Re: iSCSI - Recovery Levels

["Mallikarjun C." <cbm@rose.hp.com>]

Julian,

As per our conversation over the phone last Friday (8/28),
I thought I am doing this error recovery editing (in fact 
I am mostly done to send the doc in a few hours to you!).

Also, I thought we both agreed that I would capture what 
was presented in London in rev08 and state clearly that it's 
still under debate.  That is exactly what I summarized and sent
an email to ips on Friday afternoon!

*Please* allow us to stick to this plan, so I can manage 
to complete my editing today.  I do not want to get into 
a debate on the striation details *until* we publish rev08 -
which I thought we plan to publish by tomorrow.

Regards.
--
Mallikarjun 


Mallikarjun Chadalapaka
Networked Storage Architecture
Network Storage Solutions Organization
MS 5668	Hewlett-Packard, Roseville.
cbm@rose.hp.com


>John,
>
>It is only a placeholder - I will put in a TBD there too!
>
>Julo
>
>John Hufferd@IBMUS
>03-09-2001 22:14
>
>To:   Julian Satran/Haifa/IBM@IBMIL
>cc:   ips@ece.cmu.edu
>From: John Hufferd/San Jose/IBM@IBMUS
>Subject:  Re: iSCSI - Recovery Levels  (Document link: Julian Satran -
>      Mail)
>
>Julian,
>The way you have written it, only two levels can be specified, since level
>one contains all the currently know levels.  Perhaps, you need to leave
>some space between Zero and Everything currently known, by assigning
>Everything to be number 2, or, 3 etc.  If we define it to be the number 2
>for instance then it is possible that we only define 0 and 2 but it would
>leave room for a 1 if one was defined.
>
>.
>.
>.
>John L. Hufferd
>Senior Technical Staff Member (STSM)
>IBM/SSG San Jose Ca
>Main Office (408) 256-0403, Tie: 276-0403,  eFax: (408) 904-4688
>Home Office (408) 997-6136
>Internet address: hufferd@us.ibm.com
>
>
>Julian Satran/Haifa/IBM@IBMIL@ece.cmu.edu on 09/03/2001 06:59:38 AM
>
>Sent by:  owner-ips@ece.cmu.edu
>
>
>To:   ips@ece.cmu.edu
>cc:
>Subject:  iSCSI - Recovery Levels
>
>
>
>There seems to be consensus around the fact the recovery leves are good and
>no
>clear consensus about hpow many should there be (2 or 3).
>
>As there is no chance to settle this until 08 gets out I suggest
>intrdocucing
>a generic key=value pair (RecoveryLevel) and remove the existing keys
>(CommmandFailoverSupport and CommandReplaySupport).
>
>RecoveryLevel will be defined as follows:
>
>01   RecoveryLevel
>
>   Use: LO
>   Who can send: Initiator and Target
>
>   RecoveryLevel=<0 to x>
>
>   Default is 0.
>
>   Initiator and target negotiate the recovery level supported.
>   The minimum of the two values is selected.
>
>   Recovery levels represent a combination of recovery capabilities.
>   Each recovery level includes all the capabilities of the lower recovery
>   levels and adds to them some new ones.
>
>   In the recovery mechanisms descriptions some specific recovery
>   capabilities are used.
>
>   Those are mapped to levels as follows:
>
>      0 - SessionRecovery
>      1 - CommandFailoverSupport and CommandReplaySupport
>      [TBD]
>
>
>
>
> Comments?
>
> Julo
>
>
>
>
>
>
>