Skip to content

Latest commit

 

History

History
295 lines (250 loc) · 5.83 KB

commands.md

File metadata and controls

295 lines (250 loc) · 5.83 KB
Raw

This document is a continues WIP

#Core0/CoreX communication protocol

Core0

First process to start on bare metal. It works as a simple process manager. When starts it first configure the networking then. It also starts a local redis instance to dispatch commands to CoreX Cores.

Command structure

{
	"id": "command-id",
	"command": "command-name",
	"arguments": {}, //command arguments depends on the command itself
	"queue": "optional-queue",
	"stats_interval": 0, //optional stats gathering interval (falls to default if not set)
	"max_time": 0, //Max run time of the command, if exceeded command will be killed
	"max_restart": 0, //Max number of retries to start the command if failed before giving up
	"recurring_period": 0, //If provided command is considered recurring
	"log_levels": [int] //Log levels to store locally and not discard.
}

The Core0 Core understands a very specific set of management commands:

  • Basic Commands
    • core.ping
    • core.system
    • core.kill
    • core.killall
    • core.state
    • core.reboot
  • Info Query
    • info.cpu
    • info.disk
    • info.mem
    • info.nic
    • info.os
  • CoreX Management
    • corex.create
    • corex.list
    • corex.dispatch
    • corex.terminate
  • Bridge
    • bridge.create
    • bridge.list
    • bridge.delete
  • Disk Management
    • disk.list
    • disk.mktable
    • disk.mkpart
    • disk.rmpart
    • disk.mount
    • disk.umount
  • Btrfs Management
    • btrfs.create
    • btrfs.list
    • btrfs.subvol_create
    • btrfs.subvol_list
    • btrfs.subvol_delete

Commands arguments

core.ping

Doesn't take any arguments. returns a "pong". Main use case is to check the the core is responding.

core.system

Arguments:

{
	"name": "executable",
	"dir": "pwd",
	"args": ["command", "arguements"]
	"env": {"ENV1": "VALUE1", "ENV2": "VALUE2"},
	"stdin": "data to pass to executable over stdin"
}

Executes an arbitrary command

core.kill

Arguments:

{
    "id": "process-id-to-kill"
}

Kills a certain process giving the process ID. The process/command id is the id of the command used to start this process in the first place.

core.killall

Takes no arguments Kills ALL processes on the system. (only the ones that where started by core0 itself) and still running by the time of calling this command

core.state

Takes no arguments. Returns aggregated state of all processes plus the consumption of core0 itself (cpu, memory, etc...)

core.reboot

Takes no arguments. Immediately reboot the machine.

info.cpu

Takes no arguments. Returns information about the host CPU types, speed and capabilities

info.disk

Takes no arguments. Returns information about the host attached disks

info.mem

Takes no arguments. Returns information about the host memory.

info.nic

Takes no arguments. Returns information about the host attached nics, and IPs

info.os

Takes no arguments. Returns information about the host OS.

corex.create

Arguments:

{
    "plist": "http://url/to/plist",
    "mount": {
        "/host/directory": "/container/directory"
    },
    "network": {
        "zerotier": "zerotier network id", //options
        "bridge": [], //list of bridges names to connect to
    }
    //TODO:
    "port": {
        host_port: container_port,
    }
}

corex.list

Takes no arguments. List all available live Cores on a host.

corex.terminate

Arguemnts:

{
    "container": container_id,
}

Destroys a Core and stops the Core processes. It takes a mandatory core ID.

corex.dispatch

Arguments:

{
     "container": core_id,
     "command": {
         //the full command payload
     }
}

bridge.create

Arguments:

{
    "name": "bridge-name", //required
    "hwaddr": "MAC address" //optional
}

Creates a new bridge

bridge.list

takes no argumetns List all available bridges

bridge.delete

Arguments:

{
    "name": "bridge-name", //required
}

Delete the given bridge name

disk.list

Takes no arguments. List all block devices (similar to lsblk)

disk.mktable

Arguments:

{
    "disk": "/dev/disk", //device
    "table_type": "gpt", //partition table type
}

Creates a new partition table on device. table_type can be any value that is supported by parted mktable

disk.mkpart

Arguments:

{
    "disk": "/dev/disk", //device
    "part_type": "primary", //part_type
    "start": "1", //start sector
    "end": "100%", //end sector
}

Creates a partition on given device. part_type, start, and end values must be supported by the parted mkpart command

disk.rmpart

Arguments:

{
    "disk": "/dev/disk", //device
    "number": 1, //parition number (1 based index)
}

Removes a partition on given device with given 1 based index.

disk.mount

Arguments:

{
    "options": "auto", //mount options (required) if no options are needed set to "auto"
    "source": "/dev/part", //patition to mount
    "target": "/mnt/data", //location to mount on
}

disk.umount

Arguments:

{
    "source": "/dev/part", //partition to umount
}

btrfs.create

Create a btrfs filesystem

Arguments:

{
    "label": "FS label/name", // required
    "devices": ["/dev/sdc1", "/dev/sdc2"], // the devices, required
    "data": "data profile",
    "metadata": "metadata profile"
}

btrfs.list

List all btrfs filesystems.

Takes no argument. Return array of all filesystems.

btrfs.subvol_create

Creates a new btrfs subvolume

arguments:

{
    "path": "/path/of/subvolume" required
}

btrfs.subvol_list

List subvolume under a path

arguments:

{
    "path": "/path/of/filesystem" required
}

btrfs.subvol_delete

Delete a btrfs subvolume

arguments:

{
    "path": "/path/of/subvolume" required
}