----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/71944/#review219158 -----------------------------------------------------------
src/slave/containerizer/mesos/isolators/cgroups/subsystems/memory.cpp Lines 192-195 (patched) <https://reviews.apache.org/r/71944/#comment307265> This comment simplifies the OOM score a bit; adjustments are also made based on how long the process has been running and which capabilities it posesses, for example. Could you update the comment to explain? Maybe something like "While the OOM score of a process is a complex function of the process state and configuration, a decent approximation of the OOM score is 10 x percent of memory used..." src/slave/containerizer/mesos/isolators/cgroups/subsystems/memory.cpp Lines 199 (patched) <https://reviews.apache.org/r/71944/#comment307267> Do we really want to do this? My concern is that this will make any non-Mesos-task processes on the node (networking and security components, for example) more likely to be OOM-killed than Mesos tasks. Perhaps we should only set the OOM score adjustment for burstable tasks. What do you think? src/slave/containerizer/mesos/isolators/cgroups/subsystems/memory.cpp Lines 201-204 (patched) <https://reviews.apache.org/r/71944/#comment307266> I might recommend: "We would like burstable containers which consume more memory than their memory request to be preferentially OOM-killed first. To accomplish this, we set their OOM score adjustment as shown below, which attempts to ensure that the container which consumes more memory than its memory request will have an OOM score of 1000." - Greg Mann On Jan. 5, 2020, 2:06 p.m., Qian Zhang wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/71944/ > ----------------------------------------------------------- > > (Updated Jan. 5, 2020, 2:06 p.m.) > > > Review request for mesos, Andrei Budnik and Greg Mann. > > > Bugs: MESOS-10048 > https://issues.apache.org/jira/browse/MESOS-10048 > > > Repository: mesos > > > Description > ------- > > Set container process's OOM score adjust. > > > Diffs > ----- > > src/slave/containerizer/mesos/isolators/cgroups/subsystems/memory.hpp > 27d88e91fb784179effd54781f84000fe85c13eb > src/slave/containerizer/mesos/isolators/cgroups/subsystems/memory.cpp > 0896d37761a11f55ba4b866d235c3bd2b79dcfba > > > Diff: https://reviews.apache.org/r/71944/diff/2/ > > > Testing > ------- > > sudo make check > > > Thanks, > > Qian Zhang > >
