> On April 13, 2017, 12:45 a.m., Vinod Kone wrote: > > src/slave/containerizer/mesos/containerizer.cpp > > Lines 1386-1389 (original), 1401-1404 (patched) > > <https://reviews.apache.org/r/58262/diff/3/?file=1686498#file1686498line1404> > > > > Previously debug containers had this set in the env as well? But not > > anymore?
Correct. `environment` represents "basic" environment, e.g. agent environment. I'd argue that we should not set it for debug containers, but rather inherit parent's basic environment as part of inheritance process. Moreover, currently, debug container launch path does not set `environment` at all: https://github.com/apache/mesos/blob/aca6796b703a8925319087d33d7dc5e5539f50d3/src/slave/containerizer/mesos/containerizer.cpp#L1830 - Alexander ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/58262/#review171829 ----------------------------------------------------------- On April 7, 2017, 4:21 p.m., Alexander Rukletsov wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/58262/ > ----------------------------------------------------------- > > (Updated April 7, 2017, 4:21 p.m.) > > > Review request for mesos, Gastón Kleiman, Jie Yu, Kevin Klues, and Vinod Kone. > > > Bugs: MESOS-6782 > https://issues.apache.org/jira/browse/MESOS-6782 > > > Repository: mesos > > > Description > ------- > > See summary. > > > Diffs > ----- > > src/slave/containerizer/mesos/containerizer.hpp > 29a99f33e646593127b9dc126f398f7bca88e21d > src/slave/containerizer/mesos/containerizer.cpp > bc611a5e085de10e9953b5f942d98f2b5747fce6 > src/slave/containerizer/mesos/paths.hpp > d85fd34660faacd9c73de2ba7b87b3bbd4b6007b > src/slave/containerizer/mesos/paths.cpp > ed4bbd2491e71ad1e4a41e0575b514377d02da9b > > > Diff: https://reviews.apache.org/r/58262/diff/3/ > > > Testing > ------- > > See https://reviews.apache.org/r/58263/ > > > Thanks, > > Alexander Rukletsov > >
