*** ykarel_ is now known as ykarel | 06:13 | |
opendevreview | Slawek Kaplonski proposed openstack/neutron-tempest-plugin master: Add new scenario test for VIP address added as allowed addr pair https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/794841 | 07:50 |
---|---|---|
opendevreview | Slawek Kaplonski proposed openstack/neutron master: Make configure_for_func_testing compatible with e.g. Centos https://review.opendev.org/c/openstack/neutron/+/799625 | 07:56 |
opendevreview | Lajos Katona proposed openstack/os-ken stable/xena: Avoid logging MD5 password for BGP add neighbor https://review.opendev.org/c/openstack/os-ken/+/824299 | 07:57 |
opendevreview | Lajos Katona proposed openstack/os-ken stable/wallaby: Avoid logging MD5 password for BGP add neighbor https://review.opendev.org/c/openstack/os-ken/+/824300 | 07:58 |
opendevreview | Slawek Kaplonski proposed openstack/neutron master: Cosmetic: Change ._ovn to ._nb_ovn to match ._sb_ovn https://review.opendev.org/c/openstack/neutron/+/718767 | 08:10 |
opendevreview | Slawek Kaplonski proposed openstack/neutron master: Allow to use static Local IP openflow rules https://review.opendev.org/c/openstack/neutron/+/817399 | 08:23 |
opendevreview | Lajos Katona proposed openstack/os-ken stable/victoria: Avoid logging MD5 password for BGP add neighbor https://review.opendev.org/c/openstack/os-ken/+/824301 | 08:29 |
opendevreview | Lajos Katona proposed openstack/os-ken stable/ussuri: Avoid logging MD5 password for BGP add neighbor https://review.opendev.org/c/openstack/os-ken/+/824302 | 08:29 |
opendevreview | Lajos Katona proposed openstack/os-ken stable/train: Avoid logging MD5 password for BGP add neighbor https://review.opendev.org/c/openstack/os-ken/+/824303 | 08:30 |
opendevreview | Rodolfo Alonso proposed openstack/neutron master: Remove "PortBindingMixin" class and related DB table https://review.opendev.org/c/openstack/neutron/+/824009 | 08:42 |
opendevreview | Rodolfo Alonso proposed openstack/neutron stable/xena: [OVN] Allow only one physical network per bridge https://review.opendev.org/c/openstack/neutron/+/824344 | 08:45 |
opendevreview | Rodolfo Alonso proposed openstack/neutron stable/wallaby: [OVN] Allow only one physical network per bridge https://review.opendev.org/c/openstack/neutron/+/824346 | 08:47 |
opendevreview | Rodolfo Alonso proposed openstack/neutron stable/victoria: [OVN] Allow only one physical network per bridge https://review.opendev.org/c/openstack/neutron/+/824347 | 08:48 |
opendevreview | Rodolfo Alonso proposed openstack/neutron stable/ussuri: [OVN] Allow only one physical network per bridge https://review.opendev.org/c/openstack/neutron/+/824348 | 08:48 |
opendevreview | Lajos Katona proposed openstack/neutron master: Remove functions to enable Neutron's segments integration https://review.opendev.org/c/openstack/neutron/+/818218 | 09:05 |
opendevreview | Rodolfo Alonso proposed openstack/networking-ovn stable/train: [OVN] Allow only one physical network per bridge https://review.opendev.org/c/openstack/networking-ovn/+/824359 | 09:06 |
*** elodille1 is now known as elodilles | 09:29 | |
opendevreview | Oleg Bondarev proposed openstack/neutron master: Local IP: skip ports with invalid ofport https://review.opendev.org/c/openstack/neutron/+/824363 | 09:42 |
opendevreview | Merged openstack/neutron master: Fix placement allocation update for port with network QoS policy https://review.opendev.org/c/openstack/neutron/+/815421 | 09:42 |
EugenMayer | how is OVS networking dealing with nat-reflection right now? is https://github.com/openstack/nova/commit/b61e1ea12cd41ea507b1f6496ec1413c93bd679b even applicable for OVS at all? | 09:52 |
opendevreview | Lajos Katona proposed openstack/neutron master: BFD for Neutron, server side https://review.opendev.org/c/openstack/neutron/+/808046 | 10:00 |
opendevreview | Lajos Katona proposed openstack/networking-odl master: Fix some requirements https://review.opendev.org/c/openstack/networking-odl/+/823428 | 10:08 |
ralonsoh | EugenMayer, you have FIPs if you need to access external services from a internal IP | 10:09 |
ralonsoh | and this is available in LB, OVS and OVN | 10:10 |
opendevreview | Merged openstack/neutron stable/xena: [OVN] Add reverse DNS records https://review.opendev.org/c/openstack/neutron/+/823594 | 10:38 |
opendevreview | Merged openstack/neutron stable/wallaby: [OVN] Add reverse DNS records https://review.opendev.org/c/openstack/neutron/+/823595 | 10:38 |
opendevreview | Merged openstack/neutron stable/victoria: [OVN] Add reverse DNS records https://review.opendev.org/c/openstack/neutron/+/823596 | 10:39 |
opendevreview | Merged openstack/neutron stable/ussuri: [OVN] Add reverse DNS records https://review.opendev.org/c/openstack/neutron/+/823631 | 10:39 |
opendevreview | Elvira García Ruiz proposed openstack/networking-ovn stable/train: Always update router static route https://review.opendev.org/c/openstack/networking-ovn/+/824375 | 10:39 |
opendevreview | Merged openstack/neutron master: Disable tracebacks of eventlet.wsgi.server https://review.opendev.org/c/openstack/neutron/+/818391 | 10:39 |
EugenMayer | ralonsoh not sure what you mean, FIPs is short for floating ips right? | 10:47 |
ralonsoh | yes | 10:47 |
opendevreview | Luis Tomas Bolivar proposed openstack/neutron master: Ensure subports status is aligned with parent port https://review.opendev.org/c/openstack/neutron/+/824378 | 10:47 |
EugenMayer | ralonsoh the point here is, AFAICS the service A (client) which is trying to reach service B (target/backend) is using a gateway, which has the same ip as the external ip of service B. the external ip is used with some port forwardings for some few services, like here. So basically , the gw for service a has the same IP as the target external IP | 10:48 |
EugenMayer | this is classic nat-reflection, and it seems not to work right now. Or did i missunderstand you? | 10:49 |
ralonsoh | with FIP you'll expose a different IP per port (or client/server) | 10:49 |
ralonsoh | this is not nating through a single GW IP | 10:50 |
ralonsoh | you are exposing different IPs per service/port | 10:50 |
EugenMayer | yes, but this is NATing, since i use port forwarding in this case | 10:50 |
ralonsoh | port forwarding uses FIP | 10:51 |
ralonsoh | https://docs.openstack.org/neutron/rocky/admin/config-fip-port-forwardings.html | 10:51 |
EugenMayer | well then it is simply not working | 10:52 |
ralonsoh | port forwarding is working fine | 10:52 |
EugenMayer | i just ensure, actually the GW for service A is not the same as the external IP for the service B | 10:52 |
ralonsoh | what I don't know is the configuration you have or how are you using it | 10:52 |
EugenMayer | yes port forwarding is working fine, also here. The point is the reflection is not | 10:52 |
EugenMayer | GW IP: 1.1.1.1 | 10:53 |
EugenMayer | ServiceB IP: 2.2.2.2 on port 25 (fips port forwarding) | 10:53 |
EugenMayer | now a VM from the internal network, tries to reach 2.2.2.2:25 and fails. The reason is nat reflection. Anything else, not inside the openstack stack, can reach 2.2.2.2:25 - does this make more sense to you? | 10:54 |
EugenMayer | sorry, might be a confusion here, so i try to remove that :) | 10:54 |
ralonsoh | that should work: an internal VM using the GW should have access to service B port | 10:55 |
opendevreview | Bence Romsics proposed openstack/neutron master: Make the dead vlan actually dead https://review.opendev.org/c/openstack/neutron/+/820897 | 11:00 |
opendevreview | Bence Romsics proposed openstack/neutron master: DEBUG router ports in dead vlan https://review.opendev.org/c/openstack/neutron/+/824387 | 11:00 |
gibi | ralonsoh, slaweq: could one of you +A https://review.opendev.org/c/openstack/neutron/+/811746 ? Every dependency has been landed for it | 11:07 |
slaweq | gibi: done | 11:08 |
gibi | slaweq: thanks! | 11:08 |
gibi | and happy new year :) | 11:08 |
opendevreview | Elvira García Ruiz proposed openstack/networking-ovn stable/train: Always update router static route https://review.opendev.org/c/openstack/networking-ovn/+/824375 | 11:08 |
opendevreview | Bence Romsics proposed openstack/neutron master: Make the dead vlan actually dead https://review.opendev.org/c/openstack/neutron/+/820897 | 11:09 |
ralonsoh | gibi, sure | 11:12 |
ralonsoh | ok, I was late... | 11:12 |
gibi | ralonsoh: thanks anyhow | 11:12 |
gibi | and happy new year to you too | 11:12 |
ralonsoh | happy new year! | 11:13 |
opendevreview | Oleg Bondarev proposed openstack/neutron master: Add devstack plugin support for Local IP https://review.opendev.org/c/openstack/neutron/+/818228 | 11:33 |
EugenMayer | ralonsoh pitty that it does not. I will try to debug and get some traceroutes and tcpdumps | 11:41 |
opendevreview | Merged openstack/neutron stable/xena: Add wait event for metadataagent sb_idl https://review.opendev.org/c/openstack/neutron/+/823828 | 11:43 |
opendevreview | Mamatisa Nurmatov proposed openstack/neutron-tempest-plugin master: Add local ip scenario tests https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/823007 | 11:46 |
opendevreview | Merged openstack/neutron stable/wallaby: Add wait event for metadataagent sb_idl https://review.opendev.org/c/openstack/neutron/+/823829 | 11:56 |
opendevreview | Merged openstack/os-ken stable/xena: Avoid logging MD5 password for BGP add neighbor https://review.opendev.org/c/openstack/os-ken/+/824299 | 12:16 |
opendevreview | Merged openstack/os-ken stable/wallaby: Avoid logging MD5 password for BGP add neighbor https://review.opendev.org/c/openstack/os-ken/+/824300 | 12:23 |
opendevreview | Merged openstack/os-ken stable/victoria: Avoid logging MD5 password for BGP add neighbor https://review.opendev.org/c/openstack/os-ken/+/824301 | 12:23 |
opendevreview | Merged openstack/os-ken stable/ussuri: Avoid logging MD5 password for BGP add neighbor https://review.opendev.org/c/openstack/os-ken/+/824302 | 12:23 |
opendevreview | Merged openstack/os-ken stable/train: Avoid logging MD5 password for BGP add neighbor https://review.opendev.org/c/openstack/os-ken/+/824303 | 12:23 |
opendevreview | Oleg Bondarev proposed openstack/neutron master: Add devstack plugin support for Local IP https://review.opendev.org/c/openstack/neutron/+/818228 | 12:34 |
opendevreview | Lajos Katona proposed openstack/networking-odl master: Fix some requirements https://review.opendev.org/c/openstack/networking-odl/+/823428 | 12:40 |
opendevreview | Lajos Katona proposed openstack/neutron-lib master: Move API definitions of taas to neutron-lib https://review.opendev.org/c/openstack/neutron-lib/+/823475 | 12:52 |
opendevreview | Lajos Katona proposed openstack/neutron-lib master: Move API definitions of taas to neutron-lib https://review.opendev.org/c/openstack/neutron-lib/+/823475 | 12:56 |
opendevreview | Lajos Katona proposed openstack/neutron-lib master: Move API definitions of taas to neutron-lib https://review.opendev.org/c/openstack/neutron-lib/+/823475 | 13:03 |
opendevreview | Merged openstack/neutron stable/ussuri: Add wait event for metadataagent sb_idl https://review.opendev.org/c/openstack/neutron/+/823851 | 13:35 |
opendevreview | Merged openstack/neutron master: Ensure only the right events are processed https://review.opendev.org/c/openstack/neutron/+/823818 | 13:52 |
*** dasm|off is now known as dasm | 14:00 | |
*** ykarel_ is now known as ykarel | 14:06 | |
opendevreview | Merged openstack/neutron master: Allow to use static Local IP openflow rules https://review.opendev.org/c/openstack/neutron/+/817399 | 14:12 |
opendevreview | Elvira García Ruiz proposed openstack/networking-ovn stable/train: Always update router static route https://review.opendev.org/c/openstack/networking-ovn/+/824375 | 14:13 |
EugenMayer | ralonsoh i'am a bit rusty with debugging package flow with OVN, it would be easy for me with tcpdump and bridges, but with linux namespaces i'am not very familiar with. What would be the way you would design this? | 14:14 |
EugenMayer | ralonsoh it seems like using any non FIPs port-forwarded target service, so e.g. our FIP for k8s without port forwarding - works without issues. | 14:17 |
EugenMayer | Just the port-forwarded ones are problematic, if i understand it right | 14:18 |
EugenMayer | (we are using xena, if this is relevant) | 14:18 |
ralonsoh | EugenMayer, sorry, I don't know how you use port forwarding without FIPs in Neutron | 14:19 |
EugenMayer | i use FIPs with portfowarding - as you expected. | 14:19 |
ralonsoh | and what do you need to test? | 14:20 |
EugenMayer | are you familiar with terraform ? | 14:21 |
ralonsoh | no | 14:21 |
EugenMayer | ok, maybe it helps using the docs. I used openstack classic port forwardin via fips without defining them via tf, and later with. | 14:22 |
EugenMayer | Basically i expose a FIP using via https://registry.terraform.io/providers/terraform-provider-openstack/openstack/latest/docs/resources/networking_floatingip_v2 - so the FIP | 14:22 |
EugenMayer | then i use port forwarding https://registry.terraform.io/providers/terraform-provider-openstack/openstack/latest/docs/resources/networking_portforwarding_v2 to forward the port to an VM | 14:23 |
EugenMayer | before i had TF i did the setup using openstack cli tools. So it is all the way classic FIP + FIP forwardning. All based on vanilla OVN network, no manual hacking or custom cli adjustments. all via openstack rest API | 14:25 |
EugenMayer | I used https://docs.openstack.org/neutron/latest/admin/config-fip-port-forwardings.html as my reference | 14:26 |
EugenMayer | This is a typical cli statement i used before using terraform: https://gist.github.com/EugenMayer/f7e45d5338679ca9821ed454b85c14fb - i would expect that this matches exactly what you call NAT via FIPs | 14:27 |
EugenMayer | ralonsoh anything else i could provide to make you sense that it is setup as expected / it should be done (or not) and it might be a bug? | 14:28 |
ralonsoh | EugenMayer, what you need to provide is how you are trying to connect to this port | 14:29 |
ralonsoh | just to be able to replicate this issue | 14:29 |
EugenMayer | telnet externalip port | 14:29 |
ralonsoh | from where? | 14:30 |
EugenMayer | as simple as that right now. From the VM A to VM B via the external IP | 14:30 |
ralonsoh | are the VMs in the same host? | 14:30 |
EugenMayer | VM A and VM B are both hosted on the same openstack. | 14:30 |
ralonsoh | do you have DVR? | 14:30 |
EugenMayer | No DVR | 14:30 |
ralonsoh | by default OVN has DVR | 14:30 |
EugenMayer | but i deactivated it. | 14:31 |
EugenMayer | They are on the same compute, yes | 14:31 |
ralonsoh | ok, I'll try to replicate this issue | 14:31 |
EugenMayer | i jsut try using compute1 to compute2, so different computes to see if this can be reproduced this way | 14:32 |
opendevreview | Merged openstack/networking-ovn stable/train: Add wait event for metadataagent sb_idl https://review.opendev.org/c/openstack/networking-ovn/+/823853 | 14:33 |
EugenMayer | ralonsoh using a VM-A on compute1 and the target VM-B on compute2 does not work eithere | 14:34 |
opendevreview | Oleg Bondarev proposed openstack/neutron master: Add Local IP policy rules https://review.opendev.org/c/openstack/neutron/+/816435 | 14:34 |
opendevreview | Oleg Bondarev proposed openstack/neutron master: Add devstack plugin support for Local IP https://review.opendev.org/c/openstack/neutron/+/818228 | 14:34 |
opendevreview | Takashi Kajinami proposed openstack/neutron master: ml2 ovs: Deprecate unused [agent] veth_mtu https://review.opendev.org/c/openstack/neutron/+/824446 | 14:41 |
EugenMayer | ralonsoh let me know if you need anything else to debug here. As said, fairly new to OVN/OVS so my namespace-tooling is poor, but i'll try my best for sure | 14:42 |
*** jlibosva is now known as Guest305 | 14:45 | |
EugenMayer | ralonsoh i have some more facts which might limit the problem space. I have one 'internal' network. When the source VM is part internal, and the target VM is part of internal, it does not work. | 14:45 |
ralonsoh | EugenMayer, please, document this in a launchpad bug | 14:46 |
ralonsoh | so everyone is aware of it and can try to reproduce it | 14:46 |
EugenMayer | sure, any particular project? neutron? | 14:46 |
ralonsoh | yes | 14:46 |
EugenMayer | ralonsoh: https://bugs.launchpad.net/neutron/+bug/1957185 let me know if something is unclear or missing. Hopefully the informations are usefull | 14:54 |
ralonsoh | ok | 14:54 |
opendevreview | Lajos Katona proposed openstack/neutron-tempest-plugin master: QoS min pps API tests https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/806689 | 14:56 |
opendevreview | Terry Wilson proposed openstack/neutron master: WIP Use neutron db for ovn agents https://review.opendev.org/c/openstack/neutron/+/818850 | 15:44 |
opendevreview | Slawek Kaplonski proposed openstack/neutron stable/train: DNM Don't register workers cleanup during the module import https://review.opendev.org/c/openstack/neutron/+/824450 | 15:57 |
opendevreview | Merged openstack/neutron master: Enable min pps tempest tests https://review.opendev.org/c/openstack/neutron/+/811746 | 15:57 |
*** marlinc is now known as Guest315 | 16:30 | |
jpic__ | hi all, there's an old abandoned RFE i'd like to take over, is there a procedure for that? | 16:54 |
opendevreview | jpic proposed openstack/neutron-specs master: Update our RFE https://review.opendev.org/c/openstack/neutron-specs/+/824464 | 16:58 |
opendevreview | jpic proposed openstack/neutron-specs master: Start over from previous RFE https://review.opendev.org/c/openstack/neutron-specs/+/824465 | 16:58 |
ralonsoh | jpic__, I don't think lajoskatona was proposing to deprecate your spec | 17:48 |
ralonsoh | but just adding more info | 17:48 |
ralonsoh | you can use the old spec to amend yours, include more content, etc. | 17:48 |
lajoskatona | jpic__: For the "multiple routed provider network per compute" spec I just wanted to highlight the history of the topic | 17:53 |
lajoskatona | jpic__: perhaps that helps to write your spec, and understand all the aspects of the problem, it can be important to read even the comments/discussions under the previous spec/RFE to understand why it was written or voted that way | 17:55 |
*** marlinc is now known as Guest323 | 18:04 | |
opendevreview | Lajos Katona proposed openstack/os-ken master: Avoid missing key 'password' for neighbor_add https://review.opendev.org/c/openstack/os-ken/+/824476 | 18:17 |
opendevreview | Merged openstack/networking-ovn stable/train: Always update router static route https://review.opendev.org/c/openstack/networking-ovn/+/824375 | 19:49 |
opendevreview | Mamatisa Nurmatov proposed openstack/neutron-tempest-plugin master: Add local ip scenario tests https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/823007 | 19:51 |
*** promethe- is now known as prometheanfire | 20:42 | |
opendevreview | Merged openstack/neutron stable/train: Fix privileged create_netns function https://review.opendev.org/c/openstack/neutron/+/823963 | 21:29 |
*** dasm is now known as dasm|off | 22:45 | |
*** marlinc is now known as Guest417 | 23:00 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!