| takahashi-tsc | Hi Tacker team | 08:11 |
|---|---|---|
| shivam | Hello | 08:11 |
| takahashi-tsc | I'm not sure yasufumi-san will join... are there any Tacker members and any topics to discuss? | 08:11 |
| shivam | yes | 08:12 |
| takahashi-tsc | ok... let's start the meeting. | 08:13 |
| takahashi-tsc | #startmeeting tacker | 08:13 |
| opendevmeet | Meeting started Mon Mar 9 08:13:23 2026 UTC and is due to finish in 60 minutes. The chair is takahashi-tsc. Information about MeetBot at http://wiki.debian.org/MeetBot. | 08:13 |
| opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 08:13 |
| opendevmeet | The meeting name has been set to 'tacker' | 08:13 |
| takahashi-tsc | Hi tacker team | 08:13 |
| takahashi-tsc | #link https://etherpad.opendev.org/p/tacker-meeting | 08:13 |
| takahashi-tsc | There are several topics... Let's move forward with the discussion smoothly. | 08:14 |
| takahashi-tsc | #topic Discussion for NFV Release6 | 08:15 |
| takahashi-tsc | #link https://etherpad.opendev.org/p/tacker-meeting#L11 | 08:16 |
| takahashi-tsc | I have shared discussion topic about NFV release6. In short, the document structure will change significantly. | 08:17 |
| takahashi-tsc | Regarding SOL002 and SOL003, there are many duplicated contents, so we will propose that creating "common part", "trimmed SOL002" and "trimmed SOL003." | 08:18 |
| takahashi-tsc | I do not explain the details because we do not have much time... Please see the etherpad pages. | 08:18 |
| takahashi-tsc | And, if you have any comments, please share your opinions. Actually we have NFV meeting in this week. If I receive the comments, we can reflect them to discussion. | 08:19 |
| takahashi-tsc | Even if your comments are delayed, I think we can still revise our proposal later. | 08:20 |
| takahashi-tsc | Any comments for now? | 08:20 |
| takahashi-tsc | Good, if nothing, I'd like to move on next topic. | 08:24 |
| takahashi-tsc | OK. Shivam Please explain your topic. | 08:25 |
| shivam | Ok | 08:25 |
| shivam | #topic Community Input Required: Keep or Remove Temporary Tacker Fix for Nova scheduler_hints Issue | 08:26 |
| shivam | This discussion topic is related to Tacker side temporary fix regarding Nova bug "scheduler_hints.group Returned as List Instead of String in Get Server API (microversion 2.100)". | 08:26 |
| shivam | Nova Bug Report: https://bugs.launchpad.net/nova/+bug/2139275 | 08:27 |
| shivam | A bug in OpenStack Nova (API microversion 2.100) returns scheduler_hints.group as a list instead of a string in the Get Server API, which causes issues for OpenStack Tacker during change_vnfpkg rolling update operations. | 08:27 |
| shivam | It was discussed in IRC and following the IRC discussion outcome, the Tacker-side workaround has been implemented in the following patch, which is currently under review: https://review.opendev.org/c/openstack/tacker/+/975363 | 08:28 |
| shivam | And, the issue has been reported to the Nova team through a Launchpad bug report and email, requesting a fix from their side, but there has been no response and the issue has not been fixed yet. | 08:29 |
| shivam | Initially, the plan was to apply this workaround temporarily and remove it before the Tacker release, since the changes are in the infra driver and which is directly related to Tacker’s core functionality. | 08:30 |
| shivam | However, if the workaround is removed while the Nova bug still exists, the tacker-ft-v2-st-userdata-basic-min CI job (a voting test) will fail again, and since it is a voting CI job, the test cannot simply be removed. | 08:30 |
| shivam | This workaround allows change_vnfpkg rolling updates to complete successfully, but it also has some impacts i.e., it may cause Heat to replace servers instead of updating them in place, and VDU IDs may change after the operation. | 08:31 |
| shivam | Therefore, the current proposal is to keep the temporary workaround in Tacker until the Nova issue is fixed, to maintain CI stability despite some impacts caused by the workaround. | 08:32 |
| shivam | Hence, I would like to know community member's opinion on the proposal "whether we should keep this temporary workaround in Tacker (despite its impacts) until the Nova issue is fixed, or not." | 08:33 |
| shivam | Thank you. | 08:33 |
| takahashi-tsc | Thanks. | 08:34 |
| takahashi-tsc | My concern is that this workaround partially changes "Tacker's behavior." | 08:34 |
| takahashi-tsc | https://review.opendev.org/c/openstack/tacker/+/975363/5/tacker/sol_refactored/infra_drivers/openstack/openstack.py | 08:34 |
| takahashi-tsc | Is my understanding correct? and when Nova is fixed, might this lead to errors again? | 08:35 |
| shivam | yes, your understanding is correct. | 08:36 |
| takahashi-tsc | thanks | 08:38 |
| shivam | But, as per my unserstanding, this might not cause error when Nova fixes the issue, because current workaround removes the scheduler_hints from VDU templates. | 08:38 |
| takahashi-tsc | Ah, OK. sorry my understanding is not correct. Yes, it seems to "remove" the scheduler_hints, the error may not occur. | 08:39 |
| shivam | yes | 08:40 |
| takahashi-tsc | OK... Personally, I think merging the workaround is the simplest way. | 08:41 |
| takahashi-tsc | Just in case, I'll confirm with NTT and DOCOMO members because they are using Tacker v2. | 08:42 |
| shivam | ok, thank you. | 08:43 |
| takahashi-tsc | Any comments? Unfortunately there are few members today... I know that this week is NFV plenary week... | 08:43 |
| takahashi-tsc | OK, if nothing let's move on next topic. | 08:44 |
| takahashi-tsc | Shivam, please go ahead. | 08:45 |
| shivam | ok | 08:45 |
| shivam | #topic Tacker CI failure due to python-glanceclient removal in heat-translator 3.4.0 | 08:45 |
| shivam | This topic is related to an issue causing Tacker CI jobs failure due to python-glanceclient removal in heat-translator 3.4.0 | 08:46 |
| shivam | OpenStack Tacker CI jobs are currently failing with ModuleNotFoundError: No module named 'glanceclient', which eventually leads to retry_limit errors during execution. | 08:46 |
| shivam | The issue started after heat-translator version 3.4.0, where the dependency python-glanceclient was removed from its requirements. Earlier, Tacker use to install python-glaceclient indirectly through heat-translator. | 08:47 |
| shivam | Heat-translator commit that removed the dependency: https://github.com/openstack/heat-translator/commit/2ff9820276cc26f7c20bdbc471c84bc8e43c0c69 | 08:47 |
| shivam | A bug report has been created for this issue: https://bugs.launchpad.net/tacker/+bug/2143398 | 08:47 |
| shivam | Since Tacker still imports glanceclient, and the module is no longer installed in the environment, resulting in runtime failures in CI jobs. | 08:48 |
| shivam | The proposed fix is to add python-glanceclient as a direct dependency in Tacker’s requirements.txt, ensuring the module is always installed. | 08:49 |
| shivam | I would like to know community member's opinion on the identified root cause and proposed fix. | 08:49 |
| shivam | Thank you. | 08:49 |
| takahashi-tsc | Thanks. I think the approach is OK. | 08:50 |
| takahashi-tsc | Of course, there might be some issues if just add one line in requirements.txt, but it is not major issues. | 08:50 |
| takahashi-tsc | I'd like you to proceed with your fix. | 08:51 |
| takahashi-tsc | Any comments? | 08:51 |
| takahashi-tsc | OK, then let's move on the final topic. | 08:53 |
| shivam | ok | 08:53 |
| takahashi-tsc | This is also from Shivam? | 08:53 |
| shivam | yes | 08:53 |
| shivam | #topic Tacker CI failure due to config_format AttributeError in tox ≥ 4.48.0 | 08:54 |
| shivam | This topic is related to another issue causing Tacker CI jobs failure due to config_format AttributeError in tox ≥ 4.48.0. | 08:54 |
| shivam | OpenStack Tacker CI jobs are failing before the DevStack setup stage due to the error AttributeError: 'Parsed' object has no attribute 'config_format', which occurs during the “Get tox envlist config” pre-devstack task. | 08:55 |
| shivam | The issue started occuring after the release of tox version 4.48.0 and later, where the implementation of the show_config command expects a new configuration attribute config_format that is not present in the parsed configuration object during CI execution. | 08:55 |
| shivam | In earlier versions of tox (e.g., 4.47.3), the show_config implementation did not require the config_format attribute, which is why those versions work correctly without causing the error. | 08:56 |
| shivam | To resolve this issue, the ideal fix would be to resolve the issue within the tox project itself, since the behavior originates from changes introduced in newer tox versions. | 08:57 |
| shivam | As a temporary workaround on the Tacker side, the proposal is to add a version constraint in tox.ini to restrict tox to a stable version (e.g., 4.47.3), meaning "tox < 4.48.0". | 08:57 |
| shivam | And, once the issue is resolved from "tox", we will remove this temporary fix (tox version constraint). | 08:58 |
| shivam | A bug report has been created for this issue: https://bugs.launchpad.net/tacker/+bug/2143583 | 08:58 |
| shivam | Now, I would like to know community member's opinion on the identified root cause and proposed fix. | 08:58 |
| shivam | Thank you. | 08:58 |
| takahashi-tsc | Thanks. Ideally, we should configure config_format properly. are there any challenges with that? | 09:00 |
| shivam | yes, but since this is inside "tox" changes only. And the path "state.conf.options.config_format" where it is expecting config_format to be in, is imsdie the "tox" itself | 09:01 |
| shivam | This is the "tox" commit where it was added: https://github.com/tox-dev/tox/commit/4c728874bc724b0a386bf05283f0b553ae863f96#diff-a19af327a0ff2b2219626cd09d494cf4dc6f7c2c96e20c7ecd08d959c0c8978bR60 | 09:02 |
| shivam | so, as per my understanding, I don't think, we can configure it from Tacker side. | 09:04 |
| takahashi-tsc | OK,I'll check the details later, but I think it's OK to change the tox version. | 09:05 |
| takahashi-tsc | It will not affect Tacker development, just test environment setting. | 09:06 |
| shivam | ok | 09:06 |
| takahashi-tsc | Thank you shivam. I have added today's my comments in etherpad. Please check it and if my understanding is not correct, please let me know. | 09:08 |
| shivam | ok, thank you. | 09:08 |
| takahashi-tsc | I think it’s OK to move forward based on your proposal. Let’s gather everyone’s review comments on Gerrit, and if necessary, please refer to today’s meeting minutes. | 09:09 |
| takahashi-tsc | or etherpad notes. | 09:09 |
| shivam | ok | 09:10 |
| takahashi-tsc | OK, sorry 10 minutes over. Any other comments? | 09:10 |
| takahashi-tsc | OK, let's close the meeting. | 09:13 |
| takahashi-tsc | Thank you, bye! | 09:13 |
| shivam | Thank you, bye. | 09:13 |
| takahashi-tsc | #endmeeting | 09:13 |
| opendevmeet | Meeting ended Mon Mar 9 09:13:33 2026 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 09:13 |
| opendevmeet | Minutes: https://meetings.opendev.org/meetings/tacker/2026/tacker.2026-03-09-08.13.html | 09:13 |
| opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/tacker/2026/tacker.2026-03-09-08.13.txt | 09:13 |
| opendevmeet | Log: https://meetings.opendev.org/meetings/tacker/2026/tacker.2026-03-09-08.13.log.html | 09:13 |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!