Oracle Fail Safe and Database Upgrade(8i) Procedures in Microsoft Cluster Environment

** Credit Web
Upgrade Steps 1 & 2
1. Use Oracle Fail Safe and “Verify Cluster”
2. Microsoft Cluster Server(MSCS) remove resources
Prior to Starting Application Production Oracle Upgrade... all the resources are removed EXCEPT for the following: (e.g.)
•Disks E F H G
•IP Address 45.123.15.97 (e.g virtual IP for the cluster)
•KXP1
•Network Name <name>PRIMARY (e.g..)
•OracleOraHome81 TNS Listenerfsvsprimary
**We did this based on previous issues with resources. This may not be necessary.
Upgrade Step 3
3.Remove the Database(s) from OFS3.0.3 Place in standalone mode. (non-cluster aware).
Issue 1:
*Errors FS-10041 : Failed to delete the database resource <dbname>,FS-10074: Error Configuring Net8(or SQLNET) for requested Operation., FS-10999.
Workaround: Check the listener.ora , ensure that the IP's used for the ports are HARDWARE IP addresses for that NODE.
Issue 2:
DISK drive resources were no longer a part of PRIMARY group, but they were now part of "DISK GROUP 1" (a new resource).
Workaround: (MOVE the "DISK" resources back under PRIMARY group using Cluster Administrator.

Upgrade Steps 4 -6
4.Remove OFS (De-install 3.0.3)
5.Install Oracle Fail Safe 3.3.2 on Node1(reboot)
6.Install Oracle Fail Safe 3.3.2 on Node2(reboot)
Notes:
Oracle Universal Installer

Check Path --“remove any c:\oracle\failsafe30”
Check Registry-remove ID# and HOME#
Check Directory -rename/remove Old Oracle Home
Upgrade Steps 7 –8
7.Install Oracle 8.1.7 on Node1(reboot)
8.Install Oracle 8.1.7 on Node2(reboot)
Notes:
Available Product Components screen(custom)
Upgrade Database screen(No)
Net8 Configuration Assistant screen(typical)

Upgrade Step 9
9.Upgrade Databases to 8.1.7
Notes:
Data Tablespaces in READ ONLY mode
Manual or Automated ??
“Oracle Data Migration Assistant”(45min)
Init.ora parameters “during the upgrade”
#Job_queue_processes, #Job_queue_interval
_system_trig_enable=false
Upgrade Steps 10 -12
10.Install Oracle Patch (8.1.7.4) on node1.
11.Install Oracle Patch (8.1.7.4) on node2.
12.Patch the Databases on node1
Upgrade Steps 13 & 14
13.Remove Oracle 8.1.5 from node1.
14.Remove Oracle 8.1.5 from node2.
Notes:
Oracle Universal Installer
Check Path –Environment
Check Registry-remove ID# and HOME#
Check Services-oradim –delete …
Check Directory -rename Old Oracle Home
Upgrade Step 15
15.Configure OFS for both Databases on node1.
( Add & verify standalone DB’s back into OFS )
Notes:
Prior to Configuring OFS update your “hosts”file
Example: c:\winnt\system32\drivers\etc\hosts
125.1.1.0 localhost1
45.123.15.95 node1 #*hardware IP
45.123.15.96 node2 #*hardware IP
45.123.15.97 virtualnode #*virtual IP
45.123.15.98 cluster name #*Cluster IP
Upgrade Steps 16 -19
16.Test failover and failback
17.MSCS (Add additional resources required )
18. Configure Backup Agent Utility (Veritas Agent for Oracle Server Configuration Utility)
19. Use Oracle Fail Safe and “Verify Cluster”
*Don’t leave home without “Verifying the Cluster”
Do this periodically.