Event tests fails on py-pgsql unittest environment

[Ilya Tyaptin]

Next tests fail on py-pgsql environment:

FAIL: ceilometer.tests.api.v2.test_event_scenarios.TestEventAPI.test_get_events_filter_datetime_trait(pgsql)
tags: worker-1
FAIL: ceilometer.tests.storage.test_storage_scenarios.GetEventTest.test_generated_is_datetime(pgsql)
FAIL: ceilometer.tests.api.v2.test_event_scenarios.TestEventAPI.test_get_events_filter_op_datatime(pgsql)
FAIL: ceilometer.tests.storage.test_storage_scenarios.CounterDataTypeTest.test_storage_can_handle_large_values(pgsql)
FAIL: ceilometer.tests.storage.test_storage_scenarios.GetEventTest.test_generated_is_datetime(pgsql)

[gordon chung (chungg)]

awesome. i broke postgres

[gordon chung (chungg)]

i apparently broke mysql too. there is no intersect in mysql

[gordon chung (chungg)]

so i think there's a few different issues... your errors are datetime related i think it's because our precisetimestamp model only effects mysql

see https://github.com/openstack/ceilometer/blob/master/ceilometer/storage/sqlalchemy/models.py#L52-L78

i plan on opening separate bugs related to migration script, mysql and trait ordering

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/157836

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/157836
Committed: https://git.openstack.org/cgit/openstack/ceilometer/commit/?id=4083bfd0eab14a70ed294ea2bf7abb90b0d19cab
Submitter: Jenkins
Branch: master

commit 4083bfd0eab14a70ed294ea2bf7abb90b0d19cab
Author: Ilya Tyaptin <email address hidden>
Date: Fri Feb 20 17:55:20 2015 +0300

    Fix unittests for supporting py-pgsql env

    In CR next testcases are fixed:
    1. tests.storage.test_storage_scenarios.CounterDataTypeTest.
    test_storage_can_handle_large_values
    Currently we use for assertion 16-decimal number, but in
    postgresql double precise type used for counter_volume.
    So it supports only 15-decimal accuracy. Test volume should line
    with the practical upper bound on storable value in any backend
    that the test hits and in this CR used 15-decimal assertion number.

    2. tests.storage.test_storage_scenarios.GetEventTest.
    test_generated_is_datetime
    Issue is in traits order for asserted events. In mysql it's
    alphabetical asc and something else in postgresql.

    Partially-Bug: 1423545
    Change-Id: I0b96324eb1725a46aa80cd7d942c1b0b66ddac16

[gordon chung (chungg)]

i don't know why it didn't update status but the patch is committed

[gordon chung (chungg)]

now i know why it didn't close there's still tests failing

FAIL: ceilometer.tests.api.v2.test_event_scenarios.TestEventAPI.test_get_events_filter_datetime_trait(pgsql)
tags: worker-1
FAIL: ceilometer.tests.api.v2.test_event_scenarios.TestEventAPI.test_get_events_filter_op_datatime(pgsql)

specifically the above... relating to precisetimestamp.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/163208

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/163208
Committed: https://git.openstack.org/cgit/openstack/ceilometer/commit/?id=7e35570393b18a49a9d707e75f5782dd6f274ede
Submitter: Jenkins
Branch: master

commit 7e35570393b18a49a9d707e75f5782dd6f274ede
Author: gordon chung <email address hidden>
Date: Tue Mar 10 17:59:22 2015 -0400

    normalise timestamp in query

    when type casting query parameters, we use parse_isotime. this has
    a side effect of adding a tzinfo to the resulting datetime. no db
    currently stores tzinfo so we should normalise query to drop this
    information. as a side effect, it also adds in missing supported type

    Change-Id: Id4a5ffbb5388122b43adcf5054a7cd0cce3d275d
    Closes-Bug: #1423545