Update Harvester Configuration After Installation

Harvester's OS has an immutable design, which means most files in the OS revert to their pre-configured state after a reboot. The Harvester OS loads the pre-configured values of system components from configuration files during the boot time.

This page describes how to edit some of the most-requested Harvester configurations. To update a configuration, you must first update the runtime value in the system and then update configuration files to make the changes persistent between reboots.

note

If you upgrade from a version before v1.1.2, the cloud-init file in examples will be /oem/99_custom.yaml. Please substitute the value if needed.

DNS servers

Runtime change

1. Log in to a Harvester node and become root. See how to log into a Harvester node for more details.

2. Edit /etc/sysconfig/network/config and update the following line. Use a space to separate DNS server addresses if there are multiple servers.

   NETCONFIG_DNS_STATIC_SERVERS="8.8.8.8 1.1.1.1"

3. Update and reload the configuration with the following command:

   netconfig update

4. Confirm the file /etc/resolv.conf contains the correct DNS servers with the cat command:

   cat /etc/resolv.conf

5. Restart rke2-coredns:

   kubectl rollout restart deployment/rke2-coredns-rke2-coredns -n kube-system

6. Confirm rke2-coredns was rolled out successfully:

   kubectl rollout status deployment/rke2-coredns-rke2-coredns -n kube-system

Configuration persistence

Beginning with v1.1.2, the persistent name of the cloud-init file is /oem/90_custom.yaml. Harvester now uses a newer version of Elemental, which creates the file during installation.

When upgrading from an earlier version to v1.1.2 or later, Harvester retains the old file name (/oem/99_custom.yaml) to avoid confusion. You can manually rename the file to /oem/90_custom.yaml if necessary.

1. Backup the elemental cloud-init file /oem/90_custom.yaml as follows:

   cp /oem/90_custom.yaml /oem/install/90_custom.yaml.$(date --iso-8601=minutes)

2. Edit /oem/90_custom.yaml and update the value under the yaml path stages.initramfs[0].commands. The commands array must contain a line to manipulate the NETCONFIG_DNS_STATIC_SERVERS config. Add the line if the line doesn't exist.

   The following example adds a line to change the NETCONFIG_DNS_STATIC_SERVERS config:

   stages:
     initramfs:
       - commands:
           - sed -i 's/^NETCONFIG_DNS_STATIC_SERVERS.*/NETCONFIG_DNS_STATIC_SERVERS="8.8.8.8 1.1.1.1"/' /etc/sysconfig/network/config

   Replace the DNS server addresses and save the file. Harvester sets up new servers after rebooting.

NTP servers

We introduce the new mechanism for the NTP configuration in Harvester v1.2.0.

For more information about NTP settings in Harvester v1.2.0 and later versions, see the NTP servers.

Password of user rancher

Runtime change

1. Log in to a Harvester node as user rancher. See how to log into a Harvester node for more details.
2. To reset the password for the user rancher, run the command passwd.

Configuration persistence

1. Backup the elemental cloud-init file /oem/90_custom.yaml as follows:

   cp /oem/90_custom.yaml /oem/install/90_custom.yaml.$(date --iso-8601=minutes)

2. Edit /oem/90_custom.yaml and update the yaml path stages.initramfs[0].users.rancher.passwd. Refer to the configuration os.password for details on how to specify the password in an encrypted form.

Bonding slaves

You can update the slave interfaces of Harvester's management bonding interface mgmt-bo.

Runtime change

1. Log in to a Harvester node and become root. See how to log into a Harvester node for more details.

2. Identify the interface names with the following command:

   ip a

3. Edit /etc/sysconfig/network/ifcfg-mgmt-bo and update the lines associated with bonding slaves and bonding mode:

   BONDING_SLAVE_0='ens5'
   BONDING_SLAVE_1='ens6'
   BONDING_MODULE_OPTS='miimon=100 mode=balance-tlb '

4. Restart the network with the wicked ifreload command:

   wicked ifreload mgmt-bo
   caution

   A mistake in the configuration may disrupt the SSH session.

Configuration persistence

1. Backup the elemental cloud-init file /oem/90_custom.yaml as follows:

   cp /oem/90_custom.yaml /oem/install/90_custom.yaml.$(date --iso-8601=minutes)

2. Edit /oem/90_custom.yaml and update the yaml path stages.initramfs[0].files. More specifically, update the content of the /etc/sysconfig/network/ifcfg-mgmt-bo file and edit the BONDING_SLAVE_X and BONDING_MODULE_OPTS entries accordingly:

   stages:
     initramfs:
     - ...
       files:
       - path: /etc/sysconfig/network/ifcfg-mgmt-bo
         permissions: 384
         owner: 0
         group: 0
         content: |+
             STARTMODE='onboot'
             BONDING_MASTER='yes'
             BOOTPROTO='none'
             POST_UP_SCRIPT="wicked:setup_bond.sh"
   
   
             BONDING_SLAVE_0='ens5'
             BONDING_SLAVE_1='ens6'
   
             BONDING_MODULE_OPTS='miimon=100 mode=balance-tlb '
   
             DHCLIENT_SET_DEFAULT_ROUTE='no'
   
         encoding: ""
         ownerstring: ""
       - path: /etc/sysconfig/network/ifcfg-ens6
         permissions: 384
         owner: 0
         group: 0
         content: |
           STARTMODE='hotplug'
           BOOTPROTO='none'
         encoding: ""
         ownerstring: ""
   note

   If you didn't select an interface during installation, you must add an entry to initialize the interface. Please check the /etc/sysconfig/network/ifcfg-ens6 file creation in the above example. The file name should be /etc/sysconfig/network/ifcfg-<interface-name>.