[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-users] AoE server & client on same physical xen box?


  • To: "Tom Brown" <xensource.com@xxxxxxxxxxxxxxxxxxx>
  • From: "rabbtux rabbtux" <rabbtux@xxxxxxxxx>
  • Date: Tue, 22 Jan 2008 14:29:14 -0800
  • Cc: xen-users@xxxxxxxxxxxxxxxxxxx
  • Delivery-date: Tue, 22 Jan 2008 14:29:46 -0800
  • Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=BOLmD9WQOq3PVOoobSB5pJWV0m957T23obmEa/jJ1oWpaDpc4IgRjJCm+stpXTmuPW7b4QBZ+o5fRYncNLQ/Mw5qmsg9ZymtOGNFxnq4c6XIXYNCdPqcD5C4C17nlnpH8dfJIPygvBSy3ygmtSbBses4jziWGYZFxoTIz0lKWp4=
  • List-id: Xen user discussion <xen-users.lists.xensource.com>

Thanks for all the replies!  This is a known issue.  Correct me if I
am wrong, but I should be able to have a fully virtualized(HVM) DomU
act as the AoE server and then the clients can be either Dom0 or other
DomUs.  This will work without code modification since the problem is
in the vbladed code, right??


On Jan 22, 2008 1:56 PM, Tom Brown <xensource.com@xxxxxxxxxxxxxxxxxxx> wrote:
> >> >  A while back, I did some testing with Aoe on my xen boxes.  I found
> >> >  that I could not get an Aoe partition from Dom0 served to a DomU on
> >> >  the same box.  My conclusion is that you can't  get an Aoe server and
> >> >  Aoe client to work on the same box, even between Dom0 and DomU, or a
> >> >  pair of DomUs.  Is there something I might be missing?  Anyone else
> >> >  successful?
> >
> > I had to make a change to the vblade.c code... there's a packet size check
> > that seems to fail... I have no idea why it is only an issue in XEN
> > configurations, but it seems that it is... Here's the change:
> >
> > --- vblade-14/aoe.c     2006-11-20 09:48:05.000000000 -0800
> > +++ vblade-14.tom/aoe.c 2007-04-17 19:07:35.000000000 -0700
> > @@ -202,7 +202,7 @@
> >                         perror("read network");
> >                         exit(1);
> >                }
> > -               if (n < 60)
> > +               if (n < 20)
> >                         continue;
> >                 p = (Aoehdr *) buf;
> >                 if (ntohs(p->type) != 0x88a2)
> >
> > that seems to be sufficient to make vblade work "on the same box"
> >
> > -Tom
>
> Interesting... One of the other posts referred to
>
>     http://article.gmane.org/gmane.network.aoe.aoetools.general/606
>
> ... which is doing it "the other way", making the requester (kernel aoe
> driver) pad the request to 60 bytes. There of course is a kernel vblade as
> well, that would probably need a similar patch if you were going to fix it
> on the vblade side... of course, there's nothing wrong with fixing it on
> both sides... but if you are using the user space vblade program, that is
> by far the simplest place to fix it.
>
>
> -Tom
>
>
>
>
> _______________________________________________
> Xen-users mailing list
> Xen-users@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-users
>

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


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.