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

[nova] [neutron] multiple fixed_ip
by Arnaud Morin 04 Mar '20

04 Mar '20

Hello all, I was doing some tests to create a server using nova API. My objective is to create a server with one port but multiples IPs (one IPv4 and one IPv6). If I understand well the neutron API, I can create a port using the fixed_ips array parameter [1] Unfortunately, on nova side, it seems to only accept a string with only one ip (fixed_ip) [2] Is it mandatory for me to create the port with neutron? Or is there any trick that I missed on nova API side? Thanks! [1] https://docs.openstack.org/api-ref/network/v2/?expanded=create-port-detail#… [2] https://docs.openstack.org/api-ref/compute/?expanded=create-server-detail#c… -- Arnaud Morin

3 4

[neutron] Review priorities for RFEs
by Slawek Kaplonski 04 Mar '20

04 Mar '20

Hi neutrinos, I just went through patches related to our RFEs which we want to merge before Ussuri-3 and I set them review priority flag. You can find them at [1]. If I missed any patch related to any of those RFEs, please let me know by email or on IRC. And please try to review those patches if possible ;) Thx in advance. [1] https://tinyurl.com/vezk6n6 — Slawek Kaplonski Senior software engineer Red Hat

1 0

Re: [lists.openstack.org代发][nova] ask some questions about flavor
by Brin Zhang(张百林) 04 Mar '20

04 Mar '20

Items: [lists.openstack.org代发][nova] ask some questions about flavor Hi,all: > I have two questions about the flavor. One is the property "OS-FLV-DISABLED:disabled" defined when we created the flavor, but there is no method to change the property. And I see a patch about this [1], but it is abandoned. So I want to know our community how > to consider this function. "OS-FLV-DISABLED:disabled", this is typically only visible to administrative users. I'm not sure if it is worth supporting the modification/update. I think it should be planned in advance for managers. If the specific extra spec bound to flavor is only used as a dedicated flavor. > Another one is when we call the "list-migrations" api, why we receive the new_instance_type_id and old_instance_type_id in response of list_migrations should be internal value for the migration-type is "resize"[2]? Maybe the value should be exposed on REST > API, so that we can know which is the old flavor. In microversion 2.23 we support to show “migration-type” in the List Migrations API [1][2], I think you call List Migrations REST API is not add the “OpenStack-API-Version” or “X-OpenStack-Nova-API-Version”, right？ And you can get the list migrations filter by “migration-type” (enum in: evacuation, live-migration, migration (cold), resize), as the call such as: http://192.168.2.11/compute/v2.1/os-migrations?migration_type=resize that show what migration type what you want. [1] API changes: https://docs.openstack.org/api-ref/compute/?expanded=list-migrations-detail… [2] microverion 2.3: https://docs.openstack.org/nova/latest/reference/api-microversion-history.h… > Can you tell me more about this? Thank you very much. Ref: [1]:https://review.opendev.org/#/c/61291/ [2]:https://review.opendev.org/#/c/588481/<https://review.opendev.org/#/c/588481/4> Best Regards Rambo

2 1

[nova] ask some questions about flavor
by Rambo 04 Mar '20

04 Mar '20

Hi,all:         I have two questions about the flavor. One is the property "OS-FLV-DISABLED:disabled" defined when we created the flavor, but there is no method to change the property. And I see a patch about this [1], but it is abandoned. So I want to know our community how to consider this function.          Another one is when we call the "list-migrations" api, why we receive the new_instance_type_id and old_instance_type_id in response of list_migrations should be internal value for the migration-type is "resize"[2]? Maybe the value should be exposed on REST API, so that we can know which is the old flavor.          Can you tell me more about this? Thank you very much. Ref: [1]:https://review.opendev.org/#/c/61291/ [2]:https://review.opendev.org/#/c/588481/ Best Regards Rambo

1 0

[all] Announcing OpenStack Wallaby!
by Sean McGinnis 04 Mar '20

04 Mar '20

That's right - we have the W name already. No months on end of referring to "the W release" as we start making long term plans. Wallaby https://en.wikipedia.org/wiki/Wallaby Wallabies are native to Australia, which at the start of this naming period was experiencing unprecedented wild fires. This name has passed the legal vetting phase and we are good to go. Full results of the naming poll can be found here: https://civs.cs.cornell.edu/cgi-bin/results.pl?id=E_60b98f102c5d08d5 As a reminder, the way we choose a release name had several changes this time around. Full details of the current release process can be found here: https://governance.openstack.org/tc/reference/release-naming.html There were a lot of great names proposed. I think removing the geographical restriction has been a good move. Thank you to everyone who proposed ideas for this release cycle. Happy stacking! Sean

2 1

[glance] Different checksum between CLI and curl
by gaetan.trellu＠incloudus.com 04 Mar '20

04 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 04 Mar '20

04 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