> On Aug. 25, 2016, 9:23 p.m., Kevin Klues wrote: > > src/slave/containerizer/docker.cpp, lines 1028-1034 > > <https://reviews.apache.org/r/50523/diff/3/?file=1483054#file1483054line1028> > > > > I feel like failing to recover devices warrants more than just a LOG > > message. > > > > Consider the situation where we fail to recover them, and we just > > continue. > > > > From Mesos's perspective, these devices are now free to hand out to > > other containers. This will cause conflicts -- especially in the case of > > GPUs. > > > > We need to think through the failure scenario here a bit more.
I agree. Considering it as a failure condition in the new code. Let me know if you have other ideas. > On Aug. 25, 2016, 9:23 p.m., Kevin Klues wrote: > > src/slave/containerizer/docker.cpp, line 1118 > > <https://reviews.apache.org/r/50523/diff/3/?file=1483054#file1483054line1118> > > > > You should use `strings::remove(deviceString, prefix)` here. used string erase here - Rajat ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/50523/#review146854 ----------------------------------------------------------- On Aug. 24, 2016, 12:56 a.m., Rajat Phull wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/50523/ > ----------------------------------------------------------- > > (Updated Aug. 24, 2016, 12:56 a.m.) > > > Review request for mesos, Benjamin Mahler, Kevin Klues, Yubo Li, and Vikrama > Ditya. > > > Bugs: MESOS-5795 > https://issues.apache.org/jira/browse/MESOS-5795 > > > Repository: mesos > > > Description > ------- > > Updated docker recovery to account for GPU resources. > > > Diffs > ----- > > src/slave/containerizer/docker.hpp f2a06065cf99fed934c2c1ffc47461ec8a97f50d > src/slave/containerizer/docker.cpp 5c1ee8e467d1c54c60b67dc5275ef71e1bb90723 > src/tests/containerizer/docker_containerizer_tests.cpp > 42d4364b6fcbc94c7852721511001c103cb5a90d > > Diff: https://reviews.apache.org/r/50523/diff/ > > > Testing > ------- > > GTEST_FILTER="DockerContainerizerTest.ROOT_NVIDIA_GPU_DOCKER_LaunchWithGpuRecovery" > make -j check > > > Thanks, > > Rajat Phull > >
