Settings

This page contains a list of advanced settings which can be used in Harvester. You can modify the custom resource settings.harvesterhci.io from the Dashboard UI or with the kubectl command.

additional-ca

This setting allows you to configure additional trusted CA certificates for Harvester to access external services.

Default: none

Example

  1. -----BEGIN CERTIFICATE-----
  2. SOME-CA-CERTIFICATES
  3. -----END CERTIFICATE-----

Settings - 图1caution

Changing this setting might cause a short downtime for single-node clusters.

auto-disk-provision-paths [Experimental]

This setting allows Harvester to automatically add disks that match the given glob pattern as VM storage. It’s possible to provide multiple patterns by separating them with a comma.

Settings - 图2note

This setting only adds formatted disks mounted to the system.

Settings - 图3caution

  • This setting is applied to every Node in the cluster.
  • All the data in these storage devices will be destroyed. Use at your own risk.

Default: none

Example

The following example will add disks matching the glob pattern /dev/sd* or /dev/hd*:

  1. /dev/sd*,/dev/hd*

backup-target

This setting allows you to set a custom backup target to store VM backups. It supports NFS and S3. For further information, please refer to the Longhorn documentation.

Default: none

Example

  1. {
  2.   "type": "s3",
  3.   "endpoint": "https://s3.endpoint.svc",
  4.   "accessKeyId": "test-access-key-id",
  5.   "secretAccessKey": "test-access-key",
  6.   "bucketName": "test-bup",
  7.   "bucketRegion": "us‑east‑2",
  8.   "cert": "",
  9.   "virtualHostedStyle": false
  10. }

cluster-registration-url

This setting allows you to import the Harvester cluster to Rancher for multi-cluster management.

Default: none

Example

  1. https://172.16.0.1/v3/import/w6tp7dgwjj549l88pr7xmxb4x6m54v5kcplvhbp9vv2wzqrrjhrc7c_c-m-zxbbbck9.yaml

containerd-registry

This setting allows you to configure a private registry for the Harvester cluster. The value will be set in /etc/rancher/rke2/registries.yaml of each node. You can read RKE2 - Containerd Registry Configuration for more information.

Settings - 图4note

If you set a username and password for a private registry, the system will automatically remove it to protect the credential after the system saves it in registries.yaml.

Example

containerd-registry

  1. {
  2.   "Mirrors": {
  3.     "docker.io": {
  4.       "Endpoints": ["https://myregistry.local:5000"],
  5.       "Rewrites": null
  6.     }
  7.   },
  8.   "Configs": {
  9.     "myregistry.local:5000": {
  10.       "Auth": {
  11.         "Username": "testuser",
  12.         "Password": "testpassword"
  13.       },
  14.       "TLS": {
  15.         "InsecureSkipVerify": false
  16.       }
  17.     }
  18.   }
  19. }

csi-driver-config

Available as of v1.2.0

If you install third-party CSI drivers in the Harvester cluster, you must configure some necessary information through this setting before using Backup & Snapshot related features.

Default:

  1. {
  2.   "driver.longhorn.io": {
  3.     "volumeSnapshotClassName": "longhorn-snapshot",
  4.     "backupVolumeSnapshotClassName": "longhorn"
  5.   }
  6. }
  1. Add the provisioner for the newly added CSI driver.
  2. Configure Volume Snapshot Class Name, which refers to the name of the VolumeSnapshotClass used to create volume snapshots or VM snapshots.
  3. Configure Backup Volume Snapshot Class Name, which refers to the name of the VolumeSnapshotClass used to create VM backups.

default-vm-termination-grace-period-seconds

Available as of v1.2.0

This setting allows you to specify a default termination grace period for stopping a virtual machine in seconds.

Default: 120

http-proxy

This setting allows you to configure an HTTP proxy to access external services, including the download of images and backup to s3 services.

Default: {}

The following options and values can be set:

  • Proxy URL for HTTP requests: "httpProxy": "http://<username>:<pswd>@<ip>:<port>"
  • Proxy URL for HTTPS requests: "httpsProxy": "https://<username>:<pswd>@<ip>:<port>"
  • Comma-separated list of hostnames and/or CIDRs: "noProxy": "<hostname | CIDR>"

Settings - 图6caution

If you configure httpProxy and httpsProxy, you must also put Harvester node’s CIDR into noProxy, otherwise the Harvester cluster will be broken. If you also configure cluster-registration-url, you usually need to add the host of cluster-registration-url to noProxy as well, otherwise you cannot access the Harvester cluster from Rancher.

Example

  1. {
  2.   "httpProxy": "http://my.proxy",
  3.   "httpsProxy": "https://my.proxy",
  4.   "noProxy": "some.internal.svc,172.16.0.0/16"
  5. }

Settings - 图7note

Harvester appends necessary addresses to user configured no-proxy to ensure the internal traffic works. i.e., localhost,127.0.0.1,0.0.0.0,10.0.0.0/8,longhorn-system,cattle-system,cattle-system.svc,harvester-system,.svc,.cluster.local. harvester-system was added into the list since v1.1.2.

Settings - 图8caution

Changing this setting might cause a short downtime for single-node clusters.

log-level

This setting allows you to configure the log level for the Harvester server.

Default: info

The following values can be set. The list goes from the least to most verbose log level:

  • panic
  • fatal
  • error
  • warn, warning
  • info
  • debug
  • trace

Example

  1. debug

ntp-servers

Available as of v1.2.0

This setting allows you to configure NTP servers for time synchronization on the Harvester nodes.

Using this setting, you can define NTP servers during installation or update NTP servers after installation.

Settings - 图9caution

Modifying the NTP servers will replace the previous values for all nodes.

Default: “”

Example

  1. {
  2.   "ntpServers": [
  3.     "0.suse.pool.ntp.org",
  4.     "1.suse.pool.ntp.org"
  5.   ]
  6. }

overcommit-config

This setting allows you to configure the percentage for resources overcommit on CPU, memory, and storage. By setting resources overcommit, this will permit to schedule additional virtual machines even if the the physical resources are already fully utilized.

Default: { "cpu":1600, "memory":150, "storage":200 }

The default CPU overcommit with 1600% means, for example, if the CPU resources limit of a virtual machine is 1600m core, Harvester would only request 100m CPU for it from Kubernetes scheduler.

Example

  1. {
  2.   "cpu": 1000,
  3.   "memory": 200,
  4.   "storage": 300
  5. }

release-download-url

Available as of v1.0.1

This setting allows you to configure the upgrade release download URL address. Harvester will get the ISO URL and checksum value from the ${URL}/${VERSION}/version.yaml file hosted by the configured URL.

Default: https://releases.rancher.com/harvester

Example of the version.yaml

  1. apiVersion: harvesterhci.io/v1beta1
  2. kind: Version
  3. metadata:
  4.   name: ${VERSION}
  5.   namespace: harvester-system
  6. spec:
  7.   isoChecksum: ${ISO_CHECKSUM}
  8.   isoURL: ${ISO_URL}

server-version

This setting displays the version of Harvester server.

Example

  1. v1.0.0-abcdef-head

ssl-certificates

This setting allows you to configure serving certificates for Harvester UI/API.

Default: {}

Example

  1. {
  2.   "ca": "-----BEGIN CERTIFICATE-----\nSOME-CERTIFICATE-ENCODED-IN-PEM-FORMAT\n-----END CERTIFICATE-----",
  3.   "publicCertificate": "-----BEGIN CERTIFICATE-----\nSOME-CERTIFICATE-ENCODED-IN-PEM-FORMAT\n-----END CERTIFICATE-----",
  4.   "privateKey": "-----BEGIN RSA PRIVATE KEY-----\nSOME-PRIVATE-KEY-ENCODED-IN-PEM-FORMAT\n-----END RSA PRIVATE KEY-----"
  5. }

Settings - 图10caution

Changing this setting might cause a short downtime on single-node clusters.

ssl-parameters

This setting allows you to change the enabled SSL/TLS protocols and ciphers of Harvester GUI and API.

The following options and values can be set:

  • protocols: Enabled protocols. See NGINX Ingress Controller’s configs ssl-protocols for supported input.

  • ciphers: Enabled ciphers. See NGINX Ingress Controller’s configs ssl-ciphers for supported input.

If no value is provided, protocols is set to TLSv1.2 only and the ciphers list is ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES256-GCM-SHA384:ECDHE-ECDSA-CHACHA20-POLY1305.

Default: none

Settings - 图11note

See Troubleshooting if you have misconfigured this setting and no longer have access to Harvester GUI and API.

Example

The following example sets the enabled SSL/TLS protocols to TLSv1.2 and TLSv1.3 and the ciphers list to ECDHE-ECDSA-AES128-GCM-SHA256 and ECDHE-ECDSA-CHACHA20-POLY1305.

  1. {
  2.   "protocols": "TLSv1.2 TLSv1.3",
  3.   "ciphers": "ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-ECDSA-CHACHA20-POLY1305"
  4. }

storage-network

By default, Longhorn uses the default management network in the Harvester cluster that is limited to a single interface and shared with other workloads cluster-wide. This setting allows you to configure a segregated storage network when network isolation is preferred.

For details, please refer to the Harvester Storage Network

Settings - 图12caution

Any change to storage-network requires shutting down all VMs before applying this setting. IP Range should be IPv4 CIDR format and 4 times the number of your cluster nodes.

Default: “”

Example

  1. {
  2.   "vlan": 100,
  3.   "clusterNetwork": "storage",
  4.   "range": "192.168.0.0/24"
  5. }

support-bundle-image

Available as of v1.2.0

This setting allows you to configure the support bundle image, with various versions available in rancher/support-bundle-kit.

Default:

  1. {
  2.   "repository": "rancher/support-bundle-kit",
  3.   "tag": "v0.0.25",
  4.   "imagePullPolicy": "IfNotPresent"
  5. }

support-bundle-namespaces

Available as of v1.2.0

This setting allows you to specify additional namespaces when collecting a support bundle. The support bundle will only capture resources from pre-defined namespaces by default.

Here is the pre-defined namespaces list:

  • cattle-dashboards
  • cattle-fleet-local-system
  • cattle-fleet-system
  • cattle-fleet-clusters-system
  • cattle-monitoring-system
  • fleet-local
  • harvester-system
  • local
  • longhorn-system
  • cattle-logging-system

If you select more namespaces, it will append to the pre-defined namespaces list.

Default: none

support-bundle-timeout

Available as of v1.2.0

This setting lets you define the support bundle’s default timeout duration in minutes. Use 0 to disable the timeout feature.

Default: 10

upgrade-checker-enabled

This setting allows you to automatically check if there’s an upgrade available for Harvester.

Default: true

Example

  1. false

upgrade-checker-url

This setting allows you to configure the URL for the upgrade check of Harvester. Can only be used if the upgrade-checker-enabled setting is set to true.

Default: https://harvester-upgrade-responder.rancher.io/v1/checkupgrade

Example

  1. https://your.upgrade.checker-url/v99/checkupgrade

vip-pools

Deprecated as of v1.2.0, use IP Pool instead

This setting allows you to configure the global or namespace IP address pools of the VIP by CIDR or IP range.

Default: {}

Settings - 图13note

Configuring multi-CIDR or IP range from UI is only available from Harvester v1.1.1.

Example

  1. {
  2.   "default": "172.16.0.0/24,172.16.1.0/24",
  3.   "demo": "172.16.2.50-172.16.2.100,172.16.2.150-172.16.3.200"
  4. }

vm-force-reset-policy

This setting allows you to force reschedule VMs when a node is unavailable. When a node turns to be Not Ready, it will force delete the VM on that node and reschedule it to another available node after a period of seconds.

Default: {"enable":true, "period":300}

Settings - 图14note

When a host is unavailable or is powered off, the VM only reboots and does not migrate.

Example

  1. {
  2.   "enable": "true",
  3.   "period": 300
  4. }

UI Settings

branding

Available as of v1.2.0

This setting allows you to globally re-brand the UI by customizing the Harvester product name, logos, and color scheme.

Default: Harvester

containerd-registry

You can set the following options and values:

  • Private Label: This option replaces “Harvester” with the value you provide in most places.
  • Logo: Upload light and dark logos to replace the Harvester logo in the top-level navigation header.
  • Favicon: Upload an icon to replace the Harvester favicon in the browser tab.
  • Primary Color: You can override the primary color used throughout the UI with a custom color of your choice.
  • Link Color: You can override the link color used throughout the UI with a custom color of your choice.

ui-index

This setting allows you to configure the HTML index location for the UI.

Default: https://releases.rancher.com/harvester-ui/dashboard/latest/index.html

Example

  1. https://your.static.dashboard-ui/index.html

ui-plugin-index

This setting allows you to configure the JS address for the Harvester plugin (when accessing Harvester from Rancher).

Default: https://releases.rancher.com/harvester-ui/plugin/harvester-latest/harvester-latest.umd.min.js

Example

  1. https://your.static.dashboard-ui/*.umd.min.js

ui-source

This setting allows you to configure how to load the UI source.

You can set the following values:

  • auto: The default. Auto-detect whether to use bundled UI or not.
  • external: Use external UI source.
  • bundled: Use the bundled UI source.

Example

  1. external