Chinaunix首页 | 论坛 | 博客
  • 博客访问: 2838654
  • 博文数量: 631
  • 博客积分: 10716
  • 博客等级: 上将
  • 技术积分: 8397
  • 用 户 组: 普通用户
  • 注册时间: 2008-04-01 22:35
文章分类

全部博文(631)

文章存档

2020年(2)

2019年(22)

2018年(4)

2017年(37)

2016年(22)

2015年(1)

2013年(12)

2012年(20)

2011年(19)

2010年(20)

2009年(282)

2008年(190)

分类: Oracle

2016-06-30 09:18:11

OCR和VOTE备份与恢复

oracle 11.2 自动备份OCR时同时也会备份VOTE文件。oracle OCR自动备份存放在$CRS_HOME/cdata/$CRS_NAME下。使用restore恢复。

OCR和VOTE备份与恢复恢复步骤:

1:查看OCR自动备份

2:验证OCR

3:查看OCR、VOTE和DG信息

4:备份ASM spfile信息

5:破坏性实验--清除gridnewdg和gridmirdg下所有磁盘中信息

6:验证破坏性试验

7:以独占模式启动节点2

8:创建CRSDG和CRSMIRDG磁盘组以及spfile

9:恢复OCR和VOTE

10:重新恢复OCR和VOTE

11:重启两节点集群

12:重启之后问题解决

百度经验:jingyan.baidu.com

方法/步骤

  1. 1

    1:查看OCR自动备份

    [grid@node2 node-cluster]$ ocrconfig -showbackup

     

    node2     2015/07/23 08:08:18     /u01/app/11.2.0.4/grid/cdata/node-cluster/backup00.ocr

     

    node2     2015/07/23 04:08:18     /u01/app/11.2.0.4/grid/cdata/node-cluster/backup01.ocr

     

    node2     2015/07/23 00:08:16     /u01/app/11.2.0.4/grid/cdata/node-cluster/backup02.ocr

     

    node2     2015/07/23 00:08:16     /u01/app/11.2.0.4/grid/cdata/node-cluster/day.ocr

     

    node2     2015/07/23 00:08:16     /u01/app/11.2.0.4/grid/cdata/node-cluster/week.ocr

    PROT-25: Manual backups for the Oracle Cluster Registry are not available   无需手动备份

    -rw------- 1 root root 7204864 Jun 16 11:42 backup00.ocr

    -rw------- 1 root root 7204864 Jun 16 07:42 backup01.ocr

    -rw------- 1 root root 7204864 Jun 16 03:42 backup02.ocr

    -rw------- 1 root root 7204864 Jun 16 07:42 day_.ocr

    -rw------- 1 root root 7204864 Jun 15 07:42 day.ocr

    -rw------- 1 root root 7204864 Jun 11 03:41 week_.ocr

    -rw------- 1 root root 7204864 Jun  4 03:39 week.ocr

    [grid@db1 db-cluster]$ pwd

    /u01/app/11.2.0/grid/cdata/db-cluster

    如果需要手动备份执行下面语句:

    [root@node2 node-cluster]# ocrconfig -export /tmp/ocr_20150811.exp  可以在线备份 恢复命令:ocrconfig -import /tmp/ocr_20150811.ex


  2. 2

    2:验证OCR

    [grid@node2 node-cluster]$ cluvfy comp ocr -n all  

  3. 3:查看OCR、VOTE和DG信息

    (1)查看OCR信息

    [grid@node1 grid]$ ocrcheck

    Status of Oracle Cluster Registry is as follows :

              Version                  :          3

              Total space (kbytes)     :     262120

              Used space (kbytes)      :       3108

              Available space (kbytes) :     259012

              ID                       : 1679364394

              Device/File Name         : +gridnewdg

                                        Device/File integrity check failed

              Device/File Name         : +gridmirdg

                                        Device/File integrity check failed

    (2)查看VOTE

    [grid@node1 grid]$ crsctl query css votedisk

    ##  STATE    File Universal Id                File Name Disk group

    --  -----    -----------------                --------- ---------

     1. ONLINE   2f845578e55a4f4bbf25d0864b0bc26b (/dev/mapper/crsB13) [GRIDNEWDG]

     2. ONLINE   21c1c99ef5164faabf6186d04aeeb741 (/dev/mapper/crsC01) [GRIDNEWDG]

     3. ONLINE   69eb761aef5c4f87bfaabf9a3e2db5e4 (/dev/mapper/crsB14) [GRIDNEWDG]

    (3)查看DG信息

    SQL>  select name,path,mount_status,failgroup from v$asm_disk order by name;

     

    NAME                     PATH                             MOUNT_STAT FAILGROUP

    -------------------- ------------------------------ ---------- ------------------------------

    DATADG_0000      /dev/mapper/dataB02      CACHED     DATAFIAL08

    DATADG_0001      /dev/mapper/dataB04      CACHED     DATAFIAL08

    DATADG_0002      /dev/mapper/dataB01      CACHED     DATAFIAL04

    DATADG_0003      /dev/mapper/dataB05      CACHED     DATAFIAL09

    DATADG_0005      /dev/mapper/dataB03      CACHED     DATAFIAL04

    GRIDMIRDG_0000       /dev/mapper/crsB11        CACHED     GMIRFAIL01

    GRIDMIRDG_0001       /dev/mapper/crsB12        CACHED     GMIRFIAL02

    GRIDNEWDG_0000      /dev/mapper/crsB13        CACHED     GNEWFIAL04

    GRIDNEWDG_0001      /dev/mapper/crsC01        CACHED     GNEWFIAL05

    GRIDNEWDG_0002      /dev/mapper/crsB14        CACHED     GNEWFIAL03

    根据上面可以看到:

    OCR存放在gridnewdg中,同时存放一份镜像到gridmirdg

    VOTE存放在gridnewdg中

    数据库文件都存放在datadg中

  4. 4:备份ASM spfile信息

    SQL> show parameter spfile

    NAME     TYPE VALUE

    ------------------------------------ ----------- ------------------------------

    spfile     string +GRIDNEWDG/node-cluster/asmparame

    terfile/registry.253.885742421

    SQL> create pfile='/home/grid/asmpfile.ora' from spfile;

  5. 5:破坏性实验--清除gridnewdg和gridmirdg下所有磁盘中信息

    (1)首先停止数据库:[oracle@node2 ~]$ srvctl  stop  database -d orcl  在一个节点执行即可

    (2)停止节点2 集群:[root@node2 grid]# crsctl stop crs

    (3)停止节点1集群:[root@node1 grid]# crsctl stop crs

    (4)因为OCR和vote同时存在gridnewdg中,且ocr还做镜像到gridmirdg中

    为了做vote和ocr恢复实验,需要清除gridnewdg中磁盘和gridmirdg中磁盘:

    dd if=/dev/zero of=/dev/mapper/crsB13 bs=1M count=1000

    dd if=/dev/zero of=/dev/mapper/crsB14 bs=1M count=1000

    dd if=/dev/zero of=/dev/mapper/crsC01 bs=1M count=1000

    dd if=/dev/zero of=/dev/mapper/crsB11 bs=1M count=1000

    dd if=/dev/zero of=/dev/mapper/crsB12 bs=1M count=1000

    这里因为磁盘crs*都有100G,只需要清除1G就行。因为ocr和vote文件大小在200M左右


  6. 6:验证破坏性试验

    (1)启动其中一个节点   发下集群无法正常启动

    尝试重新启动crs: [root@node1 user_root]# crsctl start crs 发现集群无法启动

    (2)强制关闭非正常启动后的集群进程

    停止启动失败的进群相关进程:[root@node1 cssd]# crsctl stop crs -f

  7. 7:以独占模式启动节点2

    [root@node2 dbs]# crsctl start crs -excl -nocrs

  8. 8:创建CRSDG和CRSMIRDG磁盘组以及spfile

    (1) 创建CRSDG和CRSMIRDG磁盘组

    [grid@node2 ~]$ asmcmd   空的

    ASMCMD> ls

    ASMCMD>

    ASMCMD> exit

    [grid@node2 ~]$ su - oracle

    Password:

    [oracle@node2 ~]$ export ORACLE_SID=+ASM2

    [oracle@node2 ~]$ export ORACLE_HOME=/u01/app/11.2.0.4/grid/

    [oracle@node2 ~]$ export PATH=$PATH:$ORACLE_HOME/bin

    [oracle@node2 ~]$ sqlplus sys as sysasm

     

    SQL*Plus: Release 11.2.0.4.0 Production on Wed Aug 12 10:59:04 2015

     

    Copyright (c) 1982, 2013, Oracle.  All rights reserved.

     

    Enter password:

     

    Connected to:

    Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 - 64bit Production

    With the Real Application Clusters and Automatic Storage Management options

     

    SQL> show parameter spfile

     

    NAME                                        TYPE        VALUE

    ------------------------------------ ----------- ------------------------------

    spfile                                          string


    create diskgroup crsdg normal redundancy

       failgroup crsfail01 DISK '/dev/mapper/crsB13'

       failgroup crsfial02 DISK '/dev/mapper/crsB14'

       failgroup crsfial03 DISK '/dev/mapper/crsC01'

       attribute'compatible.asm' = '11.2.0.0.0';

     

    create diskgroup crsmirdg normal redundancy

       failgroup crsmirfail01 DISK '/dev/mapper/crsB11'

       failgroup crsmirfial02 DISK '/dev/mapper/crsB12'

       attribute'compatible.asm' = '11.2.0.0.0';

    (2)编辑pfile文件

    [grid@node2 ~]$ cat asmpfile.ora

    +ASM1.asm_diskgroups='CRSDG','DATADG','CRSMIRDG'#Manual Mount

    +ASM2.asm_diskgroups='CRSDG','DATADG','CRSMIRDG'#Manual Mount

    *.asm_diskstring='/dev/mapper'

    *.asm_power_limit=5

    *.diagnostic_dest='/u01/app/grid'

    *.instance_type='asm'

    *.large_pool_size=12M

    *.remote_login_passwordfile='EXCLUSIVE'

    (3)创建spfile文件

    create spfile='+crsdg' from pfile='/home/grid/asmpfile.ora';

  9. 9:恢复OCR和VOTE

    (1)编辑  /etc/oracle/ocr.loc 文件

    [root@node2 grid]# cat /etc/oracle/ocr.loc

    #Device/file  getting replaced by device +gridmirdg

    ocrconfig_loc=+crsdg

    ocrmirrorconfig_loc=+crsmirdg

    local_only=false

    (2)恢复OCR

    [root@node2 grid]# ocrconfig -restore /u01/app/11.2.0.4/grid/cdata/node-cluster/backup02.ocr

    发现 crsdg和crsmirdg下都增加了ocrfile文件夹

    [grid@node2 ~]$ ocrcheck

    Status of Oracle Cluster Registry is as follows :

              Version                  :          3

              Total space (kbytes)     :     262120

              Used space (kbytes)      :       2756

              Available space (kbytes) :     259364

              ID                       : 1679364394

              Device/File Name         :     +crsdg

                                        Device/File integrity check succeeded

              Device/File Name         :  +crsmirdg

                                        Device/File integrity check succeeded

    (3)恢复VOTE

    [grid@node2 ~]$ crsctl query css votedisk

    Located 0 voting disk(s).

    [grid@node2 ~]$ crsctl replace votedisk +crsdg  恢复失败

    CRS-4602: Failed 27 to add voting file 7a69aa1a2f094feabf02db8c75fd966b.

    CRS-4602: Failed 27 to add voting file 1bb71838797f4fb5bf25c6afcf53e5a0.

    CRS-4602: Failed 27 to add voting file 849fb6ad7e134f3bbf72f9fce25d85ea.

    Failed to replace voting disk group with +crsdg.

    CRS-4000: Command Replace failed, or completed with errors.

  10. 10:重新恢复OCR和VOTE

    (1)[root@node2 grid]# crsctl stop crs

    (2)[root@node2 grid]# crsctl start crs -excl -nocrs

    (3)重新清空磁盘

    dd if=/dev/zero of=/dev/mapper/crsB13 bs=1M count=1000

    dd if=/dev/zero of=/dev/mapper/crsB14 bs=1M count=1000

    dd if=/dev/zero of=/dev/mapper/crsC01 bs=1M count=1000

    dd if=/dev/zero of=/dev/mapper/crsB11 bs=1M count=1000

    dd if=/dev/zero of=/dev/mapper/crsB12 bs=1M count=1000

    (4)重新创建crsdg和crsmirdg

           创建失败,扫描不到磁盘,原因是asm_disktring为空

    (5)SQL> alter system set asm_diskstring = '/dev/mapper';

    (6)创建spfile

    create spfile='+crsdg' from pfile='/home/grid/asmpfile.ora';

    或者

    SQL> create pfile='/home/grid/pfile.txt' from memory;然后修改,再创建spfile。这是防止之前没有备份spfile文件

    (7)重新恢复OCR

    [root@node2 grid]# ocrconfig -restore  /u01/app/11.2.0.4/grid/cdata/node-cluster/backup02.ocr

    [root@node2 grid]# ocrcheck

    Status of Oracle Cluster Registry is as follows :

              Version                  :          3

              Total space (kbytes)     :     262120

              Used space (kbytes)      :       2756

              Available space (kbytes) :     259364

              ID                       : 1679364394

              Device/File Name         :     +crsdg

                                        Device/File integrity check succeeded

              Device/File Name         :  +crsmirdg

    (8)重新恢复VOTE

    [root@node2 grid]# crsctl query css votedisk

    Located 0 voting disk(s).

    [root@node2 grid]# crsctl replace votedisk +crsdg

    Successful addition of voting disk 2c6bf78bc9b04f0bbff6d96ced710f6e.

    Successful addition of voting disk 0d14000c1d474f12bfdd0887922f8b9d.

    Successful addition of voting disk 6da9877208eb4f9abf5d04584dc10968.

    Successfully replaced voting disk group with +crsdg.

    CRS-4266: Voting file(s) successfully replaced

    [root@node2 grid]# crsctl query css votedisk

    ##  STATE    File Universal Id                File Name Disk group

    --  -----    -----------------                --------- ---------

     1. ONLINE   2c6bf78bc9b04f0bbff6d96ced710f6e (/dev/mapper/crsB13) [CRSDG]

     2. ONLINE   0d14000c1d474f12bfdd0887922f8b9d (/dev/mapper/crsB14) [CRSDG]

     3. ONLINE   6da9877208eb4f9abf5d04584dc10968 (/dev/mapper/crsC01) [CRSDG]

    Located 3 voting disk(s).

  11. 11:重启两节点集群

    (1)编辑节点1的/etc/oracle/ocr.loc

    [root@node1 ~]# cat /etc/oracle/ocr.loc

    #Device/file  getting replaced by device +gridmirdg

    ocrconfig_loc=+crsdg

    ocrmirrorconfig_loc=+crsmirdg

    local_only=false

    (2)停掉节点2集群  ---因为之前节点2以独占方式启动

    (3)启动两节点集群 


  12. 12:重启之后问题解决

    根据11步骤图片发现,集群正常启动,但是仍存在以下问题:

    1:旧gridewdg存在OCR中  

    2:新的DG crsdg和crsmirdg不再OCR中

    3:数据库orcl不再OCR中

    问题1解决办法:

    [grid@node2 ~]$ crsctl delete res ora.GRIDNEWDG.dg

    问题2解决办法参考我的百度经验:OCR中添加ASM DG资源


    问题3解决办法参考我的百度经验:PRCR-1001 : Resource ora.orcl.db does not e.


阅读(911) | 评论(0) | 转发(0) |
给主人留下些什么吧!~~