May 2024 - openstack-discuss - lists.openstack.org

[neutron] Bug deputy report for week of May 6th
by Miguel Lavalle 13 May '24

13 May '24

Hi, I was the Neutron bugs deputy last week. Following is a brief summary of the reported bugs: Critical bugs ----------------- None High bugs -------------- https://bugs.launchpad.net/neutron/+bug/2065094 OVN revision number is bumped on noop port update. Assignee Jakub Libosvar. Released fix: https://review.opendev.org/c/openstack/neutron/+/918492 https://bugs.launchpad.net/neutron/+bug/2065459 ovn-octvia-provider: RuntimeError: cannot join current thread. In progress. Assignee Terry Wilson. Proposed fix: https://review.opendev.org/c/openstack/ovn-octavia-provider/+/919229 https://bugs.launchpad.net/neutron/+bug/2065460 ovn-octavia-provider results in a new OVSDB connection for every Octavia API request. In progress. Assignee Terry Wilson. Proposed fix: https://review.opendev.org/c/openstack/ovn-octavia-provider/+/919229 https://bugs.launchpad.net/neutron/+bug/2065566 [ovn-octavia-provider] Return DELETED on member_delete for non-existing members. In progress. Assignee Fernando Royo/ Proposed fix: https://review.opendev.org/c/openstack/ovn-octavia-provider/+/919482 Medium bugs ------------------ * https://bugs.launchpad.net/neutron/+bug/2065145 AttributeError: 'MetadataDriver' object has no attribute 'monitors' in L3 agent logs, Assignee Brian Haley. Proposed fix: https://review.opendev.org/c/openstack/neutron/+/918674 Low bugs ------------- None Wishlist bugs ------------------ https://bugs.launchpad.net/neutron/+bug/2065198. Allow ml2 MechanismDrivers to start own rpc listeners. Gathering more input from submitter at the moment

1 0

Nova filters
by Gk Gk 13 May '24

13 May '24

Hi, In the Antelope version nova scheduler, is there an option for filtering hypervisors based on available disk space ? I cant find the option "DiskFilter' in the latest documentation for antelope. How to achieve this ? Thanks G.K

2 4

[DNS] RFH: Something is answering DNS PTR queries with UDP and not EDNS
by Philipp Hahn 13 May '24

13 May '24

Hello, I'm new to this list and relatively new to OpenStack, so please be patient with me. Our infrastructure provider is hosting OpenStack for us and I observed the following irritating behavior: If my Linux VM does a reverse-IP-lookup (PTR) something™ seems to intercept those PTR queries on their way from the VM to the *configured* DNS server and answers them itself, returning the name of the VM inside some pre-configured domain: > # dig +nocmd +nocomments +nostats +noedns -x 10.201.0.211 > ;211.0.201.10.in-addr.arpa. IN PTR > 211.0.201.10.in-addr.arpa. 3600 IN PTR jenkins-ucs-5-0-ucs-5-0-7-autotestjoinopenstack-sambaversion-no.a.customers.regiocloud.tech. > > # getent hosts 10.201.0.211 > 10.201.0.211 jenkins-ucs-5-0-ucs-5-0-7-autotestjoinopenstack-sambaversion-no.a.customers.regiocloud.tech > > # python3 -c 'import DNS;req=DNS.DnsRequest("211.0.201.10.in-addr.arpa.",qtype="PTR",server=["10.201.0.211"], protocol="udp");res=req.req();print(res.answers[0]["data"])' > jenkins-ucs-5-0-ucs-5-0-7-autotestjoinopenstack-sambaversion-no.a.customers.regiocloud.tech > > # python3 -c 'import dns.resolver;r=dns.resolver.Resolver(configure=False);r.nameservers=["10.201.0.211"];r.edns=False;a=r.query("211.0.201.10.in-addr.arpa.","PTR",tcp=False);print(list(a))' > [<DNS IN PTR rdata: jenkins-ucs-5-0-ucs-5-0-7-autotestjoinopenstack-sambaversion-no.a.customers.regiocloud.tech.>] This only happens when UDP is used an EDNS is disabled: As soon as I use TCP or force EDNS my queries arrives at the configured DNS server and I get back the expected answer: > # dig +nocmd +nocomments +nostats +edns -x 10.201.0.211 > ;211.0.201.10.in-addr.arpa. IN PTR > 211.0.201.10.in-addr.arpa. 172800 IN PTR master096.autotest096.test. > 201.10.in-addr.arpa. 10800 IN NS master096.autotest096.test. > master096.autotest096.test. 80600 IN A 10.201.0.211 > > # dig +nocmd +nocomments +nostats +edns +tcp -p 53 -x 10.201.0.211 > ;211.0.201.10.in-addr.arpa. IN PTR > 211.0.201.10.in-addr.arpa. 172800 IN PTR master096.autotest096.test. > 201.10.in-addr.arpa. 10800 IN NS master096.autotest096.test. > master096.autotest096.test. 80600 IN A 10.201.0.211 > > # dig +nocmd +nocomments +nostats +noedns +tcp -p 53 -x 10.201.0.211 > ;211.0.201.10.in-addr.arpa. IN PTR > 211.0.201.10.in-addr.arpa. 172800 IN PTR master096.autotest096.test. > 201.10.in-addr.arpa. 10800 IN NS master096.autotest096.test. > master096.autotest096.test. 80600 IN A 10.201.0.211 > > # python3 -c 'import DNS;req=DNS.DnsRequest("211.0.201.10.in-addr.arpa.",qtype="PTR",server=["10.201.0.211"], protocol="tcp");res=req.req();print(res.answers[0]["data"])' > master096.autotest096.test > > # python3 -c 'import dns.resolver;r=dns.resolver.Resolver(configure=False);r.nameservers=["10.201.0.211"];r.edns=False;a=r.query("211.0.201.10.in-addr.arpa.","PTR",tcp=True);print(list(a))' > [<DNS IN PTR rdata: master096.autotest096.test.>] > > # python3 -c 'import dns.resolver;r=dns.resolver.Resolver(configure=False);r.nameservers=["10.201.0.211"];r.edns=True;a=r.query("211.0.201.10.in-addr.arpa.","PTR",tcp=True);print(list(a))' > [<DNS IN PTR rdata: master096.autotest096.test.>] > > # python3 -c 'import dns.resolver;r=dns.resolver.Resolver(configure=False);r.nameservers=["10.201.0.211"];r.edns=True;a=r.query("211.0.201.10.in-addr.arpa.","PTR",tcp=False);print(list(a))' > [<DNS IN PTR rdata: master096.autotest096.test.>] I already talked to out provider and spend multiple hours searching the internet or even having a look at the source code of Neutron myself, but did not find any clue so far. Can someone please enlighten me and tell me, how I (or our provider) can disable that "UDP DNS intercept behavior" as that breaks our current use-case, especially as I seldomly can configure our applications to (not) use UDP / EDNS, so depending on which resolver library they use they get different answers to the same query. Thank you in advance. Philipp Hahn -- Philipp Hahn Open Source Software Engineer Univention GmbH Mary-Somerville-Str. 1 28359 Bremen Germany | Deutschland Phone: +49 (0)421 22232-0 | E-Mail: info(a)univention.de https://www.univention.de | https://www.univention.com Managing Directors: Peter H. Ganten, Stefan Gohmann Local court: Amtsgericht Bremen HRB 20755 | Ust-ID: DE220051310 If you are not the intended recipient of this mail, please contact the sender and delete this message. Any unauthorized copying of this message or unauthorized distribution of the information contained herein is prohibited.Information on the processing of your personal data can be found here: https://www.univention.com/privacy-statement

2 1

Re: CAD applications on OpenStack
by Francesco Di Nucci 13 May '24

13 May '24

You're welcome have you also thought about having a chat with CAD vendor's team? Maybe they encountered it as a problem, as it's more about the CAD and less about having Windows VMs on OpenStack By the way, when a conversation is in a list, reply to all to include the list ;) Best regards Francesco Di Nucci On 13/05/24 12:20, Arne Wiebalck wrote: > Thanks, Francesco! > > Licenses are a concern and we're looking at that. > > What we're not really clear on if this approach works latency- / concurrency wise. > We will probably simply try :) > > Thanks again! > Arne > > -- > Dr. Arne Wiebalck > Group Leader, Compute and Devices Group > CERN IT > > ________________________________________ > From: Francesco Di Nucci <francesco.dinucci(a)na.infn.it> > Sent: Thursday, 9 May 2024 08:12 > To: Arne Wiebalck > Subject: Re: CAD applications on OpenStack > > Hi, > > former CAD user here - first of all I'd advice to clear with your legal > department if your license allows you to use the CAD in VMs. > > AFAIK CATIA runs on Windows, so containerization is not possible unless > you manage to get the CAD running in Wine, and you should configure PCI > passthrough > > (https://docs.openstack.org/nova/2024.1/admin/pci-passthrough.html) to > use a graphic card in the VMs. > > Regards > > Francesco Di Nucci > > On 05/05/24 12:04, Arne Wiebalck wrote: >> Dear all, >> >> We're looking into integrating the use case of CAD applications, such as CATIA V5/6 or CATIA 3DEXPERIENCE, into our OpenStack deployment, with the goals of increased resource sharing >> and to reduce the park of individual workstations we have for this at the moment. >> >> If anyone has experience running such applications on top of OpenStack (on individual VMs >> with GPU access, on bare metal servers with concurrent sessions, maybe even containerised ...) >> and would like to share what they have done, it'd be great if you got in touch with me. >> >> Thanks! >> Arne >> >> -- >> Arne Wiebalck >> CERN IT >> >> > >

1 0

Re: Nova filters
by Gk Gk 13 May '24

13 May '24

Hi, In the Antelope version nova scheduler, is there an option for filtering hypervisors based on available disk space ? I cant find the option "DiskFilter' in the latest documentation for antelope. How to achieve this ? Thanks G.K

1 0

Rabbit mq connections keep on going up, why channels and connections not closing?
by Shrey Chauhan 13 May '24

13 May '24

Hi all, I am observing a behaviour which I am not sure is right or wrong, or something I should be concerned about? Currently I am using openstack xena with rabbit-mq message broker. rabbit mq version is 3.8.3 amqp version 5.0.8 I have installed rabbit management plugin on my cluster, and I observe whenever I perform any task like a resize. My connection and channel counts go up, and with every operation it goes up. I am worried that at some point I will hit some kind of connection threshold. I just want to understand if services like nova, cinder reuse an open rabbit mq connection or reopen for every operation, and if they open msg queue connection for every operation do they close the connection once the action is done Am I missing configuration in my setup of rabbit mq with openstack? Please help me understand this scenario. Thanks for help in advance. Thanks, Shrey Chauhan

1 0

[ops] Inquiry Regarding Addition of DH Groups in OpenStack VPNaaS
by Shubham Yadav 13 May '24

13 May '24

Hello , I hope this message finds you well. I am reaching out to inquire about the possibility of adding additional DH (Diffie-Hellman) groups to the VPNaaS (VPN as a Service) project in OpenStack. Currently, our setup involves using DH groups 14, 16, 20, and 21 on Fortigate routers. However, the available DH groups at the OpenStack Web UI for creating IPSEC policies are limited to groups 2, 5, and 14. Could you please provide insights into whether there are plans to introduce more DH groups into the VPNaaS project? This addition would greatly benefit our configurations and improve compatibility with our existing infrastructure. Thank you in advance for any information or guidance you can provide on this matter. Best regards, Shubham

2 1

[oslo][release] Changing back release model of tooz ?
by Takashi Kajinami 13 May '24

13 May '24

Hello, As you may know, tooz was moved to the independent release model[1] after wallaby release because of its stable code base at that time. However, we introduced several features/improvements in tooz during a few cycles and I'm wondering if we can revisit its status now. One actual problem I'm seeing now is that, we recently got a bug report[2] about the redis sentinel support which we added during 2024.1 cycle but we can't backport the required bug fix to 2024.1 because of no stable branch. Given the situation, I'm wondering if it makes sense (and is possible) to move tooz back to the cycle-with-intermediary model, and create 2024.1 branch from 6.1.0 (the last release during 2024.1 cycle) now, so that we can maintain 2024.1 releases with future bug fixes brackported. [1] https://review.opendev.org/c/openstack/releases/+/788230 [2] https://bugs.launchpad.net/python-tooz/+bug/2063857 -- Takashi Kajinami irc: tkajinam github: https://github.com/kajinamit launchpad: https://launchpad.net/~kajinamit

4 5

[tc][all] OpenStack Technical Committee Weekly Summary and Meeting Agenda (2024.2/R-20)
by Goutham Pacha Ravi 12 May '24

12 May '24

Hello Stackers, We're officially in milestone-1 week for the 2024.2 (Dalmatian) release cycle [1]. Teams are actively planning design specification freezes, hackathons, and bug squashes. I'm delighted to announce that three Outreachy interns have been funded to work with the OpenStack community [2]. Please join me in welcoming Mohd Ariq (tspyderboy), Okeke Christian Chima (cokeke), and Abby Nduta (AbbyNduta). Mohd Ariq and Christian will be working with members of the OpenStack Manila project team, while Abby will be working with members of the OpenStack Horizon project team. Congratulations, interns! These internships begins at the end of May and will conclude at the end of August. There have been a few governance changes in the past week that I'd like to brief you about: The "Consistent and Secure Default RBAC" goal has new milestone requirements. All projects are expected to transition oslo.policy's "enforce_scope" and "enforce_new_defaults" options to True during the 2024.2 release cycle. These options will be removed from oslo.policy during the 2025.1 release cycle. [3] openstack/kuryr and openstack/kuryr-libnetwork are now under the maintainership of the openstack/zun ("Container as a service") team. The TC has adopted a house rule to mark projects that have identified no leadership as "inactive" projects [4]. We've merged slaweq's summary of the 2023 OpenStack User Survey [5]. Thanks, slaweq, and to all of you who took the survey. === Weekly Meeting === Last week's TC meeting was held simultaneously on Zoom (a recording is on YouTube [6]) and IRC on May 7th, 2024. The meeting was logged to eavesdrop [7]. Over the years, we've come to realize that monitoring project activity through objective tools wasn't representing the picture accurately. Several mature/feature-complete projects will seemingly be "inactive," and what we're interested in assessing is whether there are active maintainers tackling security and project infrastructure concerns. Artem Goncharov (gtema) will write up a proposal to tag inactive projects more prominently. We discussed a proposal to change the Distributed Project Leadership model to require DPL liaisons to re-assert their interest at the beginning of each cycle. The proposal also suggests appointing liaisons from the TC to monitor projects that have this leadership model. TC Members would be welcome to nominate themselves as liaisons for projects following DPL if this proposal is accepted. We took some AIs to continue the cleanup of maintainers on OpenStack's packages on PyPI - an ongoing community-wide effort [8]. We discussed the open goal proposal to replace eventlet. The Technical Committee wants to ensure that we're not advocating for a one-size-fits-all solution to the problem of threading in OpenStack projects. Hence, we'd prefer not to merge the proposal to replace eventlet with asyncio. Instead, we want to encourage projects to independently explore solutions that make sense to their individual architecture - which could be asyncio, native threads, or any other threading paradigm/library. We are supportive of ensuring there's no imposition of any specific library or threading mechanism imposed via oslo libraries. JayF will be summarizing this view to this mailing list soon. The meeting concluded with an open discussion on DevStack setup time and an outdated (but relevant) proposal related to OSC as a service on DevStack. Tomorrow's meeting is back on OFTC/IRC, and the call details and agenda are at [9]. We look forward to seeing you there. The TC is always honored to chat with you, so please join us, and feel free to add anything to the agenda that you'd like to chat with us about! === Goal Proposals === As stated above, this goal proposal needs your weigh-in: - Migrate eventlet usages to asyncio (https://review.opendev.org/c/openstack/governance/+/902585) === Governance Proposals === ==== Merged ==== - Remove ref of bylaw Appendix 4 | https://review.opendev.org/c/openstack/governance/+/916942 - Add timeline to remove enforce_scope in RBAC goal | https://review.opendev.org/c/openstack/governance/+/915179 - Move openstack/kuryr and openstack/kuryr-libnetwork to Zun | https://review.opendev.org/c/openstack/governance/+/917517 - 2023 User Survey TC Question Analysis | https://review.opendev.org/c/openstack/governance/+/900516 - Add option to move leaderless project to Inactive status | https://review.opendev.org/c/openstack/governance/+/914726 ==== Open for review/discussion ==== - Update to include docs and miscellaneuos repos for AC status | https://review.opendev.org/c/openstack/governance/+/915021 - Add DPL model & liaison reset policy | https://review.opendev.org/c/openstack/governance/+/916833 Thank you very much for reading! On behalf of the OpenStack TC, Goutham Pacha Ravi (gouthamr) OpenStack TC Chair [1] https://releases.openstack.org/dalmatian/schedule.html [2] https://www.outreachy.org/alums/2024-05/#:~:text=Openstack [3] https://governance.openstack.org/tc/goals/selected/consistent-and-secure-rb… [4] https://governance.openstack.org/tc/reference/emerging-technology-and-inact… [5] https://governance.openstack.org/tc/user_survey/analysis-2023.html [6] https://youtu.be/-WZT7PC3sig [7] https://meetings.opendev.org/meetings/tc/2024/tc.2024-05-07-18.00.log.html [8] https://etherpad.opendev.org/p/openstack-pypi-maintainers-cleanup [9] https://wiki.openstack.org/wiki/Meetings/TechnicalCommittee#Next_Meeting

1 0

openstack boot an instance from an IDE volume migrated from VMware
by luckydog xf 10 May '24

10 May '24

hi, list, I'm going to boot an instance from an existing volume, and it is a Windows 2016 migrated from VMware. So the bus must be set to 'ide' mode, otherwise it fails to boot up. Let's assume this volume UUID is xyz and it's bootable. I cloned it by clonezilla from VMware ESXI. since virtio drive wasn't installed in Windows, so it must be set to ide. So what's the right way to boot it from this volume and set it to IDE? I have tried lots of ways, like --- openstack server create test-windows --flavor 4 --nic net-id=7b57ee99-136f-4aa7-b3fe-a4958ec8c644 --block-device id=xxxx:disk_bus=ide and ... --block-device-mapping. But unfortunately, none of them work. And there are too many outdated docs like disk_bus=ide and bus=ide. Not sure which one works. This is fairly frustrating. I'm still struggling with this issue and need your help. Thanks a lot.

3 7