15:01:20 <ralonsoh> #startmeeting neutron_qos
15:01:21 <openstack> Meeting started Tue Dec  3 15:01:20 2019 UTC and is due to finish in 60 minutes.  The chair is ralonsoh. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:22 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:24 <ralonsoh> Hello
15:01:26 <openstack> The meeting name has been set to 'neutron_qos'
15:01:56 <slaweq> hi
15:01:57 <davidsha> o/
15:02:00 <maciejjozefczyk> \o
15:02:10 <ralonsoh> ok, let's start
15:02:11 <ralonsoh> #topic RFEs
15:02:19 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/1476527
15:02:19 <openstack> Launchpad bug 1476527 in neutron "[RFE] Add common classifier resource" [Wishlist,Triaged] - Assigned to Igor D.C. (igordcard)
15:02:31 <ralonsoh> We have already discussed this RFE in the drivers meeting
15:02:49 <ralonsoh> apart from reviewing the spec and the patches, is there something else we need to discuss?
15:02:51 <slaweq> ralonsoh: but not with quorum of drivers team :/
15:02:58 <ralonsoh> slaweq, I know
15:03:09 <ralonsoh> that's why we need to retake this conversation
15:03:13 <ralonsoh> but in the drivers meeting
15:03:16 <slaweq> so lets ask mlavalle and others to review the spec
15:03:20 <ralonsoh> with all possible cores there
15:03:46 <ralonsoh> exactly, that's what we need (once the RFE was approved in 2017 and last week again)
15:04:08 <ralonsoh> davidsha, please, try to ping Miguel (not here) to review it
15:04:20 <davidsha> ralonsoh: will do!
15:04:46 <ralonsoh> then will know if the DB migration and API definitions are OK
15:05:21 <ralonsoh> there are no more RFEs in the list
15:05:27 <ralonsoh> something else in this topic?
15:05:38 <davidsha> The os-ken dscp fix
15:05:40 <davidsha> https://review.opendev.org/#/c/695479/
15:05:55 <ralonsoh> exactly, that was in the open discussion section
15:06:10 <davidsha> kk
15:06:19 <ralonsoh> I'll add it to my review list
15:06:33 <ralonsoh> I think you addressed my comments
15:06:37 <ralonsoh> yes
15:07:03 <ralonsoh> but for now we can't remove the ofctl code
15:07:06 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/1853171
15:07:06 <openstack> Launchpad bug 1853171 in neutron "Deprecate and remove any "ofctl" code in Neutron and related projects " [Medium,In progress] - Assigned to David Shaughnessy (david-shaughnessy)
15:07:10 <davidsha> I think the only open is makeing _get_dp public for out of tree extensions
15:07:12 <ralonsoh> related bug^^
15:07:20 <ralonsoh> yes
15:07:34 <ralonsoh> IMO, you should implement this code insiode the ovs lib
15:07:43 <ralonsoh> not making _get_dp public
15:07:46 <ralonsoh> as commented in PS1
15:08:02 <ralonsoh> https://review.opendev.org/#/c/695479/1/neutron/plugins/ml2/drivers/openvswitch/agent/extension_drivers/qos_driver.py@126
15:08:13 <ralonsoh> make a method, in br-int class
15:08:25 <ralonsoh> passing only dscp mark and input
15:08:38 <ralonsoh> ok with this?
15:08:54 <ralonsoh> or maybe you prefer to make get_dp public
15:08:55 <davidsha> yup, thats in the latest PS
15:09:08 <ralonsoh> oh yes!
15:09:12 <ralonsoh> perfect
15:09:20 <ralonsoh> (I still need to review last PS)
15:09:41 <davidsha> Making _get_dp public would be more for projects like SFC which still use ofctl from the cookie bridge
15:09:52 <davidsha> NP i only pushed it a few hours ago
15:10:38 <ralonsoh> if someone else needs that, then we can make it public (or something like a property, because this is something that does not change)
15:10:46 <ralonsoh> does not USUALLY change
15:10:58 <ralonsoh> but IMO not in this patch
15:11:09 <davidsha> ack, I'll remove it
15:11:18 <ralonsoh> thanks!
15:11:22 <ralonsoh> and thanks for taking care of it
15:11:36 <davidsha> np!
15:11:52 <ralonsoh> let's move to the next section then
15:11:54 <ralonsoh> #topic Bugs
15:11:59 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/1845176
15:11:59 <openstack> Launchpad bug 1845176 in neutron "Removing of QoS queue in neutron-ovs-agent fails due to existing references" [Medium,In progress] - Assigned to Rodolfo Alonso (rodolfo-alonso-hernandez)
15:12:11 <ralonsoh> sorry for not progressing on this one
15:12:28 <ralonsoh> code is almost there https://review.opendev.org/#/c/687922/
15:12:40 <ralonsoh> but I need to change the fullstack tests
15:12:54 <ralonsoh> that's the last blocker for this patch
15:13:06 <ralonsoh> but that need to wait until next week, sorry
15:13:26 <ralonsoh> (maybe you can take a look, feel free!!!)
15:14:09 <ralonsoh> ok, next one
15:14:17 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/1845161
15:14:17 <openstack> Launchpad bug 1845161 in neutron "Neutron QoS Policy lost on interfaces" [High,In progress] - Assigned to Rodolfo Alonso (rodolfo-alonso-hernandez)
15:14:37 <ralonsoh> #link https://review.opendev.org/#/c/690098/
15:14:46 <ralonsoh> same as before, tempest tests are failing
15:15:03 <ralonsoh> this is the last blocker for this patch
15:15:12 <ralonsoh> but I didn't have time for this
15:15:40 <ralonsoh> (again, feel free to take a look if you want, thank you in advance!)
15:16:21 <ralonsoh> ok, next one
15:16:27 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/1853840
15:16:27 <openstack> Launchpad bug 1853840 in neutron "Neutron fails to create bandwidth providers if CONF.host is set" [High,In progress] - Assigned to Bence Romsics (bence-romsics)
15:16:51 <ralonsoh> bence is taking care of it
15:16:54 <ralonsoh> #link https://review.opendev.org/#/c/696600/
15:17:23 <ralonsoh> there should be two patches
15:17:33 <ralonsoh> one for master, using the latest nova API
15:18:07 <ralonsoh> and the stable branch one, using the virsh method to retrieve the host name, if I'm not wrong
15:18:44 <ralonsoh> the patch commit message is quite descriptive
15:19:40 <ralonsoh> last one I have
15:19:45 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/1851362
15:19:45 <openstack> Launchpad bug 1851362 in neutron "[RFE] ports do not inherit their associated network's policy" [Low,In progress] - Assigned to Rodolfo Alonso (rodolfo-alonso-hernandez)
15:20:12 <ralonsoh> #link https://review.opendev.org/#/c/693234/
15:20:23 <ralonsoh> the RFE is not approved yet
15:20:42 <ralonsoh> next drivers meeting I'll try to have it
15:21:16 <ralonsoh> as you can see in the first patch (neutron-lib), this extension is just adding the network QoS policy to the port
15:21:29 <ralonsoh> as a read-only parameter
15:21:48 <ralonsoh> this will allow the OS client to show it in the CLI
15:22:27 <ralonsoh> as described in the bug, user can be confused with the network QoS policy inheritance if this value is not present in the CLI
15:22:31 <ralonsoh> when showing a port
15:23:07 <ralonsoh> (btw, this should go to the previous section, sorry)
15:23:30 <davidsha> np :P
15:23:56 <ralonsoh> any other bug?
15:24:00 <ralonsoh> not listed here
15:24:27 <ralonsoh> let's move then
15:24:29 <ralonsoh> #topic Open Discussion
15:24:42 <ralonsoh> we have discussed about https://bugs.launchpad.net/neutron/+bug/1853171
15:24:42 <openstack> Launchpad bug 1853171 in neutron "Deprecate and remove any "ofctl" code in Neutron and related projects " [Medium,In progress] - Assigned to David Shaughnessy (david-shaughnessy)
15:25:01 <ralonsoh> and the patch related to the QoS extension in the OVS agent
15:25:05 <ralonsoh> for the DSCP rule
15:25:27 <ralonsoh> any other topic you want to bring here?
15:25:28 <slaweq> is this dscp only thing using old ofctl code?
15:25:40 <davidsha> In core Neutron, yes
15:25:43 <slaweq> or do we still need more work to drop it completly?
15:25:46 <slaweq> ahh, ok
15:25:53 <davidsha> But out of tree extensions like SFC are still using it
15:26:05 <ralonsoh> exactly
15:26:34 <slaweq> ok, that's bad :/
15:26:41 <ralonsoh> I know
15:26:43 <slaweq> as it't more work to do
15:26:53 <ralonsoh> but at least Neutron is not using it anymore
15:27:06 <ralonsoh> davidsha, proposed a deprecation warning
15:27:26 <ralonsoh> https://review.opendev.org/#/c/695479/4/neutron/agent/common/ovs_lib.py
15:27:58 <slaweq> ok
15:28:08 <davidsha> This was meant to be deprecated back in Pike though, could we pull in that warning for after U?
15:28:08 <ralonsoh> at least Neutron is not using the CLI code any more with this patch
15:28:35 <slaweq> yeah, that's good
15:28:38 <ralonsoh> cool
15:28:48 <slaweq> I added it to my review list for tomorrow (or tonight)
15:29:09 <davidsha> thanks!
15:29:45 <ralonsoh> something else to be discussed here?
15:29:59 <davidsha> I'm good
15:30:21 <ralonsoh> thank you all and see you online
15:30:26 <ralonsoh> #endmeeting