Fix flakey container pool test |
||||||
Issue descriptionPoolTests.testCleanup_timeout is acting flakey. See - https://uberchromegw.corp.google.com/i/chromeos/builders/daisy_skate-paladin/builds/11500/steps/UnitTest/logs/stdio - https://luci-logdog.appspot.com/v/?s=chromeos%2Fbb%2Fchromiumos.tryserver%2Fno_vmtest_pre_cq%2F144752%2F%2B%2Frecipes%2Fsteps%2FUnitTest%2F0%2Fstdout
,
Oct 13 2017
,
Oct 16 2017
pool_unittest failed in CQ: https://uberchromegw.corp.google.com/i/chromeos/builders/betty-arc64-paladin/builds/704/steps/UnitTest/logs/stdio
,
Oct 17 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/autotest/+/a1b3b1e76448f5c968dbe17ece4e524f70ae0e72 commit a1b3b1e76448f5c968dbe17ece4e524f70ae0e72 Author: Ben Kwa <kenobi@google.com> Date: Tue Oct 17 04:03:17 2017 Temporarily disable a flakey test. BUG= chromium:774534 TEST=pool_unittest.py PoolTests.testCleanup_timeout (should skip the test) Change-Id: I5293ebae524729c7f5f641e1a0c45cefcb9f0add Reviewed-on: https://chromium-review.googlesource.com/719217 Reviewed-by: Ben Kwa <kenobi@chromium.org> Tested-by: Ben Kwa <kenobi@chromium.org> Commit-Queue: Ben Kwa <kenobi@chromium.org> [modify] https://crrev.com/a1b3b1e76448f5c968dbe17ece4e524f70ae0e72/site_utils/lxc/container_pool/pool_unittest.py
,
Oct 17 2017
Issue 775355 has been merged into this issue.
,
Oct 25 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/autotest/+/785edabb9a4fe2c4c23d5fe2f53f44ce0640a0db commit 785edabb9a4fe2c4c23d5fe2f53f44ce0640a0db Author: Ben Kwa <kenobi@google.com> Date: Wed Oct 25 05:47:57 2017 Re-enable a de-flaked unittest. Add a timeout to the test to wait for all worker threads to.. time out. BUG= chromium:774534 TEST=pool_unittest.py PoolTests.testCleanup_timeout TEST=Run the unit-test repeatedly with and without the fix. Without the fix, the test fails on repeated runs. With the fix, it ran and passed 200x with no flakes. Change-Id: Ib5500ea68bb3844bad113079accef22ee9d0239a Reviewed-on: https://chromium-review.googlesource.com/736773 Commit-Ready: Ben Kwa <kenobi@chromium.org> Tested-by: Ben Kwa <kenobi@chromium.org> Reviewed-by: Dan Shi <dshi@google.com> [modify] https://crrev.com/785edabb9a4fe2c4c23d5fe2f53f44ce0640a0db/site_utils/lxc/container_pool/pool_unittest.py
,
Oct 25 2017
,
Oct 27 2017
https://uberchromegw.corp.google.com/i/chromeos/builders/tricky-paladin/builds/10014
,
Nov 9 2017
test_CleanupTimeout is no longer flaking.... but (another one is).
,
Nov 16 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromiumos/third_party/autotest/+/607e3b75aab7c6416449e50469a6abf9217219ed commit 607e3b75aab7c6416449e50469a6abf9217219ed Author: Ben Kwa <kenobi@google.com> Date: Thu Nov 16 00:07:28 2017 [autotest] Fix a flakey test. - Instead of fiddling with the error queue, use the new worker count to wait for workers to time out. - Remove the intermediate assert which was unnecessary (and flakey). The problem was that workers starting up in the interim period before the pool fully shuts down, causes this assert to be incorrect. What we really care about is that (1) something timed out, and (2) the final container creation and destruction counts are equal in spite of the timeout. BUG= chromium:774534 TEST=pool_unittest.py -v PoolTests.testCleanup_timeout TEST=Repeatedly run test under load. Observe flakes. After fix, observe no more flakes. TEST=Insert a time.sleep(1) in the top of Pool.cleanup to exacerbate the race condition. Observe now consistent failures. Apply fix, observe no more failures (don't keep the sleep in there as it is a waste of time under non-testing conditions). Change-Id: Ia35fd6a1709369729178122772880cdf6f944b11 Reviewed-on: https://chromium-review.googlesource.com/772291 Commit-Ready: Ben Kwa <kenobi@chromium.org> Tested-by: Ben Kwa <kenobi@chromium.org> Reviewed-by: Ilja H. Friedel <ihf@chromium.org> Reviewed-by: Ben Kwa <kenobi@chromium.org> Reviewed-by: Dan Shi <dshi@google.com> [modify] https://crrev.com/607e3b75aab7c6416449e50469a6abf9217219ed/site_utils/lxc/container_pool/pool_unittest.py [modify] https://crrev.com/607e3b75aab7c6416449e50469a6abf9217219ed/site_utils/lxc/container_pool/error.py [modify] https://crrev.com/607e3b75aab7c6416449e50469a6abf9217219ed/site_utils/lxc/unittest_setup.py
,
Nov 16 2017
Issue 786129 has been merged into this issue.
,
Nov 16 2017
Issue 779686 has been merged into this issue.
,
Nov 29 2017
It looks like the bug is still present in the R63 branch: Failed builds: https://luci-milo.appspot.com/buildbot/chromeos_release/novato-arc64-release%20release-R63-10032.B/49 https://luci-milo.appspot.com/buildbot/chromeos_release/guado_moblab-release%20release-R63-10032.B/53 Flakey: https://luci-milo.appspot.com/buildbot/chromeos_release/reef-android-nyc-pre-flight-branch%20release-R63-10032.B/210 https://luci-milo.appspot.com/buildbot/chromeos_release/reef-android-nyc-pre-flight-branch%20release-R63-10032.B/215 |
||||||
►
Sign in to add a comment |
||||||
Comment 1 by kenobi@chromium.org
, Oct 13 2017