p35 jobs are failing with subunit.parser ... FAILED
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Glance |
Fix Released
|
Medium
|
Unassigned | ||
OpenStack Compute (nova) |
Fix Released
|
High
|
Stephen Finucane |
Bug Description
Various py35 jobs[1] are failing recently with:
2019-01-22 00:10:44.708744 | ubuntu-xenial | {0} subunit.parser [0.000360s] ... FAILED
2019-01-22 00:10:44.708947 | ubuntu-xenial |
2019-01-22 00:10:44.709000 | ubuntu-xenial | Captured Parser Error:
2019-01-22 00:10:44.709057 | ubuntu-xenial | ~~~~~~~
2019-01-22 00:10:44.709122 | ubuntu-xenial | b'Short read - got 167 bytes, wanted 4263 bytes'
It seems that when a job fails, the log of the test case that triggered the failure is truncated. Also it feels that it is caused by extensive logging.
There was a past cinder bug[2] that was marked fixed after couple of logging reduction.
There was 11 failures in the last 24 hours[3]
[1] http://
[2] https:/
[3] http://
tags: | added: testing |
Balazs Gibizer (balazs-gibizer) wrote : | #1 |
Matt Riedemann (mriedem) wrote : | #2 |
I think I first noticed it on this change which merged on Jan 18:
https:/
I don't know if they are related.
Changed in nova: | |
status: | New → Confirmed |
Matt Riedemann (mriedem) wrote : | #3 |
This goes back to at least Jan 14 so it could be something else:
Balazs Gibizer (balazs-gibizer) wrote : | #4 |
In about half of the cases the "Captured Parser Error" is missing from the logs, but other than that the symptoms are the same. Execution is aborted and the output of the test case triggered the error is truncated.
Changed in nova: | |
importance: | Undecided → High |
Balazs Gibizer (balazs-gibizer) wrote : | #5 |
The
b'Short read - got 167 bytes, wanted 4263 bytes'
error message comes from here [0].
However recent failure[1][2][3] shows a different message:
Not a text type application/
This message comes from [4].
All of this points towards that something strange is printed to the test output that confuses ourt tooling.
I've tried to simply recreate the problem by logging extensively from a test case but that leads to a different error:
Length too long: 16814020
[0] https:/
[1] http://
[2] http://
[3] http://
[4] https:/
Sean McGinnis (sean-mcginnis) wrote : | #6 |
Cinder hit "Not a text type application/
We "fixed" it in Cinder by cleaning up the amount of output generated from unit tests. Once I got rid of warning output and unnecessary logging to the point that there wasn't a lot of stdout/stderr generated, we stopped hitting whatever threshold was causing the error to surface.
Matt Riedemann (mriedem) wrote : | #7 |
Yeah that's a good point Sean. We have a lot of warnings in the py35 console output logs:
2019-01-22 00:07:14.688023 | ubuntu-xenial | Captured stderr:
2019-01-22 00:07:14.688079 | ubuntu-xenial | ~~~~~~~~~~~~~~~~
2019-01-22 00:07:14.688266 | ubuntu-xenial | b'/home/
2019-01-22 00:07:14.688383 | ubuntu-xenial | b' if not inspect.
2019-01-22 00:08:52.909079 | ubuntu-xenial | b'/home/
2019-01-22 00:09:48.658271 | ubuntu-xenial | Captured stderr:
2019-01-22 00:09:48.658308 | ubuntu-xenial | ~~~~~~~~~~~~~~~~
2019-01-22 00:09:48.658492 | ubuntu-xenial | b'/home/
2019-01-22 00:09:48.658577 | ubuntu-xenial | b' DeprecationWarn
2019-01-22 00:09:48.658845 | ubuntu-xenial | b"/home/
2019-01-22 00:09:48.658912 | ubuntu-xenial | b' "UPDATE / DELETE expressions." % clause'
2019-01-22 00:09:48.659175 | ubuntu-xenial | b"/home/
2019-01-22 00:09:48.659242 | ubuntu-xenial | b' "UPDATE / DELETE expressions." % clause'
2019-01-22 00:09:48.659271 | ubuntu-xenial | b''
That SAWarning is probably the worst at 866 occurrences.
Matt Riedemann (mriedem) wrote : | #8 |
https:/
Matt Riedemann (mriedem) wrote : | #9 |
Here is an example of where that SAWarning comes from:
nova.tests.
-------
Captured pythonlogging:
~~~~~~~
b'2019-01-31 17:48:15,014 INFO [248_add_
b''
Captured traceback:
~~~~~~~~~~~~~~~~~~~
b'Traceback (most recent call last):'
b' File "/home/
b" db.instance_
b' File "/home/
b' return IMPL.instance_
b' File "/home/
b' return f(*args, **kwargs)'
b' File "/home/
b' ectxt.value = e.inner_exc'
b' File "/home/
b' self.force_
b' File "/home/
b' six.reraise(
b' File "/home/
b' raise value'
b' File "/home/
b' return f(*args, **kwargs)'
b' File "/home/
b' return f(context, *args, **kwargs)'
b' File "/home/
b' count = query.soft_
b' File "/home/
b' synchronize_
b' File "/home/
b' update_op.exec_()'
b' File "/home/
b' self._do_
b' File "/home/
b' self._additiona
b' File "/home/
b' expression.
b' File "/home/
b' return meth(clause)'
b' File "/home/
Matt Riedemann (mriedem) wrote : | #10 |
I opened bug 1814182 against oslo.db in case it's something in the oslo.db code, but it could be a problem in how nova's data models are done, I'm not sure, need help from Mike Bayer.
Matt Riedemann (mriedem) wrote : | #11 |
Reported sqlalchemy-migrate bug 1814288 for the inspect.getargspec warnings generated from that library.
OpenStack Infra (hudson-openstack) wrote : Related fix proposed to nova (master) | #12 |
Related fix proposed to branch: master
Review: https:/
OpenStack Infra (hudson-openstack) wrote : | #13 |
Related fix proposed to branch: master
Review: https:/
OpenStack Infra (hudson-openstack) wrote : Related fix merged to nova (master) | #14 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: master
commit 1fa2e9c3a0afcba
Author: Matt Riedemann <email address hidden>
Date: Fri Feb 1 11:34:45 2019 -0500
Ignore SAWarnings for "Evaluating non-mapped column expression"
These warnings come from oslo.db code (tracked with bug 1814199)
so there isn't much nova can do about that right now, outside of
monkey patching oslo.db which is a bad idea.
Let's ignore the warning until the bug in oslo.db is fixed to
avoid blowing up our unit/functional test console output logs
which in turn is intermittently triggering subunit.parser failures.
Change-Id: Ifdfeadae8b348d
Related-Bug: #1813147
Matt Riedemann (mriedem) wrote : | #15 |
This is probably the next one to squash:
2019-02-04 16:58:37.674764 | ubuntu-bionic | Captured stderr:
2019-02-04 16:58:37.674800 | ubuntu-bionic | ~~~~~~~~~~~~~~~~
2019-02-04 16:58:37.675041 | ubuntu-bionic | b'/home/
2019-02-04 16:58:37.675100 | ubuntu-bionic | b' DeprecationWarn
2019-02-04 16:58:37.675135 | ubuntu-bionic | b''
39 occurrences in a py35 unit test job today.
OpenStack Infra (hudson-openstack) wrote : Related fix proposed to nova (master) | #16 |
Related fix proposed to branch: master
Review: https:/
OpenStack Infra (hudson-openstack) wrote : Related fix merged to nova (master) | #17 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: master
commit cd1823195fac3d1
Author: Jan Gutter <email address hidden>
Date: Tue Feb 5 14:51:49 2019 +0200
Ignore some PendingDeprecat
os-vif got bumped to 1.14.0 in https:/
Nova is touching the VIFPortProfileO
fashion until at least Train.
However, that commit contributes to triggering bug 1813147 in check and
gate. The warning is safe to silence and should be converted to an error
in Train (once the older path is removed).
Change-Id: I9d8175c78f068c
Signed-off-by: Jan Gutter <email address hidden>
Related-Bug: #1813147
blueprint: generic-
Matt Riedemann (mriedem) wrote : | #18 |
Looks like this is our biggest issue in the functional py35 job now:
2019-02-06 21:04:13.978116 | ubuntu-xenial | b'/home/
OpenStack Infra (hudson-openstack) wrote : Fix proposed to nova (master) | #19 |
Fix proposed to branch: master
Review: https:/
Changed in nova: | |
assignee: | nobody → Matt Riedemann (mriedem) |
status: | Confirmed → In Progress |
OpenStack Infra (hudson-openstack) wrote : Related fix merged to nova (master) | #20 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: master
commit d916a045c19beb0
Author: Matt Riedemann <email address hidden>
Date: Fri Feb 1 11:45:49 2019 -0500
Ignore sqla-migrate inspect.getargspec deprecation warnings
This adds a warnings filter to ignore the "inspect.
warnings from sqlalchemy-migrate until bug 1814288 is fixed
in that library.
Change-Id: Id8c02e74574298
Related-Bug: #1813147
OpenStack Infra (hudson-openstack) wrote : Fix merged to nova (master) | #21 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: master
commit c03cc26ee04734b
Author: Matt Riedemann <email address hidden>
Date: Thu Feb 7 09:04:25 2019 -0500
Fix deprecation warning for threadgroup.
This resolves the following deprecation warning:
b'/
lib/
Deprecation
function is deprecated. Use add_timer_args() instead.'
The add_timer_args method was added in 1.34.0:
Ib2791342
So the required version of oslo.service is also updated.
Change-Id: Id54226dc926839
Partial-Bug: #1813147
Matt Riedemann (mriedem) wrote : | #22 |
Hit another failure in a unit test job that failed and dumped a ton into the console log:
I'm not exactly sure what is causing it, but there is a ton of this in the logs:
2019-02-08 12:32:26.104934 | ubuntu-xenial | b'2019-02-08 12:32:17,445 ERROR [oslo_messaging
2019-02-08 12:32:26.105103 | ubuntu-xenial | b'Traceback (most recent call last):'
2019-02-08 12:32:26.105596 | ubuntu-xenial | b' File "/home/
2019-02-08 12:32:26.105775 | ubuntu-xenial | b' res = self.dispatcher
2019-02-08 12:32:26.106238 | ubuntu-xenial | b' File "/home/
2019-02-08 12:32:26.106450 | ubuntu-xenial | b' return self._do_
2019-02-08 12:32:26.106919 | ubuntu-xenial | b' File "/home/
2019-02-08 12:32:26.107114 | ubuntu-xenial | b' ctxt = self.serializer
2019-02-08 12:32:26.107452 | ubuntu-xenial | b' File "/home/
2019-02-08 12:32:26.107689 | ubuntu-xenial | b' ctxt = super(CheatingS
2019-02-08 12:32:26.108147 | ubuntu-xenial | b' File "/home/
2019-02-08 12:32:26.108323 | ubuntu-xenial | b' return captured_
2019-02-08 12:32:26.108727 | ubuntu-xenial | b' File "/home/
2019-02-08 12:32:26.108959 | ubuntu-xenial | b" self.assertEqua
2019-02-08 12:32:26.109406 | ubuntu-xenial | b' File "/home/
2019-02-08 12:32:26.121688 | ubuntu-xenial | b' self.assertThat
2019-02-08 12:32:26.122010 | ubuntu-xenial | b' File "/home/
2019-02-08 12:32:26.122098 | ubuntu-xenial | b' raise mismatch_error'
2019-02-08 12:32:26.122297 | ubuntu-xenial | b"testtools.
It seems that those assertions don't actually caus...
Matt Riedemann (mriedem) wrote : | #23 |
That MismatchError in the fake_deserializ
https:/
So there is not really a point in having that stubbed out just to assert something that doesn't matter to the test anyway.
OpenStack Infra (hudson-openstack) wrote : Fix proposed to nova (master) | #24 |
Fix proposed to branch: master
Review: https:/
Matt Riedemann (mriedem) wrote : | #25 |
Here is another one:
oslo.versionedo
2019-02-08 14:57:03.415588 | ubuntu-xenial | Captured pythonlogging:
2019-02-08 14:57:03.415656 | ubuntu-xenial | ~~~~~~~
2019-02-08 14:57:03.415873 | ubuntu-xenial | b'2019-02-08 14:56:52,203 INFO [oslo_versioned
2019-02-08 14:57:03.416090 | ubuntu-xenial | b'2019-02-08 14:56:52,204 INFO [oslo_versioned
2019-02-08 14:57:03.416301 | ubuntu-xenial | b'2019-02-08 14:56:52,204 INFO [oslo_versioned
2019-02-08 14:57:03.416511 | ubuntu-xenial | b'2019-02-08 14:56:52,204 INFO [oslo_versioned
So either we need to change that fixture to log at debug and release it, or change logging levels for oslo.versionedo
Matt Riedemann (mriedem) wrote : | #26 |
Reported bug 1815223 against oslo.versionedo
Matt Riedemann (mriedem) wrote : | #27 |
Another giant traceback in functional test logs:
That dumps a ton of traceback from:
b'oslo_
Since there is some sort of reschedule happening in that test, I'm assuming it's coming from:
{2} nova.tests.
Which was the first test after that traceback. I'm not sure what the 'no such table' is about, maybe the test is tearing down the DB before the actual service code is done running?
It also doesn't help that we're logging all of the os-traits contents at INFO each time:
b"2019-02-13 13:33:37,217 INFO [placement.
Matt Riedemann (mriedem) wrote : | #28 |
We might want to think about defaulting OS_STDERR_CAPTURE to 0.
Matt Riedemann (mriedem) wrote : | #29 |
Although if OS_STDERR_CAPTURE=0 then we wouldn't see deprecation warnings in the logs and we want to know those.
OpenStack Infra (hudson-openstack) wrote : Related fix proposed to nova (master) | #30 |
Related fix proposed to branch: master
Review: https:/
OpenStack Infra (hudson-openstack) wrote : Related fix merged to nova (master) | #31 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: master
commit 0840bcccc0ae570
Author: Matt Riedemann <email address hidden>
Date: Thu Feb 14 08:47:53 2019 -0500
Change sqlalchemy warnings filter to an error
The warning should be gone with change
I192e84ce75
oslo.db 4.44.0 so this changes the warnings
filter from ignore to error and bumps the minimum
required version of oslo.db to include that change.
Change-Id: If7b1a9613b5847
Related-Bug: #1813147
Related-Bug: #1814199
Changed in nova: | |
assignee: | Matt Riedemann (mriedem) → Balazs Gibizer (balazs-gibizer) |
OpenStack Infra (hudson-openstack) wrote : Fix merged to nova (master) | #32 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: master
commit 102679a1cf80d7a
Author: Matt Riedemann <email address hidden>
Date: Fri Feb 8 10:16:34 2019 -0500
Trim fake_deserializ
The assertions on the user_id and project_id in the
fake_
cause tests to fail if they raise MismatchError
because oslo.messaging just swallows the exception
and logs the traceback. With enough of these getting
logged it will cause subunit parser failures in the
console output because the stream is too large.
This removes the fake method and just changes the
stub to be a lambda that returns self.context which
is what fake_deserializ
the project_id/user_id assertions.
Change-Id: I26b201b410aa1d
Partial-Bug: #1813147
Changed in glance: | |
status: | New → Triaged |
importance: | Undecided → Medium |
Thierry Carrez (ttx) wrote : | #33 |
The issue is back in Glance py37 tests:
https:/
https:/
It's probably triggered by the "Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working" deprecation warning coming from all those different Python libraries.
OpenStack Infra (hudson-openstack) wrote : Related fix proposed to nova (master) | #34 |
Related fix proposed to branch: master
Review: https:/
Matt Riedemann (mriedem) wrote : | #35 |
We seem to be hitting this in nova again with a notifications functional test:
2019-05-02 03:23:47.450578 | ubuntu-bionic | {0} nova.tests.
That test runs a bunch of actions on a server which generates a lot of logs, plus tracebacks during volume attach:
2019-05-02 03:23:47.482528 | ubuntu-bionic | b'2019-05-02 03:23:44,276 INFO [nova.api.
2019-05-02 03:23:47.482635 | ubuntu-bionic | b'2019-05-02 03:23:44,314 INFO [nova.compute.
2019-05-02 03:23:47.482747 | ubuntu-bionic | b'2019-05-02 03:23:44,355 ERROR [nova.compute.
2019-05-02 03:23:47.482796 | ubuntu-bionic | b'Traceback (most recent call last):'
2019-05-02 03:23:47.482889 | ubuntu-bionic | b' File "/home/
2019-05-02 03:23:47.482932 | ubuntu-bionic | b' do_driver_
2019-05-02 03:23:47.483021 | ubuntu-bionic | b' File "/home/
2019-05-02 03:23:47.483081 | ubuntu-bionic | b' ret_val = method(obj, context, *args, **kwargs)'
2019-05-02 03:23:47.483169 | ubuntu-bionic | b' File "/home/
2019-05-02 03:23:47.483217 | ubuntu-bionic | b' virt_driver, do_driver_attach)'
2019-05-02 03:23:47.483308 | ubuntu-bionic | b' File "/home/
2019-05-02 03:23:47.483347 | ubuntu-bionic | b' do_driver_attach)'
2019-05-02 03:23:47.483451 | ubuntu-bionic | b' File "/home/
2019-05-02 03:23:47.483508 | ubuntu-bionic | b" self['mount_
2019-05-02 03:23:47.483625 | ubuntu-bionic | b' File "/home/
2019-05-02 03:23:47.483684 | ubuntu-bionic | b' return _mock_self.
2019-05-02 03:23:47.483800 | ubuntu-bionic | b' File "/home/
2019-05-02 03:23:47.483850 | ubuntu-bionic | b' ret_val = effect(*args, **kwargs)'
2019-05-02 03:23:47.483970 | ubuntu-bionic | b' File "/home/
Matt Riedemann (mriedem) wrote : | #36 |
For the issue in comment 35 I wonder if we could flush the stdio buffer between runs of each sub-action in that test.
Matt Riedemann (mriedem) wrote : | #37 |
So one issue is we're getting a giant traceback because these tests are using the CastAsCall fixture and we're not registering an expected_exception so we get a big traceback from oslo.messaging:
b'2019-05-02 03:23:44,434 ERROR [oslo_messaging
We could (1) stub that out for the test run or (2) convert the tests to not use the CastAsCall fixture. (2) is probably better long-term but might not be very easy.
Matt Riedemann (mriedem) wrote : | #38 |
Actually the test isn't using the CastAsCall fixture so I'm not sure why we're getting the traceback from oslo.messaging since I thought that was only used if you're doing a synchronous RPC call.
Balazs Gibizer (balazs-gibizer) wrote : | #39 |
The test injects an exception here [1] that cause the stack trace.
OpenStack Infra (hudson-openstack) wrote : | #40 |
Related fix proposed to branch: master
Review: https:/
Matt Riedemann (mriedem) wrote : | #41 |
It's the notification listener in oslo.messaging that logs the traceback, that's why it wasn't the CastAsCall thing I was expecting:
So yeah we can't really do much about that since these are notification tests. We could potentially stub out LOG.exception in the specific part of the test that triggers the fault, but that's pretty hacky.
OpenStack Infra (hudson-openstack) wrote : Related fix merged to nova (master) | #42 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: master
commit 84e64a995eb4567
Author: Matt Riedemann <email address hidden>
Date: Wed May 1 08:57:39 2019 -0400
Require at least oslo.versionedo
This is to pick up change If20663ecad19f1
for our lower-constraints CI job to avoid blowing up the subunit
parser with too much log output.
Change-Id: I3c404bd650521b
Related-Bug: #1813147
OpenStack Infra (hudson-openstack) wrote : | #43 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: master
commit 93027713ab614b5
Author: Balazs Gibizer <email address hidden>
Date: Thu May 2 12:24:19 2019 -0600
Reset the stored logs at each notification test steps
This is an atempt to solve a new appearance of bug 1813147 where the
long test_instance_
This patch drops the logs of each successful test step to try to avoid
the overload of the logging system.
Change-Id: I0e5db9b2f423ab
Related-Bug: #1813147
OpenStack Infra (hudson-openstack) wrote : Related fix proposed to nova (stable/stein) | #44 |
Related fix proposed to branch: stable/stein
Review: https:/
OpenStack Infra (hudson-openstack) wrote : Related fix merged to nova (stable/stein) | #45 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: stable/stein
commit 0956539edf93736
Author: Balazs Gibizer <email address hidden>
Date: Thu May 2 12:24:19 2019 -0600
Reset the stored logs at each notification test steps
This is an atempt to solve a new appearance of bug 1813147 where the
long test_instance_
This patch drops the logs of each successful test step to try to avoid
the overload of the logging system.
Change-Id: I0e5db9b2f423ab
Related-Bug: #1813147
(cherry picked from commit 93027713ab614b5
tags: | added: in-stable-stein |
OpenStack Infra (hudson-openstack) wrote : Related fix proposed to nova (master) | #46 |
Related fix proposed to branch: master
Review: https:/
OpenStack Infra (hudson-openstack) wrote : | #47 |
Related fix proposed to branch: master
Review: https:/
OpenStack Infra (hudson-openstack) wrote : Related fix merged to nova (master) | #48 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: master
commit 2973faa07063d21
Author: Matt Riedemann <email address hidden>
Date: Fri Sep 13 14:18:25 2019 -0400
libvirt: stub logging of host capabilities
By default we capture the logging output stream at INFO level
for unit/functional tests and anything that hits the libvirt
Host.
string that is defined in the fakelibvirt module. This can cause
subunit parsing to fail because there is too much output in the
buffer.
This change simply stubs out that particular logging call in the
FakeLibvirt
detailed libvirt tests like LibvirtConnTest
to not be good enough we can try stubbing at a more global level
like in the base nova TestCase.
Change-Id: I38b350bce90800
Related-Bug: #1813147
OpenStack Infra (hudson-openstack) wrote : Related fix proposed to nova (stable/stein) | #49 |
Related fix proposed to branch: stable/stein
Review: https:/
OpenStack Infra (hudson-openstack) wrote : Related fix merged to nova (master) | #50 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: master
commit bb2b7cad294bebd
Author: Matt Riedemann <email address hidden>
Date: Fri Sep 13 14:58:49 2019 -0400
Use SpawnIsSynchron
We see that in successful tests that are usually running some
kind of negative or reschedule scenario that they will dump
a stacktrace to the subunit output buffer because
_allocate_
can eventually blow up the subunit output buffer and cause
tests to fail.
To guard against this, we use the SpawnIsSynchron
the base class of two known test classes [1][2] that show a
stacktace with _allocate_
result, one of the tests has to be fixed due to improper stubbing.
[1] UnsupportedPort
[2] PortResourceReq
Change-Id: Ib18ecaebbeccc9
Related-Bug: #1813147
OpenStack Infra (hudson-openstack) wrote : Related fix merged to nova (stable/stein) | #51 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: stable/stein
commit d6b5e81df55394a
Author: Matt Riedemann <email address hidden>
Date: Fri Sep 13 14:18:25 2019 -0400
libvirt: stub logging of host capabilities
By default we capture the logging output stream at INFO level
for unit/functional tests and anything that hits the libvirt
Host.
string that is defined in the fakelibvirt module. This can cause
subunit parsing to fail because there is too much output in the
buffer.
This change simply stubs out that particular logging call in the
FakeLibvirt
detailed libvirt tests like LibvirtConnTest
to not be good enough we can try stubbing at a more global level
like in the base nova TestCase.
Conflicts:
NOTE(mriedem): The conflict is due to not having change
I06e1f7429c
Change-Id: I38b350bce90800
Related-Bug: #1813147
(cherry picked from commit 2973faa07063d21
OpenStack Infra (hudson-openstack) wrote : Related fix proposed to nova (master) | #52 |
Related fix proposed to branch: master
Review: https:/
OpenStack Infra (hudson-openstack) wrote : Related fix merged to nova (master) | #53 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: master
commit 79ee80d159e8be0
Author: Stephen Finucane <email address hidden>
Date: Wed Sep 25 16:18:41 2019 +0100
Ignore sqla-migrate inspect.getargspec deprecation warnings on py36
The 'message' argument is a regex so you need to escape brackets and the
likes. Strange that this worked on py27.
This was coincidentally fixed by Ib9f73521677322
in Train but this change is backportable while that one is not.
Change-Id: Ia1f0fa6a46bdd0
Signed-off-by: Stephen Finucane <email address hidden>
Related-Bug: #1813147
OpenStack Infra (hudson-openstack) wrote : | #54 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: master
commit d0ca30b80a5063d
Author: Stephen Finucane <email address hidden>
Date: Wed Sep 25 15:59:12 2019 +0100
Ignore warning from sqlalchemy-migrate
This has been fixed but not released. Another warning is updated since
modules don't have .py extensions.
Change-Id: Iaa93b974321430
Signed-off-by: Stephen Finucane <email address hidden>
Related-Bug: #1813147
OpenStack Infra (hudson-openstack) wrote : Related fix proposed to nova (stable/stein) | #55 |
Related fix proposed to branch: stable/stein
Review: https:/
OpenStack Infra (hudson-openstack) wrote : Related fix proposed to nova (master) | #56 |
Related fix proposed to branch: master
Review: https:/
OpenStack Infra (hudson-openstack) wrote : Related fix merged to nova (master) | #57 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: master
commit 3c27f3cdc0538c2
Author: Matt Riedemann <email address hidden>
Date: Fri Oct 4 16:48:56 2019 -0400
Filter out alembic logs below WARNING in tests
During functional tests we see quite a bit of alembic
INFO level logging which we don't really need. This
change sets the alembic logging level to WARNING to
remove that noise.
Change-Id: I9550e1c0c93b48
Related-Bug: #1813147
OpenStack Infra (hudson-openstack) wrote : Related fix merged to nova (stable/stein) | #58 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: stable/stein
commit 58ad6aba98afe25
Author: Stephen Finucane <email address hidden>
Date: Wed Sep 25 16:18:41 2019 +0100
Ignore sqla-migrate inspect.getargspec deprecation warnings on py36
The 'message' argument is a regex so you need to escape brackets and the
likes. Strange that this worked on py27.
This was coincidentally fixed by Ib9f73521677322
in Train but this change is backportable while that one is not.
Conflicts:
NOTE(mriedem): The conflict is due to not having change
Ib9f7352167
noted above.
Change-Id: Ia1f0fa6a46bdd0
Signed-off-by: Stephen Finucane <email address hidden>
Related-Bug: #1813147
(cherry picked from commit 79ee80d159e8be0
OpenStack Infra (hudson-openstack) wrote : Related fix proposed to nova (master) | #59 |
Related fix proposed to branch: master
Review: https:/
OpenStack Infra (hudson-openstack) wrote : Related fix merged to nova (master) | #60 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: master
commit 945d662d4f0a0b1
Author: Matt Riedemann <email address hidden>
Date: Wed Dec 4 09:50:17 2019 -0500
Handle ServiceNotFound in DbDriver.
If a service record is gone from the DB the _report_state method
will log a generic traceback every time the report interval runs,
which is every 10 seconds by default. This gets pretty noisy and
the error logged isn't very useful. One could get into this state
by deleting the service record in the API before stopping the actual
process that is running.
This simply handles the ServiceNotFound error and logs a more useful
error message without the noisy traceback.
Change-Id: If0336001fbe399
Related-Bug: #1813147
OpenStack Infra (hudson-openstack) wrote : Related fix proposed to nova (master) | #61 |
Related fix proposed to branch: master
Review: https:/
Stephen Finucane (stephenfinucane) wrote : | #62 |
I have a potential "fix" for this proposed at [1], with a related change for stestr here [2].
[1] https:/
[2] https:/
OpenStack Infra (hudson-openstack) wrote : Fix proposed to nova (master) | #63 |
Fix proposed to branch: master
Review: https:/
Changed in nova: | |
assignee: | Balazs Gibizer (balazs-gibizer) → Stephen Finucane (stephenfinucane) |
OpenStack Infra (hudson-openstack) wrote : Change abandoned on nova (master) | #64 |
Change abandoned by Stephen Finucane (<email address hidden>) on branch: master
Review: https:/
Reason: This should be resolved with testtools 1.4.0 and whatever the next version of stestr ends up being
OpenStack Infra (hudson-openstack) wrote : Fix merged to nova (master) | #65 |
Reviewed: https:/
Committed: https:/
Submitter: Zuul
Branch: master
commit a2f30f9d7cad1c0
Author: Stephen Finucane <email address hidden>
Date: Wed Mar 18 10:19:07 2020 +0000
Bump python-subunit minimum to 1.4.0
This contains the fix required to avoid the 'subunit parser error'
errors we've been seeing for some time now.
Change-Id: I605005e97b58a9
Signed-off-by: Stephen Finucane <email address hidden>
Closes-Bug: #1813147
Changed in nova: | |
status: | In Progress → Fix Released |
Changed in glance: | |
status: | Triaged → Fix Released |
Added the signature to elastic-recheck https:/ /review. openstack. org/632977