April 2024 - openstack-discuss - lists.openstack.org

[all] Package reproducibility: shall we hard-wire our CI to avoid it in the future?
by Thomas Goirand 09 Apr '24

09 Apr '24

Hi, It used to be that a number of packages were not reproducible. If you do not know why reproducibility of packages is important, please read on: https://reproducible-builds.org/ This has been patched numerous times in numerous packages. However, it looks like the community is still not aware of the problem, and how to avoid it. One of the very common pitfall, is adding a default in an oslo.config string, that leaks properties of the build host, like for example the hostname, the number of CPU cores, and so on. Here's a very common example that was commited last summer: https://review.opendev.org/c/openstack/oslo.messaging/+/889425/9/oslo_messa… Yet the fix is super simple, thanks to oslo.messaging sample_default: https://review.opendev.org/c/openstack/oslo.messaging/+/915314 If I'm writing this, it is *not* to point at anyone, especially *not* the commiter of the above patch. it just happen to be the latest occurrence I noticed thanks to the Debian bug report: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1068377 This message is only to raise awareness of the problem, and avoid having it happen again. I also wonder: would it be possible to have our CI avoid this type of mistake? It wouldn't be so hard to have the doc build twice, with different hostname, CPUs, etc. configuration, would it? (note: I'm not volunteering, just pushing for the idea...) Cheers, Thomas Goirand (zigo)

2 1

[openstack-helm] PTG schedule
by Vladimir Kozhukalov 09 Apr '24

09 Apr '24

Hey team, This is just a reminder that we booked two hours on Wednesday 04/10/2024 at 13:00 UTC and 14:00 UTC in the Cactus meeting room. Etherpad is here [1]. Feel free to add your topics for the discussion. Booking table is here [2]. [1] https://etherpad.opendev.org/p/apr2024-ptg-os-helm [2] https://ptg.opendev.org/ptg.html -- Best regards, Kozhukalov Vladimir

1 0

OpenStack OVN add secondary external network
by damage.alders-0j＠icloud.com 09 Apr '24

09 Apr '24

I've setup OpenStack version 2023.2 with with devstack on a Ubuntu 22.04 baremetal. I originally had one NIC (eth1, 192.XX.XX.XX) that works fine. Now I added a second NIC (eth2, 10.XX.XX.XX) to my machine and tried to add a secondary external network through: sudo ovs-vsctl add-br br-ex2 sudo ovs-vsctl add-port br-ex2 eth2 sudo ovs-vsctl set open . external-ids:ovn-bridge-mappings=public2:br-ex2 sudo ovs-vsctl set open . external-ids:ovn-cms-options="enable-chassis-as-gw" then my whole machine (including host and all instances) lost dns service, I can only ping ip addresses such as 8.8.8.8 instead of google.com. To be precise, I found out I lose dns service after `sudo ovs-vsctl add-port br-ex2 eth2` command. (Yes, I've moved ip from eth2 to br-ex2, but didn't help) I kept on creating a network and a router: openstack network create public2 --provider-network-type flat --provider-physical-network public2 --external openstack subnet create public2-subnet --network public2 --subnet-range 10.XX.XX.0/24 --no-dhcp --gateway 10.XX.XX.1 --allocation-pool start=10.XX.XX.10,end=10.XX.XX.30 openstack router create bridge-router --external-gateway public2 openstack router add subnet bridge-router shared-public2-subnet At this point, the router is accessible from eth2 after `sudo ovs-vsctl set open . external-ids:ovn-bridge-mappings=public2:br-ex2` command, but after I tried to allocate a floating ip from public2 to any instance, the instance is just not accessible from that ip. My ultimate goal is to be able to allocate floating ip to instances from eth2, so I can access instances from both eth1 and eth2 networks, but now I lost dns service and I'm unable to access instances from the allocated eth2 floating ip. Any suggestions?

1 0

[kolla] Dalmatian vPTG reminder
by Michał Nasiadka 08 Apr '24

08 Apr '24

Hello Koalas, This year the PTG slots are following: Kolla/Kolla-Ansible: Monday, 8th April, 13:00 - 17:00 UTC Kayobe: Tuesday, 9th April, 13:00 - 15:00 UTC You can check the slots always on the PTG website - https://ptg.opendev.org/ Etherpad link: https://etherpad.opendev.org/p/kolla-dalmatian-ptg See you on Monday!

2 1

[neutron] CI Meeting cancelled this week
by Yatin Karel 08 Apr '24

08 Apr '24

Hello Neutron Team !! As discussed over the previous meeting cancelling this week CI meeting for vPTG. -- Thanks and Regards Yatin Karel

1 0

Nova filter based on VM metadata?
by Ken D'Ambrosio 08 Apr '24

08 Apr '24

Hey, all. I'm looking for a way to to filter to a particular aggregate based on arbitrary VM metadata. (E.g., maybe FACTER facts from Puppet that we pass to the VM as metadata.) Could someone point me toward documentation that might outline how to accomplish that? As I see it, there's really one thing I need to understand -- the rest I can probably glean from extant filters: how (or if) a filter could access a given VM's metadata from the API creation request (if that's even possible). Thanks! -Ken

3 3

[nova][ptg] Nova at the Dalmatian vPTG
by Sylvain Bauza 08 Apr '24

08 Apr '24

Hey folks, During this week, the Nova community will discuss on their own topics in between Tuesday-Friday 1300-1700UTC, all of the slots being in the Diablo room. You can find the related connection details and the proposed schedule in the etherpad [1] but here's a rough schedule for people interested in attending our sessions : - *Tuesday* *13:00 UTC - 17:00 UTC* - Nova sessions - retro + dalmatian schedule - nova above k8s - unified limits - *Wednesday 13:00 UTC - 17:00 UTC*: Nova sessions - vmware discussions about scheduling and metadata api - cleaning old microversions ? - ephemeral encryption status and next steps - memory encryption next steps - *Thursday 13:00 UTC - 17:00 UTC* - Nova sessions - stateless firmware support - vTPM detection by libvirt - Gap between legacy deprecated live migration options vs new live migration options - event actions missing in our API - OpenAPI schemas support - *Friday 13:00 UTC - 17:00 UTC* - Nova sessions - SPICE-direct console support + remote-consoles problem - userdata update - volume-backed instance migration requiring SSH conn - can we push removal of tunnelled libvirt migrations in dalmatian or do we need to wait for next slurp? - no validation if queued oslo.messaging messages are targeted for this specific nova-compute (based on hostname) As also said in the etherpad, you're free to add your own topics to the etherpad (at the bottom please) so I can schedule them, but fwiw our sessions will already be pretty packed with a lot of things to discuss so maybe some of the topics will be on a temptative effort. [1] https://etherpad.opendev.org/p/nova-dalmatian-ptg Thanks, -Sylvain

1 1

Fix for arguments for subnet module for Ansible
by George Shuklin 08 Apr '24

08 Apr '24

Hello everyone. I wrote a small fix for subnet Ansible module to allow use of cidr together with subnet_pool, with tests and so on. I got an approval, it passed CI, but it is not merged. What steps are required to have it merged? https://review.opendev.org/c/openstack/ansible-collections-openstack/+/9073… Sorry for nagging.

1 2

[kolla] [Wallaby] Openstack operations 2x slower after rebuilding from RHEL8/Train to RHEL9/Wallaby
by Albert Braden 08 Apr '24

08 Apr '24

After rebuilding a couple of our clusters, moving from Train on RHEL8 to Wallaby on RHEL9, we see all openstack operations taking 2x longer. Even simple database reads such as "openstack user show" average 1.5 seconds on old regions and 3 seconds on rebuilt regions. I enabled debug and poked around in the logs but nothing is jumping out at me. I was thinking maybe mariadb needed to be tuned differently from RHEL8->9 but if I log into the mariadb container and query the database it seems equally fast; I can do 100 simple queries in a second. Where should I be looking for the source of this delay?

5 10

Re:[skyline] Running a Skyline in "subdirectory"
by 吴文相 08 Apr '24

08 Apr '24

Hello, Dmitriy Rabotyagov Sorry for late response, dev team will raise a ticket and update the conclusion later. Thanks Best Regrads Wu Wenxiang Original: From：Dmitriy Rabotyagov <noonedeadpunk(a)gmail.com>Date：2024-03-16 02:45:49(中国 (GMT+08:00))To：openstack-discuss <openstack-discuss(a)lists.openstack.org>Cc：Subject：[skyline] Running a Skyline in "subdirectory"Hey folks, I was looking into a way to make both Horizon and Skyline work at the same time, and found out that I'm not really able to make Skyline to run in sub-URL. So for instance the plan was to place Horizon under /horizon (which is easily doable) and then Skyline under /skyline. However so far I haven't found any setting or anything that would allow me to do so. Basically, I've realized that static content is _really_ static, which means it does have hardcoded URIs: https://opendev.org/openstack/skyline-console/src/branch/master/skyline_con… So either I'm just not realizing the very simple technique which allows to overcome hardcoded in static files URIs (like some rewrite?), or just serving Skyline from a subdirectory is not possible by design... Would be great if someone could share their experience about Skyline and if anyone managed to do what I'm looking for. Thanks in advance

2 1