[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Direction of Row Operations (fwd)

>>>>> Sandra McLeod writes:

Sandra> I think there is an obvious need to make row-based operations
Sandra> easier and more efficient. It's not clear to me whether the
Sandra> row operations should be implemented through additions to the
Sandra> protocol with new PDUs for each type of row operation or
Sandra> whether the RowState TC is enough.

The current set operation requires to return the received varbind
(execpt in cases where you hit message size constraints). This means
that you can't return error codes specific to the semantics of the row
set operation if something goes wrong. All you have are the standard
SNMP error codes - and they are not sufficient to report why a certain
row could not be instantiated. The manager ends up with genErr or
inconsistentValue and it has to guess what might have gone wrong.

I believe that this needs to be fixed so that we can start to write
more robuts management applications. I would also prefer to have new
operations that return the current values after the set, thus making
the set/get combos we need in several cases these days go away.


Juergen Schoenwaelder      Technical University Braunschweig
<schoenw@ibr.cs.tu-bs.de>  Dept. Operating Systems & Computer Networks
Phone: +49 531 391 3289    Bueltenweg 74/75, 38106 Braunschweig, Germany
Fax:   +49 531 391 5936    <URL:http://www.ibr.cs.tu-bs.de/~schoenw/>