Re: [openstack-dev] [kuryr] Port binding query
Yes, we want to use the same naming conventions when possible. Submitted a fix here: https://review.openstack.org/#/c/315799/ Best, Mohammad From: Antoni Segura Puimedon To: "OpenStack Development Mailing List (not for usage questions)" Cc: Mohammad Banikazemi/Watson/IBM@IBMUS Date: 05/12/2016 11:51 AM Subject: Re: [openstack-dev] [kuryr] Port binding query On Thu, May 12, 2016 at 4:50 PM, Neil Jerram wrote: I'm trying Kuryr with networking-calico and think I've hit an unhelpful inconsistency. A Neutron port has 'id' and 'device_id' fields that are usually different. When Nova does VIF binding for a Neutron port, it generates the Linux device name from 'tap' + port['id']. But when Kuryr does VIF binding for a Neutron port, I think it generates the Linux device name from 'tap' + port['device_id']. Thoughts? Does that sound right, or have I misread the code and my logs? If it's correct, it marginally impacts the ability to use identical agent and Neutron driver/plugin code for the two cases (Nova and Kuryr). I think we are supposed to behave like Nova, binding wise. @Banix: Can you confirm that it is a bug and not a feature? >From a quick grepping I see hat nova sets the name to be: nova/network/neutronv2/api.py: devname = "tap" + current_neutron_port['id'] Whereas in Kuryr we use the first 8 characters of the Docker endpoint id. Thanks, Neil __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [kuryr] Port binding query
On Thu, May 12, 2016 at 4:50 PM, Neil Jerram wrote: > I'm trying Kuryr with networking-calico and think I've hit an unhelpful > inconsistency. A Neutron port has 'id' and 'device_id' fields that are > usually different. When Nova does VIF binding for a Neutron port, it > generates the Linux device name from 'tap' + port['id']. But when Kuryr > does VIF binding for a Neutron port, I think it generates the Linux device > name from 'tap' + port['device_id']. > > Thoughts? Does that sound right, or have I misread the code and my logs? > If it's correct, it marginally impacts the ability to use identical agent > and Neutron driver/plugin code for the two cases (Nova and Kuryr). > I think we are supposed to behave like Nova, binding wise. @Banix: Can you confirm that it is a bug and not a feature? >From a quick grepping I see hat nova sets the name to be: nova/network/neutronv2/api.py:devname = "tap" + current_neutron_port['id'] Whereas in Kuryr we use the first 8 characters of the Docker endpoint id. > > Thanks, > Neil > > > __ > OpenStack Development Mailing List (not for usage questions) > Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev > > __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] [kuryr] Port binding query
I'm trying Kuryr with networking-calico and think I've hit an unhelpful inconsistency. A Neutron port has 'id' and 'device_id' fields that are usually different. When Nova does VIF binding for a Neutron port, it generates the Linux device name from 'tap' + port['id']. But when Kuryr does VIF binding for a Neutron port, I think it generates the Linux device name from 'tap' + port['device_id']. Thoughts? Does that sound right, or have I misread the code and my logs? If it's correct, it marginally impacts the ability to use identical agent and Neutron driver/plugin code for the two cases (Nova and Kuryr). Thanks, Neil __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] [kuryr] Port binding query
I'm trying Kuryr with networking-calico and think I've hit an unhelpful inconsistency. A Neutron port has 'id' and 'device_id' fields that are usually different. When Nova does VIF binding for a Neutron port, it generates the Linux device name from 'tap' + port['id']. But when Kuryr does VIF binding for a Neutron port, I think it generates the Linux device name from 'tap' + port['device_id']. Thoughts? Does that sound right, or have I misread the code and my logs? If it's correct, it marginally impacts the ability to use identical agent and Neutron driver/plugin code for the two cases (Nova and Kuryr). Thanks, Neil __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev