Re: [openstack-dev] [neutron] Ocata End user and operator feedback recap
On Tue, Nov 8, 2016 at 12:57 AM, Oleg Bondarev wrote: > > > On Tue, Nov 8, 2016 at 12:22 AM, Carl Baldwin wrote: >> >> Ocata End user and operator feedback recap >> >> >> The purpose of this session was to gather feedback from end users and >> operators to help direct the efforts of developers during (at least) the >> Ocata cycle. Feedback was captured on the etherpad [1]. There was a related >> session in the operators' track [2] which may also be of interest. >> >> We began with a short discussion about client compatibility which was >> deferred to the following session specifically about the client [3]. >> >> There was a discussion about if Neutron will implement cells like Nova >> has. There is currently no plan. I have heard this come up for the past >> couple of summits but there is little concrete evidence of the scale issues >> being encountered by operators. If you are running Neutron at scale, we >> could use more of your input. >> >> There was some discussion about an issue around moving a floating ip from >> one instance to another when keepalive is in use. We needed more information >> to debug this. It should be sending a gratuitous arp. Another keepalive >> issue was discussed. Too many SIGHUPs within a period of time can cause >> failover resulting in disruption. I do not know if a bug was filed for this >> to follow up. If this was you, please follow up with a link to the bug >> report. > > > https://bugs.launchpad.net/neutron/+bug/1639315 looks related That covers keepalived not sending GARPs on SIGHUP. This will be resolved via a workaround in Neutron with a patch by Jakub Libosvar (Linked from the bug report). The other issue of keepalived hanging when receiving frequent consecutive SIGHUPs (When a floating IP is added/removed) will also be resolved via a workaround, but I don't see a bug report yet. This will be handled by John Schwarz. > >> >> >> There was a short discussion about Horizon support for security groups on >> a port. A bug was filed for this. >> >> Finally, there was some discussion about enabling end users to create >> Neutron L3 routers which are backed by hardware resources. There is no such >> thing yet but Neutron does have a new concept in development called "L3 >> flavors". This would enable a driver (to be written) which would allow this >> sort of thing. >> >> Carl Baldwin >> >> [1] >> https://etherpad.openstack.org/p/ocata-neutron-end-user-operator-feedback >> [2] https://etherpad.openstack.org/p/newton-neutron-pain-points >> [3] https://etherpad.openstack.org/p/ocata-neutron-client >> >> __ >> 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 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] [neutron] Ocata End user and operator feedback recap
On Tue, Nov 8, 2016 at 12:22 AM, Carl Baldwin wrote: > Ocata End user and operator feedback recap > > > The purpose of this session was to gather feedback from end users and > operators to help direct the efforts of developers during (at least) the > Ocata cycle. Feedback was captured on the etherpad [1]. There was a related > session in the operators' track [2] which may also be of interest. > > We began with a short discussion about client compatibility which was > deferred to the following session specifically about the client [3]. > > There was a discussion about if Neutron will implement cells like Nova > has. There is currently no plan. I have heard this come up for the past > couple of summits but there is little concrete evidence of the scale issues > being encountered by operators. If you are running Neutron at scale, we > could use more of your input. > > There was some discussion about an issue around moving a floating ip from > one instance to another when keepalive is in use. We needed more > information to debug this. It should be sending a gratuitous arp. Another > keepalive issue was discussed. Too many SIGHUPs within a period of time can > cause failover resulting in disruption. I do not know if a bug was filed > for this to follow up. If this was you, please follow up with a link to the > bug report. > https://bugs.launchpad.net/neutron/+bug/1639315 looks related > > There was a short discussion about Horizon support for security groups on > a port. A bug was filed for this. > > Finally, there was some discussion about enabling end users to create > Neutron L3 routers which are backed by hardware resources. There is no such > thing yet but Neutron does have a new concept in development called "L3 > flavors". This would enable a driver (to be written) which would allow this > sort of thing. > > Carl Baldwin > > [1] https://etherpad.openstack.org/p/ocata-neutron- > end-user-operator-feedback > [2] https://etherpad.openstack.org/p/newton-neutron-pain-points > [3] https://etherpad.openstack.org/p/ocata-neutron-client > > __ > 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] [neutron] Ocata End user and operator feedback recap
Ocata End user and operator feedback recap The purpose of this session was to gather feedback from end users and operators to help direct the efforts of developers during (at least) the Ocata cycle. Feedback was captured on the etherpad [1]. There was a related session in the operators' track [2] which may also be of interest. We began with a short discussion about client compatibility which was deferred to the following session specifically about the client [3]. There was a discussion about if Neutron will implement cells like Nova has. There is currently no plan. I have heard this come up for the past couple of summits but there is little concrete evidence of the scale issues being encountered by operators. If you are running Neutron at scale, we could use more of your input. There was some discussion about an issue around moving a floating ip from one instance to another when keepalive is in use. We needed more information to debug this. It should be sending a gratuitous arp. Another keepalive issue was discussed. Too many SIGHUPs within a period of time can cause failover resulting in disruption. I do not know if a bug was filed for this to follow up. If this was you, please follow up with a link to the bug report. There was a short discussion about Horizon support for security groups on a port. A bug was filed for this. Finally, there was some discussion about enabling end users to create Neutron L3 routers which are backed by hardware resources. There is no such thing yet but Neutron does have a new concept in development called "L3 flavors". This would enable a driver (to be written) which would allow this sort of thing. Carl Baldwin [1] https://etherpad.openstack.org/p/ocata-neutron-end-user-operator-feedback [2] https://etherpad.openstack.org/p/newton-neutron-pain-points [3] https://etherpad.openstack.org/p/ocata-neutron-client __ 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