there's a server A and server B which are configured as HA and their gateway

The preferred approach would be to follow what Sathya recommended in the last note.

If you MUST keep one server up, it is possible to do so. But you will need to eventually schedule a downtime before you bring the engines up on Server B, since the engine database files needs to be synced.

Here’s what you can do:
1)          Check-point the environment when both servers are up
2)          Stop SERVER B
3)          At this point all the gateways on SERVER A will switch to [STANDALONE] to handle all the requests/transactions
4)          Make any necessary server/configurations changes that you like for SERVER B
5)          Once you are done with the changes, use “stop-suite.bat|sh” to stop the engines on SERVER A.
6)          Stop-suite.bat|sh will ensure that the kdb is copied over to both /gw1 and /gw2 directories and both gateways will be on the same engine database id (no mismatch in kdb's).
7)          Ensure that that file was correctly copied over to the NFS that is hosting the kdb’s by looking at the /gw1 and /gw2 directories.
8)          If the result of the above step is satisfactory, start engines on SERVER A
9)          Start engines on SERVER B

This will significantly reduce the downtime in your environment, while you make the SERVER change.

Hope this helps!

The preferred approach would be to follow what Sathya recommended in the last note.

If you MUST keep one server up, it is possible to do so. But you will need to eventually schedule a downtime before you bring the engines up on Server B, since the engine database files needs to be synced.

Here’s what you can do:
1)          Check-point the environment when both servers are up
2)          Stop SERVER B
3)          At this point all the gateways on SERVER A will switch to [STANDALONE] to handle all the requests/transactions
4)          Make any necessary server/configurations changes that you like for SERVER B
5)          Once you are done with the changes, use “stop-suite.bat|sh” to stop the engines on SERVER A.
6)          Stop-suite.bat|sh will ensure that the kdb is copied over to both /gw1 and /gw2 directories and both gateways will be on the same engine database id (no mismatch in kdb's).
7)          Ensure that that file was correctly copied over to the NFS that is hosting the kdb’s by looking at the /gw1 and /gw2 directories.
8)          If the result of the above step is satisfactory, start engines on SERVER A
9)          Start engines on SERVER B

This will significantly reduce the downtime in your environment, while you make the SERVER change.

Hope this helps!