# Harbor Installation and Configuration Guide There are two possibilities when installing Harbor. - **Online installer:** The online installer downloads the Harbor images from Docker hub. For this reason, the installer is very small in size. - **Offline installer:** Use the offline installer if the host to which are are deploying Harbor does not have a connection to the Internet. The offline installer contains pre-built images so it is larger than the online installer. You download the installers from the **[official release](https://github.com/goharbor/harbor/releases)** page. This guide describes how to install and configure Harbor by using either the online or offline installer. The installation processes are almost the same. If you are upgrading from a previous version of Harbor, you might need to update the configuration file and migrate your data to fit the database schema of the later version. For information about upgrading, see the **[Harbor Upgrade and Migration Guide](migration_guide.md)**. In addition, the Harbor community created instructions describing how to deploy Harbor on Kubernetes. If you want to deploy Harbor to Kubernetes, see [Harbor on Kubernetes](kubernetes_deployment.md). ## Harbor Components The table below lists the components that are deployed when you deploy Harbor. |Component|Version| |---|---| |Postgresql|9.6.10-1.ph2| |Redis|4.0.10-1.ph2| |Clair|2.0.8| |Beego|1.9.0| |Chartmuseum|0.9.0| |Docker/distribution|2.7.1| |Docker/notary|0.6.1| |Helm|2.9.1| |Swagger-ui|3.22.1| ## Deployment Prerequisites for the Target Host Harbor is deployed as several Docker containers. You can therefore deploy it on any Linux distribution that supports Docker. The target host requires Docker, and Docker Compose to be installed. ### Hardware The following table lists the minimum and recommended hardware configurations for deploying Harbor. |Resource|Minimum|Recommended| |---|---|---| |CPU|2 CPU|4 CPU| |Mem|4 GB|8 GB| |Disk|40 GB|160 GB| ### Software The following table lists the software versions that must be installed on the target host. |Software|Version|Description| |---|---|---| |Docker engine|version 17.06.0-ce+ or higher|For installation instructions, see [docker engine doc](https://docs.docker.com/engine/installation/)| |Docker Compose|version 1.18.0 or higher|For installation instructions, see [docker compose doc](https://docs.docker.com/compose/install/)| |Openssl|latest is preferred|Used to generate certificate and keys for Harbor| ### Network ports Harbor requires that the following ports be open on the target host. |Port|Protocol|Description| |---|---|---| |443|HTTPS|Harbor portal and core API accept HTTPS requests on this port. You can change this port in the configuration file.| |4443|HTTPS|Connections to the Docker Content Trust service for Harbor. Only required if Notary is enabled. You can change this port in the configuration file.| |80|HTTP|Harbor portal and core API accept HTTP requests on this port. You can change this port in the configuration file.| ## Installation Procedure The installation procedure involves the following steps: 1. Download the installer. 2. Configure the **harbor.yml** file. 3. Run the **install.sh** script with the appropriate options to install and start Harbor. ## Download and Unpack the Installer 1. Go to the [Harbor releases page](https://github.com/goharbor/harbor/releases). 1. Download either the online or offline installer for the version you want to install. 1. Optionally download the corresponding `*.asc` file to verify that the package is genuine. The `*.asc` file is an OpenPGP key file. Perform the following steps to verify that the downloaded bundle is genuine. 1. Obtain the public key for the `*.asc` file.
gpg --keyserver hkps://keyserver.ubuntu.com --receive-keys 644FF454C0B4115CYou should see the message ` public key "Harbor-sign (The key for signing Harbor build)
gpg -v --keyserver hkps://keyserver.ubuntu.com --verify harbor-online-installer-version.tgz.asc- Offline installer:
gpg -v --keyserver hkps://keyserver.ubuntu.com --verify harbor-offline-installer-version.tgz.ascThe `gpg` command verifies that the signature of the bundle matches that of the `*.asc` key file. You should see confirmation that the signature is correct.
gpg: armor header: Version: GnuPG v1 gpg: assuming signed data in 'harbor-offline-installer-v1.10.0-rc2.tgz' gpg: Signature made Fri, Dec 6, 2019 5:04:17 AM WEST gpg: using RSA key 644FF454C0B4115C gpg: using pgp trust model gpg: Good signature from "Harbor-sign (The key for signing Harbor build) <jiangd@vmware.com> [unknown]1. Use `tar` to extract the installer package: - Online installer:
bash $ tar xvf harbor-online-installer-version.tgz- Offline installer:
bash $ tar xvf harbor-offline-installer-version.tgz## Configure Harbor You set system level parameters for Harbor in the `harbor.yml` file that is contained in the installer package. These parameters take effect when you run the `install.sh` script to install or reconfigure Harbor. After the initial deployment and after you have started Harbor, you perform additional configuration in the Harbor Web Portal. ### Required Parameters The table below lists the parameters that must be set when you deploy Harbor. By default, all of the required parameters are uncommented in the `harbor.yml` file. The optional parameters are commented with `#`. You do not necessarily need to change the values of the required parameters from the defaults that are provided, but these parameters must remain uncommented. At the very least, you must update the `hostname` parameter. **IMPORTANT**: Harbor does not ship with any certificates. In versions up to and including 1.9.x, by default Harbor uses HTTP to serve registry requests. This is acceptable only in air-gapped test or development environments. In production environments, always use HTTPS. If you enable Content Trust with Notary to properly sign all images, you must use HTTPS. You can use certificates that are signed by a trusted third-party CA, or you can use self-signed certificates. For information about how to create a CA, and how to use a CA to sign a server certificate and a client certificate, see **[Configuring Harbor with HTTPS Access](configure_https.md)**.
Parameter | Sub-parameters | Description and Additional Parameters |
---|---|---|
hostname |
None | Specify the IP address or the fully qualified domain name (FQDN) of the target host on which to deploy Harbor. This is the address at which you access the Harbor Portal and the registry service. For example, 192.168.1.10 or reg.yourdomain.com . The registry service must be accessible to external clients, so do not specify localhost , 127.0.0.1 , or 0.0.0.0 as the hostname. |
https |
Use HTTPS to access the Harbor Portal and the token/notification service. Always use HTTPS in production environments and environments that are not air-gapped. |
|
port |
The port number for HTTPS. The default is 443. | |
certificate |
The path to the SSL certificate. | |
private_key |
The path to the SSL key. | |
harbor_admin_password |
None | Set an initial password for the Harbor system administrator. This password is only used on the first time that Harbor starts. On subsequent logins, this setting is ignored and the administrator's password is set in the Harbor Portal. The default username and password are admin and Harbor12345 . |
database |
Use a local PostgreSQL database. You can optionally configure an external database, in which case you can disable this option. | |
password |
Set the root password for the local database. You must change this password for production deployments. | |
max_idle_conns |
The maximum number of connections in the idle connection pool. If set to <=0 no idle connections are retained. The default value is 50. If it is not configured the value is 2. | |
max_open_conns |
The maximum number of open connections to the database. If <= 0 there is no limit on the number of open connections. The default value is 100 for the max connections to the Harbor database. If it is not configured the value is 0. | |
data_volume |
None | The location on the target host in which to store Harbor's data. You can optionally configure external storage, in which case disable this option and enable storage_service . The default is /data . |
clair |
updaters_interval |
Set an interval for Clair updates, in hours. Set to 0 to disable the updates. The default is 12 hours. |
jobservice |
max_job_workers |
The maximum number of replication workers in the job service. For each image replication job, a worker synchronizes all tags of a repository to the remote destination. Increasing this number allows more concurrent replication jobs in the system. However, since each worker consumes a certain amount of network/CPU/IO resources, set the value of this attribute based on the hardware resource of the host. The default is 10. |
notification |
webhook_job_max_retry |
Set the maximum number of retries for web hook jobs. The default is 10. |
chart |
absolute_url |
Set to enabled for Chart to use an absolute URL. Set to disabled for Chart to use a relative URL. |
log |
Configure logging. | |
level |
Set the logging level to debug , info , warning , error , or fatal . The default is info . |
|
local |
Set the log retention parameters:
|
|
external_endpoint |
Enable this option to forward logs to a syslog server.
|
|
proxy |
Configure proxies to be used by Clair, the replication jobservice, and Harbor. Leave blank if no proxies are required. | |
http_proxy |
Configure an HTTP proxy, for example, http://my.proxy.com:3128 . |
|
https_proxy |
Configure an HTTPS proxy, for example, http://my.proxy.com:3128 . |
|
no_proxy |
Configure when not to use a proxy, for example, 127.0.0.1,localhost,core,registry . |
Parameter | Sub-Parameters | Description and Additional Parameters |
---|---|---|
http |
Do not use HTTP in production environments. Using HTTP is acceptable only in air-gapped test or development environments that do not have a connection to the external internet. Using HTTP in environments that are not air-gapped exposes you to man-in-the-middle attacks. | |
port |
Port number for HTTP | |
external_url |
None | Enable this option to use an external proxy. When enabled, the hostname is no longer used. |
storage_service |
By default, Harbor stores images and charts on your local filesystem. In a production environment, you might want to use another storage backend instead of the local filesystem. The parameters listed below are the configurations for the registry. See *Configuring Storage Backend* below for more information about how to configure a different backend. | |
ca_bundle |
The path to the custom root CA certificate, which is injected into the trust store of registry and chart repository containers. This is usually needed if internal storage uses a self signed certificate. | |
filesystem |
The default is filesystem , but you can set azure , gcs , s3 , swift and oss . For information about how to configure other backends, see Configuring a Storage Backend below. Set maxthreads to limit the number of threads to the external provider. The default is 100. |
|
redirect |
Set disable to true when you want to disable registry redirect |
|
external_database |
Configure external database settings, if you disable the local database option. Harbor currently only supports POSTGRES. | |
harbor |
Configure an external database for Harbor data.
|
|
clair |
Configure an external database for Clair.
|
|
notary_signer |
Configure an external database for the Notary signer database
|
|
notary_server |
|
|
external_redis |
Configure an external Redis instance. | |
host |
Hostname of the external Redis instance. | |
port |
Redis instance port. | |
password |
Password to connect to the external Redis instance. | |
registry_db_index |
Database index for Harbor registry. | |
jobservice_db_index |
Database index for jobservice. | |
chartmuseum_db_index |
Database index for Chart museum. |
{ "insecure-registries" : ["myregistrydomain.com:5000", "0.0.0.0"] }After you update `daemon.json`, you must restart both Docker Engine and Harbor. 1. Restart Docker Engine. `systemctl restart docker` 1. Stop Harbor. `docker-compose down -v` 1. Restart Harbor. `docker-compose up -d` ## Using Harbor For information on how to use Harbor, see the **[Harbor User Guide](user_guide.md)** . ## Managing Harbor Lifecycle You can use `docker-compose` to manage the lifecycle of Harbor. Some useful commands are listed below. You must run the commands in the same directory as `docker-compose.yml`. ### Stop Harbor: ``` sh $ sudo docker-compose stop Stopping nginx ... done Stopping harbor-portal ... done Stopping harbor-jobservice ... done Stopping harbor-core ... done Stopping registry ... done Stopping redis ... done Stopping registryctl ... done Stopping harbor-db ... done Stopping harbor-log ... done ``` ### Restart Harbor after Stopping: ``` sh $ sudo docker-compose start Starting log ... done Starting registry ... done Starting registryctl ... done Starting postgresql ... done Starting core ... done Starting portal ... done Starting redis ... done Starting jobservice ... done Starting proxy ... done ``` ### Reconfigure Harbor To reconfigure Harbor, stop the existing Harbor instance and update `harbor.yml`. Then run `prepare` script to populate the configuration. Finally re-create and start the Harbor instance. ``` sh $ sudo docker-compose down -v $ vim harbor.yml $ sudo prepare $ sudo docker-compose up -d ``` ### Other Commands Remove Harbor's containers while keeping the image data and Harbor's database files on the file system: ``` sh $ sudo docker-compose down -v ``` Remove Harbor's database and image data for a clean re-installation: ``` sh $ rm -r /data/database $ rm -r /data/registry ``` ### Managing the Harbor Lifecycle with Notary, Clair and Chart Repository Service If you want to install Notary, Clair and chart repository service together, you should include all the components in the prepare commands: ``` sh $ sudo docker-compose down -v $ vim harbor.yml $ sudo prepare --with-notary --with-clair --with-chartmuseum $ sudo docker-compose up -d ``` Please check the [Docker Compose command-line reference](https://docs.docker.com/compose/reference/) for more on docker-compose. ## Persistent Data and Log Files By default, registry data is persisted in the host's `/data/` directory. This data remains unchanged even when Harbor's containers are removed and/or recreated. You can edit the `data_volume` in `harbor.yml` file to change this directory. In addition, Harbor uses `rsyslog` to collect the logs for each container. By default, these log files are stored in the directory `/var/log/harbor/` on the target host. You can change the log directory in `harbor.yml`. ## Configuring Harbor to Listen on a Customized Port By default, Harbor listens on port 443(HTTPS) and 80(HTTP, if configured) for both Harbor portal and Docker commands. You can reconfigure the default ports in `harbor.yml` ## Configure Harbor with an External Database Currently, Harbor only supports PostgreSQL database. To user an external database, uncomment the `external_database` section in `harbor.yml` and fill the necessary information. You must create four databases for Harbor core, Clair, Notary server, and Notary signer. And the tables are generated automatically when Harbor starts up. ## Manage User Settings User settings are handled separately system settings. All user settings are configured in the Harbor portal or by HTTP requests at the command line. For information about using HTTP requests to configure user settings, see [Configure User Settings at the Command Line](configure_user_settings.md) to config user settings. ## Performance Tuning By default, Harbor limits the CPU usage of the Clair container to 150000 to avoid it using up all CPU resources. This is defined in the `docker-compose.clair.yml` file. You can modify this file based on your hardware configuration. ## Troubleshooting ### Harbor Doesn't Start or Functions Incorrectly When Harbor does not function correctly, run the following commands to find out if all of Harbor's containers in **UP** status: ``` $ sudo docker-compose ps Name Command State Ports ----------------------------------------------------------------------------------------------------------------------------- harbor-core /harbor/start.sh Up harbor-db /entrypoint.sh postgres Up 5432/tcp harbor-jobservice /harbor/start.sh Up harbor-log /bin/sh -c /usr/local/bin/ ... Up 127.0.0.1:1514->10514/tcp harbor-portal nginx -g daemon off; Up 80/tcp nginx nginx -g daemon off; Up 0.0.0.0:443->443/tcp, 0.0.0.0:4443->4443/tcp, 0.0.0.0:80->80/tcp redis docker-entrypoint.sh redis ... Up 6379/tcp registry /entrypoint.sh /etc/regist ... Up 5000/tcp registryctl /harbor/start.sh Up ``` If a container is not in the `Up` state, check the log file for that container in `/var/log/harbor`. For example, if the `harbor-core` container is not running, look at the `core.log` log file. ### Using nginx or Load Balancing When setting up Harbor behind an `nginx` proxy or elastic load balancing, look for the following line in `common/config/nginx/nginx.conf` and, if the proxy already has similar settings, remove it from the sections `location /`, `location /v2/` and `location /service/`. ``` sh proxy_set_header X-Forwarded-Proto $scheme; ``` Then re-deploy Harbor per the instructions in "Managing Harbor Lifecycle.