----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/21451/#review43144 -----------------------------------------------------------
src/slave/containerizer/isolators/cgroups/perf_event.cpp <https://reviews.apache.org/r/21451/#comment77172> Seems odd to send events, flags.perf_interval and flags.perf_duration when you are already sending 'flags'. Why not let CgroupsPerfEventIsolatorProcess deduce those from flags? You could still do the validation here but just send 'flags' over. src/slave/containerizer/isolators/cgroups/perf_event.cpp <https://reviews.apache.org/r/21451/#comment77181> So this is returning a cached value instead of collecting the stats right now like we do for cpu,memory usage stats. Is this intended? src/slave/containerizer/isolators/cgroups/perf_event.cpp <https://reviews.apache.org/r/21451/#comment77177> Why this change? src/slave/flags.hpp <https://reviews.apache.org/r/21451/#comment77167> You probably want to mention about sanitization and given an example here. src/slave/flags.hpp <https://reviews.apache.org/r/21451/#comment77169> Is there a sensible default than empty? src/slave/flags.hpp <https://reviews.apache.org/r/21451/#comment77170> This should be optional if we want a default of none. - Vinod Kone On May 14, 2014, 10:05 p.m., Ian Downes wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/21451/ > ----------------------------------------------------------- > > (Updated May 14, 2014, 10:05 p.m.) > > > Review request for mesos, Benjamin Hindman and Vinod Kone. > > > Bugs: MESOS-1278 > https://issues.apache.org/jira/browse/MESOS-1278 > > > Repository: mesos-git > > > Description > ------- > > Three new slave flags: > 1. perf_events is a list of perf events to sample. This list is validated > through perf during isolation creation. > 2. perf_interval is the time between starting new samples > 3. perf_duration is the duration of each sample > > e.g., you could sample cpu-cycles and cpu-migrations for 10 seconds every 60 > seconds. > > Isolator::usage() will initially return an empty PerfStatistics protobuf > (containing only the timing information) until the first sample is obtained. > After that it will return the most recent sample. > > > Diffs > ----- > > src/slave/containerizer/isolators/cgroups/perf_event.hpp PRE-CREATION > src/slave/containerizer/isolators/cgroups/perf_event.cpp PRE-CREATION > src/slave/containerizer/mesos_containerizer.cpp > 2a4816ec34f2298b71e8591c1e92d77f94ef4cb3 > src/slave/flags.hpp 8616817b9602f6ccf56e2af61f36c1bd295df1a1 > src/tests/isolator_tests.cpp b0eff5721b6ea4dcee8ff3748a2a11ade809e30e > > Diff: https://reviews.apache.org/r/21451/diff/ > > > Testing > ------- > > Test forthcoming. > > > Thanks, > > Ian Downes > >
