yahoo-eng-team team mailing list archive

[Jesse Keating <omgjlk@xxxxxxxxxx>]

Public bug reported:

As an operator, I want to be able to use oslo healthcheck middleware[1]
to Nova's pipeline so that I can GET the /healthcheck URI to determine
"up" or not for a given nova-api host. The healthcheck middleware allows
for manually setting the health status to offline, without actually
stopping the API service. When I can do this, I can easily "take an API
node offline" from the aspect of the load balancer, which uses the
healthcheck path for status checks, before stopping the API process
itself. This is a quick and generic way to prevent new connections to a
given API node while restarting it as part of a rolling restart.

This middleware has already been added to glance[2], and I've got an
open review to add it to keystone as well[3]. My eventual goal is to
have it in use across all the services that directly listen for client
connections.

1: http://docs.openstack.org/developer/oslo.middleware/healthcheck_plugins.html
2: https://review.openstack.org/#/c/148595/
3: https://review.openstack.org/#/c/387731/

** Affects: nova
     Importance: Undecided
         Status: New

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1637483

Title:
  Add healthcheck middleware to pipelines

Status in OpenStack Compute (nova):
  New

Bug description:
  As an operator, I want to be able to use oslo healthcheck
  middleware[1] to Nova's pipeline so that I can GET the /healthcheck
  URI to determine "up" or not for a given nova-api host. The
  healthcheck middleware allows for manually setting the health status
  to offline, without actually stopping the API service. When I can do
  this, I can easily "take an API node offline" from the aspect of the
  load balancer, which uses the healthcheck path for status checks,
  before stopping the API process itself. This is a quick and generic
  way to prevent new connections to a given API node while restarting it
  as part of a rolling restart.

  This middleware has already been added to glance[2], and I've got an
  open review to add it to keystone as well[3]. My eventual goal is to
  have it in use across all the services that directly listen for client
  connections.

  1: http://docs.openstack.org/developer/oslo.middleware/healthcheck_plugins.html
  2: https://review.openstack.org/#/c/148595/
  3: https://review.openstack.org/#/c/387731/

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1637483/+subscriptions