两台hp dl360
redhat as5.3x64
写好两台的host文件
10.59.96.65 gfs1.x.com
10.59.96.65 gfs1
10.59.96.100 gfs2.x.com
10.59.96.100 gfs2
fence设备使用的是hp ilo2
先升级hp ilo到最新1.78版本
ilo2_178
配置两台之间的ssh等同性(略)
安装rpm包:
Cluster_Administration-en-US-5.2-1.noarch.rpm
cluster-cim-0.12.1-2.el5.x86_64.rpm
cluster-snmp-0.12.1-2.el5.x86_64.rpm
luci-0.12.1-7.el5.x86_64.rpm
modcluster-0.12.1-2.el5.x86_64.rpm
rgmanager-2.0.46-1.el5.x86_64.rpm
ricci-0.12.1-7.el5.x86_64.rpm
system-config-cluster-1.0.55-1.0.noarch.rpm
两台的cluster配置文件也要一样:
#cat /etc/cluster/cluster.conf
启动cluster
/etc/init.d/cman start
/etc/init.d/rgmanager start
查看cluster状态:
clustat
cman_tool status
cman_tool services
测试fence设备
gfs2#fence_node gfs1
等gfs1起来后，观察gfs2的message,等gfs1加入cluster后,在gfs1上fence gfs2
gfs1#fence_node gfs2
如果不能互相fence对方，就考虑是否该升级fence设备的驱动
两台的drbd.conf要一样
#cat drbd.conf
global {
usage-count no;
}
resource r0 {
protocol C;
startup {
# become-primary-on both;
}
disk {
on-io-error detach;
}
net {
# allow-two-primaries;
after-sb-0pri discard-node-gfs2.x.com;
after-sb-1pri violently-as0p;
after-sb-2pri violently-as0p;
# after-sb-0pri discard-older-primary;
# after-sb-1pri call-pri-lost-after-sb;
# after-sb-2pri call-pri-lost-after-sb;
rr-conflict disconnect;
}
syncer {
rate 30M;
}
on gfs1.x.com {
device /dev/drbd0;
disk /dev/cciss/c0d0p4;
address 10.59.96.65:7788;
flexible-meta-disk internal;
}
on gfs2.x.com {
device /dev/drbd0;
disk /dev/cciss/c0d0p4;
address 10.59.96.100:7788;
meta-disk internal;
}
}
建立drbd metadata,两边同时做
drbdadm create-md all
启动drbd
/etc/init.d/drbd start,两边同时做
将gfs1设置为主节点,gfs2不要动：
drbdadm -- --overwrite-data-of-peer primary all
修改两台机器的drbd.conf
去掉配置文件里的以下两行的注释,使之成为双主
# become-primary-on both;
# allow-two-primaries;
重启drbd
cat /proc/drbd
格式化
mkfs.gfs2 -p lock_dlm -t gfs_cluster:data -j 2 /dev/drbd0
mkdir -p /mysqldata
mount -t gfs2 /dev/drbd0 /mysqldata
Ref:
来自 “ ITPUB博客 ” ，链接：https://blog.itpub.net/230160/viewspace-607812/ ，如需转载，请注明出处，否则将追究法律责任。
上一篇：
dd的使用
下一篇：
ASM process设置
请登录后发表评论
登录
全部评论
最新文章
- How to Add/Drop/Resize Redo Log with Physical Standby in place. [ID 473442.1]
- How to Change Interconnect/Public Interface IP Subnet in a 10g Cluster
- oracle rac vip oracle10g RAC更改主机public地址及vip
- Modifying the VIP of a Cluster Node
- How to Recreate OCR / Voting Disk Accidentally Deleted [ID 399482.1]
- Configuring raw devices multipath for Oracle Clusterware10g R2 (10.2.0) on RHEL5
- Oracle RAC 重建 OCR 和 Voting disk
- How to Recreate OCR / Voting Disk Accidentally Deleted [ID 399482.1]
- shrink space
- 连载4:子例程的引用
