall 64 Nov 13 15:49 11 -> /data/dmdbs/data/sd/sd/sd02.log (deleted)
lrwx------ 1 dmdba dinstall 64 Nov 13 15:49 12 -> socket:[19173]
lrwx------ 1 dmdba dinstall 64 Nov 13 15:49 13 -> socket:[19174]
lrwx------ 1 dmdba dinstall 64 Nov 13 15:49 14 -> socket:[19176]
lrwx------ 1 dmdba dinstall 64 Nov 13 15:49 15 -> /data/dmdbs/data/sd/sd/ROLL.DBF (deleted)
lrwx------ 1 dmdba dinstall 64 Nov 13 15:49 16 -> /data/dmdbs/data/sd/sd/MAIN.DBF (deleted)

此时你会发现，进入主库作一下操作，备库依旧处于同步状态

#主库操作如下：
create user test identified by test123456789;
 create table test.t1(id int);
 insert into test.t1 values(1),(2),(3);
 commit;
 #备库进行查询：
 [dmdba@Centos7-STD fd]$ disql sysdba@localhost:32142
 SQL> select * from test.t1;

LINEID     ID         
---------- -----------
1          1
2          2
3          3

还是数据库太小，打开的文件句柄不会及时被关闭，这个时候只需要杀掉备库数据库进程，高可用环境就会出现问题。

杀掉残存备库进程

[dmdba@Centos7-STD fd]$ ps -ef|grep dmserver
dmdba     1258     1  0 15:32 ?        00:00:01 /data/dmdbs/bin/dmserver /data/dmdbs/data/sd/sd/dm.ini mount
dmdba     2716  2184  0 15:59 pts/0    00:00:00 grep --color=auto dmserver
[dmdba@Centos7-STD fd]$ kill -9 1258

如上，杀掉备库残存进程后，一般来说只要备库守护进程还在运行，就会自动拉起备库
而此时，备库环境已经被人为破坏，所以集群可能面临诸多报错。

[monitor]         2020-11-13 02:59:18: Instance GRP1_SD[STANDBY, OPEN, ISTAT_SAME:TRUE] error 
                  WTIME                WSTATUS        INST_OK   INAME            ISTATUS     IMODE     RSTAT    N_OPEN   FLSN            CLSN            
                  2020-11-13 15:59:15  STARTUP        ERROR     GRP1_SD          OPEN        STANDBY   VALID    9        48311           48311

最先发现的是dbmonitor，报告备库出现异常

2020-11-13 16:03:04.664 [INFO] dmwatcher P0000000988 T0000140654251452160  [!!! Local instance restarted by dmwatcher. start info is in dmserver log!!!]

其次，备库节点上的dbwatcher进程也报告如上，这里告诉我们本地实例被守护进程自动启动，详细的需要查看本地日志

进入到日志目录下，你会发现一个新的数据库实例日志dm_unknown_202011.log
这个应该是达梦默认不知名实例日志，由于我们直接删除整个数据目录，所以相应的日志被记录在这里

[dmdba@Centos7-STD log]$ tail dm_unknown_202011.log
2020-11-13 16:06:49.906 [WARNING] database P0000003138 T0000000000000003138  open ini file /data/dmdbs/data/sd/sd/dm.ini failed!
2020-11-13 16:06:49.906 [FATAL] database P0000003138 T0000000000000003138  dmserver startup failed, code = -104 [Invalid INI file]
2020-11-13 16:06:49.906 [FATAL] database P0000003138 T0000000000000003138  nsvr_ini_file_read failed, [code: -104]

很显然，这会儿备库已经无力回天，只能重建

重建前的准备和检查

通过dmmonitor中的tip命令，我们可以获得当前数据守护集群中出现的问题汇总

tip
[monitor]         2020-11-13 03:08:37: Instance GRP1_SD[STANDBY, OPEN, ISTAT_SAME:TRUE] has no command to execute currently
[monitor]         2020-11-13 03:08:37: Instance GRP1_SD[STANDBY, OPEN, ISTAT_SAME:TRUE] error, please wait dmwatcher auto restart it

[monitor]         2020-11-13 03:08:37: Group(GRP1) has PRIMARY&OPEN instance, but still exists other instances not OK, please choose appropriate processing according to the above information!

[monitor]         2020-11-13 03:08:37: All groups' have PRIMARY&OPEN instances, but there still exist instances not OK!

在这里插入图片描述

根据提示，目前主库状态正常，处于open，依旧可以对外提供服务。但是根据官方文档，主库在没有可用备库情况下，先是处于open状态
该状态可以查询，一旦有事务发生，这时候主库会自动挂起，不具备同步落盘日志这也是为了保证主备数据一致

重建备库

这里我们备库环境已经失败，不再采用以上办法强制退出主库，让主库业务继续办下去，至于备库，采用联机备份重建方法恢复回来即可

联机备份主库

[dmdba@centos7 ~]$ disql sysdba@localhost:32141
SQL> BACKUP DATABASE FULL BACKUPSET '/data/dmdbms/data/pr_bak_20201113.bak'; 
executed successfully
used time: 00:00:01.472. Execute id is 108.

重新初始化备库

[dmdba@Centos7-STD log]$ dminit path=/data/dmdbs/data/sd db_name=sd

#将备份文件拷贝至备份机器进行恢复
[dmdba@centos7 ~]$ scp -r /data/dmdbms/data/pr_bak_20201113.bak 192.168.3.104:/data/dmdbs/data/

备库恢复

dmrman
restore database '/data/dmdbs/data/sd/sd/dm.ini' from backupset '/data/dmdbs/data/pr_bak_20201113.bak';
recover database '/data/dmdbs/data/sd/sd/dm.ini' from backupset '/data/dmdbs/data/pr_bak_20201113.bak';
recover database '/data/dmdbs/data/sd/sd/dm.ini' update db_magic;

重新配置备库的dm.ini,dmarch.ini,dmmal.ini,dmwatcher.ini文件

vim /data/dmdbs/data/sd/sd/dm.ini
#从节点
INSTANCE_NAME = GRP1_SD
PORT_NUM = 32142
ALTER_MODE_STATUS  = 0        #不允许手工方式修改实例模式/状态/OGUID 
ENABLE_OFFLINE_TS  = 2        #不允许备库 OFFLINE 表空间 
MAL_INI     = 1  #打开 MAL 系统 
ARCH_INI    = 1  #打开归档配置 

vim /data/dmdbs/data/sd/sd/dmarch.ini
[ARCHIVE_REALTIME]
ARCH_TYPE     = REALTIME  #实时归档类型
ARCH_DEST     = GRP1_PR #实时归档目标实例名     
[ARCHIVE_LOCAL1]
ARCH_TYPE     = LOCAL #本地归档类型
ARCH_DEST     = /data/dmdbs/data/sd/sd/arch  #本地归档文件存放路径
ARCH_FILE_SIZE    = 128 #单位 Mb，本地单个归档文件最大值
ARCH_SPACE_LIMIT  = 0    #单位 Mb，0 表示无限制，范围 1024~4294967294M

mkdir -p /data/dmdbs/data/sd/sd/arch

vim /data/dmdbs/data/sd/sd/dmmal.ini
MAL_CHECK_INTERVAL   = 5  #MAL 链路检测时间间隔 
MAL_CONN_FAIL_INTERVAL  = 5  #判定 MAL 链路断开的时间 
[MAL_INST1] 
  MAL_INST_NAME          = GRP1_PR #实例名，和 dm.ini 中的 INSTANCE_NAME 一致 
  MAL_HOST               = 10.10.3.103  #MAL 系统监听 TCP 连接的 IP 地址 
  MAL_PORT               = 61141     #MAL 系统监听 TCP 连接的端口 
  MAL_INST_HOST          = 192.168.3.103  #实例的对外服务 IP 地址 
  MAL_INST_PORT          = 32141 #实例的对外服务端口，和 dm.ini 中的 PORT_NUM 一致 
  MAL_DW_PORT   = 52141 #实例本地的守护进程监听 TCP 连接的端口 
MAL_INST_DW_PORT  = 33141 #实例监听守护进程 TCP 连接的端口 
 
[MAL_INST2] 
  MAL_INST_NAME          = GRP1_SD 
  MAL_HOST               = 10.10.3.104
  MAL_PORT               = 61142 
  MAL_INST_HOST          = 192.168.3.104 
  MAL_INST_PORT          = 32142 
  MAL_DW_PORT   = 52142 
MAL_INST_DW_PORT  = 33142

vim /data/dmdbs/data/sd/sd/dmwatcher.ini
[GRP1]
 DW_TYPE    = GLOBAL    #全局守护类型
 DW_MODE    = AUTO   #自动切换模式
 DW_ERROR_TIME     = 10  #远程守护进程故障认定时间
 INST_RECOVER_TIME   = 60         #主库守护进程启动恢复的间隔时间
 INST_ERROR_TIME   = 10  #本地实例故障认定时间
 INST_OGUID         = 453332 #守护系统唯一 OGUID 值
 INST_INI           = /data/dmdbs/data/sd/sd/dm.ini #dm.ini 配置文件路径
 INST_AUTO_RESTART   = 1   #打开实例的自动启动功能
 INST_STARTUP_CMD   = /data/dmdbs/bin/dmserver #命令行方式启动
RLOG_APPLY_THRESHOLD  = 0         #指定备库重演日志的时间阈值，默认关闭

以mount方式启动备库

[dmdba@Centos7-STD log]$ dmserver /data/dmdbs/data/sd/sd/dm.ini mount

#Disql登陆备库，修改OGUID以及数据库模式
[dmdba@Centos7-STD ~]$ disql sysdba@localhost:32142
SP_SET_PARA_VALUE(1, 'ALTER_MODE_STATUS', 1); 
sp_set_oguid(453332); 
alter database standby; 
SP_SET_PARA_VALUE(1, 'ALTER_MODE_STATUS', 0);

启动备库上的dmwatcher进程

[dmdba@Centos7-STD ~]$ dmwatcher /data/dmdbs/data/sd/sd/dmwatcher.ini 

#等待备库自动加入
#================================================================================#

[monitor]         2020-11-13 04:15:54: Dmwatcher process GRP1_SD status switching [STARTUP-->UNIFY EP] 
                  WTIME                WSTATUS        INST_OK   INAME            ISTATUS     IMODE     RSTAT    N_OPEN   FLSN            CLSN            
                  2020-11-13 17:15:50  UNIFY EP       OK        GRP1_SD          MOUNT       STANDBY   INVALID  9        48323           48323           

[monitor]         2020-11-13 04:15:54: Dmwatcher process GRP1_SD status switching [UNIFY EP-->STARTUP] 
                  WTIME                WSTATUS        INST_OK   INAME            ISTATUS     IMODE     RSTAT    N_OPEN   FLSN            CLSN            
                  2020-11-13 17:15:50  STARTUP        OK        GRP1_SD          OPEN        STANDBY   INVALID  9        48323           48323           

[monitor]         2020-11-13 04:15:54: Dmwatcher process GRP1_SD status switching [STARTUP-->OPEN] 
                  WTIME                WSTATUS        INST_OK   INAME            ISTATUS     IMODE     RSTAT    N_OPEN   FLSN            CLSN            
                  2020-11-13 17:15:50  OPEN           OK        GRP1_SD          OPEN        STANDBY   INVALID  9        48323           48323           

[monitor]         2020-11-13 04:15:54: Dmwatcher process GRP1_PR status switching [OPEN-->RECOVERY] 
                  WTIME                WSTATUS        INST_OK   INAME            ISTATUS     IMODE     RSTAT    N_OPEN   FLSN            CLSN            
                  2020-11-13 04:15:54  RECOVERY       OK        GRP1_PR          OPEN        PRIMARY   VALID    9        48329           48329           

[monitor]         2020-11-13 04:15:56: Dmwatcher process GRP1_PR status switching [RECOVERY-->OPEN] 
                  WTIME                WSTATUS        INST_OK   INAME            ISTATUS     IMODE     RSTAT    N_OPEN   FLSN            CLSN            
                  2020-11-13 04:15:56  OPEN           OK        GRP1_PR          OPEN        PRIMARY   VALID    9        48329           48329

ora_ckpt_Lu

关注

0
点赞
踩
0

收藏

觉得还不错? 一键收藏
0
评论
达梦数据守护（单主备）集群实践

1.0自杀实验故障自动切换模式下，必须配置确认监视器，且确认监视器最多只能配置一个。#关闭主库服务器[root@Centos7-STD root]# reboot查看监视器#捕捉到主库异常[monitor] 2020-11-13 15:22:52: Received message timeout from(GRP1_PR) WTIME WSTATUS INST_OK INAME
复制链接

扫一扫