update db cfg using SOFTMAX 100;

update db cfg using TRACKMOD ON;

update db cfg using NEWLOGPATH /home/cms/db2log/trans; ----事务日志的存储路径

update db cfg using LOGARCHMETH1 DISK:/home/cms/db2log/arch;

内存参数设定

connect to cms;

update db cfg using CATALOGCACHE_SZ AUTOMATIC;

update db cfg using PCKCACHESC AUTOMATIC;

update db cfg using STMTHEAP AUTOMATIC;

update db cfg using APPLHEAPSZ AUTOMATIC;

update db cfg using STAT_HEAP_SZ AUTOMATIC;

update db cfg using SHEAPTHRES_SHR AUTOMATIC;

update db cfg using SORTHEAP AUTOMATIC;

$ db2 disconnect all

$ db2stop

$ db2start

实施

复制数据库

备份(主)

$ db2 backup db cms online to /home/cms/111 compress

生成迁移脚本(主)

$ db2 list history backup all for cms

$ db2ckbkp -h CMS.0.cms.NODE0000.CATN0000.20140125044402.001

$ db2 "restore db cms from . taken at 20140125044402 into cms logtarget /home/cms/db2fs redirect generate script cms.clp"

----“logtarget /home/db2fs ”字段，表示rollforward时的生成的日志路径

恢复数据库 (备)

$ cd /home/cms

$ mkdir db2log

$ cd db2log

$ mkdir arch

$ mkdir trans

$ scp CMS.0.cms.NODE0000.CATN0000.20140125044402.001 cms@192.168.56.103:~

$ scp cms.clp cms@192.168.56.103:~

$ db2 -tvf cms.clp ---- ‘cms.clp’文件，需要根据实际情况，酌情修改。

配置通讯端口 (主、备)

# vi /etc/services

键入“shift+G”到最后一行

加入：db2h_cmsinst1 70000/tcp #db2 hadr service port

保存，退出！

数据库参数配置

主节点：

$ vi hadr1.sh

db2 update db cfg for cms using hadr_local_host primary-node2;

db2 update db cfg for cms using hadr_local_svc db2h_cmsinst1;

db2 update db cfg for cms using hadr_remote_host standby-node2;

db2 update db cfg for cms using hadr_remote_svc db2h_cmsinst1;

db2 update db cfg for cms using hadr_remote_inst cms;

db2 update db cfg for cms using logindexbuild on;

db2 update db cfg for cms using indexrec restart;

$ chmod +x hadr1.sh

$ ./hadr1.sh

从节点：

$ vi hadr2.sh

db2 update db cfg for cms using hadr_local_host standby-node2;

db2 update db cfg for cms using hadr_local_svc db2h_cmsinst1;

db2 update db cfg for cms using hadr_remote_host primary-node2;

db2 update db cfg for cms using hadr_remote_svc db2h_cmsinst1;

db2 update db cfg for cms using hadr_remote_inst cms;

db2 update db cfg for cms using logindexbuild on;

db2 update db cfg for cms using indexrec restart;

$ chmod +x hadr2.sh

$ ./hadr2.sh

启动主从数据库

启动从数据库

$ db2start

$ db2 start hadr on db cms as standby

启动主数据库

$ db2start

$ db2 start hadr on db cms as primary

切换HADR

监控HADR状态

$ db2pd -d cms -hadr

手工切换

在从节点上执行“takeover hadr”命令：

$ db2 takeover hadr on database cms