Mailing-List: contact cloudstack-dev-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: cloudstack-dev@incubator.apache.org
Content-Type: multipart/alternative;
 boundary="===============8790281216180461386=="
MIME-Version: 1.0
Subject: Re: Review Request: Add the ability to specify different VIF drivers
 per
 traffic type in KVM
From: "Hugo Trippaers" <htrippaers@schubergphilis.com>
To: "Wido den Hollander" <wido@widodh.nl>,
 "Marcus Sorensen" <shadowsor@gmail.com>,
 "Hugo Trippaers" <htrippaers@schubergphilis.com>
Cc: "cloudstack" <cloudstack-dev@incubator.apache.org>,
 "Dave Cahill" <dcahill@midokura.com>
Date: Fri, 01 Mar 2013 19:36:37 -0000
Message-ID: <20130301193637.17326.32218@reviews.apache.org>
Auto-Submitted: auto-generated
Sender: "Hugo Trippaers" <noreply@reviews.apache.org>
References: <20130301074638.17206.78754@reviews.apache.org>
In-Reply-To: <20130301074638.17206.78754@reviews.apache.org>
Reply-To: "Hugo Trippaers" <htrippaers@schubergphilis.com>

--===============8790281216180461386==
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/9604/#review17256
-----------------------------------------------------------


Dave,  looks good.

Couple of questions though, there is some bridge specific code in LibvirtCo=
mputingResource that checks existence of interfaces and such. Shouldn't  th=
at code also be modified to support multiple drivers per nic?

Also currently there is only the native bridge and openvswitch in here. The=
 actual recommendation is to use only one of them as far as i know. What is=
 the compelling reason to allow for this difference? In most cases where op=
envswitch is used this means configuring vswitch either via the native inte=
rface or via the brcompat layer. While possible to combine the two it could=
 lead to disastrous results when interfaces start overlapping. E.g. both op=
envswitch and bridge create an interface called br0 (i've been there ;-) )


- Hugo Trippaers


On March 1, 2013, 7:46 a.m., Dave Cahill wrote:
> =

> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/9604/
> -----------------------------------------------------------
> =

> (Updated March 1, 2013, 7:46 a.m.)
> =

> =

> Review request for cloudstack, Hugo Trippaers, Wido den Hollander, and Ma=
rcus Sorensen.
> =

> =

> Description
> -------
> =

> Feature spec: =

> https://cwiki.apache.org/confluence/display/CLOUDSTACK/Allow+specificatio=
n+of+different+VIF+drivers+per+traffic+type+in+KVM
> =

> Jira ticket: =

> https://issues.apache.org/jira/browse/CLOUDSTACK-1311
> =

> Adds the ability to specify different VIF drivers per traffic type in KVM.
> =

> =

> This addresses bug CLOUDSTACK-1311.
> =

> =

> Diffs
> -----
> =

>   plugins/hypervisors/kvm/src/com/cloud/hypervisor/kvm/resource/LibvirtCo=
mputingResource.java 99b8723 =

>   plugins/hypervisors/kvm/test/com/cloud/hypervisor/kvm/resource/LibvirtV=
ifDriverTest.java PRE-CREATION =

> =

> Diff: https://reviews.apache.org/r/9604/diff/
> =

> =

> Testing
> -------
> =

> Added unit tests to test all variations of this new configuration (no con=
figuration, defaults, override some traffic etc).
> =

> Built and deployed, spun up Advanced Isolated network with two VMs, verif=
ied internal and external connectivity.
> =

> =

> Thanks,
> =

> Dave Cahill
> =

>


--===============8790281216180461386==--