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

Re: [Xen-API] host crash when vlan is tagged in pool XCP 0.1.1


  • To: Fabio Kung <fabio.kung@xxxxxxxxx>
  • From: Marco Sinhoreli <msinhore@xxxxxxxxx>
  • Date: Thu, 8 Apr 2010 13:38:57 -0300
  • Cc: xen-api <xen-api@xxxxxxxxxxxxxxxxxxx>
  • Delivery-date: Thu, 08 Apr 2010 09:39:14 -0700
  • Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; b=CUPHM7TDY5jEuMtGc9ebIqQuPLOgnCg6r0gfJZ1MZhnzV7+qeczkIYXQyfRkyTjq59 UswlYBTgfpo2WjEWmz9iXWGTsNP3OeKMMfWZOimlHdVX6aQR5oHG2//egdhXhNMxuAkC Bd9JvbxZGsBkXJjuJprxPOLVSF/49PC02SyjE=
  • List-id: Discussion of API issues surrounding Xen <xen-api.lists.xensource.com>

Hi Fabio:

In XenServer and XCP 0.1.0 I have this processes working fine using Python XAPI. Homologating XCP 0.1.1 to map theÂinconsistencies, this issue occurred and, as I planning to use many pools managed by my application, I can't use the approach for reset all pool to have this working. I think this is a bug because I haven't this behavior in oldest releases.

Cheers,

On Thu, Apr 8, 2010 at 1:04 PM, Fabio Kung <fabio.kung@xxxxxxxxx> wrote:
I also had problems with open-vswitch and XCP 0.1.1 pools (standalone hosts work fine) while trying to use a vlan-tagged interface as a management interface (with and without bonded interfaces).

I know this is unsupported even in XenServer (before XCP), but I was able to do it with xe emergency-transition-to-master for each host inside the pool. After all hosts have the management onÂvlan-taggedÂinterfaces, I just rebuild the pool.

On XCP 0.1.1, I first created the Network/VLAN pair in all hosts inside the pool using the xe CLI (xe pool-vlan-create). Then, through pool's masterÂxsconsole, I changed master's management to the vlan-taggedÂinterface. At this moment, the master lost communication with other hosts in the pool, as expected.

Then, I went to one of the slave hosts and made it a temporary master (xe emergency-transition-to-master). The problem is that when I set the vlan-tagged interface as the management, the host somehow starts to communicate with the original master and open-vswitch cuts the link between then. I could see that the vlan-tagged port inside open-vswitch went down at one of the sides (I'm almost sure at original master's side).

I'm sorry I don't have logs anymore and I'm not really sure this is a bug. But with XenServer 5.5.x (in regular bridge mode) this works just fine.

I guess this may be not a very common requirement, but in my case the management interface must be vlan-tagged and bonded. For me, it's a blocker and I would be glad to help further investigating the issue.

Thanks in advance,

-- Fabio Kung


On Tue, Apr 6, 2010 at 1:41 PM, Marco Sinhoreli <msinhore@xxxxxxxxx> wrote:
Hi all:

I have a pool with 2 hosts XCP 0.1.1 configured and, whenÂI'm creating a VM using a tagged vlan, the second server freeze and Âonly ICMP request is responding in this server. I set the XCP to use bridge instead vswitch. I have no more information in this moment but if you need that please let me know.

Cheers,

--
Marco Sinhoreli

_______________________________________________
xen-api mailing list
xen-api@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/mailman/listinfo/xen-api





--
Marco Sinhoreli
_______________________________________________
xen-api mailing list
xen-api@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/mailman/listinfo/xen-api

 


Rackspace

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