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

[Xen-users] networking is slow/fails for more than one domU

  • To: <xen-users@xxxxxxxxxxxxxxxxxxx>
  • From: "Jeffrey Buell" <jbuell@xxxxxxxxxx>
  • Date: Wed, 3 Aug 2005 14:21:54 -0700
  • Delivery-date: Wed, 03 Aug 2005 21:21:26 +0000
  • List-id: Xen user discussion <xen-users.lists.xensource.com>
  • Thread-index: AcWYcV9iX3iDsbl1QuegKuXVLo89Cw==
  • Thread-topic: networking is slow/fails for more than one domU

Up to now I've been using only one domU and networking has been fine.  Now
I'm trying to use several domUs simultaneously (up to 14) and having many
networking problems.  In particular I'd like to fire off commands on all the
domUs at the same time (or at least within a couple seconds).  Even something
as simple as:

ssh dom1 hostname &
ssh dom2 hostname &

behaves poorly.  Typically one of the commands returns immediately, while the
other takes 10 seconds.  If I do several of these, some will not finish and
return a "Connection closed" message.  The pattern of which domU connections
are slow or fail is not repeatable.  Each of the domUs is fast when it is the
only one being used.  If I use "ssh -o ConnectionAttempts=10 ..." then some
of the connections will work after a few _minutes_.

This happens when the above is run both from a separate native Linux box
(directly connected with a gigE crossover) and dom0.  It also happens under
both SuSE 9.3 running Xen 2.0.5c and Fedora Core 4 running Xen unstable.
Even ping loses packets when I switch from pinging one domU to another.

Is this related to the "networking bug" that's been referred to on the lists?
I thought that was just for Xen unstable.  Is there a workaround, or some fix
that I should have made to my network configuration?


Xen-users mailing list



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