nova/nova/tests/functional/regressions
Lee Yarwood 3a0f26c822 conductor: Recreate volume attachments during a reschedule
When an instance with attached volumes fails to spawn, cleanup code
within the compute manager (_shutdown_instance called from
_build_resources) will delete the volume attachments referenced by
the bdms in Cinder. As a result we should check and if necessary
recreate these volume attachments when rescheduling an instance.

Note that there are a few different ways to fix this bug by
making changes to the compute manager code, either by not deleting
the volume attachment on failure before rescheduling [1] or by
performing the get/create check during each build after the
reschedule [2].

The problem with *not* cleaning up the attachments is if we don't
reschedule, then we've left orphaned "reserved" volumes in Cinder
(or we have to add special logic to tell compute when to cleanup
attachments).

The problem with checking the existence of the attachment on every
new host we build on is that we'd be needlessly checking that for
initial creates even if we don't ever need to reschedule, unless
again we have special logic against that (like checking to see if
we've rescheduled at all).

Also, in either case that involves changes to the compute means that
older computes might not have the fix.

So ultimately it seems that the best way to handle this is:

1. Only deal with this on reschedules.
2. Let the cell conductor orchestrate it since it's already dealing
   with the reschedule. Then the compute logic doesn't need to change.

[1] https://review.openstack.org/#/c/587071/3/nova/compute/manager.py@1631
[2] https://review.openstack.org/#/c/587071/4/nova/compute/manager.py@1667

Conflicts:

  nova/tests/unit/conductor/test_conductor.py

NOTE(mriedem): There was a minor conflict due to not having change
I56fb1fd984f06a58c3a7e8c2596471991950433a in Queens.

Change-Id: I739c06bd02336bf720cddacb21f48e7857378487
Closes-bug: #1784353
(cherry picked from commit 41452a5c6a)
(cherry picked from commit d3397788fe)
2018-10-22 16:08:10 -04:00
..
README.rst add a place for functional test to block specific regressions 2016-02-24 18:14:27 +00:00
__init__.py add a place for functional test to block specific regressions 2016-02-24 18:14:27 +00:00
test_bug_1404867.py fixtures: Track volume attachments within CinderFixtureNewAttachFlow 2018-10-22 15:51:59 -04:00
test_bug_1522536.py replace chance with filter scheduler in func tests 2017-08-16 14:42:51 +00:00
test_bug_1541691.py Remove unused conf 2016-09-02 09:02:23 +08:00
test_bug_1548980.py replace chance with filter scheduler in func tests 2017-08-16 14:42:51 +00:00
test_bug_1552888.py Trivial-Fix: Fix typos 2016-06-13 06:41:08 +00:00
test_bug_1554631.py Catching OverQuota Exception 2017-05-03 19:00:34 +00:00
test_bug_1558866.py conf: Remove deprecated 'null_kernel' opt 2017-10-21 00:23:51 +00:00
test_bug_1568208.py Fix generation of Guru Meditation Report 2016-04-13 17:00:32 +03:00
test_bug_1595962.py Remove unnecessary self.flags and ConfPatcher 2017-11-27 15:49:10 +00:00
test_bug_1620248.py Stop using mox stubs in cast_as_call.py 2017-07-25 00:40:41 +00:00
test_bug_1627838.py Stop overwriting thread local context in ClientRouter 2016-09-26 20:49:22 +00:00
test_bug_1670627.py Clean up ports and volumes when deleting ERROR instance 2018-02-20 14:10:26 -05:00
test_bug_1671648.py Make conductor pass and use host_lists 2017-12-21 17:02:35 -05:00
test_bug_1675570.py fixtures: Track volume attachments within CinderFixtureNewAttachFlow 2018-10-22 15:51:59 -04:00
test_bug_1678326.py Commit usage decrement after destroying instance 2017-04-05 15:19:15 -04:00
test_bug_1679750.py Delete allocations from API if nova-compute is down 2018-04-20 14:57:34 -04:00
test_bug_1682693.py replace chance with filter scheduler in func tests 2017-08-16 14:42:51 +00:00
test_bug_1689692.py Clean up ports and volumes when deleting ERROR instance 2018-02-20 14:10:26 -05:00
test_bug_1702454.py Functional regression test for evacuate with a target 2017-10-29 13:49:06 -04:00
test_bug_1713783.py Set server status to ERROR if rebuild failed 2018-01-23 11:39:45 -05:00
test_bug_1718455.py Ensure instance can migrate when launched concurrently 2017-09-28 11:40:54 -04:00
test_bug_1718512.py Don't persist RequestSpec.retry 2018-04-12 11:27:49 -04:00
test_bug_1719730.py Set group_members when converting to legacy request spec 2017-09-27 20:52:19 +00:00
test_bug_1732947.py Implement new attach Cinder flow 2017-12-07 10:29:15 -05:00
test_bug_1735407.py Add late server group policy check to rebuild 2018-02-06 13:54:12 +01:00
test_bug_1741125.py Fix TypeError in prep_resize allocation cleanup 2018-07-11 08:59:20 -04:00
test_bug_1741307.py Handle no allocations during migrate 2018-01-04 14:18:50 -05:00
test_bug_1746483.py Handle volume-backed instances in IsolatedHostsFilter 2018-02-12 14:45:39 -05:00
test_bug_1746509.py Move _make_instance_list call outside of DB transaction context 2018-03-23 20:01:15 +00:00
test_bug_1780373.py Fix server_group_members quota check 2018-07-11 15:04:34 -04:00
test_bug_1784353.py conductor: Recreate volume attachments during a reschedule 2018-10-22 16:08:10 -04:00

README.rst

Tests for Specific Regressions

When we have a bug reported by end users that we can write a full stack reproduce on, we should. And we should keep a regression test for that bug in our tree. It can be deleted at some future date if needed, but largely should not be changed.

Writing Regression Tests

  • These should be full stack tests which inherit from nova.test.TestCase directly. (This is to prevent coupling with other tests).
  • They should setup a full stack cloud in their setUp via fixtures
  • They should each live in a file which is named test_bug######.py

Writing Tests Before the Bug is Fixed

TODO describe writing and landing tests before the bug is fixed as a reproduce.