More questions/comments on the Management spec

[Ted Ross <tross@redhat.com>]

Thanks Rob for the responses to my last question.

Some more management related items:

• What is the real distinction between the name and identity of a manageable entity?  Name is used an an index for queries.  Identity appears to be assigned by the agent/container.  What is identity for and how is it used?
• The DISCOVER-* operations all specify a possible error code of 413 (Request Entity Too Large).  Is this also needed in READALL?  I would think that READALL has a greater potential for huge responses than the DISCOVER operations.
• The "limit" attribute in the DISCOVER operations seems incomplete.  If the response size is to be limited, isn't some form of pagination needed so multiple operations can be called (i.e. start-at and limit).  Of course pagination assumes that the agent has some notion of the order of the data being discovered.
• Would it be too complicated to provide an option for a selector for queries (i.e. GET all queues that have more than zero messages)?
• What is the purpose of REGISTER and DISCOVER-MGMT-NODES?  This is not spelled out in the spec.

Thanks,

-Ted