[PVE-User] Networking issues
vpadro at gmail.com
Thu Jun 25 19:50:51 CEST 2009
On Thu, Jun 25, 2009 at 12:16 PM, Andrew
Niemantsverdriet<andrew at rocky.edu> wrote:
> Jeff is exactly right, if at all possible make the number of NIC's
> match up for both boxes in the cluster. Baring that your master node's
> network configuration is the only one that is honored. You can make
> dummy interfaces on the master and have things somewhat work but your
> best bet is just to make the NIC's the same in both boxes. Other than
> the dummy interfaces the only other work around is to not have a
> cluster but you end up losing a lot if you do that.
> On Thu, Jun 25, 2009 at 10:58 AM, Jeff Saxe<JSaxe at briworks.com> wrote:
>> Aha. I think an underlying assumption of the Proxmox VE environment is
>> that if you have multiple nodes in a cluster, they all have exactly
>> the same NIC resources available to them. This is so that you can
>> smoothly (either online or offline) migrate a VE from one node to
>> another, and whatever interfaces its IP addresses were bridged to on
>> one will be able to be mapped to the same bridge places on the new
>> node to which it's moved. If you make one node with that third NIC,
>> but then another node does not have the third NIC, then you have a VE
>> that's effectively "stuck" to one node and unable to be migrated. In
>> that case, I believe Dieter would say you are violating the spirit of
>> the Proxmox clustering (not to put words in your mouth, Dieter).
>> I believe that's the crux of what you (Victor) are seeing as a user
>> interface bug or limitation.
>> -- Jeff Saxe, Network Engineer
>> Blue Ridge InternetWorks, Charlottesville, VA
>> 434-817-0707 ext. 2024 / JSaxe at briworks.com
>> On Jun 25, 2009, at 11:37 AM, Victor Padro wrote:
>>> On Thu, Jun 25, 2009 at 10:34 AM, Dietmar
>>> Maurer<dietmar at proxmox.com> wrote:
>>>> vmbr2 is available on both nodes? (you need to configure that on
>>>> both nodes).
>>> No, it's not.
>>> I just need it on the slave node(only that node has 3 NICs)
>>> I think the partial solution would be to convert the slave node into
>>> the master node and the master node into the slave node.
>> pve-user mailing list
>> pve-user at pve.proxmox.com
> /-\ ndrew Niemantsverdriet
> Academic Computing
> (406) 238-7360
> Rocky Mountain College
> 1511 Poly Dr.
> Billings MT, 59102
> pve-user mailing list
> pve-user at pve.proxmox.com
Both of you are right, I need to have the same amount of physical NICs
in both nodes.
Dummy interfaces are not letting me migrate the container also, so
temporaly I am testing the slave node as a master node until I get
another Intel NIC.
"It is human nature to think wisely and act in an absurd fashion."
"Todo el desorden del mundo proviene de las profesiones mal o
More information about the pve-user