Created attachment 1265762 [details] conf file Description of problem: /etc/octavia/octavia.conf [DEFAULT] # Print more verbose output (set logging level to INFO instead of default WARNING level). # verbose = True # Print debugging output (set logging level to DEBUG instead of default WARNING level). # debug = False # bind_host = 127.0.0.1 <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< # bind_port = 9876 # api_handler = queue_producer # # How should authentication be handled (keystone, noauth) # auth_strategy = noauth The specified i nthe cond file is host = 172.17.1.19 ---------> should be bind_host = 172.17.1.19 !!!conf file attached !!! Version-Release number of selected component (if applicable): rhos11 How reproducible: 100% Steps to Reproduce: 1. deploy ops11 with ospd + octavia yaml files 2. 3. Actual results: Octavia binds to local host Expected results: SHould bind to the networker Additional info:
Brent and I discussed this and came up with the following plan: If you can please change bind_host to the IP of the controller on the internal API network and run Tempest to see if that works. If it does, then Brent can send a patch to TripleO to ensure that octavia-api binds on normal IP address on the internal API network, which means that it will not be a post-deployment documentation step, but will happen automatically through regular deployment. If it doesn't work, then it will continue to be a post-deployment documented step (To use the IP address on the LB network) and we'll target 12 to fix it in TripleO.
(In reply to Assaf Muller from comment #1) > Brent and I discussed this and came up with the following plan: > > If you can please change bind_host to the IP of the controller on the > internal API network and run Tempest to see if that works. If it does, then > Brent can send a patch to TripleO to ensure that octavia-api binds on normal > IP address on the internal API network, which means that it will not be a > post-deployment documentation step, but will happen automatically through > regular deployment. > > If it doesn't work, then it will continue to be a post-deployment documented > step (To use the IP address on the LB network) and we'll target 12 to fix it > in TripleO. I need to execute post deployment step anyhow. I will comment out the post deployment configuration and configure manually "bind_host" in the octavia.conf file. I will get back with the result.
This bugzilla has been removed from the release and needs to be reviewed and Triaged for another Target Release.
Following up on comment 2, did we nail this down?
(In reply to Assaf Muller from comment #5) > Following up on comment 2, did we nail this down? I do not recall what happened with this one. I do recall that changing the IP to the non localhost ip did not result with issues. May be NirM remembers something regarding that.
(In reply to Alexander Stafeyev from comment #6) > (In reply to Assaf Muller from comment #5) > > Following up on comment 2, did we nail this down? > > I do not recall what happened with this one. > > I do recall that changing the IP to the non localhost ip did not result with > issues. > > May be NirM remembers something regarding that. I can't answer until both Or and Brent finalize the current Octavia:TripleO work, but maybe they'll have a better answer.
This is currently being handled in workflow driven ansible in this patch: https://review.openstack.org/#/c/487450/
According to our records, this should be resolved by openstack-tripleo-heat-templates-8.0.2-38.el7ost. This build is available now.