how to re-address Prime and MSE

[David Ritter]

I need to change segments for my Prime and MSE's.

I gather this is accomplished at the root kernel level?

I don't see anywhere to change the IP's from the GUI..

[Francesco Molino · Accepted solution]

Hi

For prime, connect using ssh,
stop the application service:
ncs stop
Then you can do a show run and usually the interface used is Gig1.
Go in conf t, then interface gig1 and change your ip using the command ip address. Finally restart the ncs service.

For mse, check the documentation guide (it's been a while i didn't implement any mse) and i believe you have to stop the service:
/etc/init.d/msed stp
Then run the wizard:
/opt/mse/setup/setup.sh
Restart the network (it should be done automatically)
/etc/init.d/network restart

and restart the service.
But again not sure 100% for mse, check the deployment guide it should be mentioned.

[David Ritter]

I found the following:
after NCS stop...
tcl345/admin# config
Enter configuration commands, one per line. End with CNTL/Z.
tcl345/admin(config)# interface
^
% incomplete or invalid command detected at '^' marker.

tcl345/admin(config)# interface gigabitEthernet 0
tcl345/admin(config-GigabitEthernet)#
and
tcl345/admin(config-GigabitEthernet)# ip address 172.25.162.100
255.255.255.0

Changing the hostname or IP may result in undesired side effects,
such as installed application(s) being restarted.
Are you sure you want to proceed? [y/n] y

and it went away... no 'save' etc
I had the systems team change the hosts segment and no ping.. after a
reboot, it was back at the old ip

[Francesco Molino]

Are you doing these using the VMware console?
Normally it should work. Have you tried doing a shutdown first then change the ip and finally no shut the interface?

[David Ritter]

Current attempts to re-address Prime have been from Prime's CLI.  There was some caveat that changing the Prime's IP from the Linux level brakes the Oracle DB because it is not informed of the change if done outside of Prime.

However, the next test will be to change the IP on the inside, then change the GW on the outside via Linux and lastly the hosting segment in VMw.

[David Ritter]

Almost fixed.  Someone forgot to say CLI into the Console rather than ssh to the management address.

Prime is back up and hopefully happy.  I don't seem to have valid credentials for the GUI anymore so I have no convenient idea what's going on.

I can still ssh cli into the management IP and observe NCS is alive and well.

I created fresh username's  but they don't work on the GUI

I suspect system is trying to AAA to tacacs+ and the strange device answering up from the new IP is being rejected.

the new usernames  do work on the SSH cli session.

[Francesco Molino]

On CPI 2, There was a command to run to change ip properties in Prime by going into root shell and running:

/opt/CSCOlumos/bin/oracleDBadmin.sh hostnameupdate x.x.x.x --> new ip

After restarting, you could check the file /opt/CSCOlumos/conf/messaging.properties and validate the old IP doesn't exists anymore.

Not done any re-ip in version 3 lately.

[David Ritter]

Problem solved.

Please to follow all the previous guidance from the VM console and NOT the live management IP.

And then insure you have someone around who can access the TACACS+ machine to ensure that your Prime, speaking from a new IP is not rejected. I am still awaiting that someone..  So my Prime is back on its original IP.

I have moved my MSE's to their addresses and reconstructed them in Prime.  Since the MSE's do not reference Tacacs+ there are no GUI issues.

[Francesco Molino]

Glad everything is working!