> On Feb. 29, 2020, 1:55 a.m., Greg Mann wrote: > > src/slave/slave.cpp > > Lines 3201-3204 (original), 3285-3291 (patched) > > <https://reviews.apache.org/r/71858/diff/10/?file=2212019#file2212019line3285> > > > > I was just thinking about this more... IIUC, we actually want to always > > set limits for executors right now, don't we? We previously discussed that > > we have several cases to handle (default, command, and custom executors), > > and I believe in the case of custom executors, we need to be sure that > > limits are always set. Wouldn't the current code allow a custom executor to > > be launched with no resource constraints if the framework specifies > > `shared_cgroups==false` and does not set limits in the tasks?
> Wouldn't the current code allow a custom executor to be launched with no > resource constraints if the framework specifies shared_cgroups==false and > does not set limits in the tasks? In that case, for memory subsystem we will set custom executor's limit to its request which is the sum of all its task's request, so custom executor will still be constrained. For CPU subsystem, it depends on the `--cgroups_enable_cfs` flag, if this flag is true, then similar with memory subsytem we will set custom executor's CFS quota to its request which is the sum of all its task's request, if this flag is false, then yes the custom executor will not be constrained which makes sense because its tasks need to be unconstrained as well (since they have no limit set). - Qian ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/71858/#review219696 ----------------------------------------------------------- On Feb. 25, 2020, 9:46 a.m., Qian Zhang wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/71858/ > ----------------------------------------------------------- > > (Updated Feb. 25, 2020, 9:46 a.m.) > > > Review request for mesos, Andrei Budnik and Greg Mann. > > > Bugs: MESOS-10046 > https://issues.apache.org/jira/browse/MESOS-10046 > > > Repository: mesos > > > Description > ------- > > Set resource limits when launching executor container. > > > Diffs > ----- > > src/slave/slave.hpp 03279db082bdba23dbfeb2d93081a908e609aec2 > src/slave/slave.cpp cce275a504effae7a6b71dd333ce8a300d1ce5be > > > Diff: https://reviews.apache.org/r/71858/diff/10/ > > > Testing > ------- > > > Thanks, > > Qian Zhang > >
