> On Feb. 23, 2015, 5:43 p.m., Ian Downes wrote: > > src/tests/containerizer_tests.cpp, line 427 > > <https://reviews.apache.org/r/31250/diff/1/?file=871371#file871371line427> > > > > This is hard-coded to true already and is soon to be removed, I > > believe. Can you check if this is all necessary?
I had to add it because otherwise resources from non-checkpointing slaves will be considered not useable by checkpointing frameworks. > On Feb. 23, 2015, 5:43 p.m., Ian Downes wrote: > > src/slave/containerizer/isolators/cgroups/cpushare.cpp, line 421 > > <https://reviews.apache.org/r/31250/diff/1/?file=871370#file871370line421> > > > > It's hitting the kernel: I'm not sure how it internally keeps track of > > the processes/threads in a cgroup, I would hope that it's not a list but.... it is a list. left a comment in the code. - Chi ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/31250/#review73591 ----------------------------------------------------------- On Feb. 24, 2015, 12:45 a.m., Chi Zhang wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/31250/ > ----------------------------------------------------------- > > (Updated Feb. 24, 2015, 12:45 a.m.) > > > Review request for mesos and Ian Downes. > > > Bugs: mesos-2103 > https://issues.apache.org/jira/browse/mesos-2103 > > > Repository: mesos > > > Description > ------- > > Ian, as discussed, I put the logic in cpu isolator for now until linux > launcher sees more and more need to report isolator-independent cgroup usage. > > > Diffs > ----- > > include/mesos/mesos.proto 9df972d750ce1e4a81d2e96cc508d6f83cad2fc8 > src/slave/containerizer/isolators/cgroups/cpushare.cpp > b6df23963f20e5f14f484d449733655d46cd6884 > src/tests/isolator_tests.cpp f7323965b94543bfda78afc4f60fc677f4900ecd > > Diff: https://reviews.apache.org/r/31250/diff/ > > > Testing > ------- > > added a new test. the test is end-to-end so that future replacement of the > actual logic won't require a change. > > > Thanks, > > Chi Zhang > >
