This command handles a directory containing a set of services.
66-tree [ -h ] [ -z ] [ -v verbosity ] [ -l ] [ -n|R ] [ -a|d ] [ -c ] [ -S after_tree ] [ -E|D ] [ -C clone ] tree
66-tree will create, destroy, or modify a tree which dynamically handles services. 66-tree will handle the tree only for the user running the process (root/user). Any tree is completely independent from another. If you want to know what trees are currently available on the system use the 66-intree tool.
-h : prints this help.
-z : use color.
-v verbosity : increases/decreases the verbosity of the command.
-l live : changes the supervision directory of service to live. By default this will be /run/66. The default can also be changed at compile time by passing the --livedir=live option to ./configure. An existing absolute path is expected and should be within a writable and executable filesystem - likely a RAM filesystem—see 66-scandir.
-n : creates a new empty tree. The new tree must not exist on the system.
-a user : allows user to deal with the given tree. Must match an existing username on the system. Several users can be allowed by separating their names with a comma. Any user not explicitly allowed is automatically denied for the given tree. By default the user issuing the command is automatically allowed when the tree is created.—This option sets the UID and GID of the service database at compilation time.
-d user : denies user to deal with the specified tree. Must match an existing username on the system. Several users can be denied by separating their names with a comma.—Useful to revoke access that has been given before with -a.
-c : makes tree the current tree to use by default with other 66 tools. The majority of commands in the 66 tools suite have a -t option to specify the tree to use. By setting tree as the current one all other 66 tools will use it by default. You can see which tree is the current one with the 66-intree tool.
-S after_tree : Start the tree after after_tree. This tells 66-all the specific order in which the start process is applied. after_tree and tree must be already enabled. You can combine this options directly with the -E option at tree creation time. Note: If after_tree and tree have the same name, this tree will be the very first tree to start.
-E : enables tree. This allows the 66-all tool to know which tree needs to be started. If the given tree is enabled, all services in that tree will be started when you use the 66-all tool.
-D : disables tree. The exact opposite of the -E option.
-R : deletes tree. Can not be undone! This will completely remove the given tree from the system! You will not be able to retrieve any information of the deleted tree after deleting it. Services currently running on tree will be not bringed down before remove it. To do so, use the -U option in conjonction e.g. 66-tree -UR tree.
-C clone : makes a strict copy of tree named clone. Clone must not exist on the system.
The usefulness of having several trees with different services can be explained with a simplified example.
Tree1 contains dhcpcd & ntpd
Tree2 contains cups & nfs
Tree3 contains xorg, notification-daemon, gvfsd & dbus
When you boot your machine and want to use it from console only, you don't care about xorg or cups, you only care about a working internet connection. So at the base you only have Tree1 enabled. At every boot this tree and all its services will now be automatically started. Then you need to print something but for this, you also need to start the nfs daemon because your document is on some other server. Normally you would need to start cups then start nfs. Using the concept of trees you start Tree2 and everything is available. When you have finished printing your awesome document instead of stopping the needed services one by one you simply stop the whole Tree2 and all its containing services are stopped automatically. This doesn't stop here. Now you want to see a video, you need a running X server and probably several other services. Tree3 was designed just for that.
(!) Caution: AVOID manipulating manually any directories and sub directories and their containing files of a tree. CORRUPTION and ERRORS may occur by doing so. The following details can help better understand the mechanics of the 66 tools. Nevertheless the 66 suite manages these paths dynamically and its tools like 66-enable, 66-disable, 66-start, 66-dbctl etc. will automatically search these directories to find information about a required service.
When creating a new tree, by default it will be created at /var/lib/66/system/tree for root and at $HOME/.66/system/tree for any regular user.
The base name of the directory, namely /var/lib/66/ and $HOME/.66/ can be changed at compile time by passing the --with-system-dir=DIR option for root and the --with-user-dir=DIR option for any regular user to ./configure.
By enabling a service its corresponding frontend service file is parsed and the result written to a sub directory of the tree's path called servicedirs/svc for classic services and servicedirs/db/source for bundle or atomic services.
As an example for root the resulting files would, by default, be found at /var/lib/66/system/tree/servicedirs/svc for services of type classic and at /var/lib/66/system/tree/servicedirs/db/source for any service of type bundle or atomic.
When you ask to start all services of the currently enabled tree at once with the 66-all tool, the directory /var/lib/66/system/tree/servicedirs/svc is opened and the command will start any service found inside of that directory.
For services of type bundle,module and atomic, instead of opening the directory /var/lib/66/system/tree/servicedirs/db/source, the corresponding compiled database found at /var/lib/66/system/tree/servicedirs/db/tree is used. The database found at this location is the result of an automatic use of the command s6-rc-compile when enabling such a service.