SORT BY:

LIST ORDER
THREAD
AUTHOR
SUBJECT


SEARCH

IPS HOME


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

    RE: iSCSI version 0.6 - Query?


    • To: "Tanjore K. Suresh" <Tanjore.Suresh@sun.com>, <ips@ece.cmu.edu>
    • Subject: RE: iSCSI version 0.6 - Query?
    • From: "Robert Griswold" <rgriswold@Crossroads.com>
    • Date: Wed, 13 Jun 2001 12:27:49 -0500
    • content-class: urn:content-classes:message
    • Content-Transfer-Encoding: 8bit
    • Content-Type: text/plain;charset="iso-8859-1"
    • Sender: owner-ips@ece.cmu.edu
    • Thread-Index: AcD0FSVoTeeGZWKdS+mc1y+wIMUJNwAF6/6Q
    • Thread-Topic: iSCSI version 0.6 - Query?

    This is a very good question.  If "silently ignore" means that at the
    time of receipt the target should not attempt to qualify the receipt of
    the command at the time it received it, but instead attempt to resolve
    this after the MaxCmdSN has been sent, then I would say it makes sense,
    but your question interests me with regards to recovery.  Since section
    6.1 requires Targets to reject PDUs in a session with format errors
    (assuming a misaligned CmdSN constitutes a format error) then it would
    seem that this section on page 12 is not in agreement with recovery.
    But, the recovery section also leaves the door open for Target mode
    implementation on recovery, so if the paragraph on page 12 over-rides
    the recovery section, it must be up to the implementer of the Target to
    attempt to recover such a misaligned command?
    
    Bob
    
    Robert Griswold
    Technologist
    Crossroads Systems, Inc.
    512-928-7272
    
     -----Original Message-----
    From: 	Tanjore K. Suresh [mailto:Tanjore.Suresh@Sun.COM] 
    Sent:	Sunday, June 10, 2001 12:26 AM
    To:	ips@ece.cmu.edu
    Subject:	iSCSI version 0.6 - Query?
    
    Hi,
    
    Page 12, Last paragraph.
    
    Regarding the command sequence number.
    
    " The target MUST silently *ignore* any command outside this range or
    duplicates with in the range that have not been flagged with the retry
    bit 
    (the X bit in the opcode)"
    
    
    Why ignore commands which are not in the expected range or duplicates
    with in the range? This type of ignoring could  cause initator to
    cause some unnecessary recovery activity which could/may end up to
    be not good for the target which is hale and healthy. 
    
    Such a  command could have been errorneously numbered or errorneously 
    checked for this condition or errnoneously delivered with  sequence 
    number change because of problems  transporting it over the internet
    fabric.
    
    Hence, my opinion is appropriate error need to be returned for such
    command
    instead of ignoring. 
    
    
    Comments?
    
    
    Thanks
    sureshtk
     
    
    I
    
    
    
    


Home

Last updated: Tue Sep 04 01:04:30 2001
6315 messages in chronological order