You should have received access to the Debian package (for example linkahead-0.11.1.deb) via your subscription or from the download page. For easier maintenance, LinkAhead uses Docker for deployment. The Debian package does however not contain the Docker image (since the image is quite large) and the image will be downloaded during the installation process from a docker registry. If you have a subscription, the installer will ask you for the credentials to the docker registry.

On a Debian or Ubuntu system, install the package with: sudo apt install ./linkahead-0.11.1.deb (replace the file name with your exact version).


apt may inform you about packages that can be removed now. Note that there is a bug in older Ubuntu versions which may lead to inadvertently removing your graphics drivers when you follow that suggestion.

The installation procedure will ask if you have a subscription and may ask for your LinkAhead subscription credentials (user name and token). It will automatically install, enable and start LinkAhead as a systemd service, so that LinkAhead will be started automatically on reboot in the future. You can also start the LinkAhead service manually using

$ systemctl start linkahead.service

To shut down LinkAhead, analogously type systemctl stop linkahead.service.


Instead of a systemd service, LinkAhead can also be started completely manually. In that case, the path to the profile (see Configuration) and further options have to be given manually. The command to start LinkAhead then is linkahead -p path/to/profile.yml start --no-build; it is stopped with linkahead -p path/to/profile.yml stop.

Typically, the required docker image will be downloaded on first startup, if it does not yet exist on the system. This may take a while. You can monitor the progress of the download and the subsequent start with:

$ journalctl -u linkahead -f

Once the log output says Starting org.caosdb.server.CaosDBServer application the installation is done and LinkAhead is running with the default configuration. You can check whether the LinkAhead server can be accessed by sending a curl request on the same machine:

$ curl -k https://localhost:10443
<?xml version="1.0" encoding="UTF-8"?>
<?xml-stylesheet type="text/xsl" href="https://localhost:10443/webinterface/1614339482/webcaosdb.xsl" ?>
<Response srid="11883904-301d-4bd9-8f74-86d95837239b" timestamp="1614339817505" baseuri="https://localhost:10443">

If the command shows a similar empty reply to the one above, the installation was successful and you can continue with the configuration.


Requirements are installed automatically. You can list them via apt-cache depends linkahead


The path to the profile with which LinkAhead is started can be set in /etc/linkahead/linkahead.conf. The default profile is located at /usr/share/linkahead/profiles/default/profile.yml. See Configuration for more details on how to configure your profile.


Unable to fetch some archives

If apt finishes with this error, run apt update and try the installation again.

Old data still exists

Data is kept in persistent docker volumes. You can remove existing data the following way:


This cannot be undone. Make sure you have backups.

Have a look at the docker volumes: docker volume ls. The name depends on your profile, but typically you want to remove the following two: docker volume rm default_caosdb-sqldata default_caosdb-caosroot

Strange characters in my browser

If you see strange characters in your browser when trying to access LinkAhead, you should explicitly specify https, like so: https://localhost:10443 If you try to access LinkAhead via plain http (http://localhost:10443), your browser will receive encrypted content and display it like this:

Screenshot of what a browser may show if you try to access LinkAhead via plain http.

docker-compose is not recent enough

You can install an up-to-date version with (at least 1.27 is required):

$ pip install --user docker-compose

Extroot from NFS mounts fails

If LinkAhead fails to start when trying to use an NFS mount as extroot, possibly with errors like Cannot create container for service caosdb-server: invalid mount config for type "bind": stat /data/foo/bar: permission denied, then restrictive access permissions may be the reason. In this example, the directory /data/foo/bar could not be accessed, because the mounted NFS resource /data/foo/ was missing permissions.

As a workaround, you could mount /data/foo/bar directly instead of relying on the /data/foo ressource:

NFS line for /etc/fstab
server:/volume/foo/bar    /data/foo/bar    nfs   defaults    0 0

Or you could create a local bind mount which exposes /data/foo/bar

Mount /data/foo/bar to /my_extroots/bar, where /my_extroots/ should be readable to root.
# mount --bind /data/foo/bar /my_extroots/bar

The screen is black after the next reboot

This is not a bug in LinkAhead, but may be because you followed apt’s recommendation to automatically remove all packages which are deemed unnecessary by the system. This may remove your graphics drivers. It is a known issue in Ubuntu 20.04 and will hopefully be fixed soon. In the meantime you may want to manually reinstall your graphics driver.