Deflaked: Guarddog_impl_test #14475
Deflaked: Guarddog_impl_test #14475
Conversation
…s flake with both threads to included in the miss / megamiss events when only one should be. Signed-off-by: Kevin Baichoo <kbaichoo@google.com>
|
/retest |
|
Retrying Azure Pipelines: |
test/server/guarddog_impl_test.cc
Outdated
| @@ -686,8 +692,14 @@ TEST_P(GuardDogActionsTest, MegaMissShouldOnlyReportRelevantThreads) { | |||
| // This will reset the loop interval timer, and should help us | |||
| // synchronize with the guard dog. | |||
| guard_dog_->forceCheckForTest(); | |||
| // Touch the second_dog in case we overslept in the real time system | |||
There was a problem hiding this comment.
should we do this extra petting only when this test is run in real-time mode?
There was a problem hiding this comment.
We can do that if you really want, but the extra petting it shouldn't affect the simulated time test much as we're really just trying to get one of the watch dogs petted and have the other not petted over the given threshold, so that when we do the final check the latter one triggers the event.
Petting the former more often doesn't change that.
There was a problem hiding this comment.
The reason I suggest that is that part of the goal of the simulated time test is that it should be completely deterministic in its behavior regardless of how loaded the testing machine is.
I think this workaround that's specific to real-time means we don't gain the confidence that the simtime version is 100% deterministic.
|
/wait |
Signed-off-by: Kevin Baichoo <kbaichoo@google.com>
…impl Signed-off-by: Kevin Baichoo <kbaichoo@google.com>
* master: (30 commits) Deflaked: Guarddog_impl_test (envoyproxy#14475) [fuzz] add fuzz tests for hpack encoding and decoding (envoyproxy#13315) [filters] Prevent a filter from sending local reply and continue (envoyproxy#14416) oauth2: improving coverage (envoyproxy#14479) owners: Change dio email address (envoyproxy#14498) macos build: Fix ninja install (envoyproxy#14495) http: use OptRef helper to reduce some boilerplate (envoyproxy#14361) doc: update test/integration/README.md (envoyproxy#14485) server: wait workers to start before draining parent. (envoyproxy#14319) api: relax inline_string length limitation in DataSource (envoyproxy#14461) oauth: properly stop filter chain when a response was sent (envoyproxy#14476) listener: deprecate use_proxy_proto (envoyproxy#14406) deps: update cel and remove a patch (envoyproxy#14473) preconnect: rename: (envoyproxy#14474) coverage: ratcheting limits (envoyproxy#14472) grpc mux: fix sending node again after stream is reset (envoyproxy#14080) [test] Replace printers_include with printers_lib. (envoyproxy#14442) tcp: nodelay in the new pool (envoyproxy#14453) test: replace mock_methodn macros with mock_method (envoyproxy#14450) tcp: extending tcp integration test (envoyproxy#14451) ... Signed-off-by: Michael Puncel <mpuncel@squareup.com>
Deflaked guarddog_impl_test; on overloaded machines it would sometimes flake under the real time system with both threads to included in the miss / megamiss events when only one should be.
Tested via 2000 runs, of which it all passed. Prior it would flake a few (~1-5)/1000 on slower / overloaded machines.
Signed-off-by: Kevin Baichoo kbaichoo@google.com
For an explanation of how to fill out the fields, please see the relevant section
in PULL_REQUESTS.md
Commit Message: Deflaked guarddog_impl_test under real time system.
Additional Description:
Risk Level: low
Testing: test change.
Docs Changes: N/A
Release Notes: N/A
Platform Specific Features: N/A
[Optional Runtime guard:]
[Optional Fixes #Issue]
[Optional Deprecated:]