> On June 19, 2014, 2:06 a.m., Vinod Kone wrote: > > src/slave/containerizer/isolators/network/port_mapping.hpp, line 210 > > <https://reviews.apache.org/r/21594/diff/4/?file=611964#file611964line210> > > > > What does this mean? Are there un-managed nonEphemeral ports?
commented to say that it's passed in by flags.resources. > On June 19, 2014, 2:06 a.m., Vinod Kone wrote: > > src/slave/containerizer/isolators/network/port_mapping.cpp, line 315 > > <https://reviews.apache.org/r/21594/diff/4/?file=611965#file611965line315> > > > > s/host// ? > > > > "host etho" would be weird in the log? looks like etho is hostname. > > > > here and everywhere else. We have 'host eth0' as well as 'eth0 in the container' that we config on. We are explicit about this through out the code. > On June 19, 2014, 2:06 a.m., Vinod Kone wrote: > > src/slave/containerizer/isolators/network/port_mapping.cpp, lines 1009-1010 > > <https://reviews.apache.org/r/21594/diff/4/?file=611965#file611965line1009> > > > > What if it's None? it means the link is not found. we have verified that earlier in the same func. > On June 19, 2014, 2:06 a.m., Vinod Kone wrote: > > src/slave/containerizer/isolators/network/port_mapping.hpp, lines 71-72 > > <https://reviews.apache.org/r/21594/diff/4/?file=611964#file611964line71> > > > > Should this not return an error if the range is already allocated? We have CHECKs inside, since it is only used by the network isolator. - Chi ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/21594/#review46101 ----------------------------------------------------------- On June 20, 2014, 12:13 a.m., Chi Zhang wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/21594/ > ----------------------------------------------------------- > > (Updated June 20, 2014, 12:13 a.m.) > > > Review request for mesos, Ian Downes, Jie Yu, Vinod Kone, and Cong Wang. > > > Bugs: https://issues.apache.org/jira/browse/MESOS-1324 > > https://issues.apache.org/jira/browse/https://issues.apache.org/jira/browse/MESOS-1324 > > > Repository: mesos-git > > > Description > ------- > > Added a network isolator using port-range based traffic redirection on Linux. > > - Containers are assigned non-ephemeral ports by the scheduler and ephemeral > ports by the network isolator. > - Virtual ethernet devices and Traffic Control filters are set up so that > network traffic in and out of the containers is isolated based on the ports > assigned to them. > - Containers run inside their own network namespaces with separate network > stacks, from which per-container network statistics can be retrieved. > > A joint work with: > - Cong Wang ([email protected]) > - Jie Yu ([email protected]) > - Ian Downes ([email protected]) > > > Diffs > ----- > > include/mesos/mesos.proto 2f6be05 > src/Makefile.am b1b7d2d > src/launcher/main.cpp b497e98 > src/slave/constants.hpp c65a62d > src/slave/constants.cpp 51f65bb > src/slave/containerizer/isolators/network/port_mapping.hpp PRE-CREATION > src/slave/containerizer/isolators/network/port_mapping.cpp PRE-CREATION > src/slave/containerizer/linux_launcher.cpp 85c74f0 > src/slave/containerizer/mesos_containerizer.cpp 61c0a8d > src/slave/flags.hpp 3b8ba08 > src/slave/main.cpp 8c2b70c > src/tests/environment.cpp 21b9d1d > src/tests/mesos.cpp 1037420 > > Diff: https://reviews.apache.org/r/21594/diff/ > > > Testing > ------- > > make check on linux. more test cases are being written. > > > Thanks, > > Chi Zhang > >
