| *** janki has joined #networking-sfc | 03:34 | |
| *** mohankumar has joined #networking-sfc | 05:38 | |
| *** mangelm has joined #networking-sfc | 08:02 | |
| *** janki has quit IRC | 09:28 | |
| mangelm | Hi everyone, when creating a port chain I am using correlation=None but I'm getting an error in logs. Is it possible to avoid MPLS correlation somehow? | 11:41 |
|---|---|---|
| mangelm | this is the error I get in case someone wants to have a look: https://hastebin.com/reloniqari.sql | 11:43 |
| mangelm | and this is the request: | 11:44 |
| mangelm | openstack sfc port chain create --port-pair-group pgClientM --port-pair-group pg_${VM_ID} --port-pair-group pgServerM --flow-classifier fc1_${VM_ID} --flow-classifier fc2_${VM_ID} --flow-classifier fc3_${VM_ID} --chain-parameters symmetric=True --chain-parameters correlation=None pc_${VM_ID} | 11:44 |
| mangelm | It does not matter if I put correlation="" or correlation=whatever, the error is still the same | 11:44 |
| bcafarel | mangelm: I have no setup around at the moment, but maybe check the python-neutronclient version (there were some sfc fixes a few months ago) | 11:57 |
| bcafarel | though they should all be in pike release | 11:57 |
| mangelm | bcafarel I'm using OPNFV XCI deployment, https://wiki.opnfv.org/display/sfc/Deploy+OPNFV+SFC+scenarios | 12:03 |
| mangelm | bcafarel, I am sure they manage to create chains but the difference is they normally use tacker, not networking-sfc cli | 12:03 |
| mangelm | bcafarel, interestingly I'm checking out ovs flows and I see "pop_nsh" actions, but no "push_nsh" in any of the rules | 12:04 |
| mangelm | bcafarel, I guess that is a sign that the rules have not been populated correctly | 12:05 |
| bcafarel | mangelm: yes tacker goes through the api itself, this is probably specific to the osc client | 12:06 |
| bcafarel | mangelm: workaround may be to try with neutron CLI | 12:06 |
| mangelm | bcafarel, ok I can try that and also let's see if using tacker helps. In any case I should see push_nsh actions in openflow, shouldn't I? | 12:09 |
| bcafarel | mangelm: tacker sets nsh correlation? (opnfv does if I remember correctly) so yep you should | 12:12 |
| mangelm | bcafarel, I am not fully sure but I remember some OPNFV colleagues saying that NSH headers can be seen by the SF | 12:17 |
| bcafarel | it depends on the correlation value for the port pair (which will create or not a proxy to strip them) | 12:21 |
| mangelm | bcafarel, hmmm I think I can have a clue for the problem. When the port pair is created with no arguments, correlation is set to "None". But when port chain is created with no parameters correlation is set to "MPLS". Do you think this could be causing problems? | 12:24 |
| bcafarel | with "None" correlation on port pair, you will not see NSH headers in the SF yes (they will be stripped before/restored after) | 12:29 |
| bcafarel | mangelm: https://docs.openstack.org/networking-sfc/latest/contributor/ietf_sfc_encapsulation.html has some details on these | 12:30 |
| mangelm | bcafarel, it seems like it is not possible to use correlation on port pair creation anyway: | 13:17 |
| mangelm | root@opnfv:~# openstack sfc --insecure port pair create --ingress=${pClientMid} --egress=${pClientMid} ppClientM --service-function-parameters correlation=nsh | 13:17 |
| mangelm | Invalid input for service_function_parameters. Reason: '[{u'correlation': u'nsh'}]' is not a dictionary. | 13:17 |
| mangelm | Neutron server returns request_ids: ['req-e10cee9b-35dc-443b-b594-2cd4127dbeef'] | 13:17 |
| mangelm | root@opnfv:~# | 13:17 |
| bcafarel | mangelm: hmm ok that one I am pretty sure is fixed in python-neutronclient (current or queens) | 13:36 |
| *** janki has joined #networking-sfc | 13:52 | |
| -openstackstatus- NOTICE: POST_FAILURE results on Tempest-based jobs since the most recent Zuul security fixes are being investigated; rechecking those won't help for now but we'll keep you posted once a solution is identified | 14:06 | |
| *** ChanServ changes topic to "POST_FAILURE results on Tempest-based jobs since the most recent Zuul security fixes are being investigated; rechecking those won't help for now but we'll keep you posted once a solution is identified" | 14:07 | |
| *** phuoc_ has joined #networking-sfc | 14:47 | |
| *** phuoc has quit IRC | 14:51 | |
| *** phuoc has joined #networking-sfc | 14:52 | |
| *** phuoc_ has quit IRC | 14:52 | |
| *** janki has quit IRC | 16:46 | |
| *** mangelm has quit IRC | 18:22 | |
| *** ChanServ changes topic to "Neutron Service Function Chaining (SFC) Extension discussions" | 19:22 | |
| -openstackstatus- NOTICE: The regression stemming from one of yesterday's Zuul security fixes has been rectified, and Devstack/Tempest jobs which encountered POST_FAILURE results over the past 24 hours can safely be rechecked now | 19:22 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!