| hyunsikyang | tojuvone | 06:02 |
|---|---|---|
| tojuvone | hyunsikyang, hi | 06:03 |
| hyunsikyang | are you there? | 06:03 |
| hyunsikyang | Ah hi | 06:03 |
| hyunsikyang | long time no see:) | 06:03 |
| tojuvone | yep :) | 06:03 |
| hyunsikyang | we uploaded first version of tacker- fenix integration | 06:03 |
| hyunsikyang | https://review.opendev.org/#/c/681157/ | 06:03 |
| hyunsikyang | But, we have some issue. | 06:03 |
| tojuvone | Awesome | 06:04 |
| tojuvone | ok, what kind of issue | 06:04 |
| hyunsikyang | When we made a blueprint, we don't think about multi VDU. Basically, tacker use VNFD and VNFD includes at least one VDU. | 06:04 |
| hyunsikyang | However, we think that we'd better make a alarm url for each VDU. | 06:05 |
| tojuvone | Is VDU something that maps to VNF (nova project_id)? | 06:06 |
| hyunsikyang | So, current implementation, it creates url for each VDU. | 06:06 |
| hyunsikyang | VNF includes several VDU in one VNF. | 06:08 |
| hyunsikyang | Actually, it is a kinds of VNFC as a logical view. | 06:08 |
| tojuvone | Sorry, I am not that aware of the "VNF" side terminology | 06:09 |
| hyunsikyang | Ah | 06:09 |
| tojuvone | just opened the page: https://docs.openstack.org/tacker/ocata/devref/vnfd_template_description.html | 06:10 |
| tojuvone | ok... | 06:11 |
| hyunsikyang | anyway, multi VDU means that we have several instance as a view of nova | 06:11 |
| tojuvone | What should matter here is that in OpenStakc spacew there is only project_id/tenant_id | 06:11 |
| hyunsikyang | But it only show one VNF as a tacker view. | 06:11 |
| tojuvone | that is the only unit that can be used in mapping then | 06:12 |
| tojuvone | so if that do not have exact mapping to single VDU, then this is a big problem | 06:12 |
| hyunsikyang | But, when we think about later, we should consider each instance. I think.. | 06:12 |
| hyunsikyang | it mapped with ecah VDU. | 06:13 |
| hyunsikyang | because we created alarm url based on VDU. | 06:13 |
| tojuvone | OpenStack side shouldn't know about VDU if it does not have any means to map it | 06:13 |
| tojuvone | Surely if one VDU could map to heat stack instance, then there could be a mapping | 06:14 |
| tojuvone | One way or the other this needs to be solved later on | 06:16 |
| tojuvone | The question remains: What is the relation of OpenStack: project_id/tenant_id and stack_id (heat), to vnfd or the VDUs | 06:18 |
| tojuvone | And does tacker even use heat to deploy | 06:18 |
| hyunsikyang | yes tacker use heat stack. heat stack id = VNF id. | 06:22 |
| tojuvone | so that is still just mapping to "vnfd" not "VDU" | 07:25 |
| hyunsikyang | Part of it. When you create VNF with one VDU, it is mapping to 1 VNF. | 08:02 |
| *** tojuvone has quit IRC | 09:42 | |
| *** tojuvone has joined #openstack-fenix | 09:43 | |
| tojuvone | hyunsikyang, Need to look into this. At the same time I am now finalizing demo for ONS Europe and preparing to start spec for ETSI FEAT03 in Fenix | 12:03 |
| hyunsikyang | ok:) | 12:04 |
| hyunsikyang | I will looj at it. | 12:04 |
| tojuvone | hyunsikyang, thx :) | 12:04 |
| hyunsikyang | Today, I talked with tacker PTL, They also aggreed with each VDU have each url for maintenance. but, we also make a another url for VNF level maintenance. | 12:05 |
| tojuvone | ok | 12:09 |
| *** hyunsikyang__ has joined #openstack-fenix | 12:30 | |
| *** hyunsikyang has quit IRC | 12:33 | |
| *** hyunsikyang has joined #openstack-fenix | 14:11 | |
| *** hyunsikyang__ has quit IRC | 14:14 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!