----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/40667/ -----------------------------------------------------------
Review request for mesos, Michael Park and Vinod Kone. Bugs: MESOS-4002 https://issues.apache.org/jira/browse/MESOS-4002 Repository: mesos Description ------- 1. Fixed some flakiness in the reservation tests around `resourceOffers` expectation being set after `driver.start()` leading to a race condition sometimes. ```GMOCK WARNING: Uninteresting mock function call - returning directly. Function call: resourceOffers(0x7fffe27c3830, @0x2ae03c9ce9d0 { 144-byte object <D0-1C B2-33 E0-2A 00-00 00-00 00-00 00-00 00-00 70-CF 00-48 E0-2A 00-00 10-D0 00-48 E0-2A 00-00 B0-D0 00-48 E0-2A 00-00 50-D1 00-48 E0-2A 00-00 A0-D1 00-48 E0-2A 00-00 40-00 01-48 E0-2A 00-00 ... 00-00 00-00 00-00 00-00 00-00 00-00 00-00 00-00 00-00 00-00 35-61 39-39 00-00 00-00 00-00 00-00 00-00 00-00 00-00 00-00 00-00 00-00 63-36 30-2D 00-00 00-00 00-00 00-00 00-00 00-00 1F-00 00-00> }) S``` 2. Added `AWAIT_READY` for the returned future from 1. 3. Removed the initial call to `reviveOffers` after `driver.start()` since that was redundant. -- Not to be committed-- Still does not fixes MESOS-4002 completely though. The test kept looping for 300 minutes meaning there is some future that was never ready and we invoked a `.get()` blocking forever or a deadlock somewhere else. Would keep digging for the root cause and for future failures on the ASF CI leading to more clues. Diffs ----- src/tests/reservation_endpoints_tests.cpp f4e332327049944000baccd3e607201240a8c5f4 Diff: https://reviews.apache.org/r/40667/diff/ Testing ------- make check + ran the tests in a loop. They did not fail for me before the fix too :-( Thanks, Anand Mazumdar