Iscsi Group Test Period - End of 2nd day

["Barry Reinhold" <bbrtrebia@mediaone.net>]

Report from the iSCSI InterOperability Front:

The second day of testing settled down significantly from the first, at
least in terms of new issues relative to the iSCSI specification. Login is
clearly a problem area but a set of simple rules has been adopted to allow
implementors to move on to other elements of iSCSI interoperability.

In terms of progress we have completed 10 rounds of testing. The new issues
identified today are:

1. Case sensistivity - Are target names case sensative (does iSCSI = ISCSI).
This issue has already been picked up as a mailing list thread.
2. The issue of what the iSCSI target represents/can represent also caused
interoperability problems. Some devices wish to be simple and return "iscsi"
in response to the SendTargets key. This is not acceptable to some
initiators.
3. Padding was an issue in a large number of interoperability failures -
causing the iSCSI stream to get out of sync. Although there is no debate
over what the standard says it may be desirable to further underscore the
padding process.
4. If a text command is sent with F=1 and the target replies with a text
command with no additional keys but does not set F=1 is this an error? (The
consensus is no, but this behavior is not desirable.)

A larger number of SCSI issues were observed in the testing process, which
is an indication that progress was being made.

The attached pdf was distributed as a guide to login.




Barry Reinhold
Principal Architect
Trebia Networks
barry.reinhold@trebia.com
603-868-5144/603-659-0885/978-929-0830 x138

target-login3.pdf

Initiator-login3.pdf