Change IP adress in 800XA system
I have one 800xa system, 5.1 rev A version, with redundant conectivity server, batch, and Im server running. in one plant (plant A)
same company have another 800xa system revA system installed in Plant B (redudant conectivity, batch, Im) , but this systems are independent.
we install system B first, and then we install system A, but we use IP adresss equal for both systems in the way that per example IM server is 172.16.4.3 in both systems.
now our customer wants to connect both networks but IP adresses have conflicts, so it will be necesary to change some IP for one of the systems,
my question is how to do this changes at ip adresses without bother the system? or how minimizing the problems that this change will produce?
Voted best answer
Is complete intra-system communication required? I.e. any node on System A must be able to communicate with any node on System B?
If not, it may be easier to connect the systems by inserting additional network interface cards on select nodes, and on these cards use a different network address and no RNRP, e.g. 192.168.1.x
Changing IP addresses in System 800xA is not a documented procedure, but basically it should work as long as the control system is shut down or put in a "safe state" (where no operator HMI interactions are required for the duration it takes to make the adjustments).
Consider the below a as a skeleton for a procedure that should work (but I can not give any 100% guarantees - additional tasks may be required on your system depending on how it is setup)
0. Adhere to Murhpy's 1st law of computers: Backup often. Backup early.
1. Check RNRP health (e.g. using the RNRP Utility - search own networks for nodes with configuration errors, etc.). RNRP is required to populate the hosts files in all nodes with correct data once the IP address has been changed.
2. Verify domain controller health (dcdiag /a). Abort or postpone the procedure if problems are found. Do not continue until domain controllers are in a good (synchronizing) state.
3. Check that Node Administration Structure Node Definition aspects lists computer names (in 800xA version 2.x it was recommended to change from names to static IPs - this is no longer required, but may linger in some systems. This need to be reverted back to default (names) before any change of IP addresses can be made.
4. Halt or put the production system in a "safe state".
5. Config Wizard->Maintenance Stop on all 800xA nodes (servers). Application servers, e.g. Batch and IM may require additional "stop actions".
6. Change IP address(es) on primary domain controller*, reboot
7. Change IP address(es) on secondary domain controller(s)*, reboot
8. Verify domain controller health again (check event log and run another dcdiag /a).
9. Make sure DNS is correct. Most records should be updated automatically once the computers have got their addresses changed, but some changes might require manual efforts (at this stage it may be a Catch 22). It may become necessary to revisit this item before completing the procedure.
10. Change IP address on first aspect server*, reboot
11. Check that aspect system is starting up OK
12. Change IP address on remaining servers*, reboot & verify system is coming up
13. Change IP address on clients*, reboot & verify connection with system.
14. Check DNS once more.
15. Change IP addresses where they have been statically defined (this might not be required at all, or may be required in some places, e.g. 800xA RAC/RAS configurations, config files for Oracle networking (TNS Listener), IM Desktop client connections, etc.)
*) If explicit addressing rule is used. RNRP parameters must be updated as well.
/Stefan
Add new comment