Re: ISCSI: ACA answer

[julian_satran@il.ibm.com]

Ralph,

That is also my work assumption - and the reason why ACA get's not
explicitly mentioned.
iSCSI per se has (almost) nothing to do with the SCSI end-node state.

However I have sympathy with David's position and I don't see why does T10
not mandate ACA for all (new) devices.

Regards,
Julo

Ralph Weber <ralphoweber@compuserve.com> on 11/11/2000 05:21:46

Please respond to ENDL_TX@computer.org

To:   ips@ece.cmu.edu
cc:
Subject:  Re: ISCSI: ACA answer




David,

I'm a little bit behind the 8 ball and not getting to do all the
reading I should.  So, I cannot claim to have read the text
behind the following exchange:

} > Reading the task management section, it sort of implies iSCSI
} > targets MUST implement ACA. Is this being mandated?
}
} Yes, ACA is being mandated, per direction from T10.
} Bridging this to non-ACA (legacy) devices is no big deal.

This is an over jealous statement of what I asked ips to do.

The iSCSI draft should clearly describe how the several elements
of ACA are handled/instanciated by the protocol so that any targets
that want to implement ACA can.

However, there is no T10 mandate to make ACA support mandatory
for targets.  The only capability that is T10 thinks is
mandatory for targets is Autosense.

Thanks.

Ralph Weber