yahoo-eng-team team mailing list archive

[Rodolfo Alonso <2103521@xxxxxxxxxxxxxxxxxx>]

Public bug reported:

The RFE proposes to create an OVN L3 router tool, providing extra
functionality to the current Neutron API.

The Neutron API must be consistent across the multiple existing
backends. It is not recommended (allowed) to implement API calls
specific for a particular backend.

In the case of ML2/OVN, it is needed to provide more functionality to the L3 service. This RFE proposes to create a new CLI tool to provide it. This tool would implement, initially, the following tools:
* Listing the available chassis and the GW LRP assignation. This list could have multiple views:
** Short: just the number of GW LRP assigned to each chassis and the priority level.
** Long: the list of GW LRP assigned to each chassis.

* A rebalancing tool. In some occasions, after deleting GW ports or routers, the chassis assignment is improperly imbalanced. This tool will check the current assignations and redo them again. Of course, that could have an impact on live traffic if we change the highest priority chassis assigned. Thus this command will have two options:
** The first one with full rescheduling. That will affect live traffic as long as the GW LRP changes the bound chassis.
** The second, without touching the highest priority chassis assigned. The reschedule won't be as efficient as the first option, but the traffic won't be affected.

NOTE: because this tool will modify the current LRP chassis assignment,
it is needed to conclude https://bugs.launchpad.net/neutron/+bug/2092271
first; that means, we need to migrate the LRP.gateway_chassis usage in
favor of HA_Chassis_Group, that is the preferred method in OVN.

** Affects: neutron
     Importance: Undecided
     Assignee: Rodolfo Alonso (rodolfo-alonso-hernandez)
         Status: New

** Changed in: neutron
     Assignee: (unassigned) => Rodolfo Alonso (rodolfo-alonso-hernandez)

** Summary changed:

- [RFE][OVN] A L3 agent tool
+ [RFE][OVN] A L3 router tool

** Description changed:

- The RFE proposes to create an OVN L3 agent tool, providing extra
+ The RFE proposes to create an OVN L3 router tool, providing extra
  functionality to the current Neutron API.
  
  The Neutron API must be consistent across the multiple existing
  backends. It is not recommended (allowed) to implement API calls
  specific for a particular backend.
  
  In the case of ML2/OVN, it is needed to provide more functionality to the L3 service. This RFE proposes to create a new CLI tool to provide it. This tool would implement, initially, the following tools:
  * Listing the available chassis and the GW LRP assignation. This list could have multiple views:
  ** Short: just the number of GW LRP assigned to each chassis and the priority level.
  ** Long: the list of GW LRP assigned to each chassis.
  
  * A rebalancing tool. In some occasions, after deleting GW ports or routers, the chassis assignment is improperly imbalanced. This tool will check the current assignations and redo them again. Of course, that could have an impact on live traffic if we change the highest priority chassis assigned. Thus this command will have two options:
  ** The first one with full rescheduling. That will affect live traffic as long as the GW LRP changes the bound chassis.
  ** The second, without touching the highest priority chassis assigned. The reschedule won't be as efficient as the first option, but the traffic won't be affected.
  
  NOTE: because this tool will modify the current LRP chassis assignment,
  it is needed to conclude https://bugs.launchpad.net/neutron/+bug/2092271
  first; that means, we need to migrate the LRP.gateway_chassis usage in
  favor of HA_Chassis_Group, that is the preferred method in OVN.

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/2103521

Title:
  [RFE][OVN] A L3 router tool

Status in neutron:
  New

Bug description:
  The RFE proposes to create an OVN L3 router tool, providing extra
  functionality to the current Neutron API.

  The Neutron API must be consistent across the multiple existing
  backends. It is not recommended (allowed) to implement API calls
  specific for a particular backend.

  In the case of ML2/OVN, it is needed to provide more functionality to the L3 service. This RFE proposes to create a new CLI tool to provide it. This tool would implement, initially, the following tools:
  * Listing the available chassis and the GW LRP assignation. This list could have multiple views:
  ** Short: just the number of GW LRP assigned to each chassis and the priority level.
  ** Long: the list of GW LRP assigned to each chassis.

  * A rebalancing tool. In some occasions, after deleting GW ports or routers, the chassis assignment is improperly imbalanced. This tool will check the current assignations and redo them again. Of course, that could have an impact on live traffic if we change the highest priority chassis assigned. Thus this command will have two options:
  ** The first one with full rescheduling. That will affect live traffic as long as the GW LRP changes the bound chassis.
  ** The second, without touching the highest priority chassis assigned. The reschedule won't be as efficient as the first option, but the traffic won't be affected.

  NOTE: because this tool will modify the current LRP chassis
  assignment, it is needed to conclude
  https://bugs.launchpad.net/neutron/+bug/2092271 first; that means, we
  need to migrate the LRP.gateway_chassis usage in favor of
  HA_Chassis_Group, that is the preferred method in OVN.

To manage notifications about this bug go to:
https://bugs.launchpad.net/neutron/+bug/2103521/+subscriptions