Network Configuration of A10 HarmonyTM Controller¶
The Harmony Controller has networking configuration at the following places:
- At Harmony Controller Appliances - if they are being used
- At nodes (mostly VMs) where Harmony Controller Software is installed
- Floating IP Address
- Appliance Networking Configuration
- When Harmony Controller appliance is powered on for the first time, an IP address needs to be configured along with its gateway on the appliance management port. This step needs to be repeated for all the appliances. For exact steps, please refer the hardware manual. The appliance IP is needed to login to the appliance and start the Harmony Controller installation. Later access to this may be needed to run any Hypervisor (KVM) level command.
- Node Networking Configuration
- The VM(s) need to be configured using public bridge networking. This requires unique IP address to be assigned to each VM, and this IP address should be accessible from the local network (on which the appliance is configured). Please note that the VM IP address should not be the same as any of the other VM IP addresses or any of the appliance IP addresses.
- Floating IP Address
- A floating IP address is assigned to the elected NodeZero. In case this node fails, the floating IP address is automatically moved to the next elected NodeZero. All external communication happens through the floating IP address. This IP address must be an unused IP address in the same subnet as nodes. This feature is not available in cloud environments (AWS, Azure, GCP, OpenStack, etc.).
Changing Network Configuration¶
In the current version, IP address of any of the Harmony Controller node (NodeZero, NodeOne and NodeTwo IP) and External IP address cannot be changed after installation i.e. the Harmony Controller will stop working if IP address of any node is changed.
Other networking configuration like subnet mask, Gateway IP address, DNS server address etc. can be changed using commands provided by operating system on which Harmony Controller is installed.
Configuring Proxy Settings¶
Incase the proxy settings are set, add no_proxy parameter and required http_proxy and https_proxy parameters under /etc/environment file and restart the system for the Harmony Controller to work properly. The no_proxy parameters are added with all the Master and node IP address along with the network that we use.
Monitoring Health of the Nodes¶
All the nodes of Harmony Controller should be monitored for the health of the nodes with the system being used by the customer for monitoring other machines. If the NodeZero goes down, please contact A10 Customer Support immediately. They will help create another installation of Harmony Controller using the data backup stored by the user.
If status of any other nodes is down, try to bring up the node with the same IP address. If it is not possible to bring back the node, login to the NodeZero and run the command to delete the node from the controller cluster:
# kubectl delete node <name-of-the-node-which-went-down>
After deleting the node please contact A10 Customer Support for procedure to replace the node.
Note: When you need to reboot the nodes in a multi-node environment, you need to bring-up the NodeOne and NodeTwo first and later reboot the NodeZero.
Configuration Data Backup¶
As of now, configuration is stored at two different places - one for Harmony Controller platform and other for Thunder Device Manager (TDM).
Hourly backup is being done using cron job and being stored on Harmony Controller nodes. Path of platform backup file is /a10harmony/data/config-datastore-0/cassandra/backups/. Path for TDM backup file is /a10harmony/data/tdm-0/backup. Date and time of backup is appended in the file name for easy identification.
Customer should make arrangement to transfer the backup regularly and store it at a safe place external to the controller, so that it can be used for restore, whenever needed. Remember to collect platform backup files from each node for a 3-node controller deployment. TDM backup is only on one node.
Restoring Configuration Data from Backup¶
Scripts for restoring configuration data are provided with installer. From the installer directory and running relevant scripts will restore the configuration. Please note that backup taken from the same version of controller should be restored. Backup taken from previous version of controller may get restored fine but may have issues in functioning of controller.
Command to restore Harmony Controller Platform configuration:
# ./onprem_restore_cds.sh <Platform backup file name>
In case of 3-node controller, all 3 backup files (collected from 3 nodes) should be passed as argument to the script.
Command to restore Thunder Device Manager (TDM) configuration:
# cd utilities # ./onprem_restore_tdm.sh <TDM backup file name>