yahoo-eng-team team mailing list archive

[Doug Wiegley <dougwig@xxxxxxxxxxx>]

The way the lbaas model works is that the object is created, the driver
called, and the driver puts it ACTIVE when it's done, or flags an error,
or it's left pending in catastrophe. In that respect, the above behavior
is working as designed.

If you want to change that to an object rollback, please submit that as
an overall rfe to how all lbaas objects work. The all-in-one creation
call may also avoid this issue.

** Changed in: neutron
       Status: New => Won't Fix

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/1515454

Title:
  In LBaaS,DB seems to be updated, even though the actual  operation may
  fail due to driver error

Status in neutron:
  Won't Fix

Bug description:
  High Level Description:
  While working on lbaas V2 , found a somewhat strange behavior.( mention below)

  Pre-conditions: Enable LBaaS v2 extension
  Step - By -Step reproduction:
  a) Verify all the members in Pool
  reedip@reedip-VirtualBox:/opt/stack/python-neutronclient/neutronclient/tests/unit/lb/v2$ neutron lbaas-member-list testpool
  +--------------------------------------+--------------+---------------+--------+--------------------------------------+----------------+
  | id                                   | address      | protocol_port | weight | subnet_id                            | admin_state_up |
  +--------------------------------------+--------------+---------------+--------+--------------------------------------+----------------+
  | 2644b225-53df-4cdf-9ab3-dea5da1d402c | 172.24.4.120 |            90 |      1 | af8b5dfb-732b-4ecd-87f5-10cd4cb0d917 | True           |
  +--------------------------------------+--------------+---------------+--------+--------------------------------------+----------------+
  b) Create a new member , it fails due to driver error

  reedip@reedip-VirtualBox:/opt/stack/python-neutronclient/neutronclient/tests/unit/lb/v2$ neutron lbaas-member-create --subnet public-subnet --address 172.24.4.121 --protocol-port 90 testpool
  An error happened in the driver

  c) List the members in the specified pool
  reedip@reedip-VirtualBox:/opt/stack/python-neutronclient/neutronclient/tests/unit/lb/v2$ neutron lbaas-member-list testpool
  +--------------------------------------+--------------+---------------+--------+--------------------------------------+----------------+
  | id                                   | address      | protocol_port | weight | subnet_id                            | admin_state_up |
  +--------------------------------------+--------------+---------------+--------+--------------------------------------+----------------+
  | 2644b225-53df-4cdf-9ab3-dea5da1d402c | 172.24.4.120 |            90 |      1 | af8b5dfb-732b-4ecd-87f5-10cd4cb0d917 | True           |
  | 39d1017e-92ca-40fd-b02d-739189a4b8df | 172.24.4.121 |            90 |      1 | af8b5dfb-732b-4ecd-87f5-10cd4cb0d917 | True           |
  +--------------------------------------+--------------+---------------+--------+--------------------------------------+----------------+
  reedip@reedip-VirtualBox:/opt/stack/python-neutronclient/neutronclient/tests/unit/lb/v2$

  Expected Output: If the driver error occurs, then the new member should not be added
  Actual Output: The new member which actually failed due to driver error was actually added to the system, which is incorrect behavior.

  Version: Ubuntu 14.04, git for Neutron Client: 3d736107f97c27a35cff2d7ed6c041521be5ab03
  git for neutron-lbaas:
  321da8f6263d46bf059163bcf7fd005cf68601bd

  Environment: Devstack installation of an All-In-One single node, with FWaaS, LBaaSv2 and octavia enabled.
  Perceived Severity: High ( this is negative behaviour, because  an inoperatable member is created and exists in the DB)

To manage notifications about this bug go to:
https://bugs.launchpad.net/neutron/+bug/1515454/+subscriptions