----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/61880/ -----------------------------------------------------------
Review request for mesos and Jiang Yan Xu. Bugs: MESOS-7714 https://issues.apache.org/jira/browse/MESOS-7714 Repository: mesos Description ------- Previously, `checkpoint(path, resources)` was overloaded such that it would automatically downgrade the resources before being checkpointed on the agent. However, `checkpoint(path, protobuf_containing_resources)` did not work correctly since we didn't recursively look within the messages to downgrade the resources. Ideally, we would use protobuf reflection to ensure that these are handled automatically. For now, we attempt to get all of the places where resources are present within a message. Diffs ----- src/slave/slave.cpp eac896c43ca0d822f94ed853107b1a9e99d7e05d src/slave/state.hpp 18c43193349ca6ec3d18967dcee69324a435e2fa src/tests/slave_recovery_tests.cpp 9aa0a510d1baad9aea13c03229816ca7c661a37c Diff: https://reviews.apache.org/r/61880/diff/1/ Testing ------- ### Before 1. 1. Ran a modified persistent volume framework that doesn't stop after the tasks finish. 2. Downgrading the agent from master to 1.3.1. 3. Observed that the agent cannot start due to resources being incompatible. ### After 1. Ran a modified persistent volume framework that doesn't stop after the tasks finish. 2. Downgrading the agent from master to 1.3.1. 3. Observed that the agent starts successfully. Thanks, Michael Park