14:00:02 <igordcard> #startmeeting network_common_flow_classifier
14:00:07 <openstack> Meeting started Tue Jul 25 14:00:02 2017 UTC and is due to finish in 60 minutes.  The chair is igordcard. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:09 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:11 <openstack> The meeting name has been set to 'network_common_flow_classifier'
14:00:16 <igordcard> hi all
14:00:54 <bcafarel> hey igordcard
14:03:34 <igordcard> give me 4 minutes, as I have to change rooms, thanks
14:04:34 <bcafarel> that's a big building :)
14:05:55 <igordcard> back
14:06:05 <igordcard> alright, agenda:
14:06:12 <igordcard> https://wiki.openstack.org/wiki/Neutron/CommonClassificationFramework#Discussion_Topic_25_July_2017
14:06:23 <igordcard> #topic CCF v0 - first wave of code
14:08:02 <igordcard> so the plan was to submit the code before pike-3
14:08:49 <bcafarel> like, in 2 days?
14:09:18 <igordcard> david and thaynara are trimming a few rough edges and will submit the code by the end of the day
14:09:59 <igordcard> one thing it will not have is consistency with the API endpoints defined in the spec, yet
14:11:16 <igordcard> but most of the rest will be aligned already
14:12:14 <igordcard> david, what other info have you got?
14:13:52 <davidsha> Ya, It has the Ethernet, IPv4, IPv6, TCP and UDP. The problem with API inconsistency was that I made a resource for each one instead of specifying a type in a "Classification" resource.
14:14:21 <igordcard> so /classifications vs. /ipv4_classification and all the rest
14:14:40 <davidsha> The other way around, but yes
14:15:28 <davidsha> I'll submit the incorrect version just so people can see and test the general idea, and try to have the correct version up asap.
14:15:35 <igordcard> yes, current state is latter example right?
14:15:41 <davidsha> yes
14:16:42 <igordcard> so most of the parts will be the same except how the resources are expressed in the API right?
14:19:01 <davidsha> Yes, I'm not sure over internal database stuff will I make a generic "classification" object and make it the same as the API or actually store them in different tables based on type.
14:20:38 <davidsha> But this probably isn't so important if everything will be retrieving classifications through the API rather than accessing it directly.
14:21:38 <igordcard> in terms of tables the schema from the spec should work with both API approaches
14:22:07 <davidsha> ack, then it's fine as it is. It's just translate the new API on top of it.
14:22:09 <igordcard> and then at the API you'll have a generic classification object
14:22:11 <igordcard> or resource*
14:22:38 <davidsha> kk, cool.
14:22:54 <igordcard> definition validation will be handled by the plugin instead of the extension
14:24:18 <igordcard> davidsha: after this patch we split in different patches with the different logical pieces
14:24:38 <davidsha> Ya, we're planning on adding a little bit of validation. but just checking that for example an IPv6 address isn't passed into a IPv4 classification.
14:24:49 <igordcard> davidsha: when would those pieces be ready? (this is including the API change work)
14:24:54 <igordcard> davidsha: cool
14:27:05 <davidsha> igordcard: We'll aim for the next meeting.
14:27:20 <igordcard> alright davidsha
14:28:01 <igordcard> meanwhile with the patch from today we can start reviewing and discussing
14:28:04 <igordcard> moving on...
14:28:12 <igordcard> #topic Open discussion
14:28:21 <igordcard> bcafarel's favourite part
14:28:25 * bcafarel gets the coffee ready
14:28:28 <bcafarel> :)
14:28:29 <igordcard> I'm actually literally going to get a coffee now
14:28:34 <davidsha> :D
14:29:16 <davidsha> If there are any pastries could someone pass me one....
14:29:47 <davidsha> Are there any Questions about v0 code?
14:30:18 <bcafarel> sadly none from my part, I did not even have the time to open the review :(
14:30:34 <bcafarel> only time to +1 igordcard's repo cleanup commit
14:31:03 <igordcard> bcafarel: the code isn't there yet :p
14:31:09 <davidsha> Someone submitted an RFE for security groups and the neutron drivers recomended they adopt neutron classifier into security groups rather than extending the existing API.
14:31:19 <davidsha> Just as a heads up.
14:31:29 <igordcard> davidsha: was just about to link it here
14:31:34 <igordcard> https://bugs.launchpad.net/neutron/+bug/1690921
14:31:34 <openstack> Launchpad bug 1476527 in neutron "duplicate for #1690921 [RFE] Add common classifier resource" [Wishlist,Triaged] - Assigned to Igor Duarte Cardoso (igordcard)
14:32:19 <igordcard> generally the CCF looks enough to achieve this
14:32:37 <bcafarel> igordcard: ah that's why :) (I was thinking about the current poc review)
14:32:43 <igordcard> but depending on the very specifics of the use case there might a shortcoming or two
14:33:01 <bcafarel> nice, that could mean more people here (with interest on security groups)
14:33:04 <igordcard> so I'll conttinue to follow this discussion
14:34:08 <igordcard> bcafarel: yeah, more people from SG here would be good, to get some info about their plans and what it would take / and how long, to migrate to CCF
14:34:14 <davidsha> igordcard: Did you take up the task to actually implement the consumption of classifications in security groups?
14:35:00 <igordcard> we can also start having conversations with them, but the focus now is to wrap up v0 and then start doing incremental changes and open the window for enhancements
14:35:50 <igordcard> davidsha: no, why?
14:36:42 <davidsha> igordcard: you marked yourself an assignee
14:37:26 <igordcard> davidsha: that's the original common classifier RFE, you're looking at the wrong link
14:37:46 <davidsha> igordcard: Oh, moving on!
14:38:44 <igordcard> actually not sure whether that origianl RFE should be closed by now...
14:38:56 <igordcard> ihar created a blueprint specifically for the CCF some time ago:
14:39:20 <igordcard> https://blueprints.launchpad.net/neutron/+spec/common-classification-framework
14:39:39 <igordcard> but armax marked the SG RFE as duplicate of the original CCF RFE
14:39:52 <davidsha> Ya
14:40:17 <igordcard> probably too early in the morning for you armax, but if you're there do you think the RFE should be closed?
14:41:15 <armax> igordcard: which one?
14:41:28 <igordcard> https://bugs.launchpad.net/neutron/+bug/1476527
14:41:28 <openstack> Launchpad bug 1476527 in neutron "[RFE] Add common classifier resource" [Wishlist,Triaged] - Assigned to Igor Duarte Cardoso (igordcard)
14:41:45 <igordcard> armax: taking into account that this was created recently: https://blueprints.launchpad.net/neutron/+spec/common-classification-framework
14:42:11 <igordcard> armax: comment from RFE link: https://bugs.launchpad.net/neutron/+bug/1476527/comments/38
14:42:51 <armax> the RFE would close as released when there’s something being released, no?
14:43:40 <igordcard> armax: makes sense, was just wondering if it was still needed with the blueprint created to track the progress
14:44:50 <igordcard> armax: thank you; we are in open discussion, let us know if you have any question about ccf
14:45:05 <armax> the state change happens on its own typically once there’s a patch that markes the bug to be closed
14:46:58 <igordcard> armax: halright.. so the plan is to have an initial version soon, but the first consumable version only by Queens - that means the final Closes-Bug patch would be by Queens only correct?
14:47:30 <armax> yeah, I would think so
14:47:39 <igordcard> armax: perfect, thanks
14:48:34 <igordcard> any other questions?
14:49:39 <davidsha> I'm good
14:50:05 <igordcard> alright folks, this is all
14:50:20 <davidsha> thanks!
14:50:24 <igordcard> I actually got tea instead of coffee, but that should count too bcafarel?
14:50:30 <igordcard> bye all
14:50:33 <bcafarel> :)
14:50:43 <bcafarel> igordcard: I will think about it until next time
14:50:56 <igordcard> okay bcafarel!
14:51:04 <igordcard> #endmeeting