Xen devel - Oct 2006 - [PATCH][RFC] do ifdown netdev before shutting down vif0

I''ve been trying to track down an occasional system hang when doing
/etc/xen/scripts/network-bridge stop.  Would it be reasonable to do the fdown
${netdev} and then set  ${vif0} down to provide ifdown with a behavior that is
as close as possible to a real network card?

Signed- off- by: Kirk Allan <kallan@novell.com>



_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

> I''ve been trying to track down an occasional system hang when
doing
> /etc/xen/scripts/network-bridge stop.  Would it be reasonable to do
the
> fdown ${netdev} and then set  ${vif0} down to provide ifdown with a
> behavior that is as close as possible to a real network card?
Does this fix the hang? 

Making changes to this script in the past has been extremely painful as
it invariably breaks one distro or another -- usually gentoo :)

Would be good to have some other eyes look at this. 

Ian

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

>>  I''ve been trying to track down an occasional system hang when
doing
>> /etc/xen/scripts/network- bridge stop.  Would it be reasonable to do
> the
>> fdown ${netdev} and then set  ${vif0} down to provide ifdown with a
>> behavior that is as close as possible to a real network card?
> 
> Does this fix the hang? 
> 
> Making changes to this script in the past has been extremely painful as
> it invariably breaks one distro or another --  usually gentoo :)
> 
> Would be good to have some other eyes look at this. 
> 
> Ian
I''ve been testing this on SLES 10 and the patch doesn''t seem
to be harmful.  But like you said, network-bridge is very fragile, so it would
be good to get input from others.
>From what I understand this may be a SLES 10 issue where the hang happens
when a) the ethernet device used for the external connection is configured for
dhcp, and b) there exists at least one process that has been started from a NFS
share.  This patch would be part one of a two part fix.  The second part,
currently being worked, would address the ifdown/ifdown-dhcp with NFS issue.
Kirk


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel