Ok, thanks 'erichill' - that very interesting. Thanks for making the Netapp 
trace available.

On my PC at work, I have two 19" monitors, so I have been able to do a 
side-by-side comparison of the Microsoft iScsi initiator taking to the Sun 
target compared to the NetApp target (and the Netbsd target).

Ok the initials conversation sequence  for a iSCSI connection is as follows:
(There are the commands from the initiator to the target)
Two 'Login' commands, then a 'Report LUNs' command, then two 'Inquiry LUN' 
commands.
It's then that the Microsoft Initiator drops the connection to the Sun Target.
The target responds to the 'Report LUNs' and 'Inquiry LUN' with 'Data In' 
packets. So there are three 'Data In' packets before it fails.

In general the Sun target and the Netapp target are behaving the same, but 
there are some subtle differences.  The questions is which one (or maybe more 
than one) is causing the Microsoft initiator to drop the line.

Ok, here's a difference that I see in these traces:
Each command from the initiator to the target is given a number in field 
'CmdSN'.
So the sequence is:
CmdSN=1, Report LUNs
CmdSN=2, Inquiry LUN
CmdSN=3, Inquiry LUN

The Target responds with a 'Data In' which has a 'ExpCmdSN' field.
With the NetApp and Netbsd targets, the corresponding values returned in 
ExpCmdSN are 2,3 and 4.
But with the Sun target the corresponding values returned in ExpCmdSN are 1,2 
and 3.

Can any one else confirm this? Is this critical? Could this be the problem?
Thanks
Nigel Smith
 
 
This message posted from opensolaris.org
_______________________________________________
storage-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/storage-discuss

Reply via email to