Marvin:
As George mentions, 6363519 may result in not getting a shutdown-complete.
However, for that to happen shutdown-ack should be treated as out of the
blue (either the association is not present OR the SCTP is in the listen
or cookie_wait/cookie_echoed state) - may not be the case here.
Is it possible for you to have dtrace probes on sctp_ootb_shutdown_ack() and
sctp_shutdown_complete()[1] on the client and let us know which gets
invoked? Alternatively, if you could check whether 'sctpOutOfBlue'
counter (via netstat -sP sctp) is incremented when a shutdown-ack arrives,
that could be helpful.
Further, could you provide the output of 'netstat -sP sctp' and
'netstat -sP ip' before and after a shutdown-ack arrives on the client.
BTW, are you running both the client and server using Solaris SCTP?
-venu
[1] It could be a simple script such as:
#!/usr/sbin/dtrace -s
fbt::sctp_ootb_shutdown_ack:entry
{
stack();
}
fbt::sctp_shutdown_complete:entry
{
stack();
}
fbt::sctp_shutdown_complete:return
{
printf("%x\n", arg0);
}
On Wed, 21 Feb 2007, George Shepherd wrote:
Hi Marvin..
I suspect that this may be due to bug 6363519
http://bugs.opensolaris.org/view_bug.do?bug_id=6363519
Synopsis: SCTP must clear any H/W checksum flag in the inbound mblk before
re-using it.
The public bug report says little or nothing to relate this to
your problem, but the internal evaluation does.
A small extract of this:
"If this mblk is sent to IP by the driver, ip_sctp_input() will not
clean the flags and the resulting SHUTDOWN-COMPLETE chunk created by
sctp_ootb_shutdown_ack() will have this flag set as well. This
effectively results in an incorrect IP hdr checksum, which means the
packet will be droppped at the other end."
Can you see the SHUTDOWN-COMPLETE in snoop/ethereal on the Sun end ??
PS. the mail attachment is missing.
George
*>Date: Wed, 21 Feb 2007 04:25:52 -0800 (PST)
*>From: Marvin Castro <[EMAIL PROTECTED]>
*>Subject: [networking-discuss] SCTP: Why is there no SCTP SHUTDOWN-COMPLETE
being sent?
*>To: [email protected]
*>MIME-version: 1.0
*>X-BeenThere: [email protected]
*>Delivered-to: [email protected]
*>X-PMX-Version: 5.2.0.264296
*>X-Original-To: [email protected]
*>X-OpenSolaris-URL:
http://www.opensolaris.org/jive/message.jspa?messageID=95440&tstart=0#95440
*>X-Mailman-Version: 2.1.4
*>List-Post: <mailto:[email protected]>
*>List-Subscribe:
<http://mail.opensolaris.org/mailman/listinfo/networking-discuss>,
<mailto:[EMAIL PROTECTED]>
*>List-Unsubscribe:
<http://mail.opensolaris.org/mailman/listinfo/networking-discuss>,
<mailto:[EMAIL PROTECTED]>
*>List-Archive: <http://mail.opensolaris.org/pipermail/networking-discuss>
*>List-Help: <mailto:[EMAIL PROTECTED]>
*>List-Id: Networking General Discussion <networking-discuss.opensolaris.org>
*>
*>Good day.
*>
*>We are using the Solaris 10 implementation of SCTP. Currently, we have
encountered a problem wherein the client(218.40.33.122) does not send a
SHUTDOWN-COMPLETE message after receiving a SHUTDOWN-ACK message.
*>
*>I have attached a capture file viewable in ethereal.
*>
*>Any comment on the problem is highly appreciated
*>
*>Regards,
*>
*>Marvin
*>
*>
*>This message posted from opensolaris.org
George Shepherd
http://clem.uk/~georges/
==============================================================================
Solaris Revenue Product Engineering: | SUN Microsystems
Core team -Internet | Guillemont Park
Email: [EMAIL PROTECTED] | Camberley GU17 9QG
Disclaimer: Less is more, more or less | United Kingdom
==============================================================================
_______________________________________________
networking-discuss mailing list
[email protected]
_______________________________________________
networking-discuss mailing list
[email protected]
