March 2020 - openstack-discuss - lists.openstack.org

[glance] Different checksum between CLI and curl
by gaetan.trellu＠incloudus.com 03 Mar '20

03 Mar '20

Hi guys, Does anyone know why the md5 checksum is different between the "openstack image save" CLI and "curl" commands? During the image creation a checksum is computed to check the image integrity, using the "openstack" CLI match the checksum generated but when "curl" is used by following the API documentation[1] the checksum change at every "download". Any idea? Thanks, Gaëtan [1] https://docs.openstack.org/api-ref/image/v2/index.html?expanded=download-bi…

12 21

[all][PTL][tc] U Community Goal: Project PTL & Contrib Docs Update #4
by Kendall Nelson 03 Mar '20

03 Mar '20

Hello! The decision has been made and the template has been merged[1] with the actual governance update well on its way to being accepted as well[2][3]. With that in mind, its time to get down to business! If You Haven't Started ================= You're in luck! Nothing has to change, just run the cookie cutter[8] and make whatever changes you need to the template and get it added to the repos for your project. Also, please keep the task tracker in StoryBoard up to date[4] as you work. If You Had 'Completed' it before Update #2 ================================ Go and check out the governance change[2] and the template[1] to make sure there isn't anything you did differently-- depending on when you 'completed' the goal, you might have some changes to make to be in line with the current (and presumably final) completion criteria. And, as always, please make sure to keep the tasks in StoryBoard up to date[4]. Previous updates if you missed them[5][6][7]. If you have any questions, please let me know. If you need help with reviewing, please feel free to add me as a reviewer. We can still get this done in time! Thanks, -Kendall (diablo_rojo) [1] Cookie Cutter Change https://review.opendev.org/#/c/708672/ [2] Governance Goal Change https://review.opendev.org/#/c/709617/ [3] Governance Goal Change Typo Fix https://review.opendev.org/#/c/710332 <https://review.opendev.org/#/c/710332/1> [4] StoryBoard Tracking: https://storyboard.openstack.org/#!/story/2007236 [5]Update #1: http://lists.openstack.org/pipermail/openstack-discuss/2020-February/012364… [6] Update #2: http://lists.openstack.org/pipermail/openstack-discuss/2020-February/012570… [7] Update 3: http://lists.openstack.org/pipermail/openstack-discuss/2020-February/012784… [8] Cookiecutter Template: https://opendev.org/openstack/cookiecutter/src/branch/master/%7b%7bcookiecu…

1 0

[openstack-helm] Virtual Midcycle - March 2020
by Gage Hugo 03 Mar '20

03 Mar '20

Hello everyone, The openstack-helm team is looking to host a virtual midcycle this month to discuss a variety of topics in a more focused group via video conferencing. We have started an etherpad to track anyone who wishes to attend as well as any topics that want to be discussed. https://etherpad.openstack.org/p/osh-virtual-ptg-2020-03 Also there is a doodle poll to determine a time-slot in the next few weeks for the best time, we will try to choose the best time-slot based on the number of available people, please check it out here: https://doodle.com/poll/g6uvdb4rbad9s8gb We will update the etherpad to track any changes in meeting schedule, as well as when we find a web conference tool that works best. - Gage

1 0

Trouble with large RAM VMs - (formerly RE: Virtio memory balloon driver)
by Albert Braden 03 Mar '20

03 Mar '20

I successfully blacklisted the "memory balloon" driver in a VM image and found that another driver is choking on the large memory. The guys in #centos think it might be the USB or PS/2 drivers. Since these are kernel drivers they cannot be blacklisted; the only way to not load them is to not call them. On the hypervisor in /etc/libvirt/qemu/<instance-name>.xml I see this: <input type='tablet' bus='usb'> <address type='usb' bus='0' port='1'/> </input> <input type='mouse' bus='ps2'/> <input type='keyboard' bus='ps2'/> I tried removing those lines, but it looks like the XML file is re-created whenever I stop and start the VM. In nova.conf I see "#pointer_model = usbtablet" If I set "pointer_model = <None>" on the HV and then restart nova, I see this in the log: 2020-03-03 11:05:17.233 228915 ERROR nova ConfigFileValueError: Value for option pointer_model is not valid: Valid values are [None, ps2mouse, usbtablet], but found '<None>' If I set "pointer_model = None" then I see this: 2020-03-03 11:06:24.761 229290 ERROR nova ConfigFileValueError: Value for option pointer_model is not valid: Valid values are [None, ps2mouse, usbtablet], but found 'None' What am I missing? # # Generic property to specify the pointer type. # # Input devices allow interaction with a graphical framebuffer. For # example to provide a graphic tablet for absolute cursor movement. # # If set, the 'hw_pointer_model' image property takes precedence over # this configuration option. # # Possible values: # # * None: Uses default behavior provided by drivers (mouse on PS2 for # libvirt x86) # * ps2mouse: Uses relative movement. Mouse connected by PS2 # * usbtablet: Uses absolute movement. Tablet connect by USB # # Related options: # # * usbtablet must be configured with VNC enabled or SPICE enabled and SPICE # agent disabled. When used with libvirt the instance mode should be # configured as HVM. # (string value) # Possible values: # <None> - <No description provided> # ps2mouse - <No description provided> # usbtablet - <No description provided> pointer_model = None -----Original Message----- From: Albert Braden Sent: Wednesday, February 5, 2020 9:34 AM To: openstack-discuss(a)lists.openstack.org Subject: RE: Virtio memory balloon driver When I start and stop the giant VM I don't see any evidence of OOM errors. I suspect that the #centos guys may be correct when they say that the "Virtio memory balloon" device is not capable of addressing that much memory, and that I must disable it if I want to create VMs with 1.4T RAM. Setting "mem_stats_period_seconds = 0" doesn't seem to disable it. How are others working around this? Is anyone else creating Centos 6 VMs with 1.4T or more RAM?

1 0

[kuryr] Job running open resolver
by corvus＠inaugust.com 03 Mar '20

03 Mar '20

Hi, The openstack-infra team received a report from one of our infrastructure donors that a gate job run by Kuryr is running a DNS resolver open to the Internet. This is dangerous as, if discovered, it can be used as part of DNS reflection attacks. The community and our infrastructure donors share an interest in avoiding misuse of our resources. Would you please look into whether this job is perhaps opening its iptables ports too liberally, and whether that can be avoided? The job is kuryr-kubernetes-tempest-containerized-ovn, and the build which triggered the alerting system is this one: https://zuul.opendev.org/t/openstack/build/166301f57b21402d8d8443bb1e17f970 Thanks, Jim

3 2

[Swift] Errno 13 Permission denied writing objects xattr while upgrading from Kilo to Queens
by Gui Maluf 03 Mar '20

03 Mar '20

Hi all, I'm struggling with something really wierd. 3 weeks ago I started upgrading my Keystone + Swift Ubuntu environment from Kilo to Queens. So I moved from Ubuntu 14.04 to 18.04. I can create new accounts and containers. But no objects. I think between Mitaka and Newton my storages started to throw Permission Denied error while writing object metadata. I saw that the piece of python code where I getting problem was changed in Rocky version and in hope of getting things fixed I've upgraded storage version. But the error persists. http://paste.openstack.org/show/790217/ I've check everything I could, user, permissions, mount options. But I still getting this error. I wrote a python script for creating files and writing metadata within the swift mount with swift user and everything works fines. Don't know what to do anymore. This is a "dev" environment with two storages only and a few disks. Since I'm planning to do in the production environment I'm quite scared if this happens again. Thanks in advance -- *guilherme* \n11 \t *maluf*

2 1

[idea] voting procedure and the choice's engineering
by Herve Beraud 03 Mar '20

03 Mar '20

Hello openstacker, I proposed a new openstack/idea to open discussion about how we should decide, when, and for what purpose, and the methods and tools to help us in this task. https://review.opendev.org/#/c/710107/ This document is not a "well finished document" but more a draft to open the debate and to help us to make the decisions better. I referenced some methods and tools. If some of you are interested by this topic then do not hesitate to leave comments and push changes on this document. Communities are driven by choices and decisions, Openstack is a community, let's decide how to choose on Openstack. -- Hervé Beraud Senior Software Engineer Red Hat - Openstack Oslo irc: hberaud -----BEGIN PGP SIGNATURE----- wsFcBAABCAAQBQJb4AwCCRAHwXRBNkGNegAALSkQAHrotwCiL3VMwDR0vcja10Q+ Kf31yCutl5bAlS7tOKpPQ9XN4oC0ZSThyNNFVrg8ail0SczHXsC4rOrsPblgGRN+ RQLoCm2eO1AkB0ubCYLaq0XqSaO+Uk81QxAPkyPCEGT6SRxXr2lhADK0T86kBnMP F8RvGolu3EFjlqCVgeOZaR51PqwUlEhZXZuuNKrWZXg/oRiY4811GmnvzmUhgK5G 5+f8mUg74hfjDbR2VhjTeaLKp0PhskjOIKY3vqHXofLuaqFDD+WrAy/NgDGvN22g glGfj472T3xyHnUzM8ILgAGSghfzZF5Skj2qEeci9cB6K3Hm3osj+PbvfsXE/7Kw m/xtm+FjnaywZEv54uCmVIzQsRIm1qJscu20Qw6Q0UiPpDFqD7O6tWSRKdX11UTZ hwVQTMh9AKQDBEh2W9nnFi9kzSSNu4OQ1dRMcYHWfd9BEkccezxHwUM4Xyov5Fe0 qnbfzTB1tYkjU78loMWFaLa00ftSxP/DtQ//iYVyfVNfcCwfDszXLOqlkvGmY1/Y F1ON0ONekDZkGJsDoS6QdiUSn8RZ2mHArGEWMV00EV5DCIbCXRvywXV43ckx8Z+3 B8qUJhBqJ8RS2F+vTs3DTaXqcktgJ4UkhYC2c1gImcPRyGrK9VY0sCT+1iA+wp/O v6rDpkeNksZ9fFSyoY2o =ECSj -----END PGP SIGNATURE-----

1 0

oslotest no longer pulls in stestr
by Ben Nemec 03 Mar '20

03 Mar '20

Hi all, We just released [0], which removes stestr as a requirement for oslotest. As a result, if you were relying on oslotest to pull in stestr you are most likely broken now. We did check for this situation before making the change, but it seems we missed at least one project so I'm sending this out in case anyone else is affected. The fix is to explicitly list stestr in your test-requirements (oslotest doesn't actually need stestr, so it's not the right place for the requirement to live). Reply here or ping us in #openstack-oslo with any questions. Thanks. -Ben 0: https://review.opendev.org/#/c/615826

1 0

[all] oslotest 4.0 drops the moxstubout and functional modules, stestr dependency
by Stephen Finucane 03 Mar '20

03 Mar '20

Per $subject, the 'oslotest.moxstubout' modules has been removed in oslotest 4.0 and we no longer include 'stestr' in our list of dependencies. I think I've resolved all issues in the 'openstack/' namespaced projects, but if not the remedies are to either switch to mock or explicitly included 'stestr' in your 'test-requirements.txt' file, respectively. We've also removed the 'oslotest.functional' module, but there do not appear to have been any users of this module. Stephen

1 0

[queens][neutron][fwaas_v2] PENDING_UPDATE
by Ignazio Cassano 03 Mar '20

03 Mar '20

Hello All, I installed firewall v2 on queens based on centos 7. I create a firewall group policy and a firewall group rulle with that policy. The firewall results ACTIVE AND UP and INACTIVE When I ttry to apply the firewall group to an instance port : openstack firewall group set --port c7c8be58-35de-47fe-87db-39bbd681db8b fwg1 It does not works and goes in pending update status. L3 agent log reports: Could not load neutron_fwaas.services.firewall.drivers.linux.iptables_fwaas_v2.IptablesFwaasDriver But the file esists I am using firewall_driver = openvswitch Please, whai is wrong ? I read Supports L2 firewalling (VM ports) was planned for ocata and I am on queens. Please, help me. Ignazio

2 2

[goals][Drop Python 2.7 Support] Week R-10 Update
by Ghanshyam Mann 02 Mar '20

02 Mar '20

Hello Everyone, Below is the progress on "Drop Python 2.7 Support" at end of R-10 week. Schedule: https://governance.openstack.org/tc/goals/selected/ussuri/drop-py27.html#sc… Highlights: ======== * We already passed the deadline but still work is not completed on this. * Few tempest plugins are failing. * I request projects again to merge the passing patches asap. Project wise status and need reviews: ============================ Phase-1 status: All the OpenStack services have dropped the python2.7. Phase-2 status: * Few Tempest plugins are still not merged. I am debugging a few failing plugins with the project team. ** tempest plugins are passing and ready to merge. *** barbican-tempest-plugin: https://review.opendev.org/#/c/704083/ *** cyborg-tempest-plugin: https://review.opendev.org/#/c/704076/ *** magnum-tempest-plugin: https://review.opendev.org/#/c/704069/ *** congress-tempest-plugin: https://review.opendev.org/#/c/694437/ *** heat-tempest-plugin: https://review.opendev.org/#/c/704225/ *** kuryr-tempest-plugin: https://review.opendev.org/#/c/704072/ ** Failing plugins which need debugging and more work: *** trove-tempest-plugin: https://review.opendev.org/#/c/692041/ *** ironic-tempest-plugin: https://review.opendev.org/#/c/704093/ * Started pushing the required updates on deployment projects. ** Completed or no updates required: *** Openstack-Chef - not required *** Packaging-Rpm - Done *** Puppet Openstack- Done ** In progress: *** Openstack Charms *** Openstackansible - In-progress. centos7 jobs are failing on few projects. ** Waiting from projects team to know the status: *** Openstack-Helm (Helm charts for OpenStack services) *** Tripleo (Deployment service) * Open review: https://review.opendev.org/#/q/topic:drop-py27-support+status:open Phase-3 status: This is audit and requirement repo work which is not started yet. I will start this once all the phase-2 work mentioned above is completed. How you can help: ============== - Review the patches. Push the patches if I missed any repo. -gmann

1 0

[tripleo] centos-8 status
by Wesley Hayutin 02 Mar '20

02 Mar '20

Greetings everyone! First off, apologies for using hackmd in this case for our status. We will be moving to a fully non-hosted open source version asap. We have been making significant progress with CentOS-8 based jobs in TripleO Ussuri, our full status is here [1]. You will see tripleo-ci-centos-8 standalone jobs hitting your reviews. If you find any issues please just treat the same as any other bug and open a launchpad w/ "alert" set to the tag. Thanks for everybody's hard work and patience as we get the rest of the jobs running smoothly in the upstream. [1] https://hackmd.io/HrQd03c9SxOMtFPFrq50tg?view

1 0

Virtualenv (and Tox) broken when run under python<3.6
by Clark Boylan 02 Mar '20

02 Mar '20

A recent release of importlib-resources (1.1.0) no longer works on python2.7 or python3.5. The issue is they import typing's ContextManager which didn't exist until python3.6 [0]. This means that python2 jobs and python3.5 jobs are currently unhappy if they need virtualenv. Unfortunately, many of our jobs use tox which uses virtualenv. One workaround being investigated [1] is to install importlib-resources==1.0.2 which does not try to use typing's ContextManager. If this is confirmed to work we will want to consider adding this change to the base job so that all jobs don't have to fix it separately. Note the version of python here is the one used to run virtualenv not the version of python being installed into the virtualenv. This means python3.6 running virtualenv to create a python2 virtualenv should be fine. But python3.5 running virtualenv to create a python3.6 env would not be fine. [0] https://gitlab.com/python-devs/importlib_resources/issues/83 [1] https://review.opendev.org/#/c/710729/ Clark

1 2

[TaskFlow] running multiple engines on a shared thread
by Sachin Laddha 02 Mar '20

02 Mar '20

Hi, We are using taskflow to execute workflows. Each workflow is executed by a separate thread (using engine.run() method). This is limiting our capability to execute maximum number of workflows that can run in parallel. It is limited by the number of threads there in the thread pool. Most of the time, the workflow tasks are run by agents which could take some time to complete. Each engine is alive and runs on a dedicated thread. Is there any way to reuse or run multiple engines on one thread. The individual tasks of these engines can run in parallel. I came across iter_run method of the engine class. But not sure if that can be used for this purpose. Any help is highly appreciated.

2 3

[neutron] bug deputy report for 2020-02-24 week
by Takashi Yamamoto 01 Mar '20

01 Mar '20

2020-02-24 week Someone familiar with these topics need to investigate DHCP agent: https://bugs.launchpad.net/neutron/+bug/1864711 DHCP port rescheduling causes ports to grow, internal DNS to be broken WSGI: https://bugs.launchpad.net/neutron/+bug/1864418 has wrong with use apache to start neutron api in docker container L3 agent: https://bugs.launchpad.net/neutron/+bug/1864963 loosing connectivity to instance with FloatingIP randomly API server performance: https://bugs.launchpad.net/neutron/+bug/1865223 [scale issue] regression for security group list between Newton and Rocky+ RFE https://bugs.launchpad.net/neutron/+bug/1864841 Neutron -> Designate integration does not consider the dns pool for zone Medium https://bugs.launchpad.net/neutron/+bug/1864822 Openvswitch Agent - Connexion openvswitch DB Broken Low https://bugs.launchpad.net/neutron/+bug/1864374 ml2 ovs does not flush iptables switching to FW ovs OVN-specific stuff High https://bugs.launchpad.net/neutron/+bug/1864620 [OVN] neutron_tempest_plugin.scenario.test_security_groups.NetworkSecGroupTest.test_multiple_ports_portrange_remote often fails Medium https://bugs.launchpad.net/neutron/+bug/1864833 [OVN] Functional tests start with OVSDB binary 2.9 instead 2.12 https://bugs.launchpad.net/neutron/+bug/1864639 [OVN] UpdateLRouterPortCommand and AddLRouterPortCommand needs to specify network Wishlist https://bugs.launchpad.net/neutron/+bug/1864640 [Ussuri] Neutron API writes to the Southbound DB

1 0

Retiring openstack/faafo repository
by Andreas Jaeger 01 Mar '20

01 Mar '20

This repo was part of the first-app document that was retired last year as part of the api-site retirement work. Thus, I'm retiring the openstack/faafo repository now. Patches are pushed with topic:retire-faafo, starting at https://review.opendev.org/710652 Andreas -- Andreas Jaeger aj(a)suse.com Twitter: jaegerandi SUSE Software Solutions Germany GmbH, Maxfeldstr. 5, D 90409 Nürnberg (HRB 36809, AG Nürnberg) GF: Felix Imendörffer GPG fingerprint = EF18 1673 38C4 A372 86B1 E699 5294 24A3 FF91 2ACB

1 0