- openstack-discuss - lists.openstack.org

[cinder] last call for ussuri spec comments
by Brian Rosmaita 01 Feb '20

01 Feb '20

The following specs have two +2s. I believe that all expressed concerns have been addressed. I intend to merge them at 22:00 UTC today unless a serious issue is raised before then. https://review.opendev.org/#/c/684556/ - support volume-local-cache https://review.opendev.org/#/c/700977 - add backup id to volume metadata cheers, brian

3 3

"encoding error when doing console log show" in Rocky
by Albert Braden 31 Jan '20

31 Jan '20

We're running openstack-ansible Rocky and seeing the "encoding error when doing console log show." I see here that there is a patch: https://bugs.launchpad.net/python-openstackclient/+bug/1747862 How can I apply this patch to our Rocky install? Do I need to just copy the file over the existing /usr/lib/python2.7/dist-packages/openstackclient/compute/v2/console.py on controllers and hypervisors? Or is there a better way?

2 13

[tripleo] no rechecks
by Wesley Hayutin 31 Jan '20

31 Jan '20

Greetings, Hey folks we need a few patches to land before it's safe to blindly recheck. A few minor patches need to land before you can expect to get +1 from zuul in a lot of cases. We appreciate your patience while we resolve the issues. Thanks * TRIPLEO GATES ARE BROKEN* * need two patches to fix - * https://review.opendev.org/#/c/705051/ - * https://review.opendev.org/#/c/704885/ - * Also need - * https://review.rdoproject.org/r/#/c/24754/

1 1

[all][requirements][qa] zipp failing py3.5 again
by Radosław Piliszek 31 Jan '20

31 Jan '20

Folks, we are running circles here. [1] merged breaking zipp again for py3.5 The script which generates these proposals needs to learn python version constraints. Reproposed [2]. [1] https://review.opendev.org/705128 [2] https://review.opendev.org/705184 -yoctozepto

3 4

[ironic] An approach to removing WSME
by Steve Baker 31 Jan '20

31 Jan '20

I've put together a set of changes for removing WSME which involves copying just enough of it into ironic, and I think we need to have the conversation about whether this approach is desirable :) This git branch[1] finishes with WSME removed and existing tests passing. Here are some stats about lines of code (not including unit tests, calculated with cloc): 4500 wsme/wsme 6000 ironic/ironic/api master 7000 ironic/ironic/api story/1651346 In words, we need 1000 out of 4500 lines of WSME source in order to support 6000 lines of ironic specific API code. Switching to a replacement for WSME would likely touch a good proportion of that 6000 lines of ironic specific API code. If we eventually replace it with a new library I think it would be easier if the thing being replaced is inside the ironic source tree to allow for a gradual transition. So this approach could be an end in itself, or it could be a step towards the final goal. My strategy for copying in code was to pull in chunks of required logic while removing some unused features (like request pass-through and date & time type serialization). I also replaced py2/3 patterns with pure py3. There is likely further things which can be removed or refactored for simplicity, but what exists currently works. If there is enough positive feedback for this approach I'll start on unit test coverage for the new code. [1] https://review.opendev.org/#/q/topic:story/1651346

3 2

[oslo] AFK this week
by Ben Nemec 31 Jan '20

31 Jan '20

I am going to be out most or all of this week. If someone wants to run the meeting on Monday then feel free to have it without me again. I may not be back right away next week either. A lot of stuff is still to be determined at this point. -Ben

1 1

[keystone] Keystone Team Update - Week of 27 January 2020
by Vishakha Agarwal 31 Jan '20

31 Jan '20

# Keystone Team Update - Week of 27 January 2020 ## News ### User Support and Bug Duty Every week the duty is being rotated between the members. The person-in-charge for bug duty for current and upcoming week can be seen on the etherpad [1] [1] https://etherpad.openstack.org/p/keystone-l1-duty ## Open Specs Ussuri specs: https://bit.ly/2XDdpkU Ongoing specs: https://bit.ly/2OyDLTh ## Recently Merged Changes Search query: https://bit.ly/2pquOwT We merged 19 changes this week. ## Changes that need Attention Search query: https://bit.ly/2tymTje There are 26 changes that are passing CI, not in merge conflict, have no negative reviews and aren't proposed by bots. ### Priority Reviews * Ussuri Roadmap Stories - Groups in keystone SAML assertion https://tree.taiga.io/project/keystone-ussuri-roadmap/us/33 https://review.opendev.org/#/c/588211/ Add openstack_groups to assertion - Add support for modifying resource options to CLI tool https://tree.taiga.io/project/keystone-ussuri-roadmap/us/53 https://review.opendev.org/#/c/697444/ Adding options to user cli * Community Goals https://review.opendev.org/#/c/699119/ [ussuri][goal] Drop python 2.7 support and testing python-keystoneclient * Special Requests https://review.opendev.org/#/c/703578/ Updating tox -e all-plugin command https://review.opendev.org/#/c/704736/ Remove six usage ## Bugs This week we opened 4 new bugs and closed 4. Bugs opened (4) Bug #1860478 (keystone:Low): fetching role assignments should handle domain IDs in addition to project IDs - Opened by Harry Rybacki https://bugs.launchpad.net/keystone/+bug/1860478 Bug #1860252 (keystone:Undecided): security problem，one user can change other user's password without admin - Opened by kuangpeiling https://bugs.launchpad.net/keystone/+bug/1860252 Bug #1861264 (keystone:Undecided): Help information should be returned when executing 'placement-status' alone - Opened by Eric Xie https://bugs.launchpad.net/keystone/+bug/1861264 Bug #1861279 (keystone:Undecided): incomplete instruction in "Install and configure" in keystone in https://docs.openstack.org/keystone/train/install/keystone-install-rdo.html - Opened by Arie Maron https://bugs.launchpad.net/keystone/+bug/1861279 Bugs closed (4) Bug #1860252 (keystone:Undecided) https://bugs.launchpad.net/keystone/+bug/1860252 Bug #1861264 (keystone:Undecided) https://bugs.launchpad.net/keystone/+bug/1861264 Bug #1856904 (keystone:Undecided): CADF Notifications are missing user name in initiator object - Fixed by Gage Hugo https://bugs.launchpad.net/keystone/+bug/1856904 Bug #1858012 (keystone:Undecided): List role assignments by role ID may leak extra system assignments outside of filter - Fixed by Colleen Murphy https://bugs.launchpad.net/keystone/+bug/1858012 ## Milestone Outlook https://releases.openstack.org/ussuri/schedule.html Milestone 2 is in two weeks, which means spec freeze is coming up followed closely by feature proposal freeze. ## Help with this newsletter Help contribute to this newsletter by editing the etherpad: https://etherpad.openstack.org/p/keystone-team-newsletter ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

1 0

[cinder][nova] Migrating servers' root block devices from a cinder backend to another
by Jean-Philippe Méthot 30 Jan '20

30 Jan '20

Hi, We have a several hundred VMs which were built on cinder block devices as root drives which use a SAN backend. Now we want to change their backend from the SAN to Ceph. We can shutdown the VMs but we will not destroy them. I am aware that there is a cinder migrate volume command to change a volume’s backend, but it requires that the volume be completely detached. Forcing a detached state on that volume does let the volume migration take place, but the volume’s path in Nova block_device_mapping doesn’t update, for obvious reasons. So, I am considering forcing the volumes to a detached status in Cinder and then manually updating the nova db block_device_mapping entry for each volume so that the VM can boot back up afterwards. However, before I start toying with the database and accidentally break stuff, has anyone else ever done something similar? Got any tips or hints on how best to proceed? Jean-Philippe Méthot Openstack system administrator Administrateur système Openstack PlanetHoster inc.

7 10

[openstack-ansible][ansible-sig][tripleo] Openstack Ansible modules - update on progress, new meeting time
by Sagi Shnaidman 30 Jan '20

30 Jan '20

Hi, all because no people have been available in the meeting, I'll write a quick summary in ML. Agenda and progress are tracked here as usual: https://etherpad.openstack.org/p/openstack-ansible-modules 1. Because people had overlapping meetings I've created a poll for better meeting time. I think we can switch now to the bi-weekly meeting if no objections to it. The poll is here: https://xoyondo.com/dp/ITMGRZSvaZaONcz 2. We have the same coverage as we had in Ansible repository, so now we can start to improve our testing. It should include a matrix with various ansible versions and openstacksdk versions. We had already some issues with versions mismatch [1] Also we should start to do a better module design. 3. In case of critical patches to 2.9 and when openstack modules will be removed from Ansible repo, please submit patches directly to the 2.9-stable branch of Ansible. Confirmed that with Ansible folks. 4. The CI job in Ansible repository now is failing with a message about moving modules. I hope it won't block from 2.9 patches to land but will do more checking. But at least we've got attention to the fact modules are moving. 5. Jobs for pushing collections to ansible-galaxy are in progress and please review patches in the repo in your time: [1] https://review.opendev.org/705054 [2] https://review.opendev.org/#/q/project:openstack/ansible-collections-openst… -- Best regards Sagi Shnaidman

1 0

[ops] Federated Identity Management survey
by Colleen Murphy 30 Jan '20

30 Jan '20

Hello operators, A researcher from the University of Kent who was influential in the design of keystone's federation implementation has asked the keystone team to gauge adoption of federated identity management in OpenStack deployments. This is something we've neglected to track well in the last few OpenStack user surveys, so I'd like to ask OpenStack operators to please take a few minutes to complete the following survey about your usage of identity federation in your OpenStack deployment (even if you don't use federation): https://uok.typeform.com/to/KuRY0q The results of this survey will benefit not only university research but also the keystone team as it will help us understand where to focus our efforts. Your participation is greatly appreciated. Thanks for your time, Colleen (cmurphy)

1 3

[neutron][drivers team] Proposing Nate Johnston as drivers team member
by Slawek Kaplonski 30 Jan '20

30 Jan '20

Hi Neutrinos, I would like to propose Nate Johnston to be part of Neutron drivers team. Since long time Nate is very active Neutron's core reviewer. He is also actively participating in our Neutron drivers team meetings and he shown there that he has big experience and knowledge about Neutron, Neutron stadium projects as well as whole OpenStack. I think that he really deservers to be part of this team and that he will be great addition it. I will wait for Your feedback for 1 week and if there will be no any votes agains, I will add Nate to drivers team in next week. -- Slawek Kaplonski Senior software engineer Red Hat

5 5

[nova] API updates week 20-5
by Ghanshyam Mann 30 Jan '20

30 Jan '20

Hello Everyone, Please find the Nova API updates of this week. Please add if I missed any BPs/API related work. API Related BP : ============ COMPLETED: 1. Add image-precache-support spec: - https://blueprints.launchpad.net/nova/+spec/image-precache-support Code Ready for Review: ------------------------------ 1. Nova API policy improvement - Topic: https://review.opendev.org/#/q/topic:bp/policy-defaults-refresh+(status:ope… - Weekly Progress: First set is merged and 4 APIs policy are ready to review. I will push more changes in this week. - Review guide over ML - http://lists.openstack.org/pipermail/openstack-discuss/2019-August/008504.h… Specs are merged and code in-progress: ------------------------------ ------------------ 1. Non-Admin user can filter their instance by availability zone: - Topic: https://review.opendev.org/#/q/topic:bp/non-admin-filter-instance-by-az+(st… - Weekly Progress: Spec is just merged. 2. Boot from volume instance rescue - Topic: https://review.opendev.org/#/q/topic:bp/virt-bfv-instance-rescue+(status:op… - Weekly Progress: Code is in progress. WIP as of now. Spec Ready for Review or Action from Author: --------------------------------------------------------- 1. Support specifying vnic_type to boot a server -Spec: https://review.opendev.org/#/c/672400/ - Weekly Progress: As discussed in spec, this is something can be done via heat or other automation instead of doing it in nova. 2. Add action event fault details -Spec: https://review.opendev.org/#/c/699669/ - Weekly Progress: This seems good to go, it has two +2. 3. Allow specify user to reset password -Spec: https://review.opendev.org/#/c/682302/5 - Weekly Progress: One +2 on this but other are disagree on this idea. More discussion on review. 4. Support re-configure deleted_on_termination in server -Spec: https://review.openstack.org/#/c/580336/ - Weekly Progress: This is old spec and there is no consensus on this till now. Others: 1. None Bugs: ==== I will track the bug data from next report. NOTE- There might be some bug which is not tagged as 'api' or 'api-ref', those are not in the above list. Tag such bugs so that we can keep our eyes. -gmann

1 0

[NOVA] instance hostname vs display_name vs dns_name
by Volodymyr Litovka 30 Jan '20

30 Jan '20

Dear colleagues, I'm using DNS Integration and just faced an issue - after renaming instance, I can't bind port to the instance using new name: 1) I've created intances with name 'devel' 2) then I renamed it to devel (openstack server set --name packager devel) 3) when binding port with dns_name using new name ('packager' in my case), the following error appear: ERROR nova.api.openstack.wsgi PortNotUsableDNS: Port c3a92cf6-b49b-4570-b69b-0c23af1d1f94 not usable for instance 6aa78bd5-099e-4878-a5ac-90262505a924. Value packager assigned to dns_name attribute does not match instance's hostname devel and yes, record in DB still uses an old hostname: mysql> select display_name from instances where hostname='devel'; +--------------+ | display_name | +--------------+ | packager | +--------------+ 1 row in set (0.00 sec) and hostname remains the same (initial) regardless of any changes to display_name. I'm on Rocky. Is it bug or feature and are there ways to work around this? Thank you. -- Volodymyr Litovka "Vision without Execution is Hallucination." -- Thomas Edison

2 2

[goals][Drop Python 2.7 Support] Week R-15 Update (# 2.5 weeks left to complete)
by Ghanshyam Mann 29 Jan '20

29 Jan '20

Hello Everyone, Below is the progress on "Drop Python 2.7 Support" for R-15 week. Schedule: https://governance.openstack.org/tc/goals/selected/ussuri/drop-py27.html#sc… Highlights: ======== * Schedule is to finish the py2-drop from every deliverable (except requirement repo) by m-2 on R13. Only 2.5 weeks left to finish the work. * There are still cases where Tempest and plugins does not work due to py2 drops. New plan & option has been sent in separate ML thread [1] and work in progress. * Few services merged the patches but not all. If your project is in the below list, I request PTLs to review it on priority. Project wise status and need reviews: ============================ Phase-1 status: The OpenStack services have not merged the py2 drop patches: NOTE: This was supposed to be completed by milestone-1 (Dec 13th, 19). * Adjutant * Karbor * Masakari * Tacker * Qinling * Tricircle Phase-2 status: This is ongoing work and I think most of the repo have patches up to review. Try to review them on priority. If any query and I missed to respond on review, feel free to ping me in irc. * Open review: https://review.opendev.org/#/q/topic:drop-py27-support+status:open How you can help: ============== - Review the patches. Push the patches if I missed any repo. [1] http://lists.openstack.org/pipermail/openstack-discuss/2020-January/012241.… -gmann

1 0

[largescale-sig] Meeting summary and next actions
by Thierry Carrez 29 Jan '20

29 Jan '20

Hi everyone, The Large Scale SIG held a meeting today. You can access the summary and logs of the meeting at: http://eavesdrop.openstack.org/meetings/large_scale_sig/2020/large_scale_si… Regarding progress on our "Documenting large scale operations" goal, we turned our collection of relevant articles more as a group background activity than an immediate TODO. The idea is more to remember to add links to the etherpad[1] when we come across relevant content. oneswig mentioned the list to the Scientific SIG, in hopes that it would trigger more entries. [1] https://etherpad.openstack.org/p/large-scale-sig-documentation On the "Scaling within one cluster, and instrumentation of the bottlenecks" goal, nobody filed a scaling story yet on the etherpad[2] set up to collect them. masahito produced a first draft of the oslo.metrics blueprint[3], up for the group review. Finally, oneswig signed up to prepare a show-and-tell about some investigation they have been doing in this area, for presentation at next meeting. [2] https://etherpad.openstack.org/p/scaling-stories [3] https://review.opendev.org/#/c/704733/ Between now and next meeting, group members should prioritize: - Reviewing oslo.metrics draft at https://review.opendev.org/#/c/704733/ and comment so that we can iterate on it - Reading page on golden signals at https://landing.google.com/sre/sre-book/chapters/monitoring-distributed-sys… Other action items: - oneswig to prepare show-and-tell for presentation at next meeting - ttx to add meeting to eavesdrop on a every-two-week cadence - all post short descriptions of what happens (what breaks first) when scaling up a single cluster to https://etherpad.openstack.org/p/scaling-stories The next meeting will happen on February 12, at 9:00 UTC on #openstack-meeting. Cheers, -- Thierry Carrez (ttx)

2 1

[cinder] driver removal policy update (draft)
by Brian Rosmaita 29 Jan '20

29 Jan '20

I've posted a patch updating our third-party driver removal policy along the lines discussed at today's cinder meeting: https://review.opendev.org/704906 Let's continue the discussion in gerrit. I'll update the wiki after we've come to a consensus on the patch. cheers, brian

1 0

[tripleo] Removing the `openstack overcloud plan *` commands
by Dougal Matthews 29 Jan '20

29 Jan '20

Hey all, While doing work on the Mistral to Ansible port I was looking at the openstack overcloud plan commands. These are; - openstack overcloud plan create - openstack overcloud plan delete - openstack overcloud plan deploy - openstack overcloud plan list - openstack overcloud plan export I believe none of these commands make sense in a post-TripleO UI world. There is no other way to interact and update a plan. Deploys are always done via "openstack overcloud deploy" and this deletes the contents of the plan container and repopulates it with the local files[1]. I am therefore proposing that we remove these commands and skip the normal deprecation process. https://review.opendev.org/#/c/704581/1 What do you think? Thanks, Dougal [1]: https://github.com/openstack/python-tripleoclient/blob/3c589979ceb05d732b3c…

3 4

horizon: Trailing spaces removed on passwords
by Orestes Leal Rodríguez 28 Jan '20

28 Jan '20

>From the dashboard openstack is removing the trailing spaces from our user's passwords. We have a modified sql.py backend, that does an ldap bind to an active directory data store. And that works almost always. I say almost because for some users it doesn't work at all. We figure out (and a co-worker also confirmed this) that openstack is removing trailing (also leading?) spaces from the password entered in the dashboard. Also, inside the dashboard trailing spaces are not accepted even when they are equal byte by byte (including the space, I get an error). So this is going on. Do anybody knows where is this removal performed? (python script location, line) So I can remove that since I have users (me included, I have the issue since the very beginning of this deployment) that cannot login. And they can use their Active Directrory passwords from other apps without problem. We are running 'stein' with the latest update for ubuntu 18.04-AMD64.

5 5

[cyborg] No IRC meeting today
by Nadathur, Sundar 28 Jan '20

28 Jan '20

Since many people are traveling on account of the Chinese New Year, there will be no IRC meeting today. Regards, Sundar

1 1

[ops] OpenStack Ops Meetups team meeting 2020-1-20
by Chris Morgan 28 Jan '20

28 Jan '20

Meeting minutes for today's meeting are linked below. This was our first meeting since the very successful (or so I am told) Ops Meetup in London earlier this month, for which the agenda is here https://etherpad.openstack.org/p/LON-2020-OPS-AGENDA and the feedback was collected here https://etherpad.openstack.org/p/LON-2020-OPS-FEEDBACK the meetups team is now turning to looking into what we can arrange for operators at the Vancouver opendev+ptg event and then a second full ops meetup in south korea in september. Watch this space! Chris Morgan on behalf of the openstack ops meetups team <•openstack> Minutes: http://eavesdrop.openstack.org/meetings/ops_meetup_team/2020/ops_meetup_tea… 10:53 AM Minutes (text): http://eavesdrop.openstack.org/meetings/ops_meetup_team/2020/ops_meetup_tea… 10:53 AM Log: http://eavesdrop.openstack.org/meetings/ops_meetup_team/2020/ops_meetup_tea… -- Chris Morgan <mihalis68(a)gmail.com>

1 0

[barbican] TPM2.0 backend
by Balázs Gibizer 28 Jan '20

28 Jan '20

Hi, Looking at the Barbican documentation I see that the secrets can be stored on disk (SimpleCrypto backend) or in a HW vendor specific HSM module. Is there a way to use a TPM 2.0 device as the backend of Barbican via something like [1]? Cheers, gibi [1] https://github.com/tpm2-software/tpm2-pkcs11

1 1

[trio2o] Current state, community interest, alternatives
by Vladimir Kozhukalov 28 Jan '20

28 Jan '20

Hi, I am currently looking for a open source solution that is able to expose a single API endpoint for a user to multiple Openstack instances. Partly the idea behind is to stop investing efforts into tuning of Openstack to make it scalable and instead to deploy a bunch of small Openstack regions (say ~100 nodes) and work with them via API frontend. I am aware of Nova Cells, but it is complicated and Nova is not the only Openstack component that needs scaling. All these ad hoc tuning like dedicated RabbitMQ instances just make the system barely maintainable and it is complicated to reproduce. Requirements for a possible solution are: - Single authentication point (I know this is simple and can be for example SAML based federation) - Single point to manage keys and quotas - Possibly L2 connectivity between VMs in different Openstack instances (I know this can be achieved using VPNs, but this is kinda too complicated for a user) - Ability to work with different Openstack versions (need for gradual upgrade) I know Huawei and others implemented PoC for Openstack cascading and then invested some efforts in Trio2o and Tricircle. Tricircle is about cascading Neutron and IMO it is better to use a scalable SDN (like TangstenFabric or OVN). Trio2o looks like what I actually need, but the community interest around Trio2o is very limited (frankly, it looks almost dead). Last few weeks I was experimenting with Trio2o and Devstack with Stein release. It looks working, but the functionality is limited. Is there anyone who is interested in reviving Trio2o and investing into it? Maybe there are other solutions that I am not aware of? -- Best regards, Kozhukalov Vladimir

1 0

[cinder] changes for sqlalchemy 1.3.13
by Brian Rosmaita 27 Jan '20

27 Jan '20

Updating to the latest release of sqlalchemy breaks a cinder unit test. This patch shows the breakage and makes some suggestions about how it could be fixed. I could use some feedback: https://review.opendev.org/704425 thanks! brian

1 0

[neutron] CI meeting date changed
by Slawek Kaplonski 27 Jan '20

27 Jan '20

Hi, Just short info, according to what we agreed on our last CI meeting, it's now changed and will be on Wednesday at 1500 UTC. Please also note that IRC channel is also changed and now it will be on "openstack-meeting-3" You can check [1] for details. [1] https://review.opendev.org/#/c/704200/ -- Slawek Kaplonski Senior software engineer Red Hat

1 0

[ceilometer] Stable releases
by Pierre Riteau 27 Jan '20

27 Jan '20

Hello, Is it planned to release new Ceilometer tarballs for rocky and stein? There are several bug fixes (for example [1]) in stable branches that would be useful to include in a release. If t can help, I can submit patches to the releases repository. Thanks, Pierre Riteau (priteau) [1] https://bugs.launchpad.net/ceilometer/+bug/1801348

2 2