Oracle Fail Safe Step-By-Step Installation and Configuration

Bookmark Fixed font  Go to End

Doc ID:  Note:74009.1
Subject:  Oracle Fail Safe Step-By-Step Installation and Configuration
Type:  BULLETIN
Status:  PUBLISHED
Content Type:  TEXT/PLAIN
Creation Date:  08-SEP-1999
Last Revision Date:  09-MAR-2001

PURPOSE

This paper describes the installation and configuration of
a complete Oracle Fail Safe system from the scratch - no
Oracle prodUCts are even installed. A step-by-step approach
is taken to simplify the eXPlanation. After each step its
the name of the node where the operation must be submitted.

SCOPE & APPLICATION

This paper is intended to people that will install and
configure the Oracle Fail Safe.
It assumes that the MSCS is already configured and
running.

RELATED DOCUMENTS

None

1.  Install Oracle 8.0.5 on <DEFAULT_HOME> (Node 1)
2.  Reboot Node 1
3.  Install Oracle 8.0.5 on <DEFAULT_HOME> (Node 2)
4.  Reboot Node 2
5.  Create a database to be added to the cluster (using Note:68720.1)
    do not configure Net8 for the newly created database. Remember
    control files, redo log files and datafiles all must be created in the
    shared disks. (Node 1)
6.  Add an ORACLE_SID variable with its value being the SID of the
    database created in Step 5. to the Oracle Hive in the Registry of both
    nodes. (Node 1, Node 2)
7.  Install FailSafe 2.1.3 on <DEFAULT_HOME> (Server and Manager
    components). (Node 1)
8.  Install FailSafe 2.1.3.1 patch on <DEFAULT_HOME> (Node 1)
9.  Reboot Node 1
10. Install FailSafe 2.1.3 on <DEFAULT_HOME> (Server and Manager
    components). (Node 2)
11. Install FailSafe 2.1.3.1 patch on <DEFAULT_HOME> (Node 2)
12. Reboot Node 2
13. Ping the Cluster Alias and each node's hostname(you could see all
    the names in the MSCS Administrator), if all reply then proceed. If
    not then ask the System Administrator to check the problem. The public
    IP should come first when pinging each node!
(Node 1)
14. Ping the Cluster Alias and each node's hostname(you could see all the
    names in the MSCS Administrator), if all reply with the same IPs of
    step 13. then proceed. If not ask the System Administrator to check
    the problem. The public IP should come first when pinging each node!
    (Node 2)
15. Edit %windir%\System32\Drivers\etc\hosts and add a unique hostname
    and a free IP in the list (Node 1)
16. Edit %windir%\System32\Drivers\etc\hosts and add the same hostname
    and IP added in 11 (Node 2)
17. Enter Fail Safe Manager(FSM): you will be prompted for four
    fields. Fill the fields username and passWord with a Windows NT
    Administrator, the cluster name is the root of the tree that
    appears in the MSCS Administrator, the domain name is the name
    of the domain that the nodes belong (it's displayed in the logon
    box of NT). (Node 1)
18. In FSM choose Verify Cluster..., if any errors are detected fix them
    before proceeding.
19. In FSM Choose Add a New Group ..., when asked for a Network Name choose
    Network Accessible by Clients and in the Name Field put the hostname
    written in the hosts file of steps 15. and 16. . You should see the
    IP field being automatically filled (with the IP specified in steps
    15. and 16.
, all other wizard steps are self explained. (Node 1)
20. Choose Verify Group ... in FSM to assure that the group was
    created ok.
21. Copy the password file and init.ora (if it is in the private
    disks) of the databases to be added to the cluster to the same path
    in the other node. (Node 1, Node 2)
22. Set all Net listeners to Manual Startup. Fail Safe will start the
    Listener automatically on failover. (Node 1)
23. In the FSM choose Verify Standalone Database... if everything is ok
    proceed to 24., if not fix the problems detected.
24. In the FSM choose Add a Standalone Database... to the previously
    created group. Be sure that no Listeners are started up when doing so.
    (Node 1)
25. Make a Move to Other Node... Operation to test the configuration
    (rigth-click the DB in the tree of FSM and choose that option).
    (Node 1)
26. Restore the DB to the original Node 1 doing another Move to Other
    Node... operation (Node 1)
.