14:00:37 <ralonsoh> #startmeeting neutron_qos
14:00:37 <openstack> Meeting started Wed Dec 14 14:00:37 2016 UTC and is due to finish in 60 minutes.  The chair is ralonsoh. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:39 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:41 <openstack> The meeting name has been set to 'neutron_qos'
14:00:45 <ralonsoh> Hi
14:01:04 <claudiub> o/
14:01:18 <ralonsoh> Let's wait one minute
14:01:22 <davidsha> Hi
14:01:45 <ralonsoh> hi davidsha
14:02:41 <ralonsoh> ok, let's start...
14:02:49 <ralonsoh> #topic RFEs
14:02:57 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/1634798
14:02:58 <openstack> Launchpad bug 1634798 in neutron "[RFE] Qos DSCP to vlan priority mapping" [Wishlist,Confirmed]
14:03:09 <ralonsoh> Any volunteer?
14:03:52 <davidsha> One sec
14:04:21 <davidsha> That's dependent on this right? https://bugs.launchpad.net/neutron/+bug/1505631
14:04:22 <openstack> Launchpad bug 1505631 in neutron "[RFE] QoS VLAN 802.1p Support" [Wishlist,Confirmed] - Assigned to Kannan Raman (kannanrc20)
14:04:55 <ralonsoh> But should be merged now
14:04:57 <ralonsoh> let me check
14:05:09 <ralonsoh> ok, not yet
14:05:18 <davidsha> It just has the spec atm
14:05:24 <ralonsoh> so maybe could be interesting to review this patch
14:05:52 <slaweq> hello
14:05:54 <slaweq> sorry for late
14:06:01 <davidsha> Hi
14:06:02 <ralonsoh> hi, we were waiting for you
14:06:08 <ralonsoh> just you, davidsha and me
14:06:19 <slaweq> ok :)
14:06:29 <ralonsoh> Second RFE
14:06:30 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/1644369
14:06:31 <openstack> Launchpad bug 1644369 in neutron "Support for DSCP marking in Linuxbridge L2 agent" [Wishlist,Fix released] - Assigned to Slawek Kaplonski (slaweq)
14:06:35 <ralonsoh> Merged!!
14:06:37 <ralonsoh> Congrats
14:06:40 <slaweq> thx
14:06:54 <ralonsoh> Very fast
14:06:56 <davidsha> I was looking at the spec but I was waiting for it to be moved from Ocata to Pike
14:07:13 <slaweq> I was even supprised how fast it goes :)
14:07:22 <ralonsoh> davidsha: yu are talking bout the DSCP patch?
14:07:23 <ralonsoh> spec
14:07:46 <davidsha> 802.1p
14:07:49 <ralonsoh> yes
14:08:04 <ralonsoh> ok, so let's follow this spec and let's review it
14:08:11 <ralonsoh> #link https://review.openstack.org/#/c/392023/
14:08:17 <slaweq> about this dscp and LB, I think it could be quite similar for sr-iov but I don't know sr-iov so I'm not sure about that
14:09:00 <davidsha> If I recall ip link could modify the ToS field but it was only for certain types of traffic
14:09:05 <ralonsoh> Yes, I'll create a bug for DSCP and SRIOV
14:09:22 <ralonsoh> I'll investigate this
14:09:22 <slaweq> ralonsoh: thx
14:09:40 <slaweq> I don't know sr-iov and I don't have dev environment to check it
14:09:42 <ralonsoh> That maybe will depend on the specific driver
14:09:47 <ralonsoh> I'll check this
14:10:10 <ralonsoh> Next RFE
14:10:11 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/1578989
14:10:13 <openstack> Launchpad bug 1578989 in neutron "[RFE] Strict minimum bandwidth support (egress)" [Wishlist,In progress] - Assigned to Rodolfo Alonso (rodolfo-alonso-hernandez)
14:10:23 <ralonsoh> Very easy
14:10:25 <ralonsoh> 1) Spec
14:10:32 <ralonsoh> #link https://review.openstack.org/#/c/396297/
14:10:46 <ralonsoh> We need to have this spec approved
14:11:06 <ralonsoh> And first patch
14:11:12 <ralonsoh> #link https://review.openstack.org/#/c/401254/
14:11:22 <ralonsoh> Please, I need reviews!!
14:11:53 <davidsha> I'll look at both asap
14:11:53 <slaweq> about this specs I had some doubts about that
14:12:00 <ralonsoh> Ok, I'll try to ping ajo later
14:12:02 <ralonsoh> Why?
14:12:11 <ralonsoh> Which doubts?
14:12:12 <slaweq> for example what with live-migration?
14:12:38 <slaweq> or updating rule which is already applied on some port
14:13:22 <ralonsoh> live-migration: it will depends on the hardware availability
14:13:28 <ralonsoh> on the second compute node
14:13:51 <ralonsoh> If you spawn new VM's on a new node, you'll update nova scheduler database
14:14:06 <slaweq> ok, what about update rule?
14:14:21 <slaweq> let's say I have 10 vms, each with 10Mbps limit
14:14:25 <ralonsoh> That's considered, please review the patch
14:14:32 <ralonsoh> continue, please
14:14:33 <slaweq> ok, I will check it again
14:14:34 <slaweq> thx
14:14:40 <ralonsoh> no no, continue
14:14:48 <slaweq> ok
14:15:12 <ralonsoh> I see were are you going....
14:15:22 <slaweq> so, then what if You will update to 20Mbps in rule?
14:15:28 <ralonsoh> Yes
14:15:31 <slaweq> You will have overall not 100Mbps but 200Mbps
14:15:42 <slaweq> and it can be too much on host
14:15:45 <ralonsoh> one option: to cancel this rule update
14:16:02 <ralonsoh> this information is in the qos extension
14:16:30 <ralonsoh> so if you extend the needed bw beyond the maximun available, there should be an error
14:16:36 <ralonsoh> that makes sense?
14:16:53 <slaweq> ok, maybe I didn't read it all, but will qos extension know about overall available/used bandwidth on host?
14:17:11 <ralonsoh> qos extension will
14:17:17 <ralonsoh> 1) read the initial bw
14:17:20 <slaweq> isn't it in Nova?
14:17:22 <ralonsoh> 2) update this info
14:17:25 <ralonsoh> 3) inform nova
14:17:37 <slaweq> ok
14:17:50 <slaweq> so forbid update can be some option
14:17:53 <ralonsoh> but, that's my design. And I need ajo to review this!
14:18:27 <ralonsoh> ok, let's move to bugs
14:18:35 <ralonsoh> anything else?
14:18:38 <slaweq> but I'm not sure if it will not be related to "improved validation" patch
14:18:48 <ralonsoh> hmmmm
14:18:49 <slaweq> and this ajo's patch to refactor notifications
14:19:12 <ralonsoh> I need to review both patches
14:19:17 <slaweq> what in case if You will have port without qos policy already on host and will do "port update" to add such QoS?
14:19:36 <ralonsoh> yes
14:19:44 <slaweq> it should be also forbiden in such case
14:19:52 <slaweq> so it can be quite hard to do now :/
14:19:55 <ralonsoh> why?
14:20:11 <slaweq> maybe not hard from technical point of view
14:20:15 <ralonsoh> if the needed bw doesn't exceed the available bw
14:20:20 <ralonsoh> it's ok
14:20:23 <slaweq> but it's similar problem like this "improved validation"
14:20:39 <slaweq> please talk with ajo about that
14:20:44 <ralonsoh> sure!
14:20:47 <ralonsoh> thanks!
14:21:16 <ralonsoh> let's continue
14:21:19 <ralonsoh> #topic Bugs
14:21:26 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/1649503
14:21:27 <openstack> Launchpad bug 1649503 in neutron "Mechanism driver can't be notified with updated network" [High,In progress] - Assigned to Hong Hui Xiao (xiaohhui)
14:21:46 <ralonsoh> #link https://review.openstack.org/#/c/410101/
14:22:09 <ralonsoh> armax move the priotity to high
14:22:21 <ralonsoh> so we should take care of this one
14:22:29 <ralonsoh> we should review the patch
14:22:33 <slaweq> ok, I will also try to review it
14:22:43 <slaweq> it's not big patch :)
14:22:54 <ralonsoh> no, it isn't
14:23:07 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/1649517
14:23:08 <openstack> Launchpad bug 1649517 in neutron "qos policy attached to network, qos_policy_id is reflecting on neutron net-show , but not on the port with neutron port-show" [Wishlist,New]
14:23:19 <ralonsoh> feedback from ajo
14:23:25 <slaweq> I talked with ajo about it yesterday
14:23:26 <ralonsoh> "to close as opinion"
14:23:40 <slaweq> yep, ajo said that it's not a bug but feature :)
14:24:22 <slaweq> but personally I think that it could be something like "network_qos_policy" and "qos_policy" (or port_qos_policy) in port details
14:24:43 <slaweq> to distinquish easily which policy is used on port
14:24:46 <ralonsoh> yes, I have the same opinion
14:24:47 <slaweq> what You think?
14:25:09 <ralonsoh> but the author should open it as a RFE
14:25:16 <slaweq> yep
14:25:38 <ralonsoh> let's wait for the author response
14:25:44 <slaweq> ok
14:25:47 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/1649503
14:25:49 <openstack> Launchpad bug 1649503 in neutron "Mechanism driver can't be notified with updated network" [High,In progress] - Assigned to Hong Hui Xiao (xiaohhui)
14:26:03 <ralonsoh> sorry
14:26:05 <ralonsoh> my fault
14:26:13 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/1649488
14:26:14 <openstack> Launchpad bug 1649488 in neutron "Duplicated revises_on_change in qos models" [Low,In progress] - Assigned to Hong Hui Xiao (xiaohhui)
14:26:30 <ralonsoh> patch in https://review.openstack.org/410065
14:26:52 <ralonsoh> Ok, I see now it's almost merged
14:27:01 <ralonsoh> Next one
14:27:02 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/1646370
14:27:03 <openstack> Launchpad bug 1646370 in neutron "QosPolicyInUse after notifying the removal to backends" [Medium,In progress] - Assigned to Miguel Angel Ajo (mangelajo)
14:27:21 <ralonsoh> patch https://review.openstack.org/#/c/405448/
14:27:45 <ralonsoh> let's wait for ajo feedback...
14:27:54 <ralonsoh> but we should review it
14:28:17 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/1640762
14:28:18 <openstack> Launchpad bug 1640762 in neutron "when I update a Qos policy, value of shared can not be changed to false from true" [Undecided,Incomplete] - Assigned to Slawek Kaplonski (slaweq)
14:28:30 <ralonsoh> Still waiting for author feedback...
14:28:42 <slaweq> yep, and I couldn't reproduce it
14:28:55 <ralonsoh> ok, so no action now
14:29:05 <slaweq> yes
14:29:21 <ralonsoh> ok, last one
14:29:22 <ralonsoh> #link https://bugs.launchpad.net/neutron/+bug/1639186
14:29:24 <openstack> Launchpad bug 1639186 in neutron "qos max bandwidth rules not working for neutron trunk ports" [Low,Confirmed] - Assigned to Luis Tomas Bolivar (ltomasbo)
14:29:38 <ralonsoh> That's a difficult one
14:29:42 <ralonsoh> qos on trunk ports
14:30:18 <ralonsoh> and I don't see a solution to fix the solution proposed https://review.openstack.org/#/c/397788/
14:30:42 <ralonsoh> It's easy to use veth ports to connect bridges
14:30:52 <ralonsoh> but the performance drops a lot
14:31:02 <ralonsoh> and it's not useful for user space ovs
14:31:12 <slaweq> yep, there is nothing new here since last meeting
14:31:46 <ralonsoh> I'll ask in launchpad if there is any movement on this
14:31:53 <slaweq> ok
14:32:06 <ralonsoh> Ok, any other bug not listed??
14:32:24 <slaweq> I don't have anything new
14:32:28 <ralonsoh> ok
14:32:29 <ralonsoh> #topic Other Changes
14:32:48 <ralonsoh> there is nothing in the agenda for this
14:33:02 <ralonsoh> any comment?
14:33:21 <slaweq> it's not related to QoS but can You have a look on https://review.openstack.org/#/c/409432/6 ?
14:33:25 <slaweq> thx in advance
14:33:51 <ralonsoh> Yes, of course. I'll put this in my todo list for today
14:33:57 <slaweq> thx a lot
14:33:58 <davidsha> Same
14:34:05 <slaweq> thx davidsha
14:34:16 <ralonsoh> #topic Open Discussion
14:34:24 <ralonsoh> OSC: #link https://review.openstack.org/#/c/352477/
14:34:32 <ralonsoh> This is for ajo and dtroyer
14:34:48 <ralonsoh> It's been a while since I uploaded the first version
14:35:38 <ralonsoh> Open question: how can a developer speed up the reviews of his patches?
14:36:09 <slaweq> IMHO ask on IRC, and try to ask specified people and ask for review
14:36:23 <slaweq> from my experience it helps :)
14:36:44 <davidsha> I agree IRC.
14:36:46 <ralonsoh> ok, but I still have patches submitted nine months ago
14:36:51 * ajo is around now, reads the backlog
14:37:04 <ralonsoh> we are almost finish
14:37:10 <ralonsoh> sorry
14:37:24 <ajo> ack :)
14:37:32 <ajo> for https://review.openstack.org/#/c/405448/ I must change a testing detail and it's done, I was pulled out to do something else, but I finished it now.
14:37:57 <ralonsoh> cool
14:38:10 <ralonsoh> slaweq, davidsha
14:38:30 <ralonsoh> do you mind if I ask ajo now about #link https://bugs.launchpad.net/neutron/+bug/1578989
14:38:32 <openstack> Launchpad bug 1578989 in neutron "[RFE] Strict minimum bandwidth support (egress)" [Wishlist,In progress] - Assigned to Rodolfo Alonso (rodolfo-alonso-hernandez)
14:38:39 <davidsha> go ahead
14:38:42 <ralonsoh> The spec, #link https://review.openstack.org/#/c/396297/
14:39:04 <ralonsoh> ajo, Can you continue with the spec?
14:39:23 <ralonsoh> ajo: and please, take a look at #link https://review.openstack.org/#/c/401254/
14:39:31 <ralonsoh> that's all
14:39:39 <ajo> I think the spec is done,
14:39:48 <ajo> I probably should answer yamamoto's question
14:40:13 <ralonsoh> thanks!
14:40:14 <ajo> which I don't fully understand
14:40:25 <ralonsoh> me neither...
14:40:36 <ajo> may be I should try to answer the question on the spec itself by making it more clear
14:40:42 <ralonsoh> no, I do
14:40:48 <slaweq> but we was talking here that it can be related to Your patch https://review.openstack.org/#/c/396651/ and "improved validation"
14:41:22 <ralonsoh> this one too!
14:41:29 <slaweq> in case if e.g. port will be updated and new qos with strict min bw limit will be set for it
14:41:38 <ajo> yeah, that one I must definitely continue working on it
14:41:42 <ajo> I restarted it yesterday night
14:41:51 <ajo> and I must send another PS between today and tomorrow
14:41:55 <slaweq> :)
14:42:19 <ajo> sorry :/  I'm being the bottleneck
14:42:41 <ralonsoh> no problem, ask for help if you need it
14:43:05 <ajo> I'll ping you for reviews on new PS for https://review.openstack.org/#/c/396651/ , it's almost ready
14:43:07 <slaweq> yeah, I can help also :)
14:43:24 <ajo> thank you ralonsoh slaweq , If I see myself stopped again I will ping you
14:43:32 <ralonsoh> This, this patch should be merged ASAP
14:43:48 <ajo> yes
14:43:53 <slaweq> or we will ping You :P
14:44:01 <ralonsoh> hehehe
14:44:05 <ralonsoh> more topics?
14:44:12 <ajo> :D
14:44:18 <ajo> slaweq ralonsoh thanks ':D
14:44:25 <ajo> slaweq ping or punch? :D
14:44:32 <ralonsoh> second option!
14:44:33 <slaweq> ajo: we will see ;)
14:44:36 <ajo> CTCP punch
14:45:01 * ajo makes irc protocol jokes, and thinks he's old now
14:45:17 <slaweq> :)
14:45:26 <ralonsoh> ok guys, I'll close the meeting
14:45:32 <ralonsoh> see you around!
14:45:34 <ajo> Thank you very much ralonsoh & slaweq
14:45:37 <slaweq> thx
14:45:42 <davidsha> Thanks ralonsoh, see ya!
14:45:43 <ajo> I'll read the log right away
14:45:44 <slaweq> ajo: and davidsha
14:45:54 <ajo> and davidsha :D
14:45:57 <slaweq> :)
14:45:59 <slaweq> see You
14:46:03 <ralonsoh> #endmeeting