| sripriya | bryan_att: ok, that implies the heat service is not setup properly | 00:01 |
|---|---|---|
| bryan_att | sripriya: OK, let me take that up with the JOID and Apex teams in OPNFV. Some difference apparently in how they install Heat. | 00:01 |
| sripriya | bryan_att: i suspect it to be with the additional steps required for heat talk to identity service correctly as pointed to earlier | 00:05 |
| bryan_att | well at least those steps are documented, right? ;-) | 00:06 |
| sripriya | bryan_att: yes | 00:06 |
| bryan_att | anywhere specific? | 00:07 |
| bryan_att | (we are not using Keystone V3) | 00:07 |
| sripriya | bryan_att: i would look into the heat user ‘heat_domain_admin’ being create in the ‘heat’ domain | 00:11 |
| sripriya | bryan_att: do you see that in the users under ‘Identity’ in dashboard? | 00:12 |
| bryan_att | checking | 00:13 |
| bryan_att | https://www.irccloud.com/pastebin/Uagaaglr/ | 00:14 |
| bryan_att | no user "heat_domain_admin" | 00:14 |
| bryan_att | https://www.irccloud.com/pastebin/gYCOnRrM/ | 00:15 |
| sripriya | bryna_att: the user is an internal user created by heat | 00:16 |
| sripriya | bryan_att: ^ | 00:17 |
| sripriya | bryan_att: can you confirm from the dashboard? | 00:17 |
| bryan_att | not sure what you mean - the user list is the same in the dashboard | 00:17 |
| bryan_att | not sure what an "internal" user is, do you mean a services project user? | 00:18 |
| sripriya | bryan_att: ‘heat_domain_admin’ user is an internal service project user unlike other users | 00:19 |
| bryan_att | OK, it's not in the list of users for the services project (or any other). How is that user created normally (automatically by Heat?) | 00:19 |
| sripriya | bryan_att: as part of heat manual setup , it needs to be setup | 00:20 |
| bryan_att | and does the creation of that user apply when the Keystone V3 API is not used? | 00:21 |
| bryan_att | the guide references the "domain" attribute which is not applicable for V2 AFAIK | 00:21 |
| bryan_att | maybe the user "heat-cfn_heat" (which is in the list) has the same purpose as the "heat_domain_admin" user (and applies to the same purpose in Keystone V2 environments?) | 00:23 |
| *** uck_ has quit IRC | 00:23 | |
| sripriya | bryan_att: yes applies to v2.0 | 00:24 |
| *** uck has joined #tacker | 00:24 | |
| sripriya | bryan_att: cfn is for the cloud formation and not related to this | 00:24 |
| bryan_att | OK, so if I create the user, and leave off the "domain" attribute maybe it will work? | 00:25 |
| sripriya | bryan_att: no, you will need the domain as well | 00:28 |
| bryan_att | but the domain is N/A to Keystone V2? | 00:28 |
| *** uck has quit IRC | 00:28 | |
| bryan_att | sripriya: "domain" parameter is N/A https://www.irccloud.com/pastebin/J412I7LR/ | 00:30 |
| sripriya | bryan_att: the heat domain is stil required, however not using the above command as obviously v2.o does not support multiple domain | 00:33 |
| sripriya | bryan_att: if i look up earlier docs from kilo, heat had a custome command to setup the domain , refer http://docs.openstack.org/kilo/install-guide/install/apt/content/heat-install-controller-node.html To install and configure the Orchestration components section 3.b | 00:34 |
| bryan_att | Ok I'll check it out | 00:36 |
| sripriya | bryan_att: keystone can talk both v2.0 and v3 on mitaka , can you enable your environment variables with v3 using: export OS_IDENTITY_API_VERSION=3 | 00:36 |
| sripriya | and export OS_AUTH_URL=http://10.18.160.13:5000/v3 | 00:36 |
| bryan_att | Ok I'll try that too | 00:37 |
| sripriya | bryan_att: that way you can run the commands specified to setup heat domain , user and roles using the openstack commands itself | 00:37 |
| *** sripriya has quit IRC | 00:47 | |
| *** gongysh has joined #tacker | 01:12 | |
| *** dkushwaha has quit IRC | 01:12 | |
| gongysh | s3wong, hi | 01:12 |
| s3wong | gongysh: sorry --- didn't get a chance over the Labor Day weekend to review your change | 01:17 |
| s3wong | gongysh: will need to check it out later today | 01:17 |
| gongysh | s3wong, you own me a review https://review.openstack.org/#/c/363398/ | 01:18 |
| gongysh | how's your vnffg feature's progress? can I test it now? | 01:18 |
| s3wong | gongysh: yeah, I know --- I also owe trozet a quick look at his modification of my abstract driver | 01:19 |
| s3wong | gongysh: not yet --- end of the week | 01:19 |
| gongysh | s3wong, fine. come on. | 01:19 |
| s3wong | gongysh: I will post a new patch (hopefully) tonight; then will work with trozet and diga on e2e testing | 01:19 |
| s3wong | gongysh: though probably mostly them than me :-) | 01:20 |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/tacker: Clean imports in code https://review.openstack.org/359023 | 01:20 |
| *** prashantD_ has joined #tacker | 01:24 | |
| *** prashantD__ has joined #tacker | 01:28 | |
| *** amotoki has joined #tacker | 01:29 | |
| *** prashantD_ has quit IRC | 01:31 | |
| *** amotoki has quit IRC | 01:33 | |
| openstackgerrit | vishwanath jayaraman proposed openstack/tacker: VNF Scaling event capture: Functional Test https://review.openstack.org/366433 | 01:35 |
| vishwanathj | gongysh hi....FYI, the functional test for https://review.openstack.org/#/c/363946/ will be captured as part of https://review.openstack.org/366433 .... the 366433 patchset builds on KanagarajM functional patchset for scaling --> https://review.openstack.org/#/c/363573/ | 01:37 |
| *** diga has quit IRC | 01:39 | |
| sridhar_ram | gongysh: can you pls merge https://review.openstack.org/364964 ? | 01:40 |
| gongysh | sridhar_ram , sure | 01:41 |
| gongysh | sridhar_ram, is the stable/newton created? | 01:41 |
| sridhar_ram | gongysh: yes | 01:41 |
| sridhar_ram | for python-tackerclient | 01:41 |
| *** amotoki has joined #tacker | 01:46 | |
| vishwanathj | gongysh hi | 01:50 |
| gongysh | vishwanathj, hi | 01:50 |
| vishwanathj | gongysh with the patchset https://review.openstack.org/#/c/366433/ uploaded will the -1 from you for https://review.openstack.org/#/c/363946/ be removed? Or is there anything else that I would need to do to remove the -1...thanks | 01:51 |
| gongysh | vishwanathj, I +4. done. | 01:54 |
| vishwanathj | gongysh thanks ... appreciate it | 01:54 |
| gongysh | vishwanathj, I still think we should try our best to do feature and companion test (unit and functional) in on shot. | 01:56 |
| vishwanathj | gongysh sounds good .... will go with your suggestion the next time around | 01:58 |
| openstackgerrit | Merged openstack/tacker: Logs events for VNF scale operations https://review.openstack.org/363946 | 01:58 |
| vishwanathj | in this particular case, I am leveraging KanagarajM work which is WIP | 01:59 |
| gongysh | vishwanathj, thanks. by the way, I am not able to join the design summit in Spain (budget). it seems we cannot enjoy the coffee or beer together. | 01:59 |
| vishwanathj | gongysh oh man ... why is that? | 02:00 |
| gongysh | vishwanathj, our company just arranges budget for two guys who have topic selected. | 02:00 |
| vishwanathj | gongysh so its a budget issue ... was looking forward to meeting you at the tacker design sessions .... I will take an extra coffee on behalf of you | 02:01 |
| *** amotoki has quit IRC | 02:05 | |
| *** s3wong has quit IRC | 02:06 | |
| *** dkushwaha has joined #tacker | 02:07 | |
| gongysh | vishwanathj, extra 10, 2 each days. | 02:10 |
| *** vishnoianil has quit IRC | 02:11 | |
| *** amotoki has joined #tacker | 02:20 | |
| *** amotoki has quit IRC | 02:28 | |
| vishwanathj | gongysh sure | 02:33 |
| tung_doan | vishwanathj: ping | 02:35 |
| dkushwaha | sridhar_ram, ping | 02:36 |
| openstackgerrit | Tung Doan proposed openstack/tacker: Alarm monitor: All in one https://review.openstack.org/365435 | 02:38 |
| dkushwaha | sridhar_ram, Regarding your comment in https://review.openstack.org/#/c/319028/ . Your comment is for all part, not only test_wsgi.py . Right? | 02:41 |
| openstackgerrit | hyunsik Yang proposed openstack/tacker: portscan driver for VNF application Monitoring https://review.openstack.org/329409 | 02:41 |
| *** amotoki has joined #tacker | 02:41 | |
| vishwanathj | tung_doan ping | 02:46 |
| tung_doan | vishwanathj: it seems like "auth_attr" param is not used in "log" and "log_and_kill" actions.. To be unified, it is fine for me.. I just wonder whether any reason is begind | 02:46 |
| tung_doan | *begind --> behind | 02:46 |
| vishwanathj | tung_doan refer https://github.com/openstack/tacker/blob/master/tacker/vm/plugin.py#L189 .... | 02:48 |
| *** amotoki has quit IRC | 02:48 | |
| vishwanathj | where execute_action is being called .... | 02:48 |
| vishwanathj | when I tried to use a template that had the execute action as 'log' and 'log_and_kill', there was an error indicating one less parameter being passed | 02:51 |
| tung_doan | ok.. I knew that.. good catch. thanks | 02:52 |
| *** vishwanathj has quit IRC | 02:54 | |
| *** lulei has joined #tacker | 02:57 | |
| sridhar_ram | dkushwaha: yes, rather do this for all parts, no just test_wsgi | 03:03 |
| sridhar_ram | dkushwaha: also, abandon this patchset as well | 03:03 |
| dkushwaha | sridhar_ram, ok, I will do. | 03:04 |
| dkushwaha | sridhar_ram, thanks | 03:04 |
| gongysh | tung_doan, hi | 03:12 |
| gongysh | ubuntu@tacker:~/devstack$ /usr/local/bin/aodh-api -d -v --config-file /etc/aodh/aodh.conf & echo $! >/opt/stack/status/stack/aodh-api.pid; fg || echo "aodh-api failed to start" | tee "/opt/stack/status/stack/aodh-api.failure" | 03:12 |
| gongysh | [1] 17206 | 03:12 |
| gongysh | usage: aodh-api [-h] [--port PORT] -- [passed options] | 03:12 |
| gongysh | aodh-api: error: unrecognized arguments: -d -v --config-file | 03:12 |
| gongysh | aodh-api failed to start | 03:12 |
| gongysh | ubuntu@tacker:~/devstack$ /usr/local/bin/aodh-api -h | 03:12 |
| gongysh | usage: aodh-api [-h] [--port PORT] -- [passed options] | 03:12 |
| gongysh | my devstack | 03:12 |
| gongysh | tung_doan, have you bumped into this problem? | 03:12 |
| *** xiayu1 has joined #tacker | 03:18 | |
| *** xiayu has quit IRC | 03:20 | |
| *** janki has joined #tacker | 03:28 | |
| *** janki has quit IRC | 03:28 | |
| *** sripriya has joined #tacker | 03:36 | |
| *** xiayu has joined #tacker | 03:42 | |
| *** janki has joined #tacker | 03:43 | |
| *** links has joined #tacker | 03:45 | |
| *** xiayu1 has quit IRC | 03:45 | |
| *** KanagarajM has joined #tacker | 03:46 | |
| *** vishnoianil has joined #tacker | 03:51 | |
| gongysh | sripriya, https://review.openstack.org/#/c/363398/ please. thanks | 03:56 |
| gongysh | KanagarajM, https://review.openstack.org/#/c/363398/ please thanks. | 03:57 |
| sripriya | gongysh: thanks for fixing the bug, reviewing it now | 04:00 |
| tung_doan | gongysh: hi gonggysh, i dont get this error.. mu devstack worked well when i enabled aodh plugin | 04:15 |
| tung_doan | dkushwaha: ping | 04:17 |
| dkushwaha | tung_doan, hello | 04:18 |
| tung_doan | dkushwaha: also, please kindly let me know if you enable aodh plugin in devstack? | 04:18 |
| dkushwaha | tung_doan, yes i did it | 04:18 |
| tung_doan | dkushwaha: did you try with curl -X POST.. does it work well to respawn? | 04:19 |
| dkushwaha | tung_doan, i could not test again yet. | 04:20 |
| *** amotoki has joined #tacker | 04:20 | |
| dkushwaha | tung_doan, i will try, and let you | 04:20 |
| dkushwaha | tung_doan, let you know | 04:21 |
| tung_doan | dkushwaha: ok.... thanks | 04:21 |
| dkushwaha | tung_doan, thanks :) | 04:22 |
| openstackgerrit | dharmendra kushwaha proposed openstack/tacker: Refactor Tacker unit tests to remove xml support https://review.openstack.org/366479 | 04:25 |
| *** hparekh has joined #tacker | 04:26 | |
| *** tbh has joined #tacker | 04:42 | |
| openstackgerrit | Janki Chhatbar proposed openstack/tacker: Deprecate warning for infra_driver and mgmt_driver at server https://review.openstack.org/363455 | 04:46 |
| *** tbh has quit IRC | 04:52 | |
| *** prashantD_ has joined #tacker | 05:03 | |
| *** prashantD__ has quit IRC | 05:06 | |
| *** prashantD_ has quit IRC | 05:07 | |
| *** neel has joined #tacker | 05:07 | |
| *** sripriya has quit IRC | 05:12 | |
| *** openstackgerrit has quit IRC | 05:18 | |
| *** openstackgerrit has joined #tacker | 05:19 | |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/tacker: Clean imports in code https://review.openstack.org/359023 | 05:37 |
| *** yifei has quit IRC | 06:11 | |
| openstackgerrit | hyunsik Yang proposed openstack/tacker: Port scan driver for VNF application Monitoring https://review.openstack.org/329409 | 06:15 |
| *** yifei has joined #tacker | 06:20 | |
| openstackgerrit | gongysh proposed openstack/tacker: Fix the monitor bug https://review.openstack.org/363398 | 06:31 |
| openstackgerrit | gongysh proposed openstack/tacker: Fix the monitor bug https://review.openstack.org/363398 | 06:34 |
| *** Ravikiran_K has joined #tacker | 06:44 | |
| openstackgerrit | gongysh proposed openstack/tacker: Fix the monitor bug https://review.openstack.org/363398 | 06:46 |
| sridhar_ram | janki: ping | 06:47 |
| openstackgerrit | Merged openstack/tacker: Refactor Tacker unit tests to remove xml support https://review.openstack.org/366479 | 06:47 |
| janki | sridhar_ram, pong | 06:47 |
| sridhar_ram | janki: are you planning to respin https://review.openstack.org/#/c/340838/ ? | 06:47 |
| janki | sridhar_ram, yes. I have included unit testcases for active and inactive vnfs. | 06:48 |
| janki | sridhar_ram, do I need to include fucntional test case for inactive vnf? already have the same for active vnf | 06:48 |
| sridhar_ram | janki: no unit test is okay for now, as this is a critical piece for VNFFG and it needs to land soon.. | 06:50 |
| sridhar_ram | janki: this is the one pending on mocking heat api ? | 06:50 |
| janki | sridhar_ram, yes it was. figured out how to mock. | 06:50 |
| janki | sridhar_ram, pushing it now. | 06:51 |
| sridhar_ram | janki: cool | 06:51 |
| sridhar_ram | gongysh: please review ^^^ and help to wrap this up | 06:51 |
| openstackgerrit | Janki Chhatbar proposed openstack/tacker: Add VNF resource details to get vnf API https://review.openstack.org/340838 | 06:52 |
| openstackgerrit | Janki Chhatbar proposed openstack/tacker: VNFFG abstract driver https://review.openstack.org/347563 | 06:52 |
| janki | sridhar_ram, gongysh ^^ | 06:52 |
| gongysh | got it | 06:52 |
| sridhar_ram | gongysh: thanks | 06:52 |
| openstackgerrit | Janki Chhatbar proposed openstack/tacker: Add VNF resource details to get vnf API https://review.openstack.org/340838 | 07:00 |
| *** manikanta_tadi has joined #tacker | 07:00 | |
| *** manikanta_tadi has quit IRC | 07:00 | |
| gongysh | janki, see comments on 30 and 31 patchset. | 07:08 |
| janki | gongysh, about the heat API not available, heat api is never actually called and mocked always. so that scenario might not be created | 07:10 |
| gongysh | then you can unit test the heat driver itself. | 07:11 |
| janki | gongysh, ohk | 07:13 |
| gongysh | sridhar_ram, the monitor test passed with config_drive. | 07:22 |
| gongysh | sridhar_ram, so this verifies we can do this with metadata service and config_drive. | 07:22 |
| sridhar_ram | gongysh: ack, though it still beats me why it used to succeed for sometime before starting to fail.. | 07:24 |
| sridhar_ram | gongysh: for e.g., it used to succeed around mid-july .. see https://review.openstack.org/#/c/329527/ | 07:25 |
| gongysh | sridhar_ram, but theoritically, config_drive and metadata service are correct way to fix monitor test case problem | 07:25 |
| sridhar_ram | gongysh: understood, but what i'm pointing out is.. without both config_driver: true and neutron router, for some reason user-data seems to have worked .. | 07:27 |
| sridhar_ram | gongysh: around, early Aug this testcase started to fail.. | 07:27 |
| sridhar_ram | gongysh: i thought it was due to a race condition inherent in the test case design | 07:27 |
| gongysh | I think maybe the infra team has changed the CI node. if 169.254.169.254 is on the zuul node, our test cases also work. | 07:28 |
| sridhar_ram | gongysh: I see | 07:29 |
| gongysh | sridhar_ram, it is the case that nova-network will put this address on node. | 07:29 |
| sridhar_ram | gongysh: again, to labor this point .. http-ping continues to work fine even without this fix | 07:29 |
| sridhar_ram | gongysh: sorry, i take the back | 07:29 |
| sridhar_ram | gongysh: you provided an explanation earlier for http-ping | 07:30 |
| gongysh | sridhar_ram, yeah | 07:30 |
| gongysh | that explanation still works. | 07:30 |
| sridhar_ram | can you try one or two rechecks so that we are convinced this result is consistent ? | 07:31 |
| janki | gongysh, there are no heat specific unit tests so far. Should I create a dir in tests/unit/infra_driver and write 1 in test_heat.py? | 07:32 |
| gongysh | janki, https://github.com/openstack/tacker/blob/master/tacker/tests/unit/vm/infra_drivers/heat/test_heat.py | 07:34 |
| janki | gongysh, oops..so sorry | 07:35 |
| gongysh | sridhar_ram, rechecking | 07:35 |
| openstackgerrit | dharmendra kushwaha proposed openstack/tacker: Allow vdu (VM) names to be specified as a parameter https://review.openstack.org/363549 | 07:37 |
| gongysh | dkushwaha, hi | 07:45 |
| gongysh | for server in servers: | 07:45 |
| gongysh | if 'test-vdu' in server.name: | 07:45 |
| gongysh | vdu_server = server | 07:45 |
| gongysh | self.assertIsNotNone(vdu_server) | 07:45 |
| gongysh | port_list = self.neutronclient().list_ports()['ports'] | 07:45 |
| gongysh | vdu_port = None | 07:45 |
| gongysh | for port in port_list: | 07:45 |
| gongysh | if 'test-cp' in port['name']: | 07:45 |
| gongysh | vdu_port = port | 07:45 |
| dkushwaha | gongysh, hello | 07:45 |
| gongysh | cannot we know exactly what is the names of CP and vdu? | 07:45 |
| dkushwaha | gongysh, we can check by listing only | 07:46 |
| gongysh | if we can, we can use 'test-cp' == port['name'] | 07:46 |
| dkushwaha | gongysh, otherwise we first need to get vm id/ port id | 07:47 |
| gongysh | what is the exact port['name | 07:48 |
| gongysh | '] | 07:48 |
| gongysh | ? | 07:48 |
| gongysh | isn't the 'test-cp'? | 07:48 |
| dkushwaha | gongysh, yes it is 'test-cp' | 07:48 |
| gongysh | then we can if 'test-cp' == port['name']: | 07:49 |
| dkushwaha | gongysh, ah, sorry for misinterpreting . yes, we can do this | 07:49 |
| *** mbound has joined #tacker | 07:51 | |
| openstackgerrit | hyunsik Yang proposed openstack/tacker: Port scan driver for VNF application Monitoring https://review.openstack.org/329409 | 07:52 |
| dkushwaha | gongysh, Thanks for addressing this. I will fix it quickly | 07:53 |
| openstackgerrit | dharmendra kushwaha proposed openstack/tacker: Allow vdu (VM) names to be specified as a parameter https://review.openstack.org/363549 | 07:56 |
| *** openstackgerrit has quit IRC | 08:34 | |
| *** openstackgerrit has joined #tacker | 08:35 | |
| gongysh | tung_doan, what's problem with your alram all in one patch? | 08:51 |
| gongysh | tung_doan, I just want to test when you -1 it. | 08:51 |
| *** diga has joined #tacker | 08:52 | |
| openstackgerrit | dharmendra kushwaha proposed openstack/python-tackerclient: Add support for multi delete https://review.openstack.org/349806 | 08:52 |
| *** links has quit IRC | 08:53 | |
| *** janki has quit IRC | 08:53 | |
| *** links has joined #tacker | 08:55 | |
| *** mbound has quit IRC | 08:56 | |
| *** janki has joined #tacker | 08:57 | |
| *** yifei has quit IRC | 09:22 | |
| tung_doan | gongysh: no worries.. I just change token v2 to token v3 as Janki'suggstion | 09:31 |
| tung_doan | gongysh: I need a bit time to modify | 09:31 |
| tung_doan | gongysh: thanks for reminding | 09:34 |
| gongysh | KanagarajM, hi | 09:35 |
| KanagarajM | gongysh, hi | 09:36 |
| gongysh | Does the scale feature in master work? | 09:36 |
| KanagarajM | gongysh, yes | 09:36 |
| gongysh | I want to test it | 09:36 |
| KanagarajM | we have a function test patch | 09:36 |
| KanagarajM | https://review.openstack.org/363573 | 09:37 |
| KanagarajM | gongysh, ^^ | 09:37 |
| gongysh | KanagarajM, I see only service group resources are generated. | 09:38 |
| gongysh | KanagarajM, the scale.yml's resources are not in heat stack. | 09:38 |
| janki | gongysh, added heat unit test case. Please review at your convinience https://review.openstack.org/#/c/340838/ | 09:39 |
| KanagarajM | gongysh, that is nested stack | 09:39 |
| KanagarajM | gongysh, will be part of scaling group | 09:39 |
| janki | sridhar_ram, submitted edited patch. Apologies. Hadn't seen your +2 and comment | 09:40 |
| gongysh | KanagarajM, does the scale feature need ceilometer? | 09:40 |
| openstackgerrit | Merged openstack/tacker: Clean imports in code https://review.openstack.org/359023 | 09:44 |
| openstackgerrit | gongysh proposed openstack/tacker: Fix the monitor bug https://review.openstack.org/363398 | 09:54 |
| *** Ravikiran_K has quit IRC | 10:08 | |
| openstackgerrit | venkatamahesh proposed openstack/tacker: [py35] Fix for jenkins-gate-py35 error https://review.openstack.org/366656 | 10:14 |
| KanagarajM | gongysh, no | 10:19 |
| KanagarajM | gongysh, it will be done once alarm monitor driver is in | 10:19 |
| openstackgerrit | venkatamahesh proposed openstack/tacker: Update the sample vnfd template in getting started guide https://review.openstack.org/363501 | 10:20 |
| *** saju_m has joined #tacker | 10:25 | |
| janki | tung_doan, ping | 10:29 |
| tung_doan | janki: pong.. hi | 10:30 |
| janki | tung_doan, hey. trying alarm patch. Could you please help me out with the curl thing you mentioned earlier | 10:30 |
| janki | tung_doan, VM is not respawning, no logs in ceilometer too | 10:31 |
| tung_doan | janki: sure.. | 10:31 |
| openstackgerrit | Kanagaraj Manickam proposed openstack/tacker: Devref for vnf scaling feature https://review.openstack.org/359206 | 10:32 |
| janki | tung_doan, what would be the exact syntax for curl? | 10:32 |
| tung_doan | janki: dkushwaha, gonggysh... fyi.. the reason it took a bit long time to trigger alarm in ceilometer is that the default interval time in ceilometer is set to 600 (10 mins) | 10:33 |
| tung_doan | janki: dkushwaha, gonggysh: we need edit this interval time to make it reduce | 10:34 |
| tung_doan | janki: wait for me :) | 10:34 |
| tung_doan | janki: http://paste.openstack.org/show/566899/ | 10:35 |
| tung_doan | janki: please change your own http | 10:35 |
| janki | tung_doan, thanks :) | 10:37 |
| tung_doan | janki: please let me know whether it works.. thanks | 10:37 |
| janki | tung_doan, sure will do | 10:40 |
| tung_doan | janki: but I am changing eith your suggestion with keystone.. it's not ready to test yet... please wait me for a bit time.. I am setting workflow to -1 | 10:42 |
| janki | tung_doan, I believe you have changed the version in PS 7. I thought "tenant" (instead of "project") in https://review.openstack.org/#/c/365435/7/tacker/vnfm/monitor_drivers/token.py L29 should throw an error. but it didnot | 10:44 |
| *** saju_m has quit IRC | 10:51 | |
| tung_doan | janki: correct.. keystone v3 should be like that.. but the error is here "token = keystone.auth_ref['token']['id']" | 10:59 |
| tung_doan | janki: i did not change totally this part | 10:59 |
| janki | tung_doan, no issue. let me know once dome :) | 11:00 |
| tung_doan | janki: could you elaborate what is better when using keystone v3... with keystone v3, It is somehow to get token | 11:01 |
| janki | tung_doan, let me try with curl and get back. what exact error are you getting? | 11:01 |
| tung_doan | *is somehow --> is somehow difficult to get token | 11:02 |
| openstackgerrit | Janki Chhatbar proposed openstack/tacker-horizon: Adds VNFFG support in Tacker-horizon https://review.openstack.org/347779 | 11:02 |
| tung_doan | janki: thanks | 11:02 |
| *** saju_m has joined #tacker | 11:04 | |
| janki | tung_doan, what is "plku8ie0" at the end of http? | 11:05 |
| tung_doan | janki: just key... each request will be added a key.. | 11:06 |
| *** amotoki has quit IRC | 11:06 | |
| janki | tung_doan, should I have a different one or this works fine? | 11:06 |
| tung_doan | janki: sure.. | 11:07 |
| tung_doan | janki: thanks | 11:07 |
| tung_doan | janki: sorry.. i think it still work fine | 11:08 |
| tung_doan | janki: but in next few days, key will be authenticated so that alarm url is requested only one time | 11:10 |
| janki | tung_doan, ack! | 11:10 |
| *** mfedosin has joined #tacker | 11:26 | |
| *** gongysh has quit IRC | 11:26 | |
| *** amotoki has joined #tacker | 11:37 | |
| *** amotoki has quit IRC | 11:50 | |
| *** amotoki has joined #tacker | 11:58 | |
| *** bobh has joined #tacker | 11:59 | |
| *** KanagarajM has quit IRC | 12:03 | |
| *** links has quit IRC | 12:04 | |
| *** amotoki has quit IRC | 12:07 | |
| *** KanagarajM has joined #tacker | 12:07 | |
| *** links has joined #tacker | 12:16 | |
| *** bobh has quit IRC | 12:20 | |
| *** bobh has joined #tacker | 12:21 | |
| *** gongysh has joined #tacker | 12:23 | |
| *** bobh has quit IRC | 12:28 | |
| *** bobh has joined #tacker | 12:28 | |
| *** bobh has quit IRC | 12:33 | |
| *** bobh has joined #tacker | 12:33 | |
| *** amotoki has joined #tacker | 12:33 | |
| *** hparekh has quit IRC | 12:37 | |
| *** bobh has quit IRC | 12:39 | |
| *** bobh has joined #tacker | 12:40 | |
| *** KanagarajM has quit IRC | 12:40 | |
| *** diga has quit IRC | 12:40 | |
| *** KanagarajM has joined #tacker | 12:41 | |
| *** tbh has joined #tacker | 12:44 | |
| *** mike_m has quit IRC | 12:44 | |
| *** bobh has quit IRC | 12:52 | |
| *** janki has quit IRC | 12:54 | |
| *** janki has joined #tacker | 12:58 | |
| *** saju_m has quit IRC | 13:01 | |
| *** gongysh has quit IRC | 13:07 | |
| *** vishwanathj has joined #tacker | 13:11 | |
| *** links has quit IRC | 13:12 | |
| *** neel has quit IRC | 13:13 | |
| *** bobh has joined #tacker | 13:32 | |
| *** tbh has quit IRC | 13:51 | |
| *** janki has quit IRC | 14:03 | |
| *** tbh has joined #tacker | 14:14 | |
| *** tbh has quit IRC | 14:23 | |
| *** yifei has joined #tacker | 14:26 | |
| *** dkushwaha__ has joined #tacker | 14:42 | |
| *** tbh has joined #tacker | 14:46 | |
| *** tung_doan has quit IRC | 14:48 | |
| *** mbound has joined #tacker | 14:54 | |
| *** amotoki has quit IRC | 14:56 | |
| *** tbh has quit IRC | 15:00 | |
| *** gongysh has joined #tacker | 15:04 | |
| gongysh | sridhar_ram, https://review.openstack.org/#/c/363398/ is ready to be reviewed | 15:05 |
| *** yifei has quit IRC | 15:16 | |
| *** tamilhce has joined #tacker | 15:19 | |
| tamilhce | Hi I'm Tamil vanan ,New to tacker | 15:19 |
| tamilhce | anything that needs a fix regarding coding , i am ready to take up | 15:20 |
| openstackgerrit | dharmendra kushwaha proposed openstack/python-tackerclient: Remove "else" branch in "create_vnfd" function https://review.openstack.org/334720 | 15:21 |
| sridhar_ram | gongysh: ack, will +2 | 15:23 |
| *** KanagarajM has quit IRC | 15:27 | |
| *** tamilhce has quit IRC | 15:43 | |
| *** gongysh has quit IRC | 15:45 | |
| *** tbh has joined #tacker | 15:46 | |
| *** tbh has quit IRC | 15:53 | |
| *** uck has joined #tacker | 15:53 | |
| *** Vijayendra has quit IRC | 15:57 | |
| *** janki has joined #tacker | 16:04 | |
| *** KanagarajM has joined #tacker | 16:05 | |
| *** Vijayendra has joined #tacker | 16:09 | |
| openstackgerrit | Kanagaraj Manickam proposed openstack/tacker: VNF scaling: Functional test https://review.openstack.org/363573 | 16:19 |
| *** sripriya has joined #tacker | 16:29 | |
| *** uck has quit IRC | 16:31 | |
| *** uck has joined #tacker | 16:32 | |
| *** tbh has joined #tacker | 16:33 | |
| *** bharatht_ has joined #tacker | 16:33 | |
| *** bharatht_ has quit IRC | 16:33 | |
| *** uck has quit IRC | 16:36 | |
| *** uck has joined #tacker | 16:38 | |
| *** mbound has quit IRC | 16:39 | |
| *** LamT_ has joined #tacker | 16:40 | |
| *** KanagarajM has quit IRC | 16:40 | |
| sridhar_ram | trozet: ping | 16:42 |
| sridhar_ram | trozet: is any *major* changes anticipated for the ffg abstract driver - https://review.openstack.org/#/c/347563 ? | 16:42 |
| sridhar_ram | janki: ping | 16:43 |
| janki | sridhar_ram: pong | 16:43 |
| *** uck has quit IRC | 16:43 | |
| sridhar_ram | janki: any reason why your vnf-resources depends on the above abstract driver patchset ? | 16:43 |
| janki | sridhar_ram: it doesnot depend on that patchset | 16:44 |
| sridhar_ram | janki: well, if you click the parent link in https://review.openstack.org/#/c/340838/ .. you can see | 16:45 |
| sridhar_ram | janki: that is one reason it hasn't merged even after it has a +A from tbh | 16:45 |
| sridhar_ram | janki: anyways, i think we are better off merging the abstract driver and move on | 16:46 |
| janki | sridhar_ram: not sure how it got dependent | 16:46 |
| sridhar_ram | sripriya: are you around to take a quick look at https://review.openstack.org/#/c/347563/ and merge ? | 16:47 |
| sripriya | sridhar_ram: yes, looking | 16:47 |
| sridhar_ram | sripriya: i anticipate there might be few fine tuning as other FFG patchsets fall in place | 16:47 |
| sridhar_ram | sripriya: at least we will have this and janki's vnf-resources out of our way | 16:47 |
| openstackgerrit | Merged openstack/python-tackerclient: Remove "else" branch in "create_vnfd" function https://review.openstack.org/334720 | 16:54 |
| sripriya | sridhar_ram: done, looks vnf-resources patch was a child of abstract vnffg driver patch, | 16:55 |
| sridhar_ram | sripriya: thanks and yes, meaning to get that unblocked | 16:56 |
| *** Vijayendra has quit IRC | 17:00 | |
| openstackgerrit | Merged openstack/tacker: VNFFG abstract driver https://review.openstack.org/347563 | 17:00 |
| openstackgerrit | Merged openstack/tacker: Add VNF resource details to get vnf API https://review.openstack.org/340838 | 17:00 |
| sripriya | sridhar_ram: https://review.openstack.org/#/c/347563/9/tacker/nfvo/drivers/vnffg/abstract_vnffg_driver.py@38 is it supposed to take multiple fc ids? | 17:01 |
| sripriya | sridhar_ram: i see update_chain taking multiple fc ids but not create_chain | 17:02 |
| sridhar_ram | sripriya: i don't think so, perhaps a typo in update method .. again, will let s3wong and trozet confirm | 17:03 |
| sridhar_ram | vishwanathj: is anything pending from your side for audit event horizon changes ? https://review.openstack.org/#/c/349755/ | 17:04 |
| vishwanathj | nope | 17:05 |
| sripriya | sridhar_ram: ok | 17:09 |
| sripriya | sridhar_ram: louis f mentioning a port chain can take multiple fc ids for a similar comment by trozet in an earlier patch. we can revisit this patch later. | 17:10 |
| *** links has joined #tacker | 17:14 | |
| *** tung_doan has joined #tacker | 17:16 | |
| sridhar_ram | sripriya: i'm not sure if we would expose that .. looking at https://review.openstack.org/#/c/344522/23/tacker/tests/unit/vm/infra_drivers/heat/data/vnffgd_template.yaml .. each FP (chain) would be associated with one FC. | 17:17 |
| trozet | sridhar_ram: hi | 17:18 |
| sridhar_ram | trozet: hi | 17:18 |
| trozet | sridhar_ram: i think for the initial implementation there should only be a single fc_id per chain | 17:20 |
| trozet | sridhar_ram: so it looks like the fc_ids in update_chain should be fc_id | 17:20 |
| sridhar_ram | trozet: sounds good.. thats what i thought | 17:20 |
| sridhar_ram | trozet: we can correct as we go... | 17:21 |
| trozet | sridhar_ram: thinking about that some more... if you have a symmetrical chain, you will need 2 classifiers (one for reverse traffic) | 17:21 |
| trozet | sridhar_ram: but iirc when I talked to s3wong we agreed that should be a driver level thing | 17:21 |
| trozet | sridhar_ram: so specifying symmetrical implies that the reverse classifier gets created and associated and the tacker layer doesn't need to know about it | 17:22 |
| sridhar_ram | trozet: but that shd be implicit, taken care by the underlying driver, correct? | 17:22 |
| trozet | sridhar_ram: exactly | 17:22 |
| trozet | sridhar_ram: you want me to upload a new patch set with the fc_ids fixed? | 17:22 |
| trozet | sridhar_ram: oh it was already merged | 17:23 |
| sridhar_ram | trozet: yet, time to iterate :) | 17:23 |
| trozet | sridhar_ram: i guess that is OK for now | 17:23 |
| sridhar_ram | *yes | 17:23 |
| trozet | sridhar_ram: if we inherit that class, and override the method, does python complain if the args arent hte exact same? | 17:24 |
| sridhar_ram | trozet: let's make necessary fixes as we go.. we are cutting too close | 17:24 |
| sridhar_ram | trozet: .. particularly with your PTO plan next week | 17:24 |
| trozet | sridhar_ram: yeah i dont mind working the evenings to finish this | 17:24 |
| sridhar_ram | trozet: appreciate your support ! | 17:25 |
| trozet | sridhar_ram: great that the patches got merged, let me rebase my patch now | 17:25 |
| sripriya | trozet: dont be surprized, but yes python wont complain on args if they are not the same as far as i have seen:) | 17:25 |
| trozet | sripriya: ok thats what i thought. we should fix it, but doesn't block us | 17:25 |
| sripriya | trozet: yes, | 17:25 |
| trozet | sridhar_ram, sripriya: s3wong said taht he had an update for hte networking-sfc driver, then i was going to help him port it to the openstack_driver | 17:26 |
| sripriya | trozet: in fact i see the symmetrical attribute being silently ignored in the n-sfc driver imlementation | 17:26 |
| trozet | sripriya^^ | 17:26 |
| trozet | sridhar_ram, sripriya: i did not see a new patch from s3wong last night though on https://review.openstack.org/#/c/347568/ | 17:26 |
| sripriya | trozet: ok | 17:26 |
| trozet | sridhar_ram, sripriya: for now i will update what I can on my patch, test it without the driver in a real openstack setup | 17:27 |
| sridhar_ram | janki: awesome job in taking this vnf-resources work, patiently working through the reviews and finishing it .. it went through 32 revisions !! | 17:27 |
| trozet | sridhar_ram, sripriya: hopefully s3wong gets online or uploads the new patch so I can take it from there | 17:27 |
| sridhar_ram | trozet: did s3wong take your offer to help him w/ his patchset ? | 17:27 |
| janki | sridhar_ram: yes. that was what I was thinking while submitting. finally it got merged after 32 (ohh :P) patches :) | 17:28 |
| trozet | sridhar_ram: yeah he said he had already finished some changes to it, but had not migrated it to the openstack driver. The plan was last night he would upload his latest patch set, then I would take it over and migrate it to the openstack_driver | 17:28 |
| janki | sridhar_ram: I learnt a lot through these reviews though. | 17:28 |
| sridhar_ram | janki: it just proves you are ready for bigger things around here :) Keep it up !! | 17:29 |
| *** links has quit IRC | 17:29 | |
| janki | sridhar_ram: thanks :) | 17:29 |
| sridhar_ram | trozet: sounds like a plan | 17:29 |
| trozet | sridhar_ram: so my plan is to at least have tested the whole thing before I leave and have the patch out of WIP | 17:30 |
| trozet | sridhar_ram: at that point maybe you or sripriya could help in reviews/bug fixes while I'm gone and co-author | 17:30 |
| trozet | sridhar_ram: because code freeze is sometime next week right? | 17:30 |
| sridhar_ram | trozet: yes, next Thursday Sept 15.. | 17:31 |
| sridhar_ram | trozet: i'd like to wrap up and land VNFFG patchsets before you are gone by this Friday .. i know this might sound aggressive.. | 17:32 |
| trozet | sridhar_ram: will do my best, let me get it out of WIP then need reviews fast | 17:32 |
| sridhar_ram | trozet: .. but the way our release window is structured we have some room post Sept 15th to push high-severity fixes | 17:33 |
| trozet | sridhar_ram: yeah fixes should be OK right? just not new features | 17:33 |
| sridhar_ram | trozet: .. it is generally easier to push point fixes even into stable/newton | 17:33 |
| sridhar_ram | trozet: exactly | 17:33 |
| trozet | sridhar_ram: so it doesnt have to be flawless, but shoudl be functional | 17:33 |
| trozet | sridhar_ram: i need functional testing too, right? | 17:34 |
| sridhar_ram | so, let's line up everything we have by end of Thursday and push things in by Friday | 17:34 |
| trozet | sridhar_ram: so far i only have unit tests | 17:34 |
| sridhar_ram | trozet: yes, it is preferred to come in w/ func tests.. but push come to shove, we can have a follow on | 17:34 |
| trozet | sridhar_ram: ok thanks | 17:35 |
| sridhar_ram | trozet: btw, are you testing this with a some off-tree ODL driver ? | 17:36 |
| trozet | sridhar_ram: no i'm not testing it with ODL, just plain networking-sfc | 17:36 |
| sridhar_ram | trozet: okay.. just checking | 17:36 |
| sridhar_ram | tbh: can you please review https://review.openstack.org/#/c/349755 ? | 17:36 |
| tbh | sridhar_ram, sure | 17:38 |
| *** janki has quit IRC | 17:41 | |
| sridhar_ram | tbh: thanks! | 17:42 |
| *** mbound has joined #tacker | 17:52 | |
| *** saju_m has joined #tacker | 17:52 | |
| *** sridhar_ram is now known as sridhar_ram_afk | 17:52 | |
| *** s3wong has joined #tacker | 17:57 | |
| openstackgerrit | Tung Doan proposed openstack/tacker: Alarm monitor: All in one https://review.openstack.org/365435 | 18:06 |
| *** yifei has joined #tacker | 18:13 | |
| *** s3wong has quit IRC | 18:15 | |
| *** vishnoianil has quit IRC | 18:16 | |
| *** yifei has quit IRC | 18:18 | |
| openstackgerrit | Tung Doan proposed openstack/tacker: Alarm monitor: All in one https://review.openstack.org/365435 | 18:22 |
| *** sridhar_ram_afk is now known as sridhar_ram | 18:37 | |
| tung_doan | sridhar_ram: ping | 18:38 |
| sridhar_ram | tung_doan: pong | 18:38 |
| tung_doan | sridhar_ram: Do you guys think it is necessary to create only one driver for monitoring? | 18:39 |
| tung_doan | sridhar_ram: logically, i thin it's not good | 18:39 |
| *** saju_m has quit IRC | 18:39 | |
| *** uck has joined #tacker | 18:40 | |
| sridhar_ram | tung_doan: let's step and see what is an ideal scenario for a monitoring driver developer... | 18:40 |
| sridhar_ram | tung_doan: .. from framework point of view we are intending to support two types of mon drivers.. | 18:41 |
| sridhar_ram | tung_doan: .. (1) tacker ---polling --> VNFs + other entities | 18:41 |
| sridhar_ram | .. (2) tacker <---receiving --- events-from VNFs + other entities (ceilometer, opnfv VES, etc) | 18:42 |
| tung_doan | sridhar_ram: thanks.. I believe that we should have a sepatate monitoring driver for (ceilometer, monasca,...) | 18:42 |
| tung_doan | sridhar_ram: make sense to me.. totally agree | 18:42 |
| sridhar_ram | tung_doan: hang on.. i'm not done yet | 18:43 |
| sridhar_ram | tung_doan: :) | 18:43 |
| sridhar_ram | tung_doan: .. i'm basically thinking out aloud | 18:43 |
| * sridhar_ram looking up something | 18:43 | |
| tung_doan | sridhar_ram: :) unserstand | 18:44 |
| sridhar_ram | tung_doan: looking at https://review.openstack.org/#/c/365435/9/tacker/vnfm/monitor_drivers/alarm_abstract_driver.py and https://github.com/openstack/tacker/blob/master/tacker/vnfm/monitor_drivers/abstract_driver.py ... | 18:46 |
| sridhar_ram | tung_doan: .. what is the difficulty in merging them into one ? | 18:46 |
| sridhar_ram | tung_doan: .. i'm leaning towards a single alarm driver of different "types" | 18:47 |
| sridhar_ram | tung_doan: ... type : { "poll", "alarm", } | 18:47 |
| sridhar_ram | ping, http-ping will be of type: poll | 18:47 |
| sridhar_ram | ceilometer, monasca, VES, etc will be of type: "alarm" | 18:47 |
| tung_doan | sridhar_ram: at the beginning, it will have problem in monitor.py | 18:48 |
| tung_doan | sridhar_ram: .. when we merge them | 18:48 |
| tung_doan | sridhar_ram: type : { "poll", "alarm", } -----> agree | 18:48 |
| sridhar_ram | tung_doan: understood, you should not add monitors of type != poll to the hosting device queue | 18:48 |
| tung_doan | sridhar_ram: actually i also follow ves in opnfv :) | 18:48 |
| sridhar_ram | tung_doan: cool.. | 18:49 |
| tung_doan | sridhar_ram: already did as you said | 18:49 |
| sridhar_ram | tung_doan: so, bottomline .. to avoid monitor driver developer confusion .. i think we should go for a *single* abstract monitor interface | 18:50 |
| sridhar_ram | tung_doan: btw, can u explain how " def call_alarm_url() " method is used ? | 18:51 |
| *** s3wong has joined #tacker | 18:52 | |
| tung_doan | sridhar_ram: the workflow is that: before creating vnf, tacker will call monitoring driver to get alarm url... it should be passed to "real" ceilometer or monasca | 18:53 |
| tung_doan | "it" --> alarm url | 18:54 |
| sridhar_ram | tung_doan: couple of things.. | 18:54 |
| sridhar_ram | tung_doan: ... you shd create alarms for a vnf ONLY after it successfully goes to ACTIVE state | 18:55 |
| sridhar_ram | .. next, i think we would be better of calling this API "create_alarm()" that returns an opaque low-level driver specific handle back to the plugin | 18:56 |
| sridhar_ram | handle == "alarm url" | 18:56 |
| tung_doan | but heat driver needs alarm url to create template.. could you elaberate it? | 18:57 |
| sridhar_ram | tung_doan: ah, i see.. | 18:58 |
| tung_doan | before ACTIVE state.. i think it make sense.. in case of monitoring templates... without alarm url, ot does not make sense | 18:58 |
| *** uck has quit IRC | 18:59 | |
| sridhar_ram | tung_doan: okay, i was worried about ceilometer tripping some CPU thresholds in the initial VM bootup (CPU usually spikes and then settles down) | 19:01 |
| sridhar_ram | tung_doan: .. agree, i don't recommend (for now) to move the create_alram_url() .. as we might need to resort to stack-update (do-able, but more effort) after a VNF goes active | 19:02 |
| tung_doan | regarding to the CPU threshold... | 19:03 |
| tung_doan | sridhar_ram: we will need to change ceilometer pipeline because the default interval in ceilometer is 600 (10mins).. I worry someonee could be confused | 19:04 |
| tung_doan | sridhar_ram: i just wonder if you need a document to explian monitor driver iin tacker? | 19:05 |
| sridhar_ram | tung_doan: can that be passed in as an argument in the create_alarm() call down to ceilometer instead of changing the default ? | 19:06 |
| sridhar_ram | tung_doan: well, tacker-spec is the place where we should've documented all this... | 19:06 |
| sridhar_ram | another question... | 19:06 |
| sridhar_ram | tung_doan: what is the different between process_alarm() and process_notification() in the abstract driver ? | 19:06 |
| tung_doan | sridhar_ram: i mean how to setup env.. becasue i got alot of questions | 19:07 |
| sridhar_ram | tung_doan: oh, sure.. absolute.. include that document as a devref into your patchset itself | 19:07 |
| sridhar_ram | *absolutely | 19:07 |
| tung_doan | "passed in as an argument in the create_alarm() call down to ceilometer instead of changing the default"... could you elaborate it? | 19:08 |
| *** vishnoianil has joined #tacker | 19:09 | |
| sridhar_ram | tung_doan: sure, my understanding is 600s is the default in ceilometer .. can that be changed on a per-alarm basis to something shorter ? | 19:09 |
| tung_doan | "different between process_alarm() and process_notification()" ---> actually it is still undevelopement and I will have to remove or develop it | 19:09 |
| sridhar_ram | tung_doan: okay.. basically i'm curious about https://review.openstack.org/#/c/365435/9/tacker/alarm_receiver.py | 19:11 |
| sridhar_ram | tung_doan: .. how it fits into the alarm driver | 19:11 |
| tung_doan | sridhar_ram: sure. just change period in alarm configuration | 19:11 |
| sridhar_ram | in ceilometer.conf ? | 19:12 |
| tung_doan | ah, regarding to alarm_receiver.. i need to change a better name | 19:12 |
| tung_doan | i mean in tosca template... for pipeline, once we enable ceilometer plugin we can change in /etc/ceilometer/pipeline.yaml | 19:14 |
| tung_doan | sridhar_ram: did you get the main idea of alarm_receiver.py? | 19:15 |
| sridhar_ram | tung_doan: okay.. i clearly need to spend more time reviewing your patchset :) | 19:16 |
| sridhar_ram | tung_doan: my understanding is, this code will receive the ceilometer callback invocation ? | 19:16 |
| *** foutatoro has joined #tacker | 19:16 | |
| tung_doan | sridhar_ram: i think so.. franky, thanks all guys for spending time with my spec :) | 19:16 |
| sridhar_ram | tung_doan: back to the original question.. can you explore merging alarm and the original monitor absract drivers into one common one ? | 19:17 |
| tung_doan | sridhar_ram: just need a bit change in monitor.py .. will do | 19:18 |
| sridhar_ram | tung_doan: thanks | 19:18 |
| sridhar_ram | tung_doan: also, if you can add that document explaining how to use this feature, it will help all of us (including me) to test this better and provide better comments | 19:18 |
| tung_doan | ok.. maybe this weekend will fine for me.. i am trying to finish my code this week | 19:19 |
| foutatoro | Hi all I would like to know if you have a wiki that describes how to build service function chaining using tacker and opendaylight-sfc-driver ? | 19:19 |
| *** tbh has quit IRC | 19:19 | |
| tung_doan | if you guys need some guides to test.. will finish ASAP | 19:22 |
| foutatoro | any suggestion guys ? | 19:28 |
| sridhar_ram | foutatoro: we don't have a write up yet.. parts of it are still in progress and will fall in place soon (in next few weeks) | 19:30 |
| *** tung_doan has quit IRC | 19:31 | |
| *** sripriya has quit IRC | 19:33 | |
| foutatoro | sridhar_ram: thanks. I've successfully create SFC in real server using opendaylight boron. Is it possible to test now tacker with the sfc-driver ? | 19:36 |
| *** sripriya has joined #tacker | 19:41 | |
| openstackgerrit | Tim Rozet proposed openstack/tacker: [WIP] Implements VNFFG into NFVO https://review.openstack.org/344522 | 19:57 |
| sridhar_ram | foutatoro: we are taking tacker --> neutron-sfc --> ODL to configure ODL-SFC.. | 19:57 |
| s3wong | foutatoro: you need to wait until ODL networking-sfc driver to be available... | 19:57 |
| sridhar_ram | foutatoro: .. we are currently waiting for https://review.openstack.org/#/c/337948/ to wrap up | 19:59 |
| sridhar_ram | vishnoianil: is working on that piece | 20:00 |
| trozet | s3wong: hi | 20:00 |
| s3wong | sridhar_ram: yeah, I was looking for that review link :-) | 20:00 |
| s3wong | trozet: pong | 20:00 |
| trozet | s3wong: are you planning to upload your new patch set soon? | 20:01 |
| s3wong | trozet: I plan to update patch sometime this afternoon | 20:01 |
| trozet | s3wong: ok cool | 20:01 |
| s3wong | trozet: last night I sinked all my time on day job stuff --- worked until 2am, in fact | 20:01 |
| trozet | s3wong: that sucks :( | 20:01 |
| s3wong | trozet: yeah, unfortunately --- the two deadlines are pretty much around the same time | 20:02 |
| vishnoianil | sridhar_ram, s3wong foutatoro i will push a v2 version of the driver sometime next week | 20:02 |
| foutatoro | Thanks for these information | 20:03 |
| *** openstackgerrit has quit IRC | 20:04 | |
| *** mbound has quit IRC | 20:04 | |
| *** openstackgerrit has joined #tacker | 20:05 | |
| sridhar_ram | vishnoianil: sounds good.. thanks | 20:27 |
| *** gongysh has joined #tacker | 20:49 | |
| gongysh | sripriya, hi | 20:50 |
| *** vishnoianil has quit IRC | 20:50 | |
| *** sripriya has quit IRC | 20:50 | |
| *** avishnoi has joined #tacker | 20:50 | |
| *** vishwana_ has joined #tacker | 20:51 | |
| s3wong | gongysh: still up? :-) | 20:51 |
| *** sripriya has joined #tacker | 20:51 | |
| gongysh | s3wong, waked up. | 20:51 |
| trozet | sripriya: ping | 20:51 |
| sripriya | trozet: pong | 20:51 |
| s3wong | trozet: is there any reason why your patch is still WIP? | 20:51 |
| gongysh | s3wong , https://review.openstack.org/#/c/363398/ please | 20:51 |
| trozet | sripriya: ran into something that I could use your input on | 20:52 |
| s3wong | gongysh: sure --- looking at it now | 20:52 |
| *** vishwanathj has quit IRC | 20:52 | |
| trozet | sripriya: so i have written the tosca validator in the openstack driver, passes unit tests just fine | 20:52 |
| trozet | sripriya: but now I go to invoke it to validate VNFFGD, and I realize in the nfvo_plugin, I don't know what vim driver to use :) | 20:53 |
| trozet | sripriya: at this point I don't have any VNFs, so i cant use those to figure out the driver type | 20:53 |
| trozet | s3wong: I just wanted to finish the TOSCA validation before i remove the WIP (working on it now) | 20:53 |
| sripriya | trozet: validation of template should be handled by the plugin itself and not by the driver | 20:54 |
| trozet | sripriya: ok i had that thought right after i finished it :) | 20:55 |
| openstackgerrit | Doug Hellmann proposed openstack/python-tackerclient: Update reno for stable/newton https://review.openstack.org/364965 | 20:55 |
| sripriya | trozet: i’m doing the same for vnfd templates now :) | 20:55 |
| trozet | sripriya: why dont we just use a common method? | 20:55 |
| trozet | sripriya: i just kind of made a single method that works for both | 20:55 |
| sripriya | trozet; i have a WIP patch to group the common tosca validations into a common module which any resource can use | 20:56 |
| sripriya | trozet: hopefully i can give some justice to it before the deadline | 20:56 |
| trozet | sripriya: yeah i saw there is some deprecated/vnf specific stuff | 20:56 |
| trozet | sripriya: so i just did this: | 20:56 |
| trozet | https://paste.fedoraproject.org/423635/ | 20:57 |
| trozet | sripriya: i guess for now i will just use my stuff so i dont depend on your common patch, then later we should converge | 20:57 |
| sripriya | trozet: does your code take the vnffd template as a string? | 20:58 |
| trozet | sripriya: no | 20:58 |
| trozet | sripriya: i saw that stuff leftover in vnfm, but i don't think its supported for vnffg | 20:59 |
| sripriya | trozet: what type of data is template on L9 in your paste link | 20:59 |
| trozet | sripriya: its just the <template> part of the content stored as {'vnfd/vnffgd': <template>} | 21:00 |
| sripriya | trozet:<template> is a dict? | 21:01 |
| trozet | sripriya: yah | 21:01 |
| sripriya | trozet: ok cool, | 21:03 |
| trozet | sripriya: i have these unit tests, which pass https://paste.fedoraproject.org/423637/raw/ | 21:03 |
| openstackgerrit | Merged openstack/tacker: Fix the monitor bug https://review.openstack.org/363398 | 21:03 |
| *** mbound has joined #tacker | 21:05 | |
| sripriya | trozet: not sure why you are converting to dict again your validate when you already have yaml input as dict | 21:05 |
| trozet | sripriya: in the tests or the method? | 21:06 |
| sripriya | trozet: in the method | 21:06 |
| sripriya | trozet: and in your tests, you are adding the template to [vnfd][ attributes][‘vnfd’] which is not defined in api extension layer for vnffgd | 21:07 |
| trozet | sripriya: thats just cause they were openstack_driver tests and i knew those vnfds would pass and thought VNFM would move to this "driver" | 21:08 |
| trozet | sripriya: https://paste.fedoraproject.org/423640/ | 21:08 |
| trozet | sripriya: i did the load because I saw the load there for vnfd | 21:08 |
| sripriya | trozet: also, you need to update vnfd as vnffd everywhere as they are two different resources | 21:09 |
| trozet | sripriya: where? | 21:09 |
| sripriya | trozet: that is exactly we “dont” want to do, take template as string and load it as dict internally, | 21:09 |
| *** mbound has quit IRC | 21:10 | |
| trozet | sripriya: oh that was only there because vnfd template could be a string | 21:10 |
| sripriya | trozet: references in your test cases | 21:10 |
| trozet | sripriya: ok let me remove that then | 21:10 |
| trozet | sripriya: yeah i will. Like i said it was just because i knew vnfd ones were correct | 21:10 |
| trozet | sripriya: i'm not so sure the vnffgd is going to pass the tosca validator | 21:11 |
| sripriya | trozet: in vnffgd (unlike vnfd) we have support now in API and db to store the template as json instead of yaml string, this is easier to work with compared to yaml string | 21:11 |
| sripriya | trozet: vnfd is goign through some refactoring to follow the same | 21:12 |
| sripriya | trozet: it is being done in smaller increments so that there is minimal impacts between different releases | 21:13 |
| trozet | sripriya: ok | 21:13 |
| trozet | sripriya: it looks like the yaml load is required, because the object is unicode | 21:25 |
| gongysh | sripriya, https://review.openstack.org/#/c/363483/ | 21:35 |
| *** bobh has quit IRC | 21:36 | |
| *** sripriya has quit IRC | 21:37 | |
| openstackgerrit | vishwanath jayaraman proposed openstack/tacker: VNF monitoring event capture: Functional test https://review.openstack.org/366226 | 21:39 |
| *** sripriya has joined #tacker | 21:40 | |
| trozet | sridhar_ram: ping? | 21:40 |
| *** LamT_ has quit IRC | 21:41 | |
| trozet | sripriya: ping? | 21:42 |
| sripriya | trozet: how are you sending the template to the vnfd api? | 21:42 |
| sripriya | gongysh: will take a look | 21:43 |
| trozet | sripriya: oh heh, it uses the _get_template function in the test utils, which reads the file as a string ;) | 21:44 |
| sripriya | trozet: :-) | 21:44 |
| trozet | sripriya: will fix tht | 21:44 |
| trozet | sripriya: so here is the deal | 21:44 |
| trozet | sripriya: i have the tosca parser running on my template | 21:44 |
| trozet | sripriya: but it turns out tosca parser is too smart :) | 21:45 |
| trozet | https://paste.fedoraproject.org/423658/raw/ | 21:46 |
| trozet | sripriya:^ that fails because the parser actually reads the forwarder VNF<num> and CP<num> and expects them to be in the definition :) | 21:47 |
| sripriya | trozet: why should the values be part of the definitions? | 21:49 |
| trozet | sripriya: because in a real NS definiton, the VNFs and everything would be in one template | 21:49 |
| trozet | sripriya: i think tacker is working towards that in the future, but for this iteration VNFD/VNFFGD is separate and not under one template def | 21:49 |
| trozet | sripriya: so i think we can either A) require a user to copy the VNFD template into the VNFFGD template, B) automatically look up the VNFD before sending it to tosca parser and include in the parser arg | 21:51 |
| trozet | sridhar_ram:^ | 21:55 |
| sripriya | trozet: so does tosca parser expect us to specify valid forwarder values in the template? | 21:57 |
| trozet | sripriya: correct | 21:58 |
| trozet | sripriya: example: KeyError: 'Node template "VNF3" was not found.' | 21:58 |
| sripriya | trozet: ah ok | 21:58 |
| sripriya | trozet: and is it the same for capability as well? | 21:59 |
| trozet | sripriya: yep | 21:59 |
| trozet | KeyError: 'Node template "CP11" was not found.' | 21:59 |
| sripriya | trozet: looking into heat translator coded | 21:59 |
| trozet | sripriya: it also doesnt like that I am using a non-tosca field "id" to hold the path ID | 21:59 |
| sripriya | code* | 21:59 |
| trozet | UnknownFieldError: Node template "Forwarding_path1" contains unknown field "id". Refer to the definition to verify valid values. | 22:00 |
| trozet | sripriya: i've got go afk for a few hours, will be back...not sure what to do about this | 22:00 |
| sripriya | trozet: yes it chokes up if you specify an unknow field, you will have to add that to tacker_nfv_defs.yaml and apply the imports | 22:00 |
| trozet | sripriya: ah ok at least that is a work around | 22:00 |
| trozet | sripriya: but let me know what you think about the VNF/CP issue | 22:01 |
| sripriya | trozet: sure i need to look into heat translator code to see that | 22:01 |
| trozet | sripriya: just tosca translator right | 22:02 |
| trozet | sripriya: no heat | 22:02 |
| *** sripriya has quit IRC | 22:03 | |
| *** gongysh has quit IRC | 22:03 | |
| *** mfedosin has quit IRC | 22:07 | |
| openstackgerrit | vishwanath jayaraman proposed openstack/tacker: VNF Scaling event capture: Functional Test https://review.openstack.org/366433 | 22:09 |
| *** gongysh has joined #tacker | 22:27 | |
| *** yifei has joined #tacker | 22:35 | |
| *** yifei has quit IRC | 22:39 | |
| *** sripriya has joined #tacker | 22:46 | |
| sripriya | trozet: right | 22:46 |
| *** sripriya has quit IRC | 22:51 | |
| *** gongysh has quit IRC | 23:08 | |
| *** sripriya has joined #tacker | 23:14 | |
| openstackgerrit | Merged openstack/python-tackerclient: Update reno for stable/newton https://review.openstack.org/364965 | 23:26 |
| sridhar_ram | sripriya: can you respin https://review.openstack.org/#/c/351431 ? | 23:35 |
| sripriya | sridhar_ram: sure, will do it right away | 23:41 |
| sridhar_ram | sripriya: thanks | 23:43 |
| openstackgerrit | Sripriya Seetharam proposed openstack/tacker: Better handle vim domain exception https://review.openstack.org/351431 | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!