yahoo-eng-team team mailing list archive

[OpenStack Infra <1257641@xxxxxxxxxxxxxxxxxx>]

Reviewed:  https://review.openstack.org/61313
Committed: http://github.com/openstack/tempest/commit/122cdf5cea8e479949f993513ec556e9b12adb59
Submitter: Jenkins
Branch:    master

commit 122cdf5cea8e479949f993513ec556e9b12adb59
Author: Ken'ichi Ohmichi <oomichi@xxxxxxxxxxxxxxxxx>
Date:   Wed Dec 11 21:32:25 2013 +0900

    Fix the scope to share a server between tests
    
    In test_server_actions.py, one server is created in setUpClass()
    and the server would be shared between all tests.
    If the server status is not 'ACTIVE' before each test, rebuild_
    server() deletes it and rebuilds a server again.
    
    The server_id scope which is passed to rebuild_server() is class
    variable not instance variable. In addition, the server_id scope
    which is stored with the return value of rebuild_server() is
    instance variable like the following:
    
      [[instance]]                         [[class]]
      self.server_id = self.rebuild_server(self.server_id)
    
    If we need to rebuild a server, the server_id, which is passed to
    rebuild_server(), cannot be changed due to its scope. As the result,
    many server creations happens in each test.
    Then it causes "Quota exceeded" exception.
    
    This patch fixes the scope of server_id for sharing the same server
    between each tests.
    
    Change-Id: I72b4aa73a6c77cce3a8e777d4f78913fb0986f66
    Closes-Bug: #1257641


** Changed in: tempest
       Status: In Progress => Fix Released

-- 
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/1257641

Title:
  Quota exceeded for instances: Requested 1, but already used 10 of 10
  instances

Status in OpenStack Compute (Nova):
  Confirmed
Status in Tempest:
  Fix Released

Bug description:
  Tempest (or nova) appears to be leaking servers.  Since we run gate
  tests with concurrency=4 I don't think we should normally hit a case
  where we are consuming 10 instances.

  http://logs.openstack.org/84/59684/4/check/check-tempest-dsvm-
  postgres-full/9e38ebb/testr_results.html.gz

  http://logstash.openstack.org/#eyJzZWFyY2giOiJcIlF1b3RhIGV4Y2VlZGVkIGZvciBpbnN0YW5jZXM6IFJlcXVlc3RlZCAxLCBidXQgYWxyZWFkeSB1c2VkIDEwIG9mIDEwIGluc3RhbmNlc1wiIiwiZmllbGRzIjpbXSwib2Zmc2V0IjowLCJ0aW1lZnJhbWUiOiIxNzI4MDAiLCJncmFwaG1vZGUiOiJjb3VudCIsInRpbWUiOnsidXNlcl9pbnRlcnZhbCI6MH0sInN0YW1wIjoxMzg2MTQ1OTIzMDMwfQ==

  query: "Quota exceeded for instances: Requested 1, but already used 10
  of 10 instances" AND filename:"console.html"

  First seen on 12/3

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1257641/+subscriptions