关于redhat advance server 2.1 的cluster问题

Question

我用advance server 2.1自带的cluster可以做到使用块设备(/dev/sdb3)上建立文件系统(ext3等)对oracle做ha。但无法使用lvm上建立得文件系统做ha。因为在第一种情况下，当node1 switch to node2时cluster会停止oralce service并umoun storge.然后在node2上up.担对于lvm不是简单的进行umount而是需要卷组在两台节点之间进行export import后在mout然后在启动oracle service。
    所以请问大家可有什么解决的办法？当然还是采用redhat advance server 2.1和lvm。

Answer 1

Thank you for  sharing,Thank you for  brother.

Answer 2

以上的这些相关的软件包到哪里找呀！

Answer 3

除了LVM以外,有没有其它的东西了?

Answer 4

很好就是没实际做过，red hat ad 2.1 cluster是不是支持ha 和nlb(负载均衡)两种模式，还是只支持ha.   

              谢谢

Answer 5

软件：   
操作系统：redhat advance server2.1
卷管理：lvm_1.0.7.tar.gz
文件系统：我用的是reiserfs
数据库：oracle 9i

硬件：
磁盘阵列。两台pc server 。两块网卡。其中一块连lan,一块互连。

1)首先两台主机都能认到磁盘阵列
2)然后安装lvm
       tar vxfz lvm_x.x.x.tar.gz
       cd lvm_x.x.x
       ./configure
       make
       make install
3)编译内核
  (将Multi-device support (RAID and LVM)->;Logical volume manager (LVM) support选中)
  （将filesystem -->; <M>; Reiserfs support选中)
  然后编译内核，改变lilo.conf文件
  当然如果你不使用lvm和reiserfs那么这步就不做了
  
4）需要注意的装了lvm在很多文档中介绍将vgscan 和vgchange -a y 加入rc.local，但在双机一定不能同时加，
   （最好不要加，在后面的脚本中来解决）。然后修改rc.sysinit也将启动vg的语句注释掉。
   
5）使两台主机能用root rsh (当然会存在安全问题，但没有办法谁叫我们用lvm呢。)

6) 磁盘分区（最少三个区）
    fdisk /dev/sdb
    sdb1和sdb2只要有100m就够了。剩下的随你怎么分配
    修改两台主机得/etc/sysconfig/rawdevices文件加入：
    /dev/raw/raw1 /dev/sdb1
    /dev/raw/raw2 /dev/sdb2
    后reboot

7）建立物理卷，建立卷组，建立卷，建立文件系统。然后手工在两台主机之间将卷export import mount umount。
   如果没有问题就成功一半了

7) 接下来配置cluster,我们分别将两台主机命名为clu1 clu2。<>;表示输入的.
   两台的hosts文件
   127.0.0.1                localhost.localdomain localhost
   192.168.1.1              clu1
   10.0.0.1                 eclu1
   192.168.1.2              clu2
   10.0.0.2                 eclu2
   192.168.1.3              culalis
   
   /sbin/cluconfig
   Red Hat Cluster Manager Configuration Utility (running on clu1)
   - Configuration file exists already.
  Would you like to use those prior settings as defaults? (yes/no) [yes]:<yes>;
  Enter cluster name:<clutest>;
  Enter IP address for cluster alias:<192.168.1.3>;
  --------------------------------
Information for Cluster Member 0
--------------------------------
Enter name of cluster member <clu1>;
Looking for host clutest2 (may take a few seconds)...
Enter number of heartbeat channels (minimum = 1) [1]: <1>;
Information about Channel 0
Channel type: net or serial [net]:<net>;
Enter hostname of the cluster member on heartbeat channel 0:<eclu1>;
Information about Quorum Partitions
Enter Primary Quorum Partition [/dev/raw/raw1]: 回车
Enter Shadow Quorum Partition [/dev/raw/raw2]: 回车
Information About the Power Switch That Power Cycles Member 'clutest2'
Choose one of the following power switches:
  o NONE
  o RPS10
  o BAYTECH
  o APCSERIAL
  o APCMASTER
  o WTI_NPS
  o SW_WATCHDOG
Power switch [NONE]: <none>;
Note: Operating a cluster without a remote power switch does not provide
maximum data integrity guarantees.
--------------------------------
Information for Cluster Member 1
--------------------------------
Enter name of cluster member :<clu2>;
Looking for host clutest1 (may take a few seconds)...
Information about Channel 0
Enter hostname of the cluster member on heartbeat channel 0:<eclu2>;
Information about Quorum Partitions
Enter Primary Quorum Partition [/dev/raw/raw1]: 回车
Enter Shadow Quorum Partition [/dev/raw/raw2]: 回车
Information About the Power Switch That Power Cycles Member 'clutest2'
Choose one of the following power switches:
  o NONE
  o RPS10
  o BAYTECH
  o APCSERIAL
  o APCMASTER
  o WTI_NPS
  o SW_WATCHDOG
Power switch [NONE]: <none>;
Heartbeat channels: 1
Channel type: net, Name: eclutest2
Power switch IP address or hostname: clutest2
Identifier on power controller for member clutest2: unused
--------------------
Member 1 Information
--------------------
Name: clutest1
Primary quorum partition: /dev/raw/raw1
Shadow quorum partition: /dev/raw/raw2
Heartbeat channels: 1
Channel type: net, Name: eclutest1
Power switch IP address or hostname: clutest1
Identifier on power controller for member clutest1: unused
--------------------------
Power Switch 0 Information
--------------------------
Power switch IP address or hostname: clutest2
Type: NONE
Login or port: unused
Password: unused
--------------------------
Power Switch 1 Information
--------------------------
Power switch IP address or hostname: clutest1
Type: NONE
Login or port: unused
Password: unused
Save the cluster member information? yes/no [yes]: <yes>;
Writing to configuration file...done
Configuration information has been saved to /etc/cluster.conf.
----------------------------
Setting up Quorum Partitions
----------------------------
Running cludiskutil -I to initialize the quorum partitions: done
Saving configuration information to quorum partitions: done
Do you wish to allow remote monitoring of the cluster? yes/no [yes]: <yes>;
----------------------------------------------------------------
Configuration on this member is complete.
To configure the next member, invoke the following command on that system:
# /sbin/cluconfig --init=/dev/raw/raw1
Refer to the Red Hat Cluster Manager Installation and Administration Guide
for details.

上面太长了，没关系。只要注意”<>;“和“回车”就可以了。接下来运行
# /sbin/cluconfig --init=/dev/raw/raw1
然后做一些监测
/sbin/cludiskutil -p
----- Shared State Header ------
Magic# = 0x39119fcd
Version = 1
Updated on Thu Sep 14 05:43:18 2000
Updated by node 0
--------------------------------
在两台主机都运行，显示的相同就可以了。
clustonith -S
null device OK.
监测完毕。

接这在两主机上安装oracle。最好在一台主机装好测试过后打包在另一台主机在解包。
oracle 安装好后,写三个脚本
(1):/home/oracle/oracleclu (拥有者是root)
#!/bin/bash

#
# Cluster service script to start/stop oracle
#

PATH=/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin;export PATH

case "$1" in
    start)
            rsh clu1 umount /data1
            rsh clu1 /sbin/vgchange -a n
            rsh clu1 /sbin/vgexport search
            /sbin/vgscan
            /sbin/vgimport -f search /dev/sdb3
            mount /dev/search/data1 /data1
            su - oracle -c /home/oracle/startdb
            su - oracle -c "lsnrctl start"
       ;;
    stop)
       su - oracle -c /home/oracle/stopdb
       su - oracle -c "lsnrctl stop"
       umount /data1
       vgchange -a n
       vgexport search   
       ;;
esac

(2)/home/oracle/startdb (拥有者是oracle)
#!/bin/bash
ORACLE_RELEASE=9.2.0
export ORACLE_SID=searchdb
export ORACLE_HOME=/home/oracle/920
export LD_LIBRARY_PATH=$LD_LIBRARY_PATHORACLE_HOME/lib
PATH=$HOME/binORACLE_HOME/bin:/usr/local/jdk/binPATH
sqlplus /nolog<<EOF
connect / as sysdba
startup pfile=/home/oracle/admin/searchdb/pfile/initsearchdb.ora
EOF
exit

(2)/home/oracle/stopdb (拥有者是oracle)
#!/bin/bash
ORACLE_RELEASE=9.2.0
export ORACLE_SID=searchdb
export ORACLE_HOME=/home/oracle/920
export LD_LIBRARY_PATH=$LD_LIBRARY_PATHORACLE_HOME/lib
PATH=$HOME/binORACLE_HOME/bin:/usr/local/jdk/binPATH
sqlplus /nolog<<EOF
connect / as sysdba
shutdown immediate
EOF
exit

在cluster中添加service
在两台主机上都执行service cluster start
/sbin/cluadmin
cluadmin>;service add oracle
Preferred member [None]: <clu1>;
Relocate when the preferred member joins the cluster (yes/no/?)
[no]: <no>;
User script (e.g., /usr/foo/script or None)
[None]: </home/oracle/oracleclu>;
Do you want to add an IP address to the service (yes/no/?): yes
IP Address Information
IP address: <192.168.1.4>;    #client connect oracle ip
Netmask (e.g. 255.255.255.0 or None) [None]: <255.255.255.0>;
Broadcast (e.g. X.Y.Z.255 or None) [None]: <192.168.1.255>;
Do you want to (a)dd, (m)odify, (d)elete or (s)how an IP address,
or are you (f)inished adding IP addresses: f
Do you want to add a disk device to the service (yes/no/?): yes #如果使用lvm这里输入no
Disk Device Information
Device special file (e.g., /dev/sda1): </dev/sda3>;
Filesystem type (e.g., ext2, reiserfs, ext3 or None): <ext3>;
Mount point (e.g., /usr/mnt/service1 or None) [None]:</oradata>;
Mount options (e.g., rw, nosuid): 回车
Forced unmount support (yes/no/?) [no]: <yes>;
Do you want to (a)dd, (m)odify, (d)elete or (s)how devices,
or are you (f)inished adding device information: <f>;

应用：
clustat 察看cluster状态
cluadmin>; cluster status
Cluster Status Monitor (eachnet)                                        15:58:26

Cluster alias: cluster

=========================  M e m b e r   S t a t u s  ==========================

  Member         Status     Node Id    Power Switch
  -------------- ---------- ---------- ------------
  clutest2       Up         0          Good        
  clutest1       Up         1          Good        

=========================  H e a r t b e a t   S t a t u s  ====================

  Name                           Type       Status   
  ------------------------------ ---------- ------------
  eclutest2    <-->; eclutest1    network    ONLINE              

=========================  S e r v i c e   S t a t u s  ========================

                                         Last             Monitor  Restart
  Service        Status   Owner          Transition       Interval Count   
  -------------- -------- -------------- ---------------- -------- -------
  oracle         started  clutest2       15:57:48 Jul 03  30       0      
cluadmin>;service relocate oracle 把oracle从一台主机切到另一台。

Answer 6

能不能共享你有成功经验，我也准备做linuxADV21的双机测试。以前我用redhat7.3加rose HA.

Answer 7

解决了：-）

Answer 8

我在写export和import的脚本。但是lvm在管理共享storge时。经常会遇到
ERROR: VGDA in kernel and lvmtab are NOT consistent。所以要经常重起机器。但在实际应用中是不可能的。
   我在测试。因为veritas的产品太贵了。除了不能用lvm外advance server 的cluster还是很不错的。advance server 是从朋友那里刻的。

Answer 9

我没用过，但一定有个 pre-service 配置文件或脚本，umount 命令和网络配置就是被这个东东执行的，把 export 卷组的命令放在里面即可。

另，你在用这个吗？多少钱买的？用于生产还是自己玩？透露一下可否