Changing Public IP, VIP and SCAN IP in 11gR2 RAC

The post list the steps for changing the public IP, VIP and SCAN IP. It must be mentioned that only the IPs are changed in this case. The subnet mask, the public and VIP interface all remain the same. Also private interconnect is not part of the change. Steps were tested on the 11gR2 RAC on RHEL 6
The IP change is as follows. The SCAN IP is set up in /etc/hosts (not supported by Oracle).

Name	Old IP	New IP
rhel6m1	192.168.0.85	192.168.0.93
rhel6m2	192.168.0.86	192.168.0.94
rhel6m-scan	192.168.0.91	192.168.0.92

The steps could be done parallel (on all nodes) or in a rolling fashion on each node. Steps executed on one node is shown here and changes were done parallel on all nodes. Since the information modified is stored in OCR it is better to take a manual backup of the OCR before proceeding.
1. First step is to change the VIP configuration. Current VIP configuration is

srvctl config nodeapps
Network exists: 1/192.168.0.0/255.255.255.0/eth0, type static
VIP exists: /rhel6m1-vip/192.168.0.89/192.168.0.0/255.255.255.0/eth0, hosting node rhel6m1
VIP exists: /rhel6m2-vip/192.168.0.90/192.168.0.0/255.255.255.0/eth0, hosting node rhel6m2

2. Stop the database instance and the VIP resource

[oracle@rhel6m1 ~]$ srvctl stop instance -d std11g2 -i std11g21
[grid@rhel6m1 ~]$ srvctl stop vip -n `hostname -s` -f

3. Verify VIP resource is down by using ifconfig on the OS and using crsctl

crsctl stat res -t 

ora.rhel6m1.vip
      1        OFFLINE OFFLINE
ora.rhel6m2.vip
      1        OFFLINE OFFLINE

4. Add the new IP information to /etc/hosts (and DNS) and proceed to modifying the VIP. The modify command must be run as root user

[root@rhel6m1 grid]# srvctl modify nodeapps -n rhel6m1 -A 192.168.0.97/255.255.255.0/eth0

Verify that changes are taken effect

[grid@rhel6m2 ~]$ srvctl config nodeapps -a
Network exists: 1/192.168.0.0/255.255.255.0/eth0, type static
VIP exists: /rhel6m1-vip/192.168.0.97/192.168.0.0/255.255.255.0/eth0, hosting node rhel6m1
VIP exists: /rhel6m2-vip/192.168.0.98/192.168.0.0/255.255.255.0/eth0, hosting node rhel6m2

If the VIP is now on a different network then do not start VIP until public IP and SCAN IP is also changed into the same network.




5. Next is to update the SCAN IP information. Current SCAN IP configuration

srvctl config scan
SCAN name: rhel6m-scan, Network: 1/192.168.0.0/255.255.255.0/eth0
SCAN VIP name: scan1, IP: /rhel6m-scan/192.168.0.91

6. Stop the SCAN IP and the scan lister and verify it is stopped.

[root@rhel6m1 grid]# srvctl stop scan_listener

[root@rhel6m1 grid]# srvctl stop scan

[root@rhel6m1 grid]# srvctl status scan
SCAN VIP scan1 is enabled
SCAN VIP scan1 is not running

[root@rhel6m1 grid]# srvctl status scan_listener
SCAN Listener MYLISTENER_SCAN1 is enabled
SCAN listener MYLISTENER_SCAN1 is not running

7. Modify the SCAN IP by executing the modify command as root user with the same SCAN name. (It is assumed that if DNS is used, the SCAN name now resolve to new IPs) This will pick up the new IP and update cluster records with the new SCAN IP.

[root@rhel6m1 grid]# srvctl modify scan -n rhel6m-scan

8. Verify the changes have taken effect

[root@rhel6m1 grid]# srvctl config scan
SCAN name: rhel6m-scan, Network: 1/192.168.0.0/255.255.255.0/eth0
SCAN VIP name: scan1, IP: /rhel6m-scan/192.168.0.92

If the SCAN IP is now on a different network than the public IP then don't start the SCAN IP until the pubilc IP is also changed.

9. Updating public IP requires no cluster related work. Shutdown the cluster stack

[root@rhel6m1 grid]# crsctl stop crs

10. Change IP address on OS (eg. editing ifcfg-eth*) files and restart the network for the changes to take effect. Once the IP changes are made it maybe worthwhile to do ssh between nodes using the new IP so it is added to known host list and no prompts are shown during the ssh between the nodes. As the final step restart the cluster stack with

# crsctl start crs

Useful metalink notes
How to update the IP address of the SCAN VIP resources (ora.scan.vip) [ID 952903.1]
Modifying the VIP or VIP Hostname of a 10g or 11g Oracle Clusterware Node [ID 276434.1]
How to Modify Public Network Information including VIP in Oracle Clusterware [ID 276434.1]

Related Posts
Changing Listener and SCAN Listener Port in 11gR2 RAC
Changing Listener and SCAN Listener Name in 11gR2 RAC

VIP, Dataguard Broker & RAC

In a RAC environment if VIP isn't used for client connection, it will defeat the whole purpose of using VIPs which give fast notification when a node fails. If not have to wait until a TCP timeout occurs which could take a while.

In RAC if the default listener is on port of 1521, then DBCA will not automatically set
the LOCAL_LISTENER. However it will default to a connection string that refers to the physical host address (not virtual address).

Oracle 11.1.0.7 and above has a new dataguard broker parameter StaticConnectIdentifier StaticConnectIdentifier configurable instance-specific property specifies the connection identifier that the DGMGRL client will use when starting database instances. The default value for this is the concatenation of The ADDRESS attribute value of the listener that is specified for the LOCAL_LISTENER initialization parameter and the value for the SERVICE_NAME attribute will be set to a concatenation of db_unique_name_DGMGRL.db_domain.

Therefore if LOCAL_LISTENER is not set explicitly for VIP address it will use the physical host address. Could be observed with

DGMGRL> show instance "TBXA04B1" StaticConnectIdentifier
StaticConnectIdentifier =
'(DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=hostname)(PORT=1521))(CONNECT_DATA=(SERVICE_NAME=TBXA04B_DGMGRL)(INSTANCE_NAME=TBXA04B1)(SERVER=DEDICATED)))'

Always better to explicitly set the LOCAL_LISTENER to VIP.

StaticConnectIdentifier is added to 11.1 with the patchset of 11.1.0.7. Documentation of it is available with 11.2