openstack team mailing list archive

[Xin Zhao <xzhao@xxxxxxx>]

Hi Nate,

Thanks for your reply. But I am not running in multi-host mode now. The 
cluster uses one nova-networking service on the controller node.

Xin

On 12/17/2012 9:31 PM, Nathanael Burton wrote:
> On Dec 17, 2012 2:05 PM, "Xin Zhao" <xzhao@xxxxxxx 
> <mailto:xzhao@xxxxxxx>> wrote:
> >
> > Hello,
> >
> > I allocate 2 public ips to instances, the first one becomes 
> accessible almost immediately, but the second one
> > always take a long time to be pingable. It doesn't matter which 
> specific IP is assigned first or second, it's always
> > the second one that is slow to be reachable, although the 
> corresponding iptable rules are all set on the controller
> > node almost immediately.   Has anyone seen similar behavior?
> >
> > I am using essex (2012.1) on RHEL6.
> >
> > Thanks,
> > Xin
> >
> >
> >
> > _______________________________________________
> > Mailing list: https://launchpad.net/~openstack 
> <https://launchpad.net/%7Eopenstack>
> > Post to     : openstack@xxxxxxxxxxxxxxxxxxx 
> <mailto:openstack@xxxxxxxxxxxxxxxxxxx>
> > Unsubscribe : https://launchpad.net/~openstack 
> <https://launchpad.net/%7Eopenstack>
> > More help   : https://help.launchpad.net/ListHelp
> >
>
> Xin,
>
> You may need to use send_arp_for_ha=true to send a gratuitous ARP if 
> you're using multi-host networking.
>
> Nate

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature