User guide
Release Notes
Redundancy / Failover
54
1. Perform a takeover of a Main Application Server running CentOS by the Backup
Application Server.
2. Reverse the takeover.
3. Perform a takeover of another Main running Red Hat Linux 9 by the Backup Application
Server.
Recommendation: It is strongly suggested to use the same operating system version on all
platforms within the same cluster (that is, Option 1 is to use the Dell PowerEdge R200 and
R210 together, Option 2 is to use the Dell PowerEdge 850 and 860 together).
Workaround:
1. Access the Application Server that has Red Hat Linux 9 by using PuTTY (available from
the Useful downloads link on iControl’s Startup page).
2. Create a temporary directory (for example, /root/tmp)
3. Copy these files from the CentOS host machine to the tmp directory of the Red Hat
Linux 9 host machine using SCP (also available from the Useful downloads link on
iControl’s Startup page):
/etc/ssh/ssh_host_rsa_key
/etc/ssh/ssh_host_rsa_key.pub
/root/.ssh/id_rsa
/root/.ssh/id_rsa.pub
4. Remove the cluster.conf files from all hosts.
5. Remove these files from all hosts:
/root/.ssh/known_hosts
/root/.ssh/authorized_keys
6. On the Red Hat Linux 9 host, copy the files by typing the following commands:
cpssh_host_rsa_keyssh_host_rsa_key.pub/etc/ssh
cpid_rsaid_rsa.pub/root/.ssh
7. Restart sshd by typing the following commands:
servicesshdrestart
8. Reconfigure the Redundancy Groups.
•
[Ref. #26389] If the eth0 and eth1 Ethernet ports of your Main Application Server are
connected to the same switch (with no VLAN to separate them) and then this Application
Server fails over, the IP address of the new Main may begin switching back and forth
between its own IP address and the IP address configured as the Free IP on the
Redundancy Configuration Form.
Workaround:
Pro-actively prevent this situation from happening by connecting your eth0
and eth1 ports to different (i.e. mutually exclusive) switches.