----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/57384/#review168222 -----------------------------------------------------------
src/slave/containerizer/containerizer.hpp Lines 155 (patched) <https://reviews.apache.org/r/57384/#comment240423> s/Destroy/Cleanup/ src/slave/containerizer/containerizer.hpp Lines 157 (patched) <https://reviews.apache.org/r/57384/#comment240424> s/destroy/cleanup/ Maybe rewrite the comment to say: ``` // NOTE: You can only cleanup artifacts from a // nested container that has been fully destroyed. ``` src/slave/containerizer/mesos/containerizer.cpp Lines 2482 (patched) <https://reviews.apache.org/r/57384/#comment240421> It's probably worth putting a TODO here that this function should check that recovery has completed before continuing. src/slave/containerizer/mesos/containerizer.cpp Lines 2486-2489 (patched) <https://reviews.apache.org/r/57384/#comment240420> `s/The nested container/Nested Container/` `s/hasn't/has not/` Also, (I know it's not consistent elsewhere) but can you put the containerID inside single quotes? Same for the error messages below too. src/slave/containerizer/mesos/containerizer.cpp Lines 2494 (patched) <https://reviews.apache.org/r/57384/#comment240422> s/terminated/destroyed/ src/tests/containerizer/nested_mesos_containerizer_tests.cpp Lines 2037 (patched) <https://reviews.apache.org/r/57384/#comment240426> Can you just have the command here be `true`? Then you don't need the explicit destroy below -- the container will exit very quickly by itself. src/tests/containerizer/nested_mesos_containerizer_tests.cpp Lines 2052-2054 (patched) <https://reviews.apache.org/r/57384/#comment240425> These two comments seem redundant. src/tests/containerizer/nested_mesos_containerizer_tests.cpp Lines 2052-2056 (patched) <https://reviews.apache.org/r/57384/#comment240427> If you follow the suggestion above, you should expect an exit status of 0 here. src/tests/containerizer/nested_mesos_containerizer_tests.cpp Lines 2068 (patched) <https://reviews.apache.org/r/57384/#comment240429> Kill this newline src/tests/containerizer/nested_mesos_containerizer_tests.cpp Lines 2071-2075 (patched) <https://reviews.apache.org/r/57384/#comment240428> I think this check is unnecessary and it desitracts from the flow of what the test is trying to verify. src/tests/containerizer/nested_mesos_containerizer_tests.cpp Lines 2083 (patched) <https://reviews.apache.org/r/57384/#comment240430> kill this newline src/tests/containerizer/nested_mesos_containerizer_tests.cpp Lines 2089 (patched) <https://reviews.apache.org/r/57384/#comment240432> Can you pull this wait out of the AWAIT_READY() call in order to be consistent with the pattern above? - Kevin Klues On March 7, 2017, 5:13 p.m., Gastón Kleiman wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/57384/ > ----------------------------------------------------------- > > (Updated March 7, 2017, 5:13 p.m.) > > > Review request for mesos, Alexander Rukletsov, Jie Yu, Kevin Klues, and Vinod > Kone. > > > Bugs: MESOS-7120 > https://issues.apache.org/jira/browse/MESOS-7120 > > > Repository: mesos > > > Description > ------- > > This new method cleans up the sandbox and runtime directories of a > terminated nested container. > > > Diffs > ----- > > src/slave/containerizer/containerizer.hpp > f65a9b9761fc254bd0778bf13aac9b256497b22f > src/slave/containerizer/mesos/containerizer.hpp > 09f94ccb3224c14a9324961b789455b119ec2257 > src/slave/containerizer/mesos/containerizer.cpp > b001d0265ec73822193eaac74c631930acce90c0 > src/tests/containerizer/nested_mesos_containerizer_tests.cpp > ea01fe55a28d17105157004d8cf0976202a49b7c > > > Diff: https://reviews.apache.org/r/57384/diff/1/ > > > Testing > ------- > > Added a test and verified that it works on Linux. > > > Thanks, > > Gastón Kleiman > >