[ovs-dev] [RFC] ovs-ctl: Allow selective start for db and switch

Aaron Conole aconole at redhat.com
Thu Mar 17 13:54:14 UTC 2016


Ben Pfaff <blp at ovn.org> writes:
> On Mon, Feb 29, 2016 at 03:40:56PM -0500, Aaron Conole wrote:
>> Currently, ``ovs-ctl start'' will attempt to start both the DB and
>> vswitchd. This is quite convenient when the database already has all of
>> the configuration values required, and when using a single services file
>> for systemd integration. The same goes for the ``ovs-ctl stop'' command.
>> 
>> However, there are some cases which are not easily covered. The case
>> where we want to set values in the database prior to starting the
>> forwarding path, as well as the case of supporting multiple service
>> files, one per daemon (which is how systemd expects services to look).
>> 
>> Signed-off-by: Aaron Conole <aconole at redhat.com>
>> ---
>> v1: Limited testing (just some quick start/stop testing). The documentation
>>     updates are rather quick.
>>     See http://openvswitch.org/pipermail/dev/2016-February/066502.html for
>>     more information.
>
> This seems pretty reasonable.  I have a few comments.
>
> We have some precedent around options for ovsdb-server and ovs-vswitchd:
> so far, we've started options for them with their program names, as in
> --ovsdb-server-wrapper and --ovsdb-server-priority.  I think that it
> would be reasonable to continue this; --control-forwarding and
> --control-ovsdb is a bit of a departure.  So I'd consider just naming
> these as --ovsdb-server and --ovs-vswitchd, so that to disable them one
> uses --no-ovsdb-server or --no-ovs-vswitchd.

Okay; I'll incorporate this.

> Second, these options make sense for "start" and "stop".  For some of
> the other commands it's necessary to be more nuanced.  So, for example,
> for "restart", we would not want to call save_flows_if_required and
> restore_flows if we're not restarting ovs-vswitchd.  And I'm not sure
> that it makes any sense to run "force-reload-kmod" if one is not
> restarting ovs-vswitchd.

Agreed. 

> Actually I think I see some opportunity for simplification now in
> ovs-ctl, because there is special-case code there for upgrading from
> obsolete versions (before OVS 1.10).  I think I'll send out a patch to
> remove those special cases, which ought to make it easier to think
> through the cases here.

Sounds good. I'll wait for your patches and respin this.

Thanks so much for the review, Ben!

-Aaron



More information about the dev mailing list