On Tue, Jun 20, 2017 at 6:09 AM, Miguel Angel Ajo Pelayo
<majop...@redhat.com> wrote:
> I'm working on the implementation changes related to the schema [1]
>
> And I thought it could be great if we discussed here over the cmdline
> changes, so this effort could happen in parallel:
>
> ovn-nbctl lrp-add-gw-chassis <port-name> <chassis-name> [priority]
> ovn-nbctl lrp-set-gw-chassis-prio <port-name> <chassis-name> priority
> ovn-nbctl lrp-del-gw-chassis <port-name> <chassis-name>
> ovn-nbctl lrp-list-gw-chassis <port-name>
>
>
> how does that look  ^  ?

Looks OK to me.

If you want to save some time, you could just use the generic DB
commands for this for now, and come back to it later as a bonus.

>
>
>
> [1] https://etherpad.openstack.org/p/ovn-l3ha-schema
>
> On Wed, Jun 14, 2017 at 10:04 PM, Miguel Angel Ajo Pelayo
> <majop...@redhat.com> wrote:
>>
>> As discussed on IRC let's meet tomorrow at 9am EST to talk about the topic
>> on IRC.
>>
>>
>>
>>
>> El 14 jun. 2017 9:02 p. m., "Russell Bryant" <russ...@ovn.org> escribió:
>>
>> On Wed, Jun 14, 2017 at 2:40 PM, Miguel Angel Ajo Pelayo
>> <majop...@redhat.com> wrote:
>> > Hi, sorry, I didn't see your comments this morning. Thanks a lot for all
>> > the
>> > feedback, I'll shuffle things around.
>> >
>> > I thought about something like adding new columns too, but I didn't
>> > manage
>> > to come up with anything that didn't make it less intuitive.
>> >
>> > The disadvantage of the current format is that you can't filter the
>> > monitoring for chassis redirect ports on an specific chassis name (with
>> > Current ovsdb API). May be two columns (one for names, another for
>> > priorities would let us do that)
>> >
>> > I'm glad to hear suggestions for column names or structure here :)
>>
>> I'm happy to collaborate on the proposed alternate schema.  Maybe we
>> can connect on IRC tomorrow to work on it?
>>
>> >
>> >
>> >
>> > El 13 jun. 2017 10:41 p. m., "Russell Bryant" <russ...@ovn.org>
>> > escribió:
>> >
>> > On Tue, Jun 13, 2017 at 4:40 PM, Russell Bryant <russ...@ovn.org> wrote:
>> >> On Fri, Jun 2, 2017 at 8:31 AM,  <majop...@redhat.com> wrote:
>> >>> From: Miguel Angel Ajo <majop...@redhat.com>
>> >>>
>> >>> This patch handles multiple gateways with priorities in
>> >>> chassisredirect
>> >>> ports, any gateway with a chassis redirect port will implement the
>> >>> rules to de-encapsulate incomming packets for such port.
>> >>>
>> >>> And hosts targetting a remote chassisredirect port will setup a
>> >>> bundle(active_backup, ..) action to each tunnel port, in the given
>> >>> priority order.
>> >>>
>> >>> Signed-off-by: Miguel Angel Ajo <majop...@redhat.com>
>> >>> ---
>> >>>  ovn/controller/binding.c        |   9 +--
>> >>>  ovn/controller/lflow.c          |   6 +-
>> >>>  ovn/controller/lport.c          | 119
>> >>> ++++++++++++++++++++++++++++++++++++++++
>> >>>  ovn/controller/lport.h          |  28 ++++++++++
>> >>>  ovn/controller/ovn-controller.c |   5 +-
>> >>>  ovn/controller/physical.c       | 114
>> >>> ++++++++++++++++++++++++++++++++------
>> >>>  6 files changed, 255 insertions(+), 26 deletions(-)
>> >>
>> >> Some high level comments to start ...
>> >>
>> >> Ideally with a patch series, each patch should be applicable on its
>> >> own.  With this patch applied, some tests are failing for me.
>> >>
>> >> Documentation should also be included with whatever patch first
>> >> introduces functionality, so I'd expect docs on the updated
>> >> redirect-chassis format here.
>> >>
>> >> Please read over
>> >> Documentation/internals/contributing/coding-style.rst.  There are some
>> >> minor style issues throughout the patch.  I can point them out in a
>> >> more detailed pass.
>> >>
>> >> The patch makes me wonder if we should introduce a more structured
>> >> format for specifying chassis associated with a router port.  It feels
>> >> like we're encoding too much in a single option string.  Maybe we
>> >> should add a new "chassis" column to Logical_Router_Port, that can
>> >> include a list of chassis, which would have to be a new record type in
>> >> OVN northbound, containing much less info than the southbound
>> >> counterpart.  We'd have to add a similar new column to the
>> >> Port_Binding table in OVN southbound.  I'm curious what you and others
>> >> think about this, or if the parsed option string is fine.
>> >
>> > Sorry, I replied to v1, but all comments apply to v2.
>> >
>> > --
>> > Russell Bryant
>> >
>> >
>>
>>
>>
>> --
>> Russell Bryant
>>
>>
>



-- 
Russell Bryant
_______________________________________________
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Reply via email to