VMware gsx 2.5 + Linux + oracle817 (OPS) Dual-host installation manual

1. install required software:

1. If you cannot find VMware gsx Server 2.5, you can download it at www.vmware.com and apply for a trial serial number. You can also search for it at www.google.com, I believe you want it!

2. RedHat Linux 6.2 is an old version. Many may ask, can this old version work? Why don't we use 2.1 or 7.1? Oh, yes, it's old, but after my tests, 7.2, 7.3, 8.0, 9.0, 2.1, and so on, the least hard disk space and the least hardware resources are saved. In front of a few G systems, 6.2 is less than 600 mb of installation space, and the short and cool performance has become an exclusive pleasure in this test, believe me, but if you think 6.2 is behind, you can use 9.0 or 2.1! It depends on the CPU, memory, and hard disk of your PC, as long as you have enough powerful hardware resources! If your hardware resources are not that adequate, you should choose 6.2 First! You can go to www.google.com to search. There will be a lot of download points!

3, oracle817 (OPS) You can go to www.oracle.com download, http://otn.oracle.com/software/products/oracle8i/index.html
This is the main software. Of course, some other software is required during the installation process, such as telnet software and virtual optical drive software. I will not introduce it here. I believe everyone has it, it will also be used! Now, let's start the installation that is fun and challenging!

Ii. Installation
1. Install and set VMware gsx Server 2.5
By default, the installation and simplicity options are acceptable. Select "yes" when the "autorun" function of the optical drive is disabled, and select "no" when searching for the virtual disk"
Then we need to restart the system. VMWare gsx Server 2.5 has been installed. There is no Chinese version yet. It is a bit annoying for e-paper, but it is nothing. Here we will teach you how to set it! After the restart, we turn on Vmware gsx for configuration,
File-> New à New virtual machine in the pop-up window, select "Custom"

In the following window, select "other" and name it "Linux"

For the next memory, I chose "64 m" because I want to run two Linux instances on one PC, so I chose 64 M. My PC only has 256 m, which is a little nervous, if your memory is large, you can select more!

Select "bridged" for the next network connection. Select the default 4 GB disk space. Select the disk space based on your hardware resources, but not less than 4 GB.

The default disk file name is used. Click "finish" to complete the VM settings.

Then we make a few changes to the configuration of this virtual machine, because the use of the soft drive is very small, we delete the soft drive from the configuration. Settings à configuration Editor

Modify the configuration and delete the drive

Okay, the first Linux Virtual Machine "Linux" has been built. Similarly, I have built a Linux Virtual Machine "linux2" here, so there will be two Linux virtual machines with the same configuration on the VMware gsx interface! In this way, it is basically defined with the cluster dual-host.

Next, we will start to create a shared PLN disk. Use a Windows command prompt to go to the installation directory of VMware gsx. Here we assume that you are installing: \ Program Files \ VMware gsx serverfiles, here you can find an executable file named “plainmaker.exe. This is the tool we use to generate a PLN shared disk. We typed the following command at the command prompt: plainmaker D: \ sys. the PLN 2000 m command generates a 2g pln disk in the "D: \ sys" directory.

After the shared disk is ready, we have to add it to the two Linux virtual machines we just created. On the VMware gsx interface, select a virtual machine, click Settings à configuration editor. In the pop-up add hardware wizard window, click "add", and then select "Hard Disk" and click "Next"

For the next "select a disk", click the second "use an existing virtual disk"

Then we find the PLN disk we just created.

When selecting a disk mode, be sure to select the first "persistent"

Then we can click "advanced", which is extremely important!

In the next "select a disk node", be sure to select "SCSI 1:0"

After the selection, click "finish" and "OK" to complete the operation of adding a shared disk, and perform the same operation on another Linux virtual machine! After completing these steps, we need to edit the configuration information file (VMware configuration file) of the two Linux virtual machines. Their names are: Linux. vmx, linux2.vmx, they are in the directory where you store Linux virtual machines. We can use NotePad to open and edit them. The following is one of the contents. Please pay attention to the content in the dark black font. These are added, mainly to tell VMware gsx scsi1: 0 is a shared bus, the disk on it is a shared disk, and these content must be added to both files. Otherwise, the dual-host cannot share the disk. This principle is equivalent to virtualizing a disk array!

Config. Version = "6"

Virtualhw. Version = 2

Localconsole = true

Displayname = "Linux"

Guestos = "Linux"

Ide1: 0. Present = true

Ide1: 0. devicetype = "atapi-CDROM"

Ide1: 0. filename = "I :"

Scsi0: 0. Present = true

Scsi0: 0. filename = "Linux. vmdk"

Scsi0.present = true

Memsize = 64

Ethernet0.present = true

Draw = "GDI"

Floppy0.present = false

Scsi1: 0. Present = true

Scsi1: 0. devicetype = "plaindisk"

Scsi1: 0. filename = "D: \ SYS. PLN"

Scsi1.present = true

Disk. Locking = false

Scsi1.sharedbus = "virtual"

Scsi1: 0. mode = "persistent"

UUID. Location = "56 4D 9C 46 C6 E7 af 51-a8 17 47 E7 E7 8C 64 F2"

Ide1: 0. startconnected = true

Tools. remindinstall = false

Tools. synctime = false

Edit and save the settings. Now we have configured the two Linux virtual machines. Now, we have started installing Red Hat Linux 6.2!

2. settings for Red Hat Linux 6.2 installation and oracle817 which OPS dual-host Installation

To be continued

1) install Red Hat in Linux 6.2

Prepare the downloaded 6.2 ISO file, install the virtual optical drive software: Daemon tools, and then load the ISO image file.

Select a Linux Virtual Machine in VMware gsx, choose settings à preferences and select 2nd "CTRL + ALT + shift" in "hot keys". Here we select this option for switching in Linux!

After that, open the selected Linux Virtual Machine and click "power on" to start 6.2 installation. After the power is turned on, click the mouse in the screen and press the F2 key, go to the "BIOS" of the VM to set the boot sequence (just like the real host we use at ordinary times, but the "BIOS" of the VM is phoenixbios, which is used by many brands ), set CD-ROM drive to the first option, and then disable the drive

Then save and restart and start the installation of linux6.2!

After the CD boot, the 6.2 installation page appears. Select "English" and click "Next"

For the next keyboard and mouse, select the default one. After the selection, go to "system installer"

Select "Custom" for the next installation type to start disk initialization.

We can see that there are two disks "SDA, SDB". We only partition "SDA". The solution is as follows:/boot 100 m, swap 128 M, /1859 M/Oracle 2000 m is as follows:

After the split, for example, "SDB" is not partitioned here, please keep it as it is!

The next "choose partition to format" option is optional by default! Choose "lilo configuration" not to create a boot disk!

Click "Next" to go to "network configuration". Disable "Configure using DHCP". You can configure it based on your actual situation.

Here, set the hostname to "dbs1" IP address to "11.0.10.12 255.255.255.0". In the next time zone, select "Shanghai"

Next, set the root password. Select "authentication configuration" as the default password. After these passwords are complete, go to the installation page.ProgramSelect page

Note that we only select the following types of "X Windows System", "gnome", "Anonymous FTP server", "Emacs", "Development", and "kernel development ", you can select the "utilities" Software Package. Do not select any other software package. Of course, you can select another software package based on your needs, but you must include the above six software packages! After the selection, click "Next" to configure "xconfiguration". Select the monitor first, select "generic monitor", and select "generic VGA compatible" and "1024k" for the video card"

After the selection, click "Next" to start the installation of 6.2, which is a little slow!

Okay, let's just relax and take a break. Of course, you can also open another Linux Virtual Machine and install it with the same configuration as the first one. Of course, this will be very slow, however, as long as your PC is strong enough, it took me about three hours to install 400 simultaneously on my P2 256 MHz 6.2 m host, I don't think everyone's machine is worse than me. It's all in the P4 age!

This is what I installed.
Handsome! Two Linux Hosts are running!

(2) Linux host trust mechanism settings

Next, we will set these two Linux host systems! First, we need these two Linux Hosts to define their "hostname". In Oracle (OPS), each host is considered as one node ), each node is differentiated by "hostname! Here, the "hostname" of the first Linux host is defined as "dbs1", and the second is "ins1", because we have defined the network settings during installation. Therefore, we only need to make some modifications to their respective "/etc/hosts" files. The following is the content of the "hosts" initial file of the dbs1 host.

127.0.0.1 dbs1 localhost. localdomain localhost

We can change it to the following content:

127.0.0.1 localhost. localdomain localhost (Note: delete dbs1 here to avoid conflict with the dbs1 defined below. This step is very important and is summed up at the cost of using blood !)

11.0.10.12 dbs1
11.0.10.13 ins1 (note: "hostname" of another Linux host ")

Similarly, the hosts of ins1 is the same as that of dbs1!

Then we Ping each other on the two hosts. For example, if we can ping ins1 on dbs1 and parse it correctly, the configuration is correct. If any problem occurs, check whether the network and hosts files are correct! Next, we need to set the security and mutual trust of the two servers. Whether Oracle (OPS) is successfully installed depends largely on whether the Sub-configuration is correct!

① Now we have removed Linux's restrictions on Root User Telnet! Run the following command under #: # cd/etc/PAM. d # vi login adds the # sign in front of "auth required/lib/security/pam_securetty.so" in line 2nd and saves and exits. In this way, the root user can use Telnet to log on.

② To ensure that each user can use rlogin and RSH. d) The rlogin file is also edited by VI, and the 2nd line "auth required/lib/security/pam_securetty.so" is also commented out so that users can use rlogin normally, of course, you must first ensure that the RSH package is installed! As long as you have installed 6.2 as I have done before, you can check it with the following command: # rpm-AQ | grep RSH

Then, we added "RSH" to rlogin by VI Editor/etc/securetty. RSH can be used normally!

③ For the root user to FTP to the host, we have to modify the "/etc/ftpusers" file and add a # number before the "root" in the file! Although this step is not necessary, I think it is very convenient to use root sometimes!

③ Next I will proceed with the Trust Mechanism Setting for two nodes (Note: two Linux Hosts, which will be used to represent the host in the future, we first create the "hosts. equiv file with the following content:

Dbs1 # hostname of node dbs1
Ins1 # hostname of node ins1

In this way, the two nodes establish a trust mechanism, which is extremely important for the installation of cluster dual-host! It ensures that the two nodes have the same permissions for users to use the "RSH, rlogin, RCP" command, that is, to trust all users of the trusted host! In this way, when the root user is used outdoors on the two nodes, they can "rlogin, RSH, RCP" each other without having to enter the password so that the root user of the two nodes can do the same, we also need to create in the main directory "/root" VI of each root user. rhosts file, whose content is:

Dbs1 Root
Ins1 Root

In this case, the root user does not need to enter the password when using "rlogin, RSH, RCP". Of course, the security is poor!

After doing this, we can execute "rlogin, RSH, RCP" on each node. I will give you several sub-classes, for example, we will execute on the node dbs1: # rlogin ins1 (rlogin logs on to node ins1, and we can see that we can successfully enter node ins1 as root without losing the password) # rsh-l root ins1 CAT/etc/hosts. equiv runs CAT/etc/hosts on the ins1 logon node as the root user. equiv command. At this time, the screen of the dbs1 node returns to the node ins1/etc/hosts. equiv file content) # RCP/tmp/test.txt ins1:/tmp (copy the file test.txt under the/tmp directory of the node dbs1 to the/tmp directory of the node ins1, the above is only a small step after the long journey, and there are a lot to set up!

3) oracle817 which OPS settings

① Linux kernel settings
Because Red Hat Linux does not have its own cluster software, it is available at 2.1, but it does not support Oracle ops, unlike Aix, HPUX, Sun Solaris, true 64 and so on all have cluster software supporting ops. Therefore, Oracle has its cluster software orcm in Oracle 817, oracle Cluster Manager is used to provide dual-host Oracle ops on Linux to implement high availability of Oracle databases on Linux, that is, ha! To use orcm, you must modify the Linux kernel and recompile the kernel again. orcm requires watchdog. Generally, after Linux is installed, watchdog is supported, however, some kernel options are not automatic but manual, or you need to adjust them when you enable them! The following three kernel parameters need to be modified:

Config_watchdog = y
Config_soft_watchdog = y
Config_watchdog_nowayout = y

Note: "Disable watchdog shutdown on close"

The following are all commands:

First, go to "/usr/src/Linux" VI to edit makefile and change "extraversion =-5.0" to "extraversion =-5.0ops" to make it easier to distinguish it from the original kernel!

Then run "make menuconfig"

# Cd/usr/src/Linux

# Make menuconfig select character devices-à from the pop-up menu to see that "[*] watchdog timer support" is marked *, this indicates that "config_watchdog = y" has been supported in the kernel! In the next watchdog card-à, we can see "[m] disable watchdog shutdown on close". Change it to "[*] disable watchdog shutdown on close", which is "config_watchdog_nowayout = y ". "Then confirm that" [*] Software Watchdog "[] is *, this is "config_soft_watchdog = Y". After all these operations are completed, select "exit" to exit. Then, ask if you want to save the new Kernel configuration. You must select "yes!

② Kernel Encoding

After completing the preceding steps, you can compile the new kernel. Then, run the following command:

# Make Dep (create the dependent files required for compilation)

# Make clean (clear the target file compiled by the kernel)

# Make bzimage (compile the kernel)

# Make modules)

# Make modules_install (after compilation, the system will generate a 2.2.14-5.0ops subdirectory under the/lib/modules Directory, which stores all the loaded modules of the new kernel .)

# Cp/usr/src/Linux/ARCH/i386/boot/bzimage/boot/vmlinuz-2.2.14-5.0ops (copy the new kernel to the/boot directory)
# Cp/usr/src/Linux/system. MAP/boot/system. map-2.2.14-50ops (copy the system. map file to the/boot directory)

#/Sbin/mkinitrd/boot/initrd-2.2.14-5.0ops.img 2.2.14-5.0ops (generate new imgfile)

In this way, the new kernel can be used. You can delete the old kernel and sysmap. Of course, you can also stay!

③ Configure the/etc/Lilo. conf file.

Modify the following behavior in the file:

Image =/boot/vmlinuz-2.2.14-50.ops
Label = Linux
Initrd =/boot/initrd-2.2.14-50ops.img

Save and restart! Although the kernel is rebuilt, the device file watchdog is not generated.
Run the following command to create the watchdog device file:

# Mknod/dev/watchdog C 10 10 130

After oracle users and DBAs are set up, change their owner to Oracle: DBAs are acceptable!

④ Create a disk partition for Oracle Cluster Manager to manage partitions and store Oracle tablespace, control files, and redo log files.

Mongom needs a shared raw device to store real-time data dramas. Oracle OPS also needs to store its database files on the shared raw device, so we partition the shared disk, create a raw device file.
First, we partition the shared disk and perform the same operation on one of the two nodes. The solution is as follows:

We have three primary partitions and one extended partition with 10 partitions ).

/Dev/sdb1 10 m (used for Oracle Cluster Manager partition management)
/Dev/sdb2 500 m (used to store the Oracle rollback data table space rollback)
/Dev/sdb3 300 m (used to store the Oracle System tablespace System)
/Dev/sdb5 100 m (used to store Oracle tool tablespace tools)
/Dev/sdb6 100 m (used to store temporary Oracle tablespace temp)
/Dev/sdb7 200 m (used to store the Oracle user tablespace user)
/Dev/sdb8 100 m (used to store the Oracle index tablespace indx)
/Dev/sdb9 10 m (used to store the control file of the Oracle Control System)
/Dev/sdb10 10 m same as above
/Dev/sdb11 8 m (used to store the redo log file of the Oracle redo log file of 1st nodes, 2 redo log files for each node)
/DE/sdb12 8 m same as above
/Dev/sdb13 8 m (used to store Oracle redo log files on 2nd nodes)
/Dev/sdb14 8 m same as above

Note that the system tablespace must be no less than 275 MB,
The control file and redo log file are all multiples of 8 KB.

Based on this scheme, we partition, # fdisk/dev/SDB such:

Because it is a shared disk, you only need to partition it on one node! After dividing the partition, enter "W" to save the partition!

⑤ Create raw device files for each node

Create a raw device file according to the following scheme and use the corresponding partition bound.

/Dev/raw/raw1/dev/sdb1
/Dev/raw/rbs01/dev/sdb2
/Dev/raw/sys01/dev/sdb3
/Dev/raw/tools01/dev/sdb5
/Dev/raw/temp01/dev/sdb6
/Dev/raw/user01/dev/sdb7
/Dev/raw/indx01/dev/sdb8
/Dev/raw/ctl1/dev/sdb9
/Dev/raw/ctl2/dev/sdb10
/Dev/raw/log1_1/dev/sdb11
/Dev/raw/log1_2/dev/sdb12
/Dev/raw/log2_1/dev/sdb13
/Dev/raw/log2_2/dev/sdb14

Raw file creation method

# Mknod/dev/raw/rbs01 C 162 2; mknod/dev/raw/sys01 C 162 3; mknod/dev/raw/tools01 C 162 4; mknod/dev/raw/temp01 C 162 5; mknod/dev/raw/user01 C 162 6; mknod/dev/raw/indx01 C 162 7; mknod/dev/raw/ctl1 C 162 8; mknod/dev/raw/ctl2 C 1629; mknod/dev/raw/log1_1 C 162 10; mknod/dev/raw/log1_2 C 162 11; mknod/dev/raw/log2_1 C 162 12; mknod/dev/raw/log2_2 C 162 13

After creating an oracle user and DBA group, change the raw file owner to Oracle: DBA! These are also required on the other node, and must be the same! Modify the "/etc/rc. d/rc. Local" file of the two nodes and add the following content at the end:

Raw/dev/raw/raw1/dev/sdb1
Raw/dev/raw/rbs01/dev/sdb2
Raw/dev/raw/tools01/dev/sdb5
Raw/dev/raw/sys01/dev/sdb3
Raw/dev/raw/temp01/dev/sdb6
Raw/dev/raw/user01/dev/sdb7
Raw/dev/raw/indx01/dev/sdb8
Raw/dev/raw/ctl1/dev/sdb9
Raw/dev/raw/ctl2/dev/sdb10
Raw/dev/raw/log1_1/dev/sdb11
Raw/dev/raw/log1_2/dev/sdb12
Raw/dev/raw/log2_1/dev/sdb13
Raw/dev/raw/log2_2/dev/sdb14

After the device is restarted, run # raw-AQ on the two nodes to check the health status of the raw device!

⑥ Create an oracle user, DBA, and oinstall group on each node

# Groupadd DBA (create DBA group)
# Groupadd oinstall (to create an oinstall group, you must have an oinstall group to install Oracle)
# Useradd-D/home/Oracle-g dba-G oinstall-M Oracle (create an oracle user whose home directory is/home/Oracle)
# Passwd Oracle (password for Oracle Users)

After the creation, log on to the Oracle user and create the ora directory to store the Oracle Installation File. $ Mkdir-P ora, and then use the root user to change the owner of the "/oracle" directory to Oracle: dBA, # chown ORACLE: DBA/Oracle, the/Oracle directory is the main directory installed in Oracle!

7. Configure Oracle environment variables and Linux Kernel Parameters

VI edit the "/home/Oracle/. bash_profile" file for each node and add the following content:

Export oracle_base =/Oracle

Export ORACLE_HOME = $ oracle_base/product/8.1.7

Export oracle_sid = ops1

Export ora_nls33 = $ ORACLE_HOME/ocommon/NLS/admin/Data

Export nls_lang = "simplified Chinese" _ China. zhs16cgb231280

Export LD_LIBRARY_PATH = $ ORACLE_HOME/lib

Export classpath =.: $ ORACLE_HOME/jdbc/lib/classes111.zip

Export Path = $ path: $ ORACLE_HOME/bin: $ ORACLE_HOME/DBS:/etc

Export ld_assume_kernel = 2.2.5

VI: edit the "/etc/services" file and add the following content at the end:

Set shmsys: shminfo_shmmax = 134217728 (Note: 256mx0. 5x1024x1024)
Set shmsys: shminfo_shmmin = 1
Set shmsys: Sh minfo_shmmni = 128
Set shmsys: shminfo_shmseg = 16
Set semsys: seminfo_semmni = 128
Set semsys: seminfo_semmns = 2048
Set semsys: seminfo_semms = 2048
Set semsys: seminfo_semume = 64
Set semsys: seminfo_semmnu = 768
Set semsys: seminfo_semmap = 20
Set semsys: seminfo_semopm = 100
Set semsys: seminfo_semvmx= 32767

After these steps are completed, you can restart them!

4. Oracle ops Installation

(1) install vmware-Tools for Linux. On which node do you install Oracle, you can install vmwa on which node.
Re-Tools for Linux, here I select the node dbs1 to install. This is mainly to make X Windows display in 16 or 24 colors, that is, the so-called installation video card! The installation instructions are as follows:

Log on to dbs1 as the root user and run the following command: # startx. The X Windows interface appears, of course, a 16-color interface. In this case, click the virtual machine settings-à VMware Tools instlall.

Select "Install" in the pop-up window"

The Installation File vmware-linux-tools.tar.gz appears in the "/mnt/CDROM" directory on the X Windows interface,
This is automatic. If a window does not appear, go to the "/mnt/CDROM" directory!

Press the four keys Alt + Ctrl + Shift + F2 to switch Linux to another logon window and log on as the root user, copy the "vmware-linux-tools.tar.gz" file under "/mnt/cdrom.pdf" to the "/tmp" directory # cp/mnt/CDROM /*. GZ/tmp, and then run # tar xvzf * In the "/tmp" directory *. GZ. After the file package is unwrapped, the "vmware-Linux-Tools" directory is generated under the "/tmp" directory, and the directory is executed. /install. PL "installs vmware-Linux-tools. After installation, exit xwindows and restart. Run startx as the root user again. You will see a beautiful xwindows interface!

(2) handling of installation files

Copy the oracle817 which OPS file package linux81701.tar from the Oracle website to the "/oracle" directory of the dbs1 node. Use Win2000 for FTP commands and use Oracle for logon, select the bin format for the transmission mode, and switch the contents to the linux81701.tar directory, and use "mput. GZ "transfers the file to the/Oracle directory of dbs1!

After uploading, run # tar xvf Linux * In the "/oracle" directory *. tar-C/home/Oracle/ora, if you use the root user for execution, note that the owner of "disk1" in the "/home/Oracle/ora" directory must be: oracle: DBA. If not, change it to Oracle: dba. Delete the linux81701.tar file to free up space for the "/oracle" directory and store the Oracle file!

(3) Installation

We enter the "/oracle" directory from each node as an oracle user, create the product directory, and then
Create the 8.1.7 directory

$ Mkdir-P Product
$ Cd product
$ Mkdir-P 8.1.7

After these steps are completed, run startx on the dbs1 node to go to the X Windows interface.
Open a terminal window, go to the "/home/Oracle/ora/disk1" directory, and run "./runinstaller" to start Oracle installation! (Do not worry about the map failed in the terminal window. This will not affect the installation)

Click "Next" to display the installation file information, indicating the source file and target file. If the file is correct, click "Next"

Next, if you want to enter the Unix group for Oracle Installation, enter "oinstall"

In the displayed window, Run "orainstroot" as the root user. sh "script file, you can open another terminal window, execute $ su-root to become the root user, and execute". /orainstroot. after SH is executed, click Retry to proceed to the next step.

Next, select the installation document, and select 1st "Oracle8i Enterprise Edition 8.1.7.0.1"

Select "Custom" as the installation type"

Select "Oracle Parallel Servier 8.1.7.0.1" for the next installation file. This is what we call ops, that is, the Oracle Parallel service, remove "Oracle HTTP Server 1.3.12.0.1a". For "Product Versions", select "simplified Chinese!

Select the default value for the next Java environment.

The following "osdba, osoper" is a DBA group in a Unix group.

Click "Next" to see "cluster node selection". Because it is a dual-host 2 node, only the name of the other node is added to ins1 (that is, the hostname of the other node)

Click "Next". If this parameter is not selected for security authentication, click "Next"

Click "Install" to start installation!

The installation and link will be smooth, and the link to 62% will be slow. Please wait for a while.

After the installation and link are successfully installed, it will take a long time for the installer to Start copying the installation file to the ins1 node of another node.
After a long wait, you have a lot of patience! After another node is installed, a window will pop up asking you to run the root command under the "/Oracle/product/8.1.7" directory as the root user. sh file. Run "orainstroot. sh. before the sh file, modify the root of each node. two errors in SH, one of which is: please put

RMF =/bin/Rm-F to RMF = "/bin/Rm-F". The other is to set row 156 ruid = '/usr/bin/ID | $ awk-f \ ('{print}' | $ awk-f \) '{print}' to ruid = '/usr/bin/ID | $ awk-f \ (' {print} '| $ awk-f \) '{print} ''is the last multiple' symbol. After these changes, all the files under "/Oracle/product/8.1.7/lib" of the RCP node dbs1 will go to the same directory of node ins1.

$ RCP/Oracle/product/8.1.7/lib/* ins1:/Oracle/product/8.1.7/lib

Then, the root. Sh file is executed on both nodes. Click "OK" on the installation screen after execution"
In this way, the installation is complete, and we quit the installation!

4) Oracle Cluster Manager (MB)

After oracle is installed, we can see "M" in "/Oracle/product/8.1.7" on each node"
Directory, which is the Oracle cluster Software Directory. We enter it as an oracle user and create a new file named nmcfg. ora under the "admin" directory under this directory. As follows:

$ VI nmcfg. ora

The content is as follows:

Definednodes = dbs1 ins1
Cmhostname = dbs1 (note: the node is ins1)
Cmdiskfile =/dev/raw/raw1

After the two nodes are edited, save and exit. Then we enter "/Oracle/product/8.1.7/Orac

The M/bin directory VI creates a file named test. conf with the following content:
Dbs1:/dev/raw/raw1
Ins1:/dev/raw/raw1

Save! To ensure smooth running of ipvm, we also edit the "/etc/rc. d/rc. Local" file for each node and add:
Insmod softdog soft_margin = 60

Add the following at the end of the "/etc/profile" file:
Export ORACLE_HOME =/Oracle/product/8.1.7
Export Path = $ path: $ ORACLE_HOME/Taobao M/bin

Please restart after all these steps are completed! Then run the wdttest1 command (executed by the root user) in the "/Oracle/product/8.1.7/mongom/bin" Directory of each node. If it takes 60 seconds, if each node is restarted, it is indicated in "/etc/rc. d/RC. the "insmod softdog soft_margin = 60" added by "local" is in use and can prove that the watchdog device works properly! Then, execute rdevtest test. conf in the same directory of each node. If "OK" is returned, it indicates that nmcfg. ora is correctly configured and the M-managed partition works normally! Here is 1. conf

If these are all passed, we can execute "ocmstart. Sh" to run the Oracle Cluster Manager software under the same directory (/Oracle/product/8.1.7/mongom/bin) of the two nodes!

(5) Dual-host Oracle ops net8 and initialized local database

When the worker m of the two nodes is running, we log on to the dbs1 node as an oracle user and open X Windows. Then, in the terminal window, run ". /netca "to start the configuration of net8 under the dual-machine Cluster

Select "cluster configuration" and click "Next" to select both nodes!

Click "Next" and select "Listener Configuration" to configure listener.

Click "Next" and select "add"

Select the default listener name "listener" on the screen"

Select "TCP" for the communication protocol"

The default port number is "1521"

Whether to select "no" for configuring another listener"

In this case, the Listener Configuration is complete.

Other options do not need to be configured. If you want to configure them, it is very easy. I will not talk about them here. Next we start to create a database. In fact, Oracle ops is a service (server) that runs two instances ).

Similarly, run "./dbassist" in the same directory as dbs1 to start database creation.

Select "Oracle Parallel Servier configuration" when selecting a Database Type"

Select "Create a database", click "Next", and select "Custom" for the two nodes in the next node selection"

Select "multipurpose" for the database application type"

Select the default value for the maximum number of connected database users. Of course, you can also select a value of 1000.

Select "dedicated server mode" for the database service mode"

Please do not select database installation options. Otherwise, you will go crazy during the extremely abnormal installation process. I tried to select all of them and installed them for more than 28 hours !!!! Unless you like self-abuse !!!

Add "Sid prefix" under "ops" to "Global Database Name ".
However, you can also add any name you like, which will generate two real-class (Instance) Names ops1 and OPS2. Select zhs16cgb231280 for "change character set"

Click "Next", "Control File 1:, control file 2:", and fill in the raw file we created earlier. "/dev/raw/ctl1,/dev/raw/ctl2"

The following system tablespace, tool tablespace, user tablespace, rollback, index tablespace, and temporary tablespace) fill in the raw file we created earlier: "/dev/raw/sys01,/dev/raw/tools01,/dev/raw/user01,/dev/raw/rbs01, /dev/raw/indx01,/dev/raw/temp01"

After filling out the information, click "Next" and fill in the raw file "/dev/raw/log1_1" created earlier in "redo log file" on the two nodes dbs1 and ins1 respectively, /dev/raw/log1_2,/dev/raw/log2_1,/dev/raw/log
2_2"

Next, select the default option. If you do not select enable archive log"

Select the default SGA parameter next!

Next, select the default

After the selection, click "Next" to display the database creation interface. Select "create database now" and click "finish" to start the database creation.
Create a database.

When it reaches 95%, a window will pop up saying that the "/etc/oratab" file cannot be modified. Don't worry about it. Click "OK" to continue! Finally, the dialog box is displayed, indicating that the database has been created. The global name is Ops, the SID is ops1, (the other node is ins1 is OPS2), and the Sys password is change_on_install, the system password is Manager. Click "OK ".
Install!

Then we start listener on two nodes, and Oracle on both nodes is started!

To verify whether ops is running, run the SQL command:
SQL> connect/As sysdba
SQL> select * from V $ active_instances;
We will see:
Inst_number
------------------
Inst_name
-----------------------------------------------------
1
Dbs1: ops1
2
Ins1: OPS2

This proves that our OPS has been fully installed!