Re: iSCSI: response to second login (with same ISID)

[Stephen Bailey <steph@cs.uchicago.edu>]

Matt,

> In this case, the initiator is at fault (it performed a non-orderly
> shutdown).  Hence, when it logs in with the target, the target comes back and
> says, "bad initiator, you're already logged in with me".  The initiator can
> then say "oh yeah, ok, let's clean up" and perform an orderly shutdown of the
> session.

What if the initiator lost its mind (hence the non-orderly shutdown)?
It doesn't seem to make sense to require a new instance of an
initiator to clean up for an old one.

What's WRONG with the FCP-style behavior (PLOGI establishes a new
session, PDISC/ADISC reconnects)?

I don't see that sanity checking the initiator is really necessary in
this case.  I know sanity checking is NICE, but getting the implicit
logout seems more important.  If Login with known SID results in an
implicit logout, the target will terminate the existing connections
for the session, and the initiator will certainly hear about this!

Steph