utilities.py 文件源码

def generate_analysis_target_id(analysis_target, name):
    # this function generates 'machine-id's for analysis target's that
    # might not be hosts.
    #
    # 'machine_id' is what Insights uses to uniquely identify
    # the thing-to-be-analysed.  Primarily it determines when two uploads
    # are for the 'same thing', and so the latest upload should update the
    # later one Up till now that has only been hosts (machines), and so a
    # random uuid (uuid4) machine-id was generated for the host as its machine-id,
    # and written to a file on the host, and reused for all insights
    # uploads for that host.
    #
    # For docker images and containers, it will be difficult to impossible
    # to save their machine id's anywhere.  Also, while containers change
    # over time just like hosts, images don't change over time, though they
    # can be rebuilt.  So for images we want the 'machine-id' for an 'image'
    # to follow the rebuilt image, not change every time the image is rebuilt.
    # Typically when an image is rebuilt, the rebuilt image will have the same
    # name as its predicessor, but a different version (tag).
    #
    # So for images and containers, instead of random uuids, we use namespace uuids
    # (uuid5's).  This generates a new uuid based on a combination of another
    # uuid, and a name (a character string).  This will always generate the
    # same uuid for the same given base uuid and name.  This saves us from
    # having to save the image's uuid anywhere, and lets us control the created uuid
    # by controlling the name used to generate it.  Keep the name and base uuid) the
    # same, we get the same uuid.
    #
    # For the base uuid we use the uuid of the host we are running on.
    # For containers this is the obvious choice, for images it is less obviously
    # what base uuid is correct.  For now we will just go with the host's uuid also.
    #
    # For the name, we leave that outside this function, but in general it should
    # be the name of the container or the name of the image, and if you want to
    # replace the results on the insights server, you have to use the same name

    if analysis_target == "host":
        return generate_machine_id()
    elif (analysis_target == "docker_image" or
            analysis_target == "docker_container" or
            analysis_target == "compressed_file" or
            analysis_target == "mountpoint"):
        return generate_container_id(name)
    else:
        raise ValueError("Unknown analysis target: %s" % analysis_target)