[Openstack-operators] OpenStack node redundancy

James Dempsey jamesd at catalyst.net.nz
Mon Dec 21 22:30:40 UTC 2015


Hi Serguei,

We have pairs of Top of Rack switches in MLAG configurations.  On the
Server-side, we are using OVS LACP bonds in balance-tcp mode.  We
haven't had any kernel issues with bonding on Ubuntu kernels 3.13, 3.19,
and 4.2.

I'm not sure what you mean by "Openstack will remove the isolated node".
 Can you explain further?



For example, this is what our physical connectivity bridge looks like
when viewed with ovs-vsctl


    Bridge br-default
        Port bond-default
            Interface "eth4"
            Interface "eth6"
        Port phy-br-default
            Interface phy-br-default
                type: patch
                options: {peer=int-br-default}
        Port br-default
            Interface br-default
                type: internal



Cheers,
James

On 19/12/15 13:22, Wade Holler wrote:
> Yes and, I prefer mode=4 MLAG bond/team.  On Cent we had to use
> xmit_hash_policy layer2+3 because layer3+4 caused kernel panics @ Cent
> 7.1.1503
> 
> Cheers
> Wade
> 
> 
> On Fri, Dec 18, 2015 at 12:17 PM Serguei Bezverkhi (sbezverk) <
> sbezverk at cisco.com> wrote:
> 
>> Hello team,
>>
>>
>>
>> I would appreciate if you could share approach you use for OpenStack node
>> redundancy from network failure perspective. Example in case of a failure
>> of an upstream switch and if a node does not have redundant link to a
>> second upstream switch, it gets isolated. I understand that OpenStack will
>> deal with this situation by removing this node, but in my case there is a
>> strict requirement to be able to prevent a single switch failure.
>>
>>
>>
>> Here is the solution I came up with, it is nothing new but appreciate your
>> critic/comments/suggestions.
>>
>>
>>
>> Redundant OpenStack is connected to two upstream switches, interfaces are
>> teamed into group and bound to respective OpenStack bridges. So br-int
>> would have physical link interface team0 and not eth0 or eth1 connected. In
>> this case if one of the upstream links fails, the connectivity would be
>> preserved.
>>
>>
>>
>> Thank you
>>
>>
>>
>> Serguei
>>
>>


-- 
James Dempsey
Senior Cloud Engineer
Catalyst IT Limited
--



More information about the OpenStack-operators mailing list