SORT BY:

LIST ORDER
THREAD
AUTHOR
SUBJECT


SEARCH

IPS HOME


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

    RE: A couple of iFCP questions



    > It is safe to discard a translation when there are no active N_PORT
    > sessions for a remote N_PORT.  In that case, the iFCP gateway SHOULD
    > invalidate and remove the mapping to that remote N_PORT.
    
    I don't believe that is always correct.  I know of Fibre Channel systems
    that do something approximating a mainframe-style disconnect/reconnect
    based on dropping the N_PORT session and opening up a new one if the
    gap in I/O activity is sufficient.  I suspect things may be even worse in
    the
    tape world where the gap between sessions may be much longer (e.g.,
    when the tape device is being sequentially shared among a set of
    systems); is anyone on this list a tape expert who knows for certain? 
    Prematurely discarding a translation will cause unexpected results in
    these cases.
    
    OTOH, there should be a solution along the lines of accumulating a
    large number of "should be invalid" translations, discarding them in a
    single batch and issuing the appropriate state change notifications.
    Care is still required, as this "correct" operation of the (logical/virtual)
    fabric may still have undesirable higher-level consequences (e.g., backup
    application aborts).  I should note that this approach is unique to
    Fibre Channel, as the networking analogs (e.g., running UDP through
    a NAT) don't have a facility like state change notification available, and
    hence generally can't recover from incorrectly discarding a translation.
    
    > Of course,
    > even if it doesn't, a device establishing a new N_PORT session has the
    > responsibility of validating any pre-existing mapping by checking WWPN's
    > of the remote device.
    
    If "device" refers to the Fibre Channel device containing the N_PORT, I'm
    not
    sure this check is (always) done.  OTOH, if a mapping has changed, the state
    change notification mechanism described in the response to the second
    question will prevent this sort of problem.
    
    The explanation of iSNS's role in change propagation seems fine, although
    it makes connectivity to iSNS a requirement for operation of an iFCP
    gateway -- in other words, if the iFCP gateway loses contact with iSNS,
    it loses the ability to detect that its translations have become invalid.
    I mostly wanted to note the contrast in availability requirements with DNS -
    DNS does not have to be available after a connection has been set up.
    
    --David
    ---------------------------------------------------
    David L. Black, Senior Technologist
    EMC Corporation, 42 South St., Hopkinton, MA  01748
    +1 (508) 435-1000 x75140     FAX: +1 (508) 497-8500
    black_david@emc.com       Mobile: +1 (978) 394-7754
    ---------------------------------------------------
    
    


Home

Last updated: Tue Sep 04 01:06:00 2001
6315 messages in chronological order