Hi Roman
I think Ross has now explained those large packets
we are seeing on the initiator.

If you were capturing the packets 'on the wire'
or on the target, you would see them broken
down into many normal size packets.

Ok, so back to the original problem...

I still suspect the amount of data being sent,
in effectively one lump, may be overwhelming the
network buffering capability at the target.

I would try increasing the /dev/tcp tuneables values,
as described here:

http://www.solarisinternals.com/wiki/index.php/Networks

It would still be useful to see a snoop capture from the target size.

And if you could capture an initial iscsi login,
(say the first hundred packets from the TCP SYN packet)
we could see what iscsi parameters are being negotiated.
The value of those parameters could affect performance.

Regards
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