*** ralonsoh_out is now known as ralonsoh\ | 05:46 | |
*** ralonsoh\ is now known as ralonsoh | 05:47 | |
opendevreview | Rodolfo Alonso proposed openstack/neutron-tempest-plugin master: Bump advance image to Ubuntu Jammy Server 22.04 https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/949140 | 05:57 |
---|---|---|
opendevreview | Rodolfo Alonso proposed openstack/neutron-tempest-plugin master: WIP -- Bump advance image to Ubuntu Noble 24.04 https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/949522 | 06:00 |
opendevreview | Rodolfo Alonso proposed openstack/neutron stable/2025.1: [OVN] Create a HA_Chassis_Group without raising an exception https://review.opendev.org/c/openstack/neutron/+/949699 | 06:05 |
haleyb | ralonsoh: hey rodolfo, i wanted to ask your opinion on remote address group SG rules. would you expect the OVN driver to work without this fix? https://review.opendev.org/c/openstack/neutron/+/851509 | 08:54 |
haleyb | i actually don't see any tempest tests excercising it except for the openvswitch firewall driver either | 08:55 |
haleyb | i'm in an EMEA timezone this week so it's not the middle of the night :) | 08:57 |
opendevreview | Brian Haley proposed openstack/neutron-tempest-plugin master: DNM: Test remote address groups with OVN https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/949713 | 09:05 |
opendevreview | Brian Haley proposed openstack/neutron master: DNM: Test remote address groups with OVN https://review.opendev.org/c/openstack/neutron/+/949715 | 09:05 |
opendevreview | Merged openstack/neutron master: [OVN] Create a HA_Chassis_Group without raising an exception https://review.opendev.org/c/openstack/neutron/+/945837 | 09:17 |
*** ralonsoh_ is now known as ralonsoh | 09:21 | |
ralonsoh | haleyb, hi! I was looking at this today because of https://bugs.launchpad.net/neutron/+bug/2110306 | 09:34 |
haleyb | o/ | 09:34 |
ralonsoh | but my testing env is not correctly pinging to the FIPs | 09:34 |
ralonsoh | so I can't test it | 09:34 |
ralonsoh | I need to figure out what is happening with my env | 09:35 |
ralonsoh | haleyb, do you happen to know how to configure a devstack all in one node with external connectivity? | 09:35 |
ralonsoh | because I can ping from the VM to the external IP, but not in the other way around | 09:36 |
ralonsoh | btw, I have port without port security | 09:36 |
haleyb | ralonsoh: it's been a while since i've done that, i think i remember adding a route on the external system but can't recall | 09:37 |
ralonsoh | I don't really need to have external connectivity to internet | 09:37 |
ralonsoh | but at least connectivity from a local IP (in the host) and a VM | 09:37 |
ralonsoh | I can ping VM -> ext IP, but not in the other way | 09:38 |
ykarel | ralonsoh, you deployed default devstack settings? and ping to FIP network 172.24.5.X is not working? | 09:38 |
ralonsoh | ykarel, sorry, the other way | 09:38 |
ralonsoh | hmmmm | 09:38 |
ralonsoh | I can ping from local IP --> FIP | 09:38 |
ralonsoh | but from inside the VM, I can't ping to the local IP in the host | 09:39 |
haleyb | i'm guessing it doesn't have a return route? | 09:39 |
ralonsoh | what I trying is https://docs.openstack.org/neutron/latest/admin/config-subnet-onboard.html | 09:40 |
ralonsoh | no sorry | 09:40 |
ralonsoh | arggggg (sorry, I didn't sleep a minute today) | 09:40 |
ralonsoh | this: https://docs.openstack.org/neutron/latest/admin/config-address-scopes.html | 09:40 |
ralonsoh | so, in the host, I'm adding a route to the internal IP | 09:41 |
ralonsoh | FIP: 10.50.0.27, int IP 10.100.0.71 | 09:41 |
ralonsoh | so I'm setting this | 09:41 |
ralonsoh | ip r add 10.100.0.64/26 via 10.50.0.10 | 09:41 |
ralonsoh | (10.50.0.10 is the host IP address, that is in the external subnet range) | 09:42 |
ralonsoh | and then I'm trying to ping to the internal IP address | 09:42 |
ralonsoh | # ping 10.100.0.71 | 09:42 |
ralonsoh | so, at this point, I can ensure address scope is working | 09:43 |
haleyb | right, i'm not sure i've been able to ping the internal IP in a devstack, just the FIP. except when using the ovnmeta namespace that is | 09:46 |
ykarel | me too only accessed those via ovnmeta namespaces | 09:52 |
ralonsoh | so I think this kind of routing was possible with the L3 agent, but not with OVN | 09:56 |
ralonsoh | maybe some nat rule is missing, but by default we don't allow this | 09:56 |
ralonsoh | and makes sense: we use FIPs to hide the internal IP to the external network | 09:57 |
ykarel | yes it seems true | 10:02 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: Initialize the network segment ranges only in first WSGI worker https://review.opendev.org/c/openstack/neutron/+/948200 | 10:56 |
*** QG1 is now known as QG | 11:09 | |
ralonsoh | haleyb, slaweq hello! if you have 1 min: https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/926508 | 11:25 |
ralonsoh | finally all neutron patches are merged now | 11:25 |
ralonsoh | --> https://review.opendev.org/q/I3728aa3932931320d9ccbc24fb9f8b3533dffff7_ | 11:26 |
ralonsoh | https://review.opendev.org/q/I3728aa3932931320d9ccbc24fb9f8b3533dffff7 | 11:26 |
opendevreview | Brian Haley proposed openstack/neutron-tempest-plugin master: DNM: Test remote address groups with OVN https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/949713 | 11:26 |
opendevreview | Merged openstack/neutron-tempest-plugin master: Pin "designate-tempest-plugin" to zed-last in Zed branch https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/949643 | 12:06 |
opendevreview | Lajos Katona proposed openstack/os-vif master: OVS Trunk: Add bridge_name to external_ids https://review.opendev.org/c/openstack/os-vif/+/949736 | 12:43 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: Initialize the network segment ranges only in first WSGI worker https://review.opendev.org/c/openstack/neutron/+/948200 | 12:49 |
opendevreview | Rodolfo Alonso proposed openstack/neutron-tempest-plugin master: Bump advance image to Ubuntu Jammy Server 22.04 https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/949140 | 12:57 |
stephenfin | Could some who knows something about the neutron-dynamic-routing plugin explain the purpose of the 'POST /v2.0/agents/{bgp-agent-id}/bgp-drinstances' and 'DELETE /v2.0/agents/{bgp-agent-id}/bgp-drinstances/{bgp-speaker-id}' APIs? | 13:00 |
stephenfin | Context is https://github.com/gophercloud/gophercloud/issues/3380 | 13:00 |
stephenfin | We're seeing frequent failures in a Gophercloud test. We create a BGP speaker, ensure that it is attached to an agent, and then attempt to delete that association via the DELETE API above | 13:01 |
stephenfin | I think something in the backend is _reassigning_ it but I can't find logs to prove that | 13:01 |
stephenfin | So I wonder if we're supposed to be calling that API, or if it's only for situations where e.g. you have multiple agents and speakers can be attached to only some of them (but it must be attached to one)? | 13:02 |
frickler | stephenfin: I think I can answer that after the kolla meeting | 13:04 |
frickler | stephenfin: so essentially it depends on which scheduler you have configured https://docs.openstack.org/neutron-dynamic-routing/latest/admin/agent-scheduler.html#scheduler | 13:16 |
frickler | the "modern" suggestion is to use the static scheduler and do all associations manually | 13:16 |
frickler | (now let me read the gh link and find out the real question is something completely different ;) | 13:17 |
stephenfin | Oh, interesting. So it sounds like we are holding it wrong (TM). I assume the agent in use is not discoverable? | 13:18 |
stephenfin | (via the API) | 13:18 |
frickler | I don't think it is. but it looks like you are using devstack to deploy it? that should usually do the right thing | 13:21 |
frickler | but I also don't rule out a bug. though I'd assume that n-d-r should have proper test coverage | 13:22 |
stephenfin | neutron-dynamic-routing defaults to the chance scheduler https://opendev.org/openstack/neutron-dynamic-routing/src/branch/master/neutron_dynamic_routing/db/bgp_dragentscheduler_db.py#L38-L44 | 13:22 |
stephenfin | the devstack plugin allows overriding that with 'BGP_SCHEDULER_DRIVER' https://opendev.org/openstack/neutron-dynamic-routing/src/branch/master/devstack/lib/dr#L33 | 13:23 |
stephenfin | and that's being overridden in a job to test it with the static scheduler https://opendev.org/openstack/neutron-dynamic-routing/src/branch/master/.zuul.yaml#L6 | 13:24 |
stephenfin | so I guess we can start setting 'BGP_SCHEDULER_DRIVER' in our CI jobs, but it'll fail for people running it locally unless they've done the same | 13:25 |
stephenfin | assuming, yes, there are no other bugs | 13:25 |
stephenfin | :) | 13:25 |
frickler | but yes, I think that would explain the issue you are seeing. and we didn't want to change the default to avoid breaking existing deployments, maybe we should revise that decision | 13:26 |
frickler | tobias-urdin: ^^ not sure how much you are still interested in n-d-r, but do you have an opinion on this? | 13:27 |
frickler | first step could be just switch the default in the devstack plugin, not in the code directly | 13:28 |
tobias-urdin | frickler: i don't have any strong opinion, i'm fine changing the default to static in either or both devstack+n-d-r as i think static makes more sense than chance by default, if we write a good release note i dont see a problem changing the default | 13:37 |
opendevreview | Lajos Katona proposed openstack/os-vif master: OVS Trunk: Add bridge_name to external_ids https://review.opendev.org/c/openstack/os-vif/+/949736 | 13:38 |
opendevreview | Stephen Finucane proposed openstack/neutron-lib master: api-ref: Add warning about DRA scheduler https://review.opendev.org/c/openstack/neutron-lib/+/949744 | 13:41 |
opendevreview | Stephen Finucane proposed openstack/neutron-lib master: api-ref: Remove crud from conf.py https://review.opendev.org/c/openstack/neutron-lib/+/949745 | 13:41 |
stephenfin | frickler: tobias-urdin: Can you sanity check ^ for me? | 13:41 |
stephenfin | (the first one, naturally) | 13:41 |
opendevreview | Bence Romsics proposed openstack/neutron master: Limit trunk ACTIVE state hack to OVN https://review.opendev.org/c/openstack/neutron/+/949217 | 14:16 |
opendevreview | Brian Haley proposed openstack/neutron-tempest-plugin master: DNM: Test remote address groups with OVN https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/949713 | 15:10 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: WIP == DNM == Add provider network resources filter for "network" https://review.opendev.org/c/openstack/neutron/+/949771 | 15:11 |
opendevreview | Stephen Finucane proposed openstack/neutron-lib master: api-ref: Add warning about DRA scheduler https://review.opendev.org/c/openstack/neutron-lib/+/949744 | 15:46 |
opendevreview | Stephen Finucane proposed openstack/neutron-lib master: api-ref: Remove crud from conf.py https://review.opendev.org/c/openstack/neutron-lib/+/949745 | 15:46 |
opendevreview | Stephen Finucane proposed openstack/neutron-lib master: zuul: Don't run neutron-functional for api-ref changes https://review.opendev.org/c/openstack/neutron-lib/+/949775 | 15:46 |
opendevreview | Merged openstack/neutron-tempest-plugin master: Unset "router:external" in subnet tests https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/926508 | 15:46 |
opendevreview | Merged openstack/neutron unmaintained/zed: [stable-only] Remove "devstack-tobiko-neutron" job https://review.opendev.org/c/openstack/neutron/+/949559 | 15:46 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!