https://bugzilla.novell.com/show_bug.cgi?id=415607
User nfbrown@novell.com added comment
https://bugzilla.novell.com/show_bug.cgi?id=415607#c35
--- Comment #35 from Neil Brown 2009-01-29 19:17:46 MST ---
I've looked at this a few times now and haven't made much progress.
There is not NFS traffic between client and server at all. Assuming
the capture is correct, that definitely shows something is wrong on the
client.
Based on the rpc tracing messages, it is conceivable that XPRT_CLOSING
has been set on the transport, but it isn't getting closed completely.
However I'm not really familiar enough with the code to know if that is
even possible.
If you feel up to rebuilding a kernel (which I suspect is unlikely)
it would be helpful to put a printk in xs_tcp_connect in net/sunrpc/xprtsock.c
like this:
/* Exit if we need to wait for socket shutdown to complete */
if (test_bit(XPRT_CLOSING, &xprt->state)) {
printk("Still closing\n");
return;
}
and run the same experiment and see if that message gets printed out.
If it does, I would know where to start looking.
If not.. I'm still lost.
A less intrusive thing that might be helpful is to run
netstat -nt
while the client is hanging and see if it mentions the connection
to the server at all, and if so, what state the connection is in.
Do this a few times over several minutes to see if it changes at all.
That is all I can suggest at the moment.
--
Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.