| *** santoshk has quit IRC | 00:00 | |
| openstackgerrit | Kanagaraj Manickam proposed openstack/tacker: Adds unit test case for scaling https://review.openstack.org/371368 | 00:04 |
|---|---|---|
| *** sripriya has quit IRC | 00:10 | |
| openstackgerrit | Sridhar Ramaswamy proposed openstack/tacker: VNFFG: neutron networking-sfc driver and plugin fixes https://review.openstack.org/347568 | 00:18 |
| sridhar_ram | trozet: ping | 00:22 |
| s3wong | sridhar_ram: on my IRC client, his nick is being faded out --- meaning that he has been inactive (or the computer went into sleep) | 00:22 |
| sridhar_ram | s3wong: possible.. i was trying to track down the portion of his code that grabs neutron ports (resource ids) | 00:23 |
| s3wong | sridhar_ram: I believe that was the get_detail that was done by someone else (forgot who) | 00:24 |
| sridhar_ram | s3wong: it was done by janki | 00:24 |
| s3wong | sridhar_ram: yeah, so look for that in trozet's code | 00:25 |
| *** KanagarajM_ has quit IRC | 00:26 | |
| sridhar_ram | s3wong: looking | 00:31 |
| *** yifei has joined #tacker | 00:38 | |
| *** n-harada has joined #tacker | 00:40 | |
| *** dkushwaha has joined #tacker | 00:44 | |
| s3wong | sridhar_ram: leaving early today to attend some tech talk at Facebook --- please send email if you encounter any further issue(s) | 00:47 |
| sridhar_ram | s3wong: okay, thanks | 00:47 |
| sridhar_ram | s3wong: will you be online later today ? | 00:47 |
| s3wong | sridhar_ram: probably --- during these times, I usually work at night after my daughter goes to bed | 00:57 |
| s3wong | sridhar_ram: so something like 10pm I should be online | 00:57 |
| sridhar_ram | s3wong: okay, thanks | 00:58 |
| s3wong | sridhar_ram: (that was the time I was trying to work with diga, but for that 4-5 nights I didn't see him online) | 00:58 |
| *** s3wong has quit IRC | 00:59 | |
| *** sripriya has joined #tacker | 01:00 | |
| *** xiayu has joined #tacker | 01:01 | |
| *** sripriya has left #tacker | 01:04 | |
| *** sripriya has joined #tacker | 01:04 | |
| openstackgerrit | Sripriya Seetharam proposed openstack/tacker: Fix respawn action for alarm monitor https://review.openstack.org/375181 | 01:05 |
| *** xiayu has left #tacker | 01:05 | |
| *** bobh has joined #tacker | 01:08 | |
| *** bobh has quit IRC | 01:19 | |
| openstackgerrit | Sripriya Seetharam proposed openstack/tacker: update vnf monitor to use vim type https://review.openstack.org/372802 | 01:49 |
| *** sripriya has quit IRC | 01:53 | |
| trozet | hi sridhar_ram, s3 | 02:03 |
| openstackgerrit | Sridhar Ramaswamy proposed openstack/tacker: VNFFG: neutron networking-sfc driver and plugin fixes https://review.openstack.org/347568 | 02:22 |
| sridhar_ram | trozet: hi | 02:23 |
| trozet | sridhar_ram: hi trying to test out your patch now | 02:23 |
| sridhar_ram | trozet: good...i just pushed a new one.. | 02:24 |
| trozet | sridhar_ram: heh | 02:24 |
| sridhar_ram | trozet: look at https://review.openstack.org/#/c/347568/21/tacker/nfvo/nfvo_plugin.py L233 | 02:25 |
| sridhar_ram | trozet: classifier needs logical_source_port .. | 02:25 |
| trozet | sridhar_ram: hmm | 02:26 |
| sridhar_ram | trozet: .. for now i'm picking the first CP in the chain .. | 02:26 |
| sridhar_ram | trozet: finally, finally.. after more than a week of tweaking.. VNFFG now goes to ACTIVE state | 02:26 |
| trozet | sridhar_ram: i think that is an issue | 02:26 |
| trozet | sridhar_ram: i think it is a constraint of the current networking-sfc design | 02:27 |
| trozet | sridhar_ram: they have to have flows that match on source port where traffic will come from? | 02:27 |
| trozet | well good it went active :) | 02:27 |
| sridhar_ram | yep.. frankly i didn't care for the "correctness" .. just tried to keep the neutron sfc API happy.. | 02:28 |
| sridhar_ram | trozet: at least .. it is a check-point | 02:28 |
| trozet | sridhar_ram: so then due to this constraint, it means we need to remove that line 233, and we need to require a user to put the network_src_port_id in their match criteria | 02:29 |
| sridhar_ram | trozet: i'm frustrated with lack of useful samples to create an "hello world" networking-sfc chain + classifier.. | 02:29 |
| trozet | sridhar_ram: yeah not good docs | 02:29 |
| trozet | sridhar_ram: are you ok with that approach? | 02:30 |
| sridhar_ram | trozet: hmm.. first i need to understand what is expected to be populated for logical_source_port in n-sfc.. | 02:31 |
| sridhar_ram | trozet: .. is it the first port in the chain .. or is it the port just before the first port (neutron router port?) where the pkts emerge towards the chain | 02:31 |
| sridhar_ram | trozet: .. next I'd rather not have user to make this decision if the system can "figure it" out | 02:32 |
| trozet | sridhar_ram: no i do not think it is the first port in the chain | 02:32 |
| trozet | sridhar_ram: the restriction is the very problem that NSH solves | 02:32 |
| trozet | sridhar_ram: its because they dont carry chain information in the packet, they ahve to reclassify every time a packet comes into br-int | 02:32 |
| trozet | sridhar_ram: so the easiest way is to always include the neutron port in that classifier | 02:32 |
| sridhar_ram | trozet: which neutron port ? there would be many in a multi-VNF chain | 02:33 |
| trozet | sridhar_ram: yeah so if you send packet to a VNF and it ocmes back into br-int, how do you know where that packet is in the chain? | 02:35 |
| trozet | sridhar_ram: since the packet carries no chain information (unlike NSH) | 02:35 |
| trozet | sridhar_ram: it is a regular packet | 02:35 |
| trozet | sridhar_ram: so the only way to know is to know what port it came back into the bridge on | 02:35 |
| sridhar_ram | trozet: based on what s3wong said (which u also mentioned about) the pkt get reclassified on every ingress into the br-int and placed on a mpls tunnel towards the next VNF port | 02:36 |
| trozet | sridhar_ram: yeah the mpls tunnel only applies across br-ints though | 02:37 |
| trozet | sridhar_ram: so I remember having this conversation with Cathy actually | 02:37 |
| trozet | sridhar_ram: they require the neutron port as part of hte match, and are planning on removing it | 02:37 |
| sridhar_ram | trozet: oh.. then we need multiple flow-classifiers .. one per VNF port ? | 02:37 |
| trozet | sridhar_ram: i think we talked about this a networking-sfc meeting | 02:37 |
| trozet | sridhar_ram: no. so the VNFs are all part of the chain, the match source port is the port where traffic will come from that you want to divert to the chain | 02:38 |
| sridhar_ram | trozet: well, the sad part is we are out of time for newton... i doubt we can take it as far as a traffic test in next couple of days.. we can try | 02:38 |
| trozet | sridhar_ram: http client--(port A)----VNF1---VNF2-- http server(port B) | 02:38 |
| trozet | sridhar_ram: you have to include port A in the classifier | 02:38 |
| trozet | sridhar_ram: right now you are using VNF1 ingress port, which isnt right | 02:39 |
| sridhar_ram | trozet: to me that make sense.. logical source is the "source" of all pkts toward the chain | 02:39 |
| trozet | sridhar_ram: so try removing that | 02:40 |
| trozet | sridhar_ram, | 02:40 |
| sridhar_ram | trozet: i understand.. again, i was trying to keep the API happy.. atleast now it proceeds and create port-pairs, port-pair-groups and flow-classifier (even if is incorrect) | 02:40 |
| trozet | sridhar_ram: create a cirros instance, find its neutron port, and include it in the match criteria of the NFP | 02:40 |
| sridhar_ram | trozet: no, without logical_source_port flow-classifier API fails | 02:40 |
| *** amotoki has quit IRC | 02:41 | |
| sridhar_ram | trozet: ah, thats punting the problem to the user.. hmm. i like it ;-) | 02:41 |
| trozet | sridhar_ram: yes | 02:41 |
| trozet | sridhar_ram: let me just make sure that is in the classifier stuff | 02:41 |
| trozet | sridhar_ram: when is hte last day we can get the changes in? | 02:42 |
| sridhar_ram | trozet: hmm.. this also mean this logical src port could be even a neutron port associated with a neutron router for a particalar network the VNFs are hanging off.. yeah ? | 02:42 |
| trozet | sridhar_ram: sure it could be | 02:42 |
| trozet | sridhar_ram: yeah so just create a router, then specify network_src_port_id in your NFP classifier | 02:43 |
| sridhar_ram | trozet: sounds good.. is this something you can pile into this same patchset ? https://review.openstack.org/347568 | 02:44 |
| trozet | sridhar_ram: yeah, i think jenkins verify was failing due to the find_vnf_resource by name thing | 02:44 |
| trozet | sridhar_ram: so we need to fix that too | 02:45 |
| sridhar_ram | trozet: yes.. | 02:45 |
| trozet | sridhar_ram: when is the last day we can merge this? | 02:45 |
| sridhar_ram | trozet: i'd like to merge this intial iteration by tomorrow.. | 02:46 |
| trozet | sridhar_ram: is the deadline monday though? | 02:46 |
| trozet | sridhar_ram: i need to do that dev ref doc as well | 02:46 |
| sridhar_ram | trozet: planning to request a release by Monday EOD | 02:46 |
| trozet | sridhar_ram: ok can you link me an example dev ref doc in the repo please? | 02:46 |
| sridhar_ram | trozet: yes, devref.. and just realized i also need a Reno releasenote | 02:46 |
| sridhar_ram | trozet: here it is .. https://github.com/openstack/tacker/tree/master/doc/source/devref | 02:47 |
| trozet | sridhar_ram: thanks just booted up tacker iwth your patch, let me see what happens | 02:47 |
| sridhar_ram | trozet: again, devref can be inspired heavily from your spec write up.. | 02:48 |
| sridhar_ram | trozet: we have room to fix up devref in a follow-on newton dot release | 02:48 |
| sridhar_ram | trozet: do you know how to create Reno release notes ? | 02:49 |
| trozet | sridhar_ram: yeah i think i did it once before | 02:49 |
| sridhar_ram | trozet: i can take it up if you are not comfortable | 02:49 |
| trozet | sridhar_ram: sure...i think i only did one line for hte client or something | 02:49 |
| sridhar_ram | trozet: okay, i'll push one | 02:50 |
| trozet | policy: | 02:50 |
| trozet | type: ACL | 02:50 |
| trozet | criteria: | 02:50 |
| trozet | - network_src_port_id: 640dfd77-c92b-45a3-b8fc-22712de480e1 | 02:50 |
| trozet | sridhar_ram: ^ i just created a cirros lets see | 02:50 |
| *** achatterjee has joined #tacker | 02:52 | |
| *** amotoki has joined #tacker | 02:53 | |
| trozet | sridhar_ram: failed https://paste.fedoraproject.org/433066/raw/ | 02:53 |
| trozet | sridhar_ram: but looks liek it ssomething to do with neutron client heh | 02:53 |
| trozet | sridhar_ram: i'm using stable/mitaka neutron, will the neutron auth changes you made not work for mitaka? | 02:53 |
| sridhar_ram | trozet: did u install networking-sfc ? | 02:54 |
| trozet | sridhar_ram: oh yeah it is 404... let me see | 02:54 |
| trozet | sridhar_ram: yeah i did, but maybe i didnt restart neutron..sec | 02:54 |
| sridhar_ram | trozet: neutron shd come up w/ networking-sfc extensions loaded | 02:55 |
| trozet | sridhar_ram: i dont see the extension loaded, do you have to specify it in the neutron.conf | 02:58 |
| trozet | sridhar_ram: it doesnt look like its trying to load it, but it is installed | 02:59 |
| trozet | sridhar_ram: i see https://ask.openstack.org/en/question/91643/using-networking-sfc-with-liberty-neutron-not-a-devstack/ | 03:00 |
| trozet | sridhar_ram: have to specify the plugins | 03:00 |
| trozet | 2016-09-23 03:02:32.466 12253 INFO neutron.api.extensions [req-fa2af36e-6aed-4d97-a50f-0bf3774f503d - - - - -] Loaded extension: flow_classifier | 03:03 |
| *** amotoki has quit IRC | 03:09 | |
| trozet | sridhar_ram: still there? | 03:14 |
| sridhar_ram | trozet: sorry, was afk.. | 03:15 |
| trozet | sridhar_ram: so i got it to kind of work.. | 03:15 |
| trozet | sridhar_ram: the classifier gets created, but port chain create fails | 03:15 |
| trozet | sridhar_ram: https://paste.fedoraproject.org/433073/raw/ | 03:15 |
| trozet | sridhar_ram: seen that one before? this was an OpenDaylight setup, so the instances were setup with ODL as the ML2 driver, i just switched it to be openvswitch as the ML2 plugin | 03:16 |
| sridhar_ram | trozet: no.. i used devstack master with that enable_plugin line for networking-sfc | 03:17 |
| sridhar_ram | trozet: https://wiki.openstack.org/wiki/Neutron/ServiceInsertionAndChaining | 03:18 |
| sridhar_ram | trozet: it has some sample .conf / .ini file | 03:18 |
| trozet | sridhar_ram: yeah i think i got all that worked out now | 03:18 |
| sridhar_ram | trozet: cool | 03:19 |
| *** amotoki has joined #tacker | 03:19 | |
| trozet | sridhar_ram: but anyway the classifier got created...so putting it into the vnffgd works | 03:19 |
| trozet | sridhar_ram: so we can remove that line 233 | 03:19 |
| sridhar_ram | trozet: can u post ur ffgd template ? | 03:19 |
| trozet | sridhar_ram: https://paste.fedoraproject.org/433075/raw/ | 03:20 |
| trozet | sridhar_ram: https://paste.fedoraproject.org/433076/ | 03:21 |
| sridhar_ram | trozet: hmm.. using an exact uuid makes VNFFGD template specific to one chain :( | 03:22 |
| trozet | sridhar_ram: no it makes it specific to one port to match traffic on | 03:22 |
| trozet | sridhar_ram: which is not good...but its a networking-sfc limitation- i wonder if we can list multiple ports there | 03:23 |
| trozet | networking_sfc/tests/unit/services/sfc/agent/test_agent.py: 'local_endpoint': '10.0.0.2' | 03:23 |
| trozet | sridhar_ram: looks like my setup is a little screwed cause i switched from ODL | 03:27 |
| trozet | sridhar_ram: anyway i will clean up your patch in the morning my time tmrw, then when you get online you can review. I think we are close | 03:27 |
| sridhar_ram | trozet: i guess so.. if we can get some labels to indicate for logical-source-port .. it wuld be better | 03:28 |
| sridhar_ram | trozet: lets think a bit more | 03:28 |
| *** uck has joined #tacker | 03:29 | |
| trozet | sridhar_ram: labels? | 03:33 |
| trozet | sridhar_ram: i mean we could allow port "name" and then do the translation like the resource name to ID function | 03:34 |
| sridhar_ram | trozet: more like params... | 03:34 |
| trozet | sridhar_ram: oh you mean like give an arg when you create vnffg --source-port or something? | 03:34 |
| sridhar_ram | trozet: which we can fill during vnffg-create time instead of vnffgd | 03:34 |
| trozet | sridhar_ram: i dont like that because then we are hacking around a design problem in networking-sfc | 03:34 |
| trozet | sridhar_ram: lets just do it like this, and I will talk to Cathy and see if we can proceed with getting that restriction removed | 03:35 |
| sridhar_ram | trozet: params are general Tosca constructs | 03:35 |
| trozet | sridhar_ram: oh | 03:35 |
| sridhar_ram | trozet: basically ur approach .. but the value will be substituted during vnffg-create | 03:36 |
| sridhar_ram | trozet: again this is optional for non n-sfc backend | 03:36 |
| trozet | sridhar_ram: how would it know what to substitute in? | 03:37 |
| trozet | sridhar_ram: would a user input it on the vnffg-create line? | 03:37 |
| sridhar_ram | trozet: --param file is passed to the vnffg-create cmd | 03:37 |
| sridhar_ram | trozet: damn, the client is already released | 03:37 |
| trozet | sridhar_ram: so that would be an extra patch to support that argument right? | 03:38 |
| trozet | sridhar_ram: yeah heh | 03:38 |
| sridhar_ram | trozet: yes | 03:38 |
| sridhar_ram | trozet: that's a super elegant solution... | 03:38 |
| trozet | sridhar_ram: i think we can fix this in networking-sfc as a bug fix | 03:38 |
| trozet | sridhar_ram: but i do like the params idea, for declaring the classifier at vnffg create time | 03:38 |
| sridhar_ram | trozet: that need to be supported from general Tosca point of view anyway | 03:39 |
| trozet | sridhar_ram: would be nice to use with matching network names, etc | 03:39 |
| *** amotoki has quit IRC | 03:39 | |
| sridhar_ram | trozet: exactly | 03:39 |
| trozet | sridhar_ram: ok well you functionally got it working. So I will clean up this patch in the morning, make the dev ref, and ping Cathy and see about that restriction | 03:39 |
| trozet | sridhar_ram: then we should be good for having something experimental that works | 03:40 |
| sridhar_ram | trozet: sounds good | 03:40 |
| sridhar_ram | trozet: definitely a progress, since last many days | 03:40 |
| trozet | sridhar_ram: thanks for driving the testing...sorry i was on vacation/not around | 03:40 |
| sridhar_ram | trozet: no worries at all... | 03:40 |
| sridhar_ram | trozet: try tomorrow | 03:40 |
| trozet | sridhar_ram: yep. talk to you tmrw | 03:41 |
| *** uck has quit IRC | 03:48 | |
| *** amotoki has joined #tacker | 03:49 | |
| *** hparekh has joined #tacker | 03:58 | |
| *** hparekh has quit IRC | 04:30 | |
| *** hparekh has joined #tacker | 04:31 | |
| *** tung_doan has joined #tacker | 04:43 | |
| *** hparekh has quit IRC | 04:57 | |
| *** tung_doan has quit IRC | 05:04 | |
| *** Vijayendra has joined #tacker | 05:08 | |
| *** tung_doan has joined #tacker | 05:09 | |
| *** s3wong has joined #tacker | 05:22 | |
| *** tung_doan has quit IRC | 05:25 | |
| *** hparekh has joined #tacker | 05:26 | |
| openstackgerrit | Kanagaraj Manickam proposed openstack/tacker: Adds release notes for VNF scaling feature https://review.openstack.org/373871 | 05:59 |
| *** neel has joined #tacker | 06:10 | |
| openstackgerrit | Himanshu Kumar proposed openstack/tacker: Changes in post_test_hook.sh to run dsvm-functional tests successfully https://review.openstack.org/372399 | 06:21 |
| openstackgerrit | Himanshu Kumar proposed openstack/tacker: changes in file post_test_hook.sh https://review.openstack.org/372897 | 06:25 |
| openstackgerrit | Kanagaraj Manickam proposed openstack/tacker: Adds release notes for VNF scaling feature https://review.openstack.org/373871 | 06:40 |
| openstackgerrit | Kanagaraj Manickam proposed openstack/tacker: Adds limitations to scaling dev ref https://review.openstack.org/373820 | 06:43 |
| *** neel has quit IRC | 06:46 | |
| *** neel has joined #tacker | 06:47 | |
| *** KanagarajM has joined #tacker | 06:49 | |
| KanagarajM | sridhar_ram, ping | 06:49 |
| sridhar_ram | KanagarajM: pong | 06:49 |
| KanagarajM | sridhar_ram, can you please review other two patches in scaling | 06:50 |
| *** Ravikiran_K has joined #tacker | 06:50 | |
| sridhar_ram | KanagarajM: let me lookup | 06:50 |
| KanagarajM | sridhar_ram, sure. | 06:51 |
| sridhar_ram | KanagarajM: I see this.. https://review.openstack.org/#/c/373820/ | 06:51 |
| sridhar_ram | KanagarajM: .. and unit testing.. | 06:51 |
| KanagarajM | sridhar_ram, yes pls | 06:52 |
| sridhar_ram | KanagarajM: for unit testing, my understanding is we are waiting for coverage for scaling changes in heat driver | 06:53 |
| sridhar_ram | KanagarajM: .. for the code https://github.com/openstack/tacker/blob/master/tacker/vnfm/infra_drivers/openstack/openstack.py#L291-L464 | 06:54 |
| openstackgerrit | Sridhar Ramaswamy proposed openstack/tacker: Add release notes for VNF Forwarding Graph https://review.openstack.org/375276 | 07:01 |
| openstackgerrit | dharmendra kushwaha proposed openstack/tacker-specs: Network Service Descriptor support in Tacker https://review.openstack.org/304667 | 07:06 |
| sridhar_ram | dkushwaha: ping | 07:07 |
| dkushwaha | sridhar_ram, hello | 07:07 |
| sridhar_ram | dkushwaha: any progress on the ceilometer triggering alarm mon ? | 07:08 |
| dkushwaha | sridhar_ram, not yet. I am still looking on it. | 07:09 |
| dkushwaha | sridhar_ram, waiting for Tung for discuss | 07:09 |
| sridhar_ram | dkushwaha: okay, just checking as we are close the deadline ... i'm glad ur on this subteam :) | 07:10 |
| sridhar_ram | dkushwaha: retiring for the day, will check back in the morning | 07:10 |
| dkushwaha | sridhar_ram, sure, will try to fix it by today | 07:11 |
| sridhar_ram | dkushwaha: thanks | 07:12 |
| *** s3wong_ has joined #tacker | 07:15 | |
| *** s3wong has quit IRC | 07:16 | |
| KanagarajM | sridhar_ram, its already taken care | 07:18 |
| KanagarajM | sridhar_ram, got the test, https://github.com/openstack/tacker/blob/master/tacker/tests/unit/vm/infra_drivers/openstack/test_openstack.py#L421-L427 | 07:20 |
| KanagarajM | sridhar_ram, as part of the current unit testing patch, the plugin part is taken care | 07:21 |
| sridhar_ram | KanagarajM: cool, thanks | 07:21 |
| KanagarajM | sridhar_ram, :) i think its very late night for you ... still online !! | 07:22 |
| sridhar_ram | KanagarajM: yeah, need to shutdown... long day fighting vnf ffg feature | 07:23 |
| KanagarajM | sridhar_ram, ok. take good rest ... good night ! | 07:24 |
| *** s3wong_ has quit IRC | 07:40 | |
| *** dave-johnston has joined #tacker | 07:46 | |
| *** amotoki_ has joined #tacker | 08:00 | |
| *** amotoki has quit IRC | 08:03 | |
| *** Ravikiran_K has quit IRC | 08:19 | |
| *** ksantoshk has quit IRC | 08:24 | |
| *** neel has quit IRC | 08:24 | |
| *** amotoki_ is now known as amotoki | 08:24 | |
| *** Ravikiran_K has joined #tacker | 08:26 | |
| *** openstackgerrit has quit IRC | 09:03 | |
| *** openstackgerrit has joined #tacker | 09:04 | |
| *** amotoki has quit IRC | 09:32 | |
| *** janki has joined #tacker | 09:37 | |
| janki | trozet: ping | 09:41 |
| *** saju_m has joined #tacker | 09:41 | |
| *** neel has joined #tacker | 09:45 | |
| *** KanagarajM has quit IRC | 09:46 | |
| *** amotoki has joined #tacker | 09:54 | |
| *** _achatterjee_ has joined #tacker | 09:54 | |
| *** amotoki has quit IRC | 09:55 | |
| *** achatterjee has quit IRC | 09:57 | |
| *** janki has quit IRC | 09:59 | |
| *** tung_doan has joined #tacker | 10:15 | |
| *** Ravikiran_K has quit IRC | 10:39 | |
| *** neel has quit IRC | 11:03 | |
| *** _achatterjee_ has quit IRC | 11:30 | |
| *** amotoki has joined #tacker | 11:38 | |
| openstackgerrit | Manikantha Srinivas Tadi proposed openstack/tacker: Store the right description of VNF given from Horizon/CLI. https://review.openstack.org/331039 | 12:10 |
| *** dkushwaha has quit IRC | 12:12 | |
| openstackgerrit | Manikantha Srinivas Tadi proposed openstack/tacker: Remove the service id from vnfd-create command output https://review.openstack.org/348874 | 12:12 |
| *** hparekh has quit IRC | 12:16 | |
| *** manikanta_tadi has quit IRC | 12:50 | |
| openstackgerrit | vishwanath jayaraman proposed openstack/tacker: Devref for event logging feature https://review.openstack.org/370428 | 13:04 |
| openstackgerrit | venkatamahesh proposed openstack/tacker: Added py34, py35 to tox.ini https://review.openstack.org/375468 | 13:20 |
| *** dkushwaha has joined #tacker | 13:53 | |
| *** Ravikiran_K has joined #tacker | 13:54 | |
| *** tung_doan is now known as tung_doan_afk | 14:05 | |
| *** saju_m has quit IRC | 14:11 | |
| *** Liuqing has joined #tacker | 14:19 | |
| *** tung_doan_afk is now known as tung_doan | 14:34 | |
| *** amotoki has quit IRC | 14:58 | |
| *** santoshk has joined #tacker | 15:00 | |
| *** santoshk has quit IRC | 15:04 | |
| *** vishwanathj has joined #tacker | 15:05 | |
| *** amotoki has joined #tacker | 15:14 | |
| *** amotoki has quit IRC | 15:17 | |
| *** Liuqing has quit IRC | 15:19 | |
| *** lamt has quit IRC | 15:39 | |
| openstackgerrit | Tim Rozet proposed openstack/tacker: VNFFG: neutron networking-sfc driver and plugin fixes https://review.openstack.org/347568 | 15:55 |
| trozet | sridhar_ram:^ i think it is all good now, should pass CI | 15:55 |
| sridhar_ram | trozet: cool.. | 15:56 |
| sridhar_ram | trozet: I'm thinking to land this first.. but we should continue to test for other issues.. | 15:57 |
| sridhar_ram | trozet: i haven't tried vnf_mapping .. | 15:57 |
| *** lulei has quit IRC | 16:03 | |
| *** gongysh has joined #tacker | 16:03 | |
| gongysh | dkushwaha, hi | 16:03 |
| gongysh | https://review.openstack.org/#/c/373820/3/doc/source/devref/scale_usage_guide.rst | 16:03 |
| gongysh | dkushwaha, what does mean by "Scaling feature does not support to selectively choose the VDU as part | 16:03 |
| gongysh | of scaling.?" | 16:03 |
| gongysh | dkushwaha, does we support to choose vdu as the target of scale policy? | 16:04 |
| gongysh | policies: | 16:04 |
| gongysh | sp1: | 16:04 |
| gongysh | type: tosca.policies.tacker.Scaling | 16:04 |
| gongysh | description: Simple VDU scaling | 16:04 |
| gongysh | properties: | 16:04 |
| gongysh | min_instances: 1 | 16:04 |
| gongysh | max_instances: 3 | 16:04 |
| gongysh | default_instances: 2 | 16:04 |
| gongysh | increment: 1 | 16:04 |
| gongysh | targets: [vdu1, vdu2] | 16:04 |
| gongysh | sridhar_ram, do you have idea? | 16:04 |
| *** lamt has joined #tacker | 16:12 | |
| openstackgerrit | Merged openstack/tacker: Adds unit test case for scaling https://review.openstack.org/371368 | 16:26 |
| *** gongysh has quit IRC | 16:32 | |
| *** dave-johnston has quit IRC | 16:41 | |
| *** saju_m has joined #tacker | 16:47 | |
| openstackgerrit | Merged openstack/tacker: Add release notes for VNF Forwarding Graph https://review.openstack.org/375276 | 17:04 |
| *** vishnoianil has quit IRC | 17:05 | |
| trozet | sridhar_ram: yeah | 17:13 |
| *** mbirru has quit IRC | 17:50 | |
| *** uck has joined #tacker | 17:53 | |
| *** vishnoianil has joined #tacker | 17:56 | |
| *** uck has quit IRC | 18:01 | |
| *** vishnoianil has quit IRC | 18:01 | |
| *** s3wong has joined #tacker | 18:12 | |
| *** vishnoianil has joined #tacker | 18:14 | |
| *** Ravikiran_K has quit IRC | 18:25 | |
| openstackgerrit | Tung Doan proposed openstack/tacker: Fix alarms stuck in "insufficient state" and update usage guide https://review.openstack.org/375686 | 19:26 |
| *** sripriya has joined #tacker | 19:37 | |
| *** vishnoianil has quit IRC | 19:38 | |
| *** uck has joined #tacker | 19:43 | |
| openstackgerrit | Merged openstack/tacker: Adds limitations to scaling dev ref https://review.openstack.org/373820 | 19:50 |
| openstackgerrit | Tim Rozet proposed openstack/tacker: Adds devref guides for VNFFG https://review.openstack.org/375691 | 19:50 |
| trozet | sridhar_ram:^ | 19:52 |
| openstackgerrit | Sripriya Seetharam proposed openstack/tacker: Fix respawn action for alarm monitor https://review.openstack.org/375181 | 19:52 |
| sripriya | trozet: s3wong: sridhar_ram: https://review.openstack.org/#/c/347568/ i see a green build now, can you help me understand where exactly it stands now in terms of creating a successful vnffg? | 19:53 |
| *** vishnoianil has joined #tacker | 19:56 | |
| *** sripriya has quit IRC | 20:23 | |
| *** sripriya has joined #tacker | 20:24 | |
| *** saju_m has quit IRC | 20:24 | |
| sridhar_ram | sripriya: as of now, vnffg is going to ACTIVE state, the restriction is logical-src-port id is mandatory in VNFD. We remove this restriction either working w/ n-sfc team or introducing param support for vnffg (in Ocata) | 20:29 |
| sripriya | sridhar_ram: thanks for the update | 20:34 |
| *** sripriya has quit IRC | 20:46 | |
| *** sripriya has joined #tacker | 20:47 | |
| *** sripriya has quit IRC | 20:48 | |
| openstackgerrit | Tony Xu proposed openstack/python-tackerclient: Remove pdb https://review.openstack.org/375718 | 20:49 |
| s3wong | sridhar_ram: so the OVS side is relatively pain-free? | 20:50 |
| sridhar_ram | s3wong: well, one of the gotchas was logical_source_port id is mandatory and need to be specified by the user in the VNFFGD TOSCA template.. | 20:51 |
| *** uck has quit IRC | 20:51 | |
| *** tung_doan is now known as tung_doan_afk | 20:52 | |
| *** sripriya has joined #tacker | 20:52 | |
| s3wong | sridhar_ram: (putting on my networking-sfc hat) I think that is needed so we know which OVS data path's flow table to add the fc => flow entries | 20:52 |
| sridhar_ram | s3wong: i didn't do any traffic test yet.. but the above restriction is bad user-experience.. | 20:52 |
| sridhar_ram | s3wong: understood, but trozet mentioned there was a thought to remove this restriction in the n-sfc scheme | 20:53 |
| s3wong | sridhar_ram: in fact, I would guess once symmetrical is supported, even the destination logical port needs to be resolved at the time we deploy the chain | 20:53 |
| sridhar_ram | s3wong: do you know more about that? | 20:53 |
| s3wong | sridhar_ram: actually no :-) | 20:53 |
| s3wong | sridhar_ram: I do, however, agree that the chain can be a logical construct, and we can have another API to deploy (when the chain needs to be realized from source -> destination) | 20:54 |
| *** TonyXu has joined #tacker | 21:13 | |
| *** lamt has quit IRC | 22:01 | |
| dkushwaha | sridhar_ram, sripriya kindly review the patch https://review.openstack.org/#/c/375686/ | 22:13 |
| sridhar_ram | dkushwaha: sure... what is the current state of things? Do you still have respawn issue ? | 22:13 |
| sripriya | dkushwaha: will do, per your comment, looks ceilometer alarm is triggering the alarm with the new fix. | 22:14 |
| dkushwaha | sridhar_ram, regarding respawn, i suspect its might be environmental issue. Because alarm is triggering now | 22:15 |
| sridhar_ram | dkushwaha: okay, and basic respawn from ping / http_ping works fine based on dsvm results | 22:16 |
| sripriya | dkushwaha: i have a related fix for respwan using alarm monitor, not sure if you are facing the same one https://review.openstack.org/#/c/375181/ | 22:17 |
| dkushwaha | sridhar_ram, will check for basic ping | 22:17 |
| dkushwaha | sripriya, ok, let me check | 22:18 |
| *** sripriya has quit IRC | 22:34 | |
| *** sripriya has joined #tacker | 22:42 | |
| openstackgerrit | Sridhar Ramaswamy proposed openstack/tacker: VNFFG: neutron networking-sfc driver and plugin fixes https://review.openstack.org/347568 | 22:47 |
| dkushwaha | sridhar_ram, I am accessing my office system remotely so unable to login on horizon. Could you please let me know how can i disable ping on cirros, or how can shutdown it? | 22:48 |
| sripriya | dkushwaha: you can ssh to the cirros instance and access the nic | 23:00 |
| dkushwaha | sripriya, yes, but permission issue. No sudo allowed on cirros. btw I did it with nova delete :) | 23:01 |
| sripriya | dkushwaha: ‘cirros’ user has root access, you can just swap to the root user to bring down the nic. anyways :-) | 23:02 |
| dkushwaha | sripriya, ok. Thanks | 23:04 |
| *** vishwanathj has quit IRC | 23:05 | |
| openstackgerrit | Sridhar Ramaswamy proposed openstack/tacker: VNFFG: neutron networking-sfc driver and plugin fixes https://review.openstack.org/347568 | 23:07 |
| sridhar_ram | s3wong: trozet: hi guys.. just tried vnffg-delete. Looks only the underlying flow-classifier is correctly delete. The port-pair-group and pair-pairs are not deleted.. See http://paste.openstack.org/show/582830/ | 23:12 |
| sridhar_ram | sripriya: looks you already pointed this out in the last comment.. some how i missed it | 23:14 |
| sripriya | sridhar_ram: yeah, the methods just need to be invoked as part of delete_chain call in https://review.openstack.org/#/c/347568/24/tacker/nfvo/drivers/vim/openstack_driver.py@416 | 23:16 |
| sridhar_ram | sripriya: ack. looks the port-pairs / group id is not stored in tacker.. need to do a getter of the leaf objs, delete them and unwind back | 23:18 |
| sripriya | sridhar_ram: we should be able to delete them based on names as well right? | 23:20 |
| sripriya | sridhar_ram: since you now handle the name based on vnf[‘name], we could pass these down probably? | 23:21 |
| sripriya | sridhar_ram: vnf['name'] + '-port-pair-group' and vnf['name'] + '-connection-points' | 23:22 |
| sridhar_ram | sripriya: yes, name shd work given then are now unique, thanks to u :) | 23:23 |
| sridhar_ram | sripriya: i need to touch up delete_chain() abstract interface.. but doable | 23:23 |
| sripriya | sridhar_ram; np. i just thought about the same, about ‘name’ now being unique :) | 23:24 |
| sripriya | sridhar_ram: ur we had to base it off vnf[‘id’] :p | 23:24 |
| sripriya | s/ur/or | 23:25 |
| sridhar_ram | sripriya: no thanks.. will stick to 'name' ;-) | 23:25 |
| sripriya | sridhar_ram: lol | 23:25 |
| dkushwaha | sripriya, i tried your patch. With this patch respawning started, but failed with error: http://paste.openstack.org/show/582831/ and the vm got deleted itself | 23:35 |
| dkushwaha | sridhar_ram, I also tested basic respawn with ping, and it is working fine | 23:37 |
| *** sripriya has quit IRC | 23:39 | |
| sridhar_ram | dkushwaha: great, thanks for checking.. sigh of relief | 23:49 |
| *** sripriya has joined #tacker | 23:49 | |
| sripriya | dkushwaha: is this on top of tung_doan’s new patch? | 23:51 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!