This is a non-trivial operation. There is a constant flow of communication between different components of the platform, and a change in the IP address of one, some or all of them must be correctly coordinated. Here we present two scenarios: changing the IP address of just the flexVDI Manager, or changing the address of both the Manager and the hosts (e.g. to move them to a different subnetwork).
Changing the IP address of the flexVDI Manager instance
This change does not require to stop the currently running guests, but we recommend you disable the access to the platform to new users until you finish this procedure.
Stop the flexvdi-agent service in every host:
This is needed so they do not try to restart the flexVDI Manager instance once its address is changed and the communication is lost.
- Log into your flexVDI Manager instance (with ssh) and edit the file
/etc/sysconfig/network-scripts/ifcfg-eth0. Modify it to your needs, changing the IPADDR field, and possibly the NETMASK and GATEWAY fields too.
Restart the network service:
You will lose communication with the manager once the new IP address is made effective. You may ssh into it again to check that the change has been successfully done.
- Fix the flexVDI Manager IP address in the files
/etc/flexvdi/flexvdi-gateway.confin every host. Also, in every other flexVDI Gateway instance you have.
Finally, restart the flexvdi-agent service in every host:
Changing the IP address of the flexVDI hosts
Changing the IP address of the flexVDI hosts is a bit more delicate process. In particular, if you have configured the OCFS2 cluster and mounted one or more OCFS2 Volumes, you will have to unmount them and stop the cluster in order to change the hosts' IP addresses. This means that you must first stop all the Guests that are using any of these Volumes. In fact, we recommend that you stop all the Guests and disable the access to the platform before continuing.
- Stop the flexvdi-agent service in every host.
- If you also want to change the IP address of the flexVDI Manager, this is the moment. Just do not restart the flexvdi-agents yet.
- In the flexVDI Dashboard, you should see all the hosts in an unreachable state. Disable all of them.
- Also with the Dashboard, change the IP address (and VDI address) of every host to its new value, and enable them again.
- If you have configured the OCFS2 cluster:
In every host, stop the o2cb service:
Your flexVDI Manager instance will be stopped if its images are being stored in an OCFS2 Volume.
- Edit the file
/etc/ocfs2/cluster.confand replace all the IP addresses with their new values. You can edit this file once and then copy it to the other hosts.
- Using flexvdi-config, change the IP address of every host. When asked to, restart the network service. If you are logged in with ssh, you will lose connection when the new IP address is made effective.
- In every host, check the /etc/hosts file to see if it needs to be updated with the new IP addresses. If you have a DNS service, you may need to change it too.
If you have reconfigured the OCFS2 cluster, restart it again in every host:
- Restart the flexvdi-agent service in every host.
- If the flexVDI Manager instance was stopped during the process (e.g. if its images were stored in an OCFS2 Volume), wait until one of the agents restarts it.
- Check that the platform is correctly configured again. You may need to fix the IP address of your Media Storages if any of them was being served from one of the hosts.